|º G 9604. UNCLASSIFIED i- B 727,375 " AD THE FUNDAMENTALS OF HUMAN AND MACH | NE TRAN SLA T | ONS 633 6.82 | . . . Revz in , et a . For eign Technology Division - Wright-Patter son Air Force Base, Ohio March 966 Processed for . . . DEFENSE DOCUMENTATION CENTER DEFENSE SUPPLY AGENCY _-T— (CILIE/ADRDENIGHG)(USE FOR FEDERAL SCIENTIFIC AND TECHNICAL INFORMATION I U. S. DEPARTMENT OF COMMERCE / NATIONAL BUREAU OF STANDARDS / INSTITUTE FOR APPLIED TECHNOLOGY UNCLASSIFIED THE UNIVERSITY OF MICH!GAN L'EFARES Notice to DEFENSE Documentation center users This document is being distributed by the Clearinghouse for Federal Scientific and Technical Information, Department of Commerce, as a result of a recent agreement between the Department of Defense (DOD) and the Department of Commerce (DOC). * The Clearinghouse is distributing unclassified, unlimited documents which are or have been announced in the Technical Abstract Bulletin (TAB) of the Defense Documentation Center. . The price does not apply for registered users of the DDC services. . . . - - FTD-TI- 65-1597/1+2 • - + ****** * UNEDITED ROUGH DRAF TRANSLATION THE FUNDAMENTALS OF HUMAN AND MACHINE TRANSLATIONS BY: I. I. Revzin and v. Yu. Rozents veyg anº English pages: 299 This translation was made to provide the users with the basic essentials of the original document in the shortest possible time. It has not been edited to refine or improve the grammatical accuracy, syntax or technical terminology. r - TM500 1509 - - * 4. - & - * "... - Twis TRANslarlow Is A as Moirion of Twe onio. NAL Foreign ruxy witviour ANY Analytical on spironial commany. starsugurs on Trigories. PQ[PAQED DY; ADVOCATED OR IMPLºp ARC TMOSU Oſº. Tº $oºtC. AMD 90 NOT MICL83ARILY QUIA. ICT TM POSITION TAN3LATION DIV:30% Qº OPºtºGM OF THE FOREIGN TECMMOLOGY 54- possign Technology Division Wºº. WP-APD, OMiO. - . - ſ | - FTD-TI- 65–1597/1+2 º Date 17 March 19 6i5 AFLc-weare-app 66 8 4 S sº sº I. I. Revzin, V. Yu. Rozents veyg TABLE OF CONTENTS Preface . . . . . . . . . . . . . . . . . . . . . . . . . . l Chapter l. The Theory of Translation and Its Position Among the Other Linguistic Disciplines . . . . . . . . . l! § 1. Traditional Translation Theory . . . . . . . . . . . l, §2. The Position of Translation in Modern Culture . . . 12 §3. The Origin of the Idea of Machine Translation. The Significance of Machine Translation for the General Theory of Translation . . . . . . . . . . . . . . . 13 §4. The Subject and the Method of the Theory of Transla– tion © º tº ſe gº e e & º e e * & gº wº tº º g te º © e ſº ſº 2O OSNOVY OBSHCHEGO §5. The Fundamental Definitions for the Theory of Trans- I lation e sº e tº tº tº º e e tº © g e e e e º & e wº & e e 28, MASHINNOGO §6. The Theory of Translation and Linguistics . . . . . 3O PEREVODA §7. The Theory of Translation and the Comparative Study of Languages . . . . . . tº & 35 §8. The Theory of Translation and Problems of Training in Foreign Languages . . . . . . . . . . . . . . . . 39 Chapter 2. The Process of Translation From the Linguistic Point of View. . . . . . . . . . . . tº e 51 $9. Language and Metalanguage. . . . . . . . . . . . . . 5l §10. The Intermediate Language. . . . . . . . . . . . . . 53 - * a §ll. Certain Fundamental Concepts of Structural Linguis-" 6 * ... • . tics . . . . . . . . . . . . . . . . . . . . . . . . 5 Izdatel'stvo "Vysshaya Shkola" - § 12. The Basic Process of Communication and Certain Con – • * - - cepts From the Theory of Information . . . . . . . . 62 Moskva 1964 § 13. Translation and Interpretation . . . . . . . . . . . 67 § 14. Meaning as an Invariant of Translation . . . . . . . 76 243 pages § 15. The Problem of Translatability . . . . . . . . . . . 81 § 16. On the Problem of "Interpretability" . . . . . . . . 91 ſ Jhapter 3. Analysis and Synthesis . . . . . . . . . . . . . . 99 § 17. Breaking Down the Process of Translation Into Its Two Basic Stages . © e º º Gº tº º 99 § 18. Models for the Creation of a Text. . . . . . . . . . 103 $19. Transformation as Intralingual Translation . . . . . 119 $20. Semantics in the Formation Model. The Concept of Modulation . . . . . . . . . . . . . . . . . . . . . 124 $21. Units of Translation. Context . . . . . . . . . . . 136 FTD-TT–65–1597/1+2 - i – TD-TT–65–1597/1+2 U.S. DEFARTMENT OF COM MERCE POSTAGE AND FEES PAl D E. NATIONAL BU REAU OF STANDARDS U. S. DEPARTMENT OF COMMERC CLEARING HOUSE FOR FEDERAL SCIENTIFIC AND TECHNICAL |NFORMATION SPRINGFIELD, VIRGIN IA 2.2151 OFFICIAL BUSINESS : 79"> slater's BooksroRE Inc. $36 S. $tate St. Ann Arbor, Mich. 4810& PARTIAL SHIPMºMº PRINTED MATTER !º- | control No. w YOUR REFERENCE CODE A MOUNT NO. NS. DEPOSIT ACCOUNT #WED $ |TEMS CODE NO. CODES 2374.04K | 8/10/66 i8 F 771.97 SOLD TO (IF DIFFERENT THAN RECIPIENT) zes ACTION CODE ACCESSION NUM BER QUAN- | TOTAL $ TITY PRICE : C O D E S PURCHASE ORDER LETTER PHONE TELEGRAM CASH DEPOSIT ACCOUNT GPO COUPONS CHECK MONEY ORDER SPECIAL -1 SPECIAL -2 A|RMAIL SPECIAL DELIVERY REGISTERED MAIL CERTIFIED MAIL SHELF STOCK PHOTOCOPY MICROFICHE 35 MM MICROFILM L|BRARY OF CONGRESS SUPERINTENDENT OF DOCUMENTS JOINT PUBLICATIONS RESEARCH SERVICE COMMERCE DEPARTMENT Lº: **** i º - * * * : e 3: - jº 3.09. - º RN . . . $ ºf º * 9 * - *-* - - - —- - .. - ######## ... nº. § º *-*. *Tºº , sº º - Hºº: - - a & s = **ms º wrº w *: SS IN THE INTEREST OF EXPEDIENCY, EACH ITEM ORDERED MAY BE SHIPPED SEPARATELY. ITEMS MAILING $ MF MARKED LC, SD, JP, CD, ETC. HAVE BEEN ENDORSED TO ANOTHER AGENCY FOR SUPPLY ACTION. CHARGES O : * THIS IS NOT AN INVOICE, PAYMENT NOT REQUIRED TOTAL CD - - - - - s FORM NO. NBS-81O (1/21/65) Chapter H. Ways of Accomplishing the Translation Process. º 1118 *-*. $22. Certain Important Concepts in the Traditional Theory §2 of Translation. * * is e 3. :*::...ºpe. Of Correspondences Between units' -148. $21, ; * 6 e e *se and Units of the Target Lan- • Fundamental D . . . . . . . . . . . . . . . 150 $25 tion . . . . periºn for the Theory of Transla– e The Portion Common to Two i. º º tº º º e o e e º e o 158 e anguages and I * ; Borrowing of Words and #: sºle in PREFACE #. # Translation . . : * * * * * * * . . . . . . . 162 This text seeks to familiarize potential specialists in the field ~ . º mplifying T ' ' ' ' ' . . . . . . . . 173 - - - - #. ; #ºtion . . . . . . . . . . . . . . ić of machine translation and related disciplines with the fundamentals of e equate T - " * * * * * > . . . . . . l - $30. tº: Of ..º. - 2 m..... . . . . . . . . . . . . . 196 the theory of translation. the #..."; the Translation Process and - tion . . . anguage in the Act of Communica- A course in general translation theory for individuals specializ- $31. The Time Factor in the pºol....' ... ...:... . . . 2O7 n the Process of (Written and Oral Translation). • Translation 2ll. ing in the field of machine translation is necessary, since further de- Chapter 5. .*.*.*i; Problems in the Theory of Human velopment of machine translation would be senseless without consider- C §32 ne Translation. . . . . . . . . .”. • . . . 22O ably more thorough understanding of the mechanism of translation than #3 • The Problems of the Dictionary. . . . . - 33. Configurational Analysis. . . . . . . * * * . . . 22O is currently the case with the experiments being conducted. §3!!. Analysis Through Synthesis. * - . * * * * * * , 226 #: gº; Problems of Syntactic Text Analysis : ; ; ; #! The present text makes an attempt to discuss the problems of tra – § e emantic Problems of Synthesis. . * * * * tº e 3 37. Syntactic Problems of §nt • . . . . 255 ditional translation theory in the terms of structural linguistics. § 38 - ynthesis . . . . 6 $38. The Independence of Analysis and Synthesis. . . . . # s - References • e s e * This modification of traditional theory is necessary because of the - - * * * * * * * . . . . . . . . . . 275 fact that in the form in which this theory has been discussed up to the - present time it has been of little benefit to the specialists in the area of applied linguistics and has been entirely inapplicable in prac- t ice. Indicative of the situation is the fact that not a single book on ! machine translation even refers to books and articles on the tradition- - al theory of translation. However, we should like to demonstrate that there exists a number of valuable and rather easily formulated concepts within the tradition- al theory of translation. This approach, in our opinion, will make it possible to accustom to the problems of applied linguistics those individuals who up to this time preferred to work in the area of the traditional theory of trans- lation. From this standpoint the book may prove to be useful for all - ii – – 1 – FTD-TT–65-1597/1+2 - FTD-TT–65–1597/1+2 those concerned with problems of translation. With the exception of chapter 5, as well as several sections (§§l2, 18), assuming specific comprehension of formalized text, the book may be used by advanced philology Students, degree candidates and teachers of foreign languages desirous of understanding those phenomena of translation which they en- counter in their practical work. It is clear that despite the effort of the author's to remain with- in the framework of a purely reformulated traditional theory, they felt themselves compelled to introduce a number of new concepts (and conse - quently, new terminology) and at Several points to alter the theory in essence, not to speak of the sequence in which certain aspects of trans – lation are discussed, nor the position which is assigned to each of these aspects. Apparently there is no other solution and no contact with machine-translation practice, nor consideration of translation - theory in the general context of structural linguistics, can alter this theory. It is therefore obvious that this course in translation the ory, as > envisaged by us , cannot be regarded as being established in final form; numerous problems are in need of further study and refinement, but to the extent to which the given course is taken as part of the current - scholastic program by Students who have been trained in the fundamen- tails of mathematics and general linguistics - and, what is most impor- tant , who must exhibit the habits Of independent Scient ific effort — it is a SSumed that many concepts and even Sections of this book will be refined by these students themselves during the course of seminars, in course papers and in theses. - This book is based on lectures read by the authors in 1959–1961 at the Translators School of the First Moscow State Pedagogical Institute of Foreign Languages. This book is devoted to the first graduating - 2 – se FTD-TT–65–1597/1 -2 class of the Machine —Translation Department of this institute. The Authors FTD-TT–65–1597/1+2 Chapter 1 THE THEORY OF TRANSLATION AND ITS POSITION AMONG THE OTHER LINGUISTIC DISCIPLINES Śl. TRADITIONAL TRANSLATION THEORY This book is devoted to the description of what occurs when a text in a given language is transmitted through the facilities of another language , i.e. , the process generally referred to as translation. Trans- lation arose out of practical requirements and productive human activ – ity in connection with the role of language as a means of Social inter- Course. The first historically established translations involved offi- Cial text : political documents, religious material, etc. With expanding international relations translation begins to play a greater role in trade, manufacture and, finally, in the cultural exchange between Peo- ples (translation of literary works) (Thieme, page 58). However, it was not until recently that a scientific study of the phenomena of transla - U 1 on was undertaken, and this involved primarily the translation of literary texts. "The history of translation (at least in recent times, i.e. , be — sinning with the 16th-17th centuries) - a S A. W. Fedorov has noted cor- rectly – has been studied primarily if not eXclusively from the stand- point of the history of literary translation" (Fedorov, 1958, page 25). It is therefore not surprising that the basic concepts traditionally part of translation theory are attributed to writers and poet S. In ad- dition - and this is significant — the fundamental problems of literary translation have undergone little change for centuries. Indeed, even – 4 - *... *- ::::::::::::::::::::gºez.--------~~~~<-- " - "…-----... --> ----ºr: ***********-ºs.º.º.º.º.º.º.º->5.3 tº: the ancient writers had dealt with the question of translation quality. In speaking of his translation of the speeches of Demosthenes, cicero noted that he had treated these not purely from the standpoint of a translator (ut interpres), but from the standpoint of a writer (sed ut orator): he found it impossible to reproduce a Greek with a Latin Word , working under the assumption that the words of the original should be rendered in terms of intent rather than in terms of literal meaning. Horatio repeats this thought half a century later, warning the transla - tor against a word-for-word translation. The concepts of translation established by the classicists of an- tiquity regained their validity during the Renaissance. The people of this remarkable period in the history of culture were confronted to an even greater extent than were the Roman translators of Greek texts with the problem of discerning and transmitting "the spirit" of Hellenic and Roman literature. At the same time, the Renaissance writers and poets became interested in the development of uniquely national writings : let us not forget that the Renaissance coincides with the period of the formation of national states and national literatures. As a result it. is understandable that there existed two trends in the pract ice of translation. One called for translations faithfully reflecting the en- tire linguistic Structure , fabric and form of the original, while the other Sought to transmit only the sense of the original. A typical representative or the first school is the German Renais- sance translator (15th Century), Nicklaus von Wyle, the author of a special treatise on the problems of translation. He required the exact transferrence into German of any Latin construction and called for the exact duplication of the original 's grammatical peculiarities. Here is a small example of his translation : - 5 - - Euryalis aber gabeien, ging a voidannen. . . . ["Euryalis having been asked, departed from that place. 'I This is a deliberate reproduction of the absolute participal con- struction of Latin that is foreign to the German language. Apparently the instructions of Martin Luther were more productive for that period of time and in his translation of the Bible he showed himself to be a champion of translations styled in the Spirit of the German language. This is what Luther wrote in his famous message on translation ("Von der Kunst des Dolmetschens," 1530): - Man muss nicht die Buchstaben in der lateinischen Sprache fragen, wie man soll Deutsch reden ... sondern man muss die Mutter im Hause, die Kinder auf der Gasse, den gemeinen Mann auf dem Markt darum fragen, und denselbigen auf das Maul Sehen, wie Sie reden und danach dolmetsch- en, so verstehen sie es denn und merken, dass man Deutsch mit ihnen re- det. [ . One cannot ask the letters of the Latin language how German should be spoken, when this question should be put to a housewife, children on the street, and to the peasant in the marketplace; one has to look at their mouths and how they speak, and to translate in this manner so that they will understand and note the fact that they are be- ing spoken to in German. ' ) At another point Luther speaks of how he translated the Latin ex- pression: Ex abundantia cordis os loquitur- [ . From an excess of the heart Speaks the mouth. " ) He write S : Wenn ich den Eseln soll folgen, ..sie werden mir die Buchstaben Vor- legen und also dolmetschen: Aus dem Uberfluss des. Herzens redet der Mund. Sage mir, ist das deutsch geredet? Was ist "Uberfluss des Herz- ens" für ein Ding? Das kann kein Deutscher sagen . . . Sondern also redet die Mutter im Haus und der gemeine Mann : Wes das Herz voll ist, des gehet der Mund über. Das heisst gut deutsch geredet, dessen ich mich geflis Sen und lie — der nicht immer erreicht noch getroffen habe. Denn die lateinischen Buchstaben hindern sehr gut deutsch zu reden. [ . If I were to take after these donkeys, they would present the letters to me and translate as follows: Aus dem Uberflºss des Herzens redet der Mund. Tell me, is this German? What is this Uberfluss, des Herzens ... no German would ever speak this way . . . This is what a housewife or an average man would say: Wes das Herz Voll 1st , des gehet der Mund über. This is German and it is toward this end that I have al- ways striven, although not necessarily attaining it. It is because La – – 6 - *- tin interferes with speaking very good German. 'I An interesting point is the fact that the expression Wes das Herz voll ist, des gehet der Mund über, as hundreds of others created by Lu– ther, have come into the German language and are in use today (see $25). Luther correctly understood his problem and found the only solution that was both valid for his time and for the problems with which he was C Onfronted. Accordingly, it was Luther and his translation of the Bible that "developed contemporary German prose. "% The effort to maintain accuracy, on the one hand, and standards of national language, on the other hand, can be noted in all theoreticians of translation of the 16th century. This is expressed most completely in the work of the French humanist Etienne Dolet "On the Method of Translating Well from One Language into Another" ("De la maniere de bi- en traduire d'une langue en l'autre"), published in 1540. "To translate well from one language to another," wrote Dolet, "the following five conditions must be satisfied. The translator must , first of all , tho- roughly understand the individual style of the author whose work he is translating; secondly, he must know the language of the author who he is translating and he must know the language into which he is translat- ing; thirdly, he must not translate literally but so as to express the intent of the author; fourthly, in translating into languages which have not yet developed their own art of words ("... non reduictes enc cre en art certain et repees"), such as French, Spanish, English, etc. , the translator should not misuse borrowed words, but should employ the gen– erally accepted language ("le commun langage"); fifthly, the translator should convey the Sound and harmony of the original" ("Babel," 1955, No. 1). As we can see , Dolet approaches translation from the esthetic point of view - translation and literary translation are identical con– - 7 - ---------------~ *--------------->m::Rear-º-º-º-ºr-to------ ---- --------- ~ : -----> --& “.---- cepts to him. With this concept of translation doubts naturally arose in connec- tion with the question of translatability. Indeed, during the 15th Cen- tury Dante maintains that "nothing harmoniously linked by the muses (. . . nulla cosa per legame musaico armonizzata...) can be transformed from the original language to another without destroying all of its charm." (see Dante Alighieri, "Il convito"). Cervantes expresses a simi- 11 lar opinion through the words of Don Quixote: ". . . it is my opinion," says DCn Quixote, "that translation from one language to another... is like a Flemish carpet with flaws : the pattern, to be sure , can be made out , o it the abundance of threads relegates the patterns to the back- ground and we fall to find the smoothness and the colors which we ad- mire sc:33 h on the right side. . . "% A slightly different concept of translation is characteristic for the Fre: 2:... literature of the 17th–18th Centuries. For the writers of the 17: K. Century, setting rigid esthetic standards for literature, and for he ir 3 accessors – the French and English writers of the 18th cen- tury - translation is permissible only to the extent to which it cor- res p Cr. C C : C the literary and linguistic standards of the 17th and 18th Centuri=3. The first translator of Homeric poetry into French (1681) wrote: "Tº the extent to which it was possible for me to do so, I ap- proximated the moods of the Ancients" ("J'ai rapproché les moeurs des Ancienz autant qu'il m'a 6té permis"). In the 18th Šentury only an abridged translation of the "Iliad" is published in France. Moreover, the Homeric text includes something new : "It became necessary," wrote the translator of this abridged edition, "to replace ideas which pre- vailed 32 ring the time of Homer with ideas which are current today" (see Mc ar. In , i955, page 90ff). This limitation on the C Oncept of liter- ary trar:32.2 t ion had an historical basis. Even such a progressive writer – 8 - **** *.*- : . . . . . . . . . . . . . . * * : **-*r.- ... --Tº-Fºr-. Tºº-- - - - - - ---------. :::------------ :-. 3.; ; ;... ." .. 2::::":-:::: ... : *:::::::::::::::::-------.:*::::::::::::--------------, as Voltaire was unable to comprehend the spirit of the ancient litera- ture. "I am convinced," he wrote, "that we have at our disposal in France two or three poets capable of an outstanding translation of Ho- mer; but at the same time I am firmly convinced that these translations will remain unread unless they are modified and cleansed virtually in their entirety. This is a result d' the fact that one must write, for his time, and not for times gone by." (see Mounin, 1955, page 90). Voltaire adhered to this precept in his translations of Shakespeare. With this approach the linguistic aspect of translation was not taken into consideration. It was required of the translator that he be capable of adopting the text to the existing esthetic norms. It is only in connection with the tremendous changes taking place in Europe at the end of the 18th and the beginning of the 19th Centurie S and with the development of the historical method in philosophy, history, literature, etc. , in the practice and then in the theory of translation, that there develops a trend opposed to the one enunciated by French clas Sicism. This trend was expressed in the effort of the romantic writers to con- vey the historic and national distinctive characteristics of the trans- lated work. The effort to convey the historic and national flavor again placed the possibility of translation in doubt. This is most fully expressed in the well-known statement by Baron Wilhelm von Humboldt : "It seen. t O me that any translation is simply an attempt to solve an impossible task" ("On Translation," page 275). - It is important for us to take note of the fact that translation in these statement S is regarded exclusively as an artistic literary ac- tivity. Such an approach is characteristic of the Russian writers as well. In Russia translation always occupied an important position. The significance of translation increased considerably during the 18th and - 9 – at the beginning of the 19th Centuries, 1.e., during the epoch of vig- erous development on the part of the Russian state and Russian culture. Whether the writers tailored a foreign apthor to their own measurements, as was the case with Derzhavin, or whether they were followers of the great Pushkin and Sought to reproduce in the Russian language the Spe- º i. * cific national characteristics of the original, in either case they were dealing with problems of 11terary translation. It is well worth while, for example, to turn to Belinsky's pronouncement S on transla - tion in order to make it clear that Russian thought on translation, just as the thinking in Western Europe, was concentrated on literary translation. The labor of the individuals translating scientific works and establishing Russian Scientific terminology was of great Signifi- canºe for the development of Russian culture and the Russian language , but it remained unnoticed and did not attract the attention of the the - crevisians of translation. Thus it may be said that thoughts on translation in the past dealt primarily with the study of esthetic problems in literary translation. * . . .” How is one to explain this development in thought on translation? can cite two factors operating and continuing to operate in *, s this direction. The first and fundamental factor is the need, in gener- ial practice, only for artistic literary translations. Indeed, transla - ions of literary texts occupied a leading position in the field Of t ** * translation. The ancient world is virtually unaware of Scientific transiations. Latin was the language of Science during the Middle Age S. It is true that the scientists of the Middle Ages had the problem of translating into Latin those scientific works written in Arabic - it is a well known fact that Arab science attained a high level during this perios. Joseph Ernest Renan put it well when he said of these transla - tions: "The Latin word covers the Arab word much like a chesspiece cov- - 10 - ers a Square On the board. The structure of a phrase is much more like - ly to be Arabic than Latin. The majority of the technical terms and those words unfamiliar to the translators are presented in the most rudimentary forms of transliteration. . . the translator, relying on the ambiguities of the text, imposes on the reader the problem of making out its sense" (see Olschki, page 289). - The reliance on the part of the translators of Arabic scientific books during the Middle Ages on the textual ambiguity of the original is quite understandable. Even the greatest and most conscientious of these translators did not always have adequate knowledge of the subject. "If to judge by their translations, the problems of terminology pre- Sented them with great and even insurmountable difficulties. Encounter- ing Special expressions, on reading these they committed fatal error; and one can sense how they attempted gropingly and approximately to £uess at the meanings of these terms. If they were unfamiliar with the subject matter, they were compelled to render the words of their auth- ors by means of Latin expressions, word for word" (ibid.). To this we should add that some of the Arabic scientific writings were not Original but translations of Hindu and Greek texts that had not always been translated Satisfactorily. If we take into considera- tion that written Arabic does not indicate an Vowels , we can see the eXtent to which proper Hindu or Greek names (or Special terms) would be dist Orted in the transition to the Arabic , and from this to the Latin •. text of the Middle Ages. If this was the quality of the language em- ployed in the Latin scientific translations, it is not surprising that they did not come to the attention of the humanists and philologist G of the Renaissance who were interested in problems of language and tran: – lation. Apparently they regarded the Study of problems associated with the translation of scientific literature as a matter of little Sign 1' 1 – — ll - cance. Although in actual practice interest in problems of translation Continuously increase, its position in the Ory as a whole, however, un- derwent little change; down to the 20th Century translation was regard – ed primarily as a creative activity, i.e., from the esthetic point of View. - §2. THE POSITION OF TRANSLATION IN MODERN CULTURE The translation of nonliterary texts (scientific, technical, com— mercial, military, etc.), the translation of the Written word, as well as the interpretation of the spoken Word , have assumed tremendous Sig- nificance in the 29th Century. Of course, this is understandable. While in the past there existed international languages of science and diplo- macy (Latin, French), or if there were one Or two national languages in which the majority of scientific works were Written, now this situation has changed radically. The development of science and international cultural relationships would be unthinkable at the present time without the translation of scientific and technical literature into tens and even hundreds of languages. Unlike past periods when the translation of this literature did not yet involve an independent profession, in the 29th Century (particularly during the second half of this century) it has become necessary to train translation specialists; translation of - fices and professional associations Of translators are being organized, international congresses are being convened and there are being pub – lished special journals devoted not to translation as an art but to translation as a specific profession requiring 1ts own special tools. The following data are indicative : prior to the 16th Century Gight — tenths of all publications were printed in Latin. It is only during the l6th Century that there appear, for example, in France, translations from the Spanish; in the 17th Century we have translations from the En- - 12 - glish and in the l8th Century we have translations from the German. At the present time the number of translations is constantly on the rise. Our country [the Soviet Union] holds first place in the number of pub- lished translations. For example, of the 29,661 titles translated in 1959 in 63 countries, the Soviet Union published 525l. * However, in oth— er nations, even those which are not multinational in character, the number of translations continues to rise. For example, in France the per- centage of translations relative to the total quantity of published ma – terial amounted to: 11% in 1929; 8% in 1935; 9% in 1950; 9% in 1951; and more than 10% in 1956 (Cary, 1956). Of the 675l books printed in Ital- ian in 1951, 1153 had been translated from foreign languages ("Babel," II, l, page 32). Nevertheless, at the present time only a small fraction of the scientific and technical literature which is being published in the various languages of the world is being translated. In this connection there has arisen the idea of mechanizing the translation process. ** §3. THE ORIGIN OF THE IDEA OF MACHINE TRANSLATION. THE SIGNIFICANCE OF MACHINE TRANSIATION FOR THE GENERAL THEORY OF TRANSIATION The pioneer in the field of machine translation is apparently the Soviet invent or , Engineer P. P. Smirnov-Troyanskiy. This is how Profes- s or L. I. Zhirkov describes the history of this development : "In 1939, Engineer P. P. Smirnov-Troyanskiy, an inventor, appeared at the headquarters of the USSR Academy of Science S and' reported that he had developed a method of machine translation from one language to another: the developer requested a consultation on this new process from the linguistic standpoint. It must be stated that P. P. Smirnov — Troyanskiy 's development was greeted at that time by linguists with considerable scepticism; the development was regarded both as unattain- able and quite useless. Only a few individuals accepted the possibility - 13 - - -- or machine translation. During the course of the subsequent consulta- tions in which I too participated it was shown that his machine-trans- lation rethod made possible, for example, the translation of a Russian text in Moscow and the print out of this translation into French, let us say, ºr Paris. . . The project dragged on for a rather long time and came to a conclusion on 31 June 1944 with the convening of a meeting at the Institute of Automation and Telemechanics of the USSR Academy of Scien- ces, with the participation of specialists from the fields of mechanics, electrical engineering and linguistics. It should be pointed out that the specialists in mechanics and engineering addressed the conference to a greater extent with proofs of the "impossibility" of machine trans- lation and, upon entering the unfamiliar area of linguistics, spoke of synonyms, precision in their shades of meaning, i.e. , they spoke of things wrich had no bearing on their specialities. As a result the ex- words) was not built. The developer P. P. Smirnov–Troyanskiy, to the best of my krcºiledge, quickly left Moscow and I have since been informed that P. P. Smirnov-Troyanskiy has passed away" (Zhirkov, page 122). - A pºwerful stimulus to expanded efforts in the field of machine translation was the development and extensive use after the Second World War of electronic computers designed to convert information in- put in atcordance with specific rules or , in the terminology of the discipline, in accordance with a spºrts "algorithm." "With these ma– chines it is in principle possible to set up any algorithm, if only the volume ºf the memory is sufficiently large" (Lyapunov, 1958, page 7). In connection with a discussion of the problem as to the limits of application for electronic computers, Warren Weaver in 1946 expressed the opiniºn that machine translation and the analysis of a language by means ºf a computer were possible. In 1949 Weaver discussed his thoughts – 14 – in the noteworthy memorandum "Translation" which drew the interest of a wide circle of scientists. The first conference on the theoretical problems of machine translation was convened in the spring of 1952, at the Massachusetts Institute of Technology (MIT) and in September 1952 HO linguists partici- pated in the discussion of problems relat- ing to machine translation at the VII In- ternational Congress of Linguists. The first public demonstration of machine trans- P. P. Smirnov–Troyanskiy lation (the so-called "Dostert experiment") was conducted in 1954. In March 1952 the Massachusetts Institute of Technology published the first issue of the special journal "Mechanical Translation. "* The first sample translations in the USSR were accomplished in 1955–1956. An English scientific-engineering text was translated into Russian on an experimental basis at the In- stitute of Precision Mechanics and Computer Engineering (see Panov, page 3). The first experiment with translation of a French Diagram of the P. P. Smirnov–Troyanskiy ma- * * chine. out at the Steklov Institute of Mathematics mathematics text into Russian was carried (Kulagina and Mel 'chuk, 1956). The Department of Machine Translation was established in Moscow in 1956 and proceeded to distribute its publication "Bulletin of the De- partment on Problems of Machine Translation," subsequently renamed "Machine Translation and Applied Linguistics," subsequently playing an important role in the consolidation of all activities in this field - 15 - (Rozents veyg, 1958). The 1st All-Union conference on Machine Translation was held 1n Moscow in the summer of 1958, attended not Only by scientists from Mos- °ow and Leningrad, but from Kiev, Yerevan, Tbilisi, Gor'kiy and other cities* ("Theses," 1958). - Two trends were observed in machine translation both here and abroad (Yngve, 1956, pages 99-101). We can agree to refer to one of these trends as theoretical, referring to the other as the practical trends. ** The pr’Op Onents of the practical trend strive toward the most rapid practical a CComplishment of machine translation. Translation quality need not be high; it is enough if the greater portion of the text is understandable to specialists. As a *ule, the advocates of this trend regard the development of Structural methods of describing a lan- £uage with scepticism. For example, D. Yu. Panov, citing numerous €2.3.In – Ples of differences in Phrase types between two languages, primarily examples from literature, draws the °onclusion that "here abstract Structural analysis is Completely pointless" (Panov, page 52). On the Other hand, D. Yu. Panov, analyzing the Concluding lines of the Tyutchev Poem "On the way," states that the "difficulty in the translation of these phrases lies not in their structure but in the seeking out of equivalents precisely for those individual words which the poet, em- Ployed and which have formed into extraordinarily specific and yet pre- Clse relationships." D. Yu. Panov draws the further conclusion: "All of the foregoing, it seems to *S , explains why in our point of view the analysis of linguistic Structures cannot solve the problem of automat – ing translation, although a number of Scientists both in the USA and eVen in the USSR are attempting to employ this method. " For the moment we will not dwell on the views of D. Yu. Panov, Since their limitation and , therefore, their &roundlessness will become - l6 - - *x ºf fººwº-, -g- ºr-., - º dºº-º-º-º-...--~~~. -...--~~ - • :**ś 2:::: -ºvº irº 3 ºr “lººs-r •rº ***:::::::::::::::::::::g#º*::::3# §2.3.73:5. Iºx=&e. --- -. **śzz. **T*-----------------.r- - ºs--->4:… º.º.º.3sº ºº:::::::::::::::::::::::: evident from the following. We will note only the following: the suc- C & SS with the first tests of machine translation gave S One researchers the illusion of ease. It seemed as if the majority of all problems had been solved and that it was possible to speak exclusively of expanding dictionaries, improving programs, etc. However , the true situation is considerably more complex. When we analyze the first tests we immediate — ly See the extent of the work which must be accomplished in order to Set up even limited programs, and the extent to which today the work required for the Setting up of rules and programs is not commensurate with the practical results which can be attained. The advocates of the theoretical trend hold that for progress in machine translation it is necessary to develop new methods of language description, exhaustive on the one hand and of sufficient generality on the other. As properly noted by W. Yngve, "this work is laborious and time consuming, because today not nearly enough is known about langua- ges and how we translate" (Yngve, 1956 b , page 100). The proponents do not regard machine translation from the empiri- cal standpoint, but as one of the aspects of the general cybernetic problem of studying processes of information modification. Hence fol- lows "the importance of enabling the machine to employ human speech" and the need to regard "machine translation as the first stage in training a machine to work with language" (Lyapunov, Kulagina, 1958, . / page 18). As is well known, "the basic problem in cybernetics is the problem of the interrelationship between the potentials of a computer and thought" (Lyapunov, 1958, page 15). In this case, the feature of the cybernetic approach is that the validity of advanced hypotheses Such as , for example, the hypotheses of the knowledge of languages, can be ver' l- fied experimentally by simulating certain processes in the C Olſiput. Cr'. - l'7 - - In this connection the possibility of experimentation is made available to the theory of translation. Dreams of experimentation in linguistics date to L. V. Shcherba who introduced the concept of "sty - listic experiment." He wrote: ". . . the principle of experimentation is introduced into linguistics. With any assumption of meaning for a given word, for a given form, for a given rule of word formation, etc. , an attempt should be made to test whether or not various phases can be linked (and these can be multiplied infinitely), employing this rule for this purpose . . . , without waiting for some writer to use the given procedure or a given combination, to test whether it is possible arbi- trarily to combine words and, Systematically replacing one word with s another, changing their sequence, intonation, etc. , to observe the re- Sulting differences in meaning, which is what we continuously do when we write anything" (Shcherba, 1931, pages lzl–123). Let us note that Shcherba's "stylistic experiment" is a necessary integral element of the process of translation, but unfortunately this experiment is subjective and suffers all of the resulting consequences. Machine translation represents the first step toward objective experimentation in the theory of translation and in linguistics in gen - eral. Machine translation presupposes certain hypotheses regarding the process of translation, which will Subsequently be checked by simula – tion of the translation process in the computer. It is clear that with such a formulation of the problem machine translation serves not only a practical role, but opens paths for the objective study of linguistic processes which occur in translation. This in turn must facilitate the pract lcal realization of machine trans- lation and will also aid in the construction of other computers related to language : machines for the conversion of spoken speech into written for m, data processing equipment, etc. - 18, - occasionally there arises the question: Will the computer replace the translator We can make the following remarks. It would seem that the development of electronic computers should have led to the "unem- ployment" of mathematicians, since the machine today solves problems which earlier were the province of mathematicians. In actual fact, how – ever, the exact opposite took place : mathematics has become a specialty with a shortage of personnel. It may be reliably maintained that the same will be the case in the field of translation. First of all, the greater the number of ma— chine-translated texts, as well as the range of languages from which these translations are being accomplished, the greater the number of specialists familiar with the practice and theory of translation will be needed to set up and consistently improve the machine programs , to edit the finished translations, etc. (Cary, 1956 b , Yngve, 1956 b ). Secondly, if there is no longer any need for mass translations of Sci- entific and engineering literature (this work will doubtlessly be ac- complished successfully by means of computers) specialists will become available to handle the translation of literary works and advertising and publicity material in a considerably greater volume than is cur- rently the case. Of course, there arises the following question: will not the machine be capable of literary translation? At the moment it is difficult to answer this question. If some day it becomes possible to build such a machine , it will be a tremendous achievement of human gen- ius. But in that event one might ask which is more advantageous: trans – lation by means of machine or to turn such an assignment over to Spe- cialists in literary translation, retaining a more important applica- tion for the machine and namely , the simulation with the machine of the process of artistic creativity or , in other words, to understand this process in its true sense. Even now the significance of machine tran G – - 19 — sº-º,-- . . .” - . . - — ;-- ::...º.:-3.36:.. . . . . . . - § *::sº “s.…--> *::: #º**::::::::::: - w- lation lies not so much in the obtaining of the finished product as in the possibility of simulation and experimentation, i.e., the movement - - toward ever deeper understanding of thought processes. § 1. THE SUBJECT AND THE METHOD OF THE THEORY OF TRANSLATION The change in the position of translation in the contemporary world, as well as the internal requirements of linguistics itself, have caused theoreticians to become increasingly interested in the Specific problems of language relationships and other purely linguistic aspect S of translation. It is therefore not surprising that even prior to the appearance of machine —translation problems a number of linguists attempted to jus- tify the principles of the general theory of translation based on lin- guistic grounds, and with particular solutions in which these princi- ples are attained. An interesting attempt at a systematic outline of the linguistic approach to translation was the work of Ya. I. Retsker "On the Quantita- tive Relationships in Translation into (ne's Native Language" (Retsker, 1950). An even more significant event in the development of the linguis- ic the cry of translation is the book by A. V. Fedorov "Introduction to the Theory of Translation" (lst edition, 1953; 2nd edition, 1958). Unlike his previous works in which he treated translation from the literary standpoint, here A. V. Fedorov proceeds from the following posi- t i on : - "Since translation always deals with language, that is , always in- dicates work with language, and since translation always requires great — er study from the linguistic viewpoint with respect to the problem of the nature of the relationships between two languages and their Stylis- t i C facilities,... the theory of translation as a special branch of philological science primarily represents a linguistic discipline" - 20 - (1953, page 14). A.V. Fedorov did not limit himself to the contention that it was necessary to construct a theory, but made an attempt to describe a number of specific lexical and grammatical relationships be — tween languages. Unfortunately, having stated the need for the construction of a fundamentally linguistic theory, Fedorov at the same time operates with literary concepts, looking upon translation as the result of a creative process and he speaks of the inseparable unity of content and form, of the rendering of the national and artistic-stylistic unique character- istics of the original, of the recreation of the model, etc. In addi- tion to these categories he examines such completely linguistic pheno- mena as the rendering of sender, mood, etc. With this dichotomy it is difficult to attain theoretical integrity. But this is not a necessary requirement of the initial attempt at a linguistic basis for the theory of translation. It is to the author's basic credit that he drew atten – tion precisely to the linguistic aspect of the problem of translation. In speaking of the inconsistent theoretical conception of transla - tion, as presented in Fedorov's book, it is not our Intention to main- tain that literary translation cannot be considered in the context of a linguistic description of translation. * Moreover, it is important to recognize that the problem of translation stands out more clearly in a 3 tudy of literary translationS. ! It is for this reas on that consideration of problems in 11terary translation is to the point and necessary in connection with any theory of translation, including the theory of machine translation. However, it is important in a consideration of the pertinent phenomena clearly to delineate those which may and must be described from the linguistic standpoint from the phenomena whose description would call for resort to the methods and concepts of poetry. – 21 - ** In this connection, we cannot agree with A.A. Reformatskiy that 1t is impossible to devise a untried linguistically valid theory "which would resolve all problems in translation, Starting with problems in the translation of tourist phrase books and ending with the complex problems of literary translation" (Reformatskiy, 1952). The problem of the possibility or impossibility (feasibility or unfeasibility) of a linguistic theory of translation is related to the problem of whether or not it is possible to isolate a certain object of Study which is present in any translation and whether or not this ob- ject can be studied with the facilities of linguistics. The very process of translation represents such an object (das fibersetzen, the translating) in which is accomplished the transition from one system of signs to another and which may be described in the terminology of semeiotics. * The differentiation of the process of translation and its results (die tibersetzung, the translation) is extremely important from the fol- lowing standpoint. The science of translation, including the aforemen- tioned works of Ya. I. Retsker and A.v. Fedorov, is built traditionally as a normative discipline whose main goals are the determination of re- Sults emanating from the process of translation and the establishment of criteria to evaluate the auality of a translation. (Compare, for ex- ample, the following statements: "The purpose of the theory of transla - tion. . . is to generalize in the light of scientire data conclusions drawn from observations of individual particular cases of translation and to render assistance in the practice of translation, such that it might be used in the search for needed means of expression and to de- rive from this theory arguments and proof favoring some given solution of specific problems" (Fedorov, 1958, page 15). The normative theory of translation was derived empirically on the basis of a comparative analy – - 22 - sis of the originals and their translations. The theoretical positions in this case were taken over from lexicology, grammar, stylistics and, in other cases, from the history of literature and literary criticism. It is clear that with this approach no theoretical interpretation Of the process of translation can be achieved. . A Science which seeks to describe translation as a process must not be normative but theoretical. * This science should describe not that which should be , but that which is contained within the very nature of the phenomenon. This is Self-evident if the theory of translation is to be built as a linguis- tic Science , because contemporary linguistics, by the way, differs in its attempt to describe language as it functions in reality, rather than to establish standards for the use of speech (Peshkovskiy, pages 231–212). It is natural that the normative directions can and should be formulated on the basis of theory. It is , however, the fact of the mat – ter that the normative discipline cannot exist in the developed stage without the oretical description. The theory Studying the process of translation leads to conclu- Sidns that have not been formulated subjectively by comparis on and evaluation of the original text and the final translation, but have been arrived at objectively, through a description of all stages in- volved in, the accomplishment of the translation. Moreover , such a the o- / ry is &bjective in the sense that it does not proceed from the specific Structure of two comparable languages, but seeks to understand transla - tion as a natural process contained within the very nature of Spoken intercourse, independent of human will and regardless of whether it is people or machines , or a combination of man and machine , who are part 1– cipating in this exchange. With this concept of theory it is not cut of the question – as a matter of fact, it is assumed – that a computer can — 23 – º º tº: - sº play a role in the development of the theory and in the disclosure and rigorous formalization of the phenomena which take place auring the process of translation. The following comment, made with respect to machine translation, holds for the general theory of translation: "Stud- ies of machine translation assume that all participants are prepared to Subscribe to the following belief : languages are subject to laws as Constant and as natural as the laws governing the motion of the heaven- ly bodies" (Oettinger, 1959, page 246). Thus the object of the theory of translation is much like the pur- pose of linguistic structural analysis in general, because it is the purpose Cf. Structural analysis to study the rules or laws governing the Organization of spoken material in terms of the system of language. It should be pointed out that in the normal use of the Word, rules or laws in language refer to something other than the laws which are encoun- tered in the natural Sciences: compare, for example, the following Statement of Mendeleyev : "Laws of nature do not tolerate exceptions and thus are clearly different from such rules and axioms a S, for example, grammatical and similar human inventions, methods and relationships." (Mendeleyev, page 617). It is precisely this system of norms for proper speech in a given lar.guage that is the important and perhaps basic object of traditional descriptive linguistics * 1t is not without reason that grammar has frequently been defined as the science of how properly to express One 'S the lºghts. . The system of norms for a given language, i. e. , the spoken word of the more cultural elements of society and, in particular , lit – erary language have always been preferred to other forms of the spoken language. Structural linguistics studies language in another a Spect ; Struc – tural linguistics is concerned with the natural laws governing the or - – 24 – - º šº tºº- # §º: .* ...Y. ºº- ganization of spoken material, these laws associated not with a given specific language, but rather with the very essence or the process of communication in its most general form. It is for this reason that it is totally irrelevant from the Standpoint of Structural analysis whether a linguist is studying a lan – guage rich in literary tradition or some dialect having no written form of expression; the methods of structural analysis are applicable to the analysis of any Spoken language, to any grouping of phrase S. While the statements of such highly cultured representatives of a people as its scientists, poets and writers with regard to a given lan– guage assume great Significance in traditional linguistics, structural linguistics limits the area of the study to that sphere in which strict - ly objective methods are applicable. Reduction of qualitative relationships in translation to the pure — ly structural aspect of the problem may work against simplification Of extraordinarily complex and contradictory phenomena. Are we not trying to simplify the process of translation? We must give a positive answer to this question: yes, we are Sim- plifying the process of translation. It seems to us, however, that it. is only through such simplification that we can approach the Scient ific understanding of the phenomena of translation. If we examine the his – tory of contemporary, science, we will see that its progress is associº ated with the refusal to consider complex processes prior to the quali- fication of the simpler concepts and the determination of the relation- ships between these. This principle of scientific thought was estab- lished in the struggle against the Scholasticism and dogmatism of the Middle Ages. In his "Rules for the Direction of the Mind" (1628) Des – cartes wrote: ". . . it is better entirely to remain unoccupied, than to occupy oneself with the study of such difficult things that , unable to ! i t' < . . . distinguish the true from the false, we are forced to accept the doubt- ful as reliable..." (Rule II). And later on, in Ruie v "... I... have decided in my search for knowledge stubbornly to hold to the following procedure : always to begin with the Very simplest and easiest of things and never to move on to others until I see that I am completely incap- able of drawing out from these things anything else." And, finally, in Rule IX: "everyone must be firmly convinced that the most secret truths are derived not from the Significant but hidden Sources , but from those things which are Simplest and most easily attainable." If we approach the theory of translation and its traditional de – velopment with this point of View, We immediately note the tendency to resolve the most complex of problems in literary translation even prior to the qualification of the basic and elementary concepts on which these problems are based. Thus, it is stated on the very first page of the first chapter in Fedorov's book that "... translation means to ex- press exactly and fully in the words of one language that which has been expressed in the words of another language, maintaining the insep- arable unity of content and form. A true translation differs from edit – in5, Paraphrasing or abstracting, from any kind of so-called "adapta- tion" in the completeness and accuracy of the final product (Fedorov, 1958, page ll). This definition contains much that is valuable and lºn agreement with practice, but from the Standpoint of building a scientific theory the definition suffers from Serious drawbacks. Indeed, the cited defini- tion assumes that definitions exist for the implicit concepts "to express exactly and fully," "language," "unity of content and form." At the same time, these concepts are introduced into the definition without any stipulation. Moreover, it is stated on page 129 that "...the word 'accuracy' in its application to literary translation * being used – 26 - less frequently in our theoretical literature. In this is expressed the basically true principle Of rejecting attempts to establish any abso- lute relationships between texts from various languages, to operate with any quantities whatever, to weigh and to measure." However, if the feasibility of even attempting the establishment of precise relation- Ships is negated, there is no possibility of establishing any distinc- tive property which, by the author's definition, characterizes transla - tion. The Situation is not altered by the attempt to Substitute the word "adequacy" for the word "accuracy": this basic concept in the theory of translation has also not been defined. Here is a definition given by A. A. Smirnov in the "Literary Encyclopedia," Vol. 8: "We must recognize a S adequate that translation which conveys all of the author's inten– tions (whether consciously or subconsciously developed) in the sense of their specific ideational-emotional artistic effect on the reader, maintaining to the extent possible exact equivalents or satisfactory substitutes for all of the author's forms of expression, shading of meaning, rhythm, etc.; this last must not be regarded, however, as a goal in itself, but Simply as a means of attaining a general effect. There is no doubt that something will have to be sacrificed here." This definition is subject to Fedorov's valid criticism: the prob- lem of translatabylity is treated in a contradictory manner here — the J indications of unavoidable losses (sacrifices) in the translation con- tradicts the categoric requirement to convey all of the author's inten– tions, including even those subconsciously derived. Fedorov proposes another definition: "The integrity of a translation is expressed in its exhaust ive accuracy in the transmiss 1 on of the semantic content of the original and in its complete functional-stylistic correspondence with the original." (Fedorov, 1958, page 132). - 27 - It is, however, not difficult to see that both Smirnov's and Fed- orov's definitions do not provide uniform criteria of definition, but subjectively describe the phenomena of translation; it goes without saying that this is extremely useful, but these definitions cannot be included directly in the theory. A theory describing the process of translati On must proceed from simpler and more elementary concepts. Of c Our Se Such a description of translation will be incomplete, but it can be compared with a complex process , given the obvious condition that the elements and relationships described in the simplified situation correspond to the elements and relationships existing in the transla - tion, i.e. , if the model reflects certain essentially important aspects of the true process. It is clear that no description fully exhausts its Subject. + §5. THE FUNDAMENTAL DEFINITIONS FOR THE THEORY OF TRANSLATION In the derivation of a theory describing the process of transla - tion we must : a) derive a number of initial concepts which are not de – fined within the theory of translation and are regarded as rather sim– ple and clear; and b) define all remaining concepts in terms of the in- it ial concept S. The initial concepts for the theory of translation will be enumer- ated in Šll. Here we will dwell only on several of the requirements which are expediently imposed on the definitions for the theory of -- ** .- e translation. It is frequently maintained that in Scientific theory the choice of definition (as well as of the terminology which arises out of the given definitions) is purely a convention and that the definitions are simply "agreements establishing the meaning to be applied to an expres — sion not encountered in a given discipline and whose meaning may not be self-evident" (Tarskiy, page 66). From this stand point there are no – 28 - "good" ("correst") In Orº "poor" ("incorrect") definitions; there are only aerºnitions which are "conventent" for the subsequent derivation of the theory and those which are "inconvenient" for this purpose. This View - point may be treated as a reaction to the scholastic disputes as to whether a given definition corresponds to some essential point. This viewpoint, however, hold S a certain danger, particularly for Such a young and not yet firmly established branch of science as the theory of translation. The danger lies in the fact that We lose sight of the ba- sic methodological criterion, and namely the criterion of experience (of course, when we speak of "convenience" and "productivity" of a def- inition, etc. , this is essentially a subconsciously or implicitly ex- pressed criterion of experience, whereas for a young science this cr1- terion must be present in explicit form). We might take note Of the fact that in the initial stage of the development of structural and mathematical linguistics there appeared many definitions of no deep linguistic significance. Occasionally one refers to experience fºr Om mathematics where apparently all definitions represent conventions and nothing more. However, this is not quite so. As an example, let us cite the following statements of the well-known French mathematician G. Le- besgue : "we hold that all definitions are subject to the condition that there exist both good and bad definitions, despite the fact that it iS generally maintained withouf hesitation that definitions are free '." I have never been able tº understana this phrase: I do not understand what is meant by freedom here, nor do I understand the meaning Of the term "definition" as used here. If this word is used in the **** Of "designation," then indeed everyone is free to introduce his own lan- guage, although at the risk of not being understood. If the word is used in the sense of "determination" and implies that everyone is free to do whatever he wishes with the subject matter of his thoughts, the – 29 - cited phrase is also valid, but in this case there is the danger of re- maining totally isolated with one's thoughts, without advantage to the development of the solence. Whichever the case, to those of us who re- gard mathematics as an applied science definitions are not free : at least some definitions are not free, and namely those which serve to refine practical concepts" (Lebesgue, pages 122–123). * One would think that all of this has a direct bearing on linguis- tic definitions, including the definitions from the linguistic theory of translation. Of course, it occasionally is necessary to devise a number of auxiliary definitions which are nothing more than convent ions and we will be dealing with definitions of this type in our book as well. However, the value of these auxiliary definitions is established in relation to those basic definitions in whose formulation they play a part , but in this case they will reflect certain of the most essential features of the phenomena encountered in the practical use of language. It is therefore extremely important not to reject concepts and definitions developed in writings on the science of translation because of inaccuracies , but rather to refine these concepts and definitions , , , retaining those essentials which reflect the achievements of practical work. $6. THE THEORY OF TRANSLATION AND LINGUISTICS As was Stated at the beginning of this chapter, the new approach to the theory of translation is a result of contemporary practical re- quirements. In addition, it should be stated that contemporary linguis- tics is interested in the development of such a theory. The following statement of the English linguist Firth is indicative: "The position of translation in linguistics has not been studied in the proper manner. . . , do we know how we translate? Moreover, do we know what it is that we are translating? If we were able to answer these questions in the term- - 30 - inology of science, we would be undertaking the formulation of a new and clearer theory Of language..." (Firth, page 139). The interest of language specialists in a theory of translation 15 governed primarily by the position currently occupied in linguistics by the problem of meaning. * Gaining wider and wider acceptance is the idea, based on the semeiotic conception of language, that the "meaning of any language sign is its translation into another alternative sign. . . " . (Ja- kobson, 1959, page 232). This translation occurs in three forms: l) in- tralingual translation, i.e. , the interpretation of word Signs by means of other signs in the same language ; 2) interlingual translation, i.e. , translation in the exact meaning of this word, that is , the interpreta- tion of the word signs by means of word signs from another language ; and 3) intersemeiotic translation, i.e. , the interpretation of word signs by means of signs from sign systems not involving words (ibid. , page 233). The theory of translation, as it is understood in this book, is concerned with the process of interlingual translation and is closely associated, as We Shall see , with translation within a language. ** With regard to intersemeiotic translation (for example, the translation of a dialogue text into the language of the motion-picture art), its de – velopment is in an initial stage, but even now we note certain point S of contact between this form of translatiºn and the literal definition of the concept of translation. p In this connection considerable interest has been expressed in the semeiotic analysis of the problems involved in establishing the rela- tionships between the vocal and visual forms, as begun by Eisenstein, who wrote: "in this case we hold in our hands the two combined fields in the form of two independent strips, with one carrying Images and the other recorded sound" (cited from Ivanov, 1961, page 17). - 31 – What is the role of translation in contemporary linguistics? In the linguistic definition of meaning any description of a language one- n Omen on involves the determination of similarities and differences be - tween this phenomenon and another, * it being possible to establish equivalence between units of all levels of language. Whether it is some language phenomenon that is being described in comparative-historical terms or whether it is being considered in Synchrony with the typologi- Cal Viewpoint, any description of two Systems presupposes the formation of a third system which contains a complete set of differential Signs of comparable units and it further pre Supposes the equivalence between these. Consequently, linguistics may be regarded as the science of lin- £uistic relationships (Ivanov, 1961), of which translation is a part 1 – cular case. The theory of translation is of particular interest to the study of language contacts, i.e. , the relationships established between lan- guages alternately by the same people (Weinreich, Haugen). The pheno- mena which arise in language contacts, as well as the various types and Stages of evolution in these linguistic relationships are conveniently described in the terminology of the theory of translation (Rozents veyg, 1962). We spoke earlier (see $4) of the fact that the description of translation as a process presupposes the application of a number of concepts developed as part of so-called structural linguistics. We will dwell in greater detail on these concepts in §§9–ll. For the moment we will devote but a few words to the major aspect of the problem. What is meant by structuralism? Structuralism is that trend in linguistics which must be regarded as having been started by Ferdinand de Saussure, who in his work "A Course in General Linguistics" put forth a number of important Statements which exerted considerable influence On the Subse- - 32 – quent development of the science of language, but simultaneously € Il- gendered numerous disputes and bewilderment. Much is said of the con- parison which De Saussure conducted between history and the contempor- ary State of language , of the comparis on between synchrony and diachro- ny. Carried to the extreme, this comparison is of course senseless, but historically it was quite valid, and necessary for the following rea— sons. Prior to De Saussure the statement "theoretical linguistics" was equivalent to "historical linguistics," i.e., it was held (and many continue to hold) that the theoretical study of a language is the same as its historical study. Prior to De Saussure the only scientific study of a language was the historical and only those linguists concerned With the history of a language could lay claim to having studied lan- guage Scientifically, Since they had a method of study at their dispos – al, and namely the comparative-historical method, and for this reas on they were able to generalize and explain facts, an attribute without which science would be unthinkable. However, De Saussure drew attention to the fact that the synchro- nous Study of a language was at least as scientific as the diachronic. Here the very principles of synchronous analysis may differ from the principles of diachronic analysis. 4. There developed in comparative-historic linguistics a tendency of extreme importance, and namely, the study of material, facts. The signs or language were treated and studied as material units and, for exam- ple, it was regarded as important that the form of one of the German suffixes in Old High German was rendered as -nassi, -nessi, -nussi and that this suffix changed into -nis, i.e. , here it is important that the change in signs is purely material in nature. This was brought about by the purpose of the investiga t + Cºn, but certain linguists hold that it is precisely this a pººr oach that § 3 the – 33 - only correst and Only materialistic procedure. as a meter or rest, this rendering absolute a single form of matter has natiºns in common with philosophical materialism. On a purely linguistic level this abso- lutism takes us away from the essence of language. After all, every hu- man being conceives intuitively that Danish (or Russian), written ac- cording to the rules of Danish (or Russian) orthography, or transmit – ted in Morse code, or written in some other code, remains Danish (Rus– sian) (Hjelmslev, page #19). - Let us present another example. There exists in Russian the para- digm ‘cron’, ‘crona’, ‘crony' etc. Let us now agree that we will replace the null ending [of the first word l with the letter c, in the place of the ending a we will, write m, and in the place of the y we will write III, etc. If we extend this system through the entire Russian language , we will obtain something that at first glance seems entirely unlike the Russian language. The declension paradigm will have the form: “ctoric', ‘ctoniu', ‘ctoniu' etc. Let us now ask ourselves the question: did this lan- ºne remain Russian or not? Prior to De Saussure linguists apparently responded to this question with the statement that it represented abra- cadabra rather than Russian. It is precisely a characteristic of the Russian language that its genetive case , masculine Singular, for exam— ple, ends in a, etc. Now, on the other hand, linguists hold that in our example we are dealing with two possible versions of a single language. What various manifestations of a single phenomenon have in common is objective reality as it pertains to the essence of the phenomenon Stud- ied by the linguist. The comprehension of this circumstance and the derivation of all resulting consequences is historically credited to De Saussure and his followers. * Another important position taken by De saussure involves the ne - - 34 - cessity of carefully differentiating between internal linguistics, i. e. , the racts touching upon the internal organization of a language and all that pertains to the structure of that language, from external linguis- tics, i.e. , the facts which pertain to the relationship between lan- guage and society, language and the history of the people speaking that language, between language and culture, et C. As in the case of the comparison between synchrony and diachrony, this comparison is important in that it vindicates the specifics of studying the internal structure of a language, but , of course 2 it should not be interpreted to mean that only internal linguistics has a right to existence. We have already spoken of the fact that a the Ory describing the process of translation must be based on structural principles. In the light of the foregoing, however , it should be clear that this theory is incompatible both with the abstract structural approach which complete - ly separates the study of internal from external linguistics and reduce J the entire problem of describing a language to the isolation of a set of abstraet formulas (we are referring primarily to the Copenhagen school) and with the empirical structural or descriptive approach which C. Oſſ: - pletely rejects the study of semantic S > thu S making impossible any the - Cretical interpretation of translation (certain extreme representatives Of descriptive linguistics). As we will see later on 2 it is possible for the theory of translation to borrow much from both schools." .” The theory of translation presupposes a development of linguistics in which the structural approach is extended to the study of situations involving language 2 relationships between language and society , between language and thought and between language and culture. * §7. THE THEORY OF TRANSLATION AND THE comparaTIVE STUDY OF LANGUAGEs The relationship between the theory of translation and general – 35 - linguistics was recognized comparatively recently. However, there ex- ists one linguistic discipline which traditionally 1S cºnsidered in conjunction with the description of translation, and namely, Stylistics (see, for example, Gal 'perin, pages 127-1||O). It goes without saying that this relationship is not one of chance, particularly if we under- Stand Stylistics to refer to the science of synonomy, k i. e. , the vari- 99.5 means of expressing uniform content. It is Well known that it was Precisely in this manner that the founder of Contemporary scientific stylistics, the Swiss scientist Charles Bally, formulated the Subject of Stylistics. It is therefore natural that Bally considered the rela - tionship between Stylistics and the theory of translation. Charles Bally is also responsible for the idea of describing one language by C Omparing it with another', as he did in his work "General Linguistics and Problems in the French Language" (Bally, 1955). This is the origin of the new trend in stylistics, i.e., stylistique comparée, developed in recent years in France (Malblanc, Vinay & Darbelnet). The prop One nt S of this trend recognize the relationship between Stylistics and the theory of translation, a fact borne out by the titles of their Works. Expanding on the methods Of comparative linguistic description developed by Bally, they seek to isolate and define the basic concepts of stylistique comparée [comparative Stylistics ). Recognizing the value of the work done by this school in connec – tion with the theory of translation (in the following we will have fre- quent occasion to refer to these), it is nevertheless impossible not to notice the confusion in these works between the concepts of Stylistics and the theory of translation. In this connection the trend being con- Sidered here deviates somewhat from the comparative methods of Bally who was concerned with the building of a theory of the French language (in C Oſiparis on with German), but who did not propose to describe the pheno- – 36 - mena of one language by the method of comparative translations. Bally . . characterizes the systematic features of a language (specifically, French), comparing these with the properties of the other language Sy S - tem (German). In works on comparative Stylistics, as well as in many - other books dealing with the comparative grammar of two or more langua- ges, we find that individual phenomena are regarded outside of the sys- tem of the given language, and are interpreted only by means of trans- lation into another language , thus causing the interpretation to tend toward Standardization. This method of comparative language description has been employed in recent years in Several Soviet Works, including monographs and texts, With primary concentration on various types of textbooks on translation. The general principles and methods of comparative description are form- ulated in the following manner: We cannot regard as a specific feature of a given language the expressed meaning that it has in common with another language , but rather the formal categories in which this mean– ing is expressed , i.e. , the structural features. From the formal cate – gory or from objective facts (for example, from a specific type of word, word-forming models , from word order , from double-compound clauses with a specific type of Subject , etc.) the investigator proceeds to the de – termination of it S meaning in one language and then to the formal fa- cilities of the other language to express these meanings (coincident Or” / other types of words, identical or different word order , compound or * - simple clauses, etc.). It is natural that with such an approach to the analysis the material of the translations is quite worthwhile. . . " (Fed- orov, l961, page 19). This method of comparis on yields little for the theory of transla - tion treated as a particular case of the theory of interlingual rela - tionships in which equivalence is not established directly between un- - 37 – its or systems of two languages but through comparison with a third sys - ten in which all or the differential signs or comparable units are con- tained and where, consequently, the units are broken down into more - - - elementary units and treated Systematically. With the method referred to in the above quotation an attempt is made to establish direct equiv – alence between comparable units , i.e. , the same procedure is followed as in the case of a two-language dictionary in which, as is well known, no attempt is made at a comparative description of the vocabulary of the languages in question. Here it has not been taken into considera - tion that it is not the method of expressing a given category that is specific for the language in question, but rather the presence of the category itself, i.e. , the method of articulating facts. Moreover, if it is assumed that the purpose of comparative language study ". . . is the determination of Similarities and differences as well as features of coincidence and divergence for all elements. . . " (ibid., page 18), this problem can hardly be resolved (in particular, in the field of vocabul- ary). In addition, it is impossible to build a theory of language in this manner: it te impossible to build any theory proceeding exclusive - iſ from the fact that a given phenomenon (lexical, grammatical, etc.) can be translated into another language in a variety of ways in view of the infinity of possibilities of verbal expression. The theory of translation, as a science with unique problems and unique categories and methods must be built primarily on a deductive ba- sis (although, naturally, that which is obtained inductively in compari- son of languages must be employed). If this is not the case the theory cf translation dissolves into stylistics, grammar, lexicology, thus surrendering its subject. However, there exists an area of linguistic phenomena in which the Subject of Stylistics - synonymous methods of expression - and the Sub- - 38 - ject of the theory of translation are in close contact, 1.e., the phe- nomena of architectonios' (or, as 1s sometimes said, syntax), the phe- nomena of a text as a whole, and the associated problems of Selecting the units of speech from among a series of synonymous units. We know that the conjunction of clauses and larger syntactical segments varies for different functional styles and its significance in the act of Speech depends on the intent of the communication. As will become evident from the following, the necessity of conjunction and the various means by which it can be accomplished define the manner and means of carrying out the translation proces S. Conjunction and the relationships of architectonics (syntax) of text in various spoken styles and, we repeat , the resulting rules f'Or, the selection of linguistic units must apparently be studied by Stylistics. On the other hand, as will be demonstrated in the following, stylistics may also employ certain concepts from the theory of translation to des – cribe phenomena with which it, stylistics, is concerned. §3. THE THEORY OF TRANSLATION AND PROBLEMS OF TRAINING IN FOREIGN LAN- GUAGES . . - considerable attention has always been devoted in the method of teaching foreign languages to the problem of using or excluding one's native language. A parallelism is noted between recognition of the ad- vantages to be gained in utilizing translation in the teaching of a foreign language and the building up of d passive vocabulary, i.e. , the ability to read and to be able to comprehend a written text to Some ex- tent. Thus, "translation methods" were used in the teaching of Latin during the Middle Ages and during the Renaissance. These methods were retained until living languages began to be studied in Europe. The importance of translation in teaching foreign language S was keenly felt by M. W. Lomonosov, ** who wrote : "teach not in the ordinary - 39 - manner of home tutors by practice alone, but demonstrate the rules of grammar as well. These, however, should not be excessively burdened, and particularly at the beginning practice should be employed diligent - ly, words and conversations should be repeated, and there should be ex- ercises in translations and in writing" (Lomonosov, Vol. 7, pages 767– 768). During the middle of the 19th Century, an age of flourishing in- struction in the classical languages , the basic goal was the fluent (productive) mastery of Latin: here the main concentration was centered on the mastery of the written forms of speech. Translation into Latin was one cf. the main methods of teaching the language and it was treated with at least the same importance as composition and exposition , etc. And although the goals of this teaching method came into direct con- flict at that time with the needs of society (see Shcherba, 1947), thus spelling the doom of the method, it must be recognized that it was pre- cisely during the middle of the 19th Century that the methodists of this school contributed much that was valuable, part icularly in connection with the development of translation methods and in laying the ground- × Sºk for its importance in achieving a general rise in the level of f cr'e ign speech. In this connection, an example of one of the last works of this type is the Naegelsbach book "The Stylistics of the Latin Lan- guage for Germans" which appeared in eight editions (from 1846 through 1888). Works of this type based their teachings on a number of qualita- tive relationships between two languages, an attempt being made to find living fullblooded equivalents in the foreign language , rather than purely formal "grammatical" translations (it was apparently this cir- cumstance that led Naegelsbach to entitle his work "Stylistics"). For this reas on it became necessary to devote attention to the language be – ing studied in those cases in which a noun in one language may corre S- – 110 - gº pond to an adjective or a verb in another, in which a verb may corres- pond to a participle, etc. The entire Naegelsbach book is based on this principle and, for example, in the chapter on "Nouns" he presents the following cases: l) generic concepts rather than specific; 2) specific concepts in the place of generic ; 3) abstract concepts in the place of Specific, and vice versa; l!) the substitution of German nouns by Latin substantivized and nonstantivized adjectives, etc. - The carefully developed and detailed system of agreements, backed up by examples from the best Roman authors, doubtlessly represented 3.11 important Spur to the teaching of foreign languages by the analytical method, Since with this method a conscious rejection of one 's native language is achieved without carrying over the peculiarities of the lat- ter' into the foreign language. The ideas contained in Cauer's book "On the Art of Translation" have much in common with the ideas contained in the book cited above. For example, it points to Such important steps as : l) the need to determine the basic meaning (compare $30); 2) the importance of Such means of conjunction as particles; 3) the a1stinction between interpretation and translation (iber- setzen oder erklären?) [sic]; • * i It is interesting to"note the fact that the above-cited books an- ticipated certain ideas from "comparative stylistics" (stylistiane C OIII - parée) of which we had occasion to speak in the previous Section, with their contemporaries sensing that works of the Naegelsbach type of Sty- listic S were in fact not stylistics but comparative grammar (Otto, page 281). - - Unfortunately , all of the efforts by the advocates of this method – 41 - . were directed exclusively toward the goal of breathing life into the study or classical languages ºnton were doomed to extinction by the de- velopment of society. Not a single one of these people thought of ap- plying the accumulated methodological methods to the study of new lan- guage S. At the end of the 19th Century, when the need for active mastery of living foreign languages became a mass phenomenon, translation meth- ods of instruction were rejected and replaced by methods not based OI) translation. In addition, there was an essential reform in the teaching of foreign languages in Western Europe at the end of the 19th Century which was continued in Our country after the October Revolution. The most productive approach in the Soviet method of teaching for- eign languages was the One proposed in the works of L. W. Shcherba – the first Russian linguist particularly concerned with the problems of the methods involved in the teaching of languages. He was the first to con- ceive of the idea that what was required was an efficient combination of the translation methods and those methods based on demonstrative nontranslational methods. "In teaching a foreign language," said L. V. Shcherba, "we are con- fronted with at least two different goals:" "a) our purpose is to enable the student to understand a foreign- languag; text so that the function of each form in the foreign text is "b) our purpose is to enable the student to express his thoughts in a foreign language , i.e. , that he knows all forms which express a given thought and the conditions under which each of these forms may be employed." The value of Shcherba 's disclosures lie in the following. He dem- onstrated that a unique grammar must correspond to each of these pur- * & clear to him; " - f - l;2 - poses, i. e. , a "passive grammar" to the first goal and an "active gram- mar" to the Second goal. - "Passive grammar teaches the functions and values of the structur- al elements of a given language flowing out of their forms , i.e. , from their external aspects. The active £raſſima" teaches the utilization of these forms" (Shcherba, 1947, page 8||). It would be wrong to assume that traditional grammar fails Com- pletely to take these goals into consideration; there does not exist al "pure grammar" that is written without some form of appendix material and primarily written out of the context of linguistic teaching. To the contrary, as Shcherba demonstrated, such a gramſhal' seeks simultaneously to satisfy both requirements, and this is impossible. Despite the ingenious simplicity of Shcherba 'S ideas, they Iºe – mained virtually without application in the teaching of language. A number of discussions were held, # and then these ideas were forgotten. Apparently, the idea of nontranslational teaching of language is moving to the forefront. In this connection, Our methodological literature and practical operations continue to contrast translation as a conscious creative activity in the field of language to the automation of Spoken speech. - - When translation methods (permitted only in the passive mastery Of a language) are contrasted with the nontranslational methods (feasible in the act 1 ve teaching Of speech) one proceeds from the fact that "trans- lation from a foreign language into one 's native language is the pl’O- duct of a creative act and , consequently, appropriate only during the late I’ stages of language training. "Translation is a remarkable art," wrote the French linguist of the Reform period, Paul Passy,” "but ev- ery subject must be put in its place , and translation's place is not at the beginning, but at the end" (see Rakhmanov, page 106). - 113 - This approach to the translational methods is quite understandable so long as translation is treated as a creative rue tºon, 1.e., the - C Omprehension of facts of reality. At the same time, as we shall see later, in connection with the over-all development of the theory of translation, as well as in connection with machine translation, a new Conception of translation has arisen in recent times , which in no way Contradicts the development of customs of automation in Spoken language. In connection with the rise of machine translation in general there has appeared the need for a new approach to the problems of language in- Struction (see Bull's report to the 1954 Conference of Machine Transla - tion, see "MT," 1954, Vol. 1, No. 3). Machine translation is important to the methodology of teaching foreign languages primarily because it can be regarded as a model of the translation Process. Moreover, in a certain sense it also simulates the process of teaching a foreign language. As we shall see later On , machine translation confirms the ideas of a number of linguists (Yes – persen, Shcherba) to the effect that the passive and active aspects in language must be delineated. It seems to us that Shcherba's ideas on the division of active and passive grammar, Supported and enriched by experience in machine trans- lation, may be applied in the teaching of a foreign language. FOn example, the use of Certain elements of passive grammar in Secondary Schools would make it possible to coordinate the teaching of a foreign language with the general expansion and development in the teaching of many technical arts. Indeed, the methods of passive grammar provide the most effective analysis of the formal structure of a Sentence, which is particularly important when the meaning of individual words is unclear, and this is frequently the case in the translation of even the Simplest scientific - - lik - technical texts in which a school must train its students. Also impor’— tant – as we shall see in Chapter 5 – are the methods of formal analy- sis of a text which are very close to the precise methods of deductive derivation which are characteristic of mathematics and physics. Of great practical assistance in the development of a sensible method Of passive grammar may be the analysis of existing computer-analysis al- gorithms which on the basis of purely formal criteria contains a com— plete listing of rules which are required for the translation of a rather simple text. It would be naïve to assume that it is precisely these rules in their "computer form" that should be employed in the teaching of a language. The advantage here lies in the fact that each rule of convent ional grammar breaks down in the case of machine trans- lation into a number of more elementary rules , which is important from the standpoint of the machine , but completely unnecessary for a human being. The rules of "passive grammar" derived on the basis of an analy- sis program may even be simplified. The important thing here is that these rules, in accordance with the principle laid down by Shcherba, present a complete listing of all fºrms of the given language with an indication of the possible functions of each form and the conditions under which a given form serves a given function (compare Zaliznyak). • . The application of formal methods, it goes without saying, given the condition of a sensible combination Of these methods with others u Sed in grammatical analysis, will serve only to increase' the impor’— tance of grammar from the general educational standpoint. All of this would seem to be a significant contribution to the expansion and devel- opment of teaching of many technical arts ["polytechnization"]. "Passive grammar" might find extensive application in the teaching of language by the natural science departments of universities and in- St. ltutes of higher technical education where language teachers as yet : t| : *.| - i i . ... do not make use of the widely developed formal-logical habits of their students in the teaching of 8. foreign language; at the same time, the analysis of a text in accordance with the principles described later on (Chapter 5) is even more appropriate here than in the Secondary school. The development of a scientific theory of "active grammar" would enhance the teaching of languages in an institution of higher linguis- tic learning where grammar continues to be taught without differentia- tion. However, in this case active grammar must be Studied in its full Scope, including, if possible, a complete review of grammatical synonyms and conditions of their utilization. Earlier we spoke of the aid which might be rendered by machine translation to the method of teaching for- eign languages. However, even more important is the fact that the more thorough comprehension of the processes which are involved in the teaching of foreign languages is a necessary prerequisite for the de- velopment of rachine translation, and thus for the general theory of translation. - The comprehension of the process of machine translation is most easily accomplished in the terminology of teaching language to a compu- ter. The vc cac alary and rules fed into a computer correspond to the vocabulary ar. d rules which a Student memorizes. The point here , however , is that the co-puter must be provided with all the habits which enable a student to understand and construct phrases, in a foreign language. In this connection, the approach to machiné translation proposed by a group of Japanese scientists is of interest. They devised a program de – signed to accomplish a machine translation of simple phrases and voca – bulary, with b cth the grammar and the vocabulary at the level of stu- dents of English in the lst and 2nd classes of the Japanese secondary school system (the 1st class of the Japanese secondary school system in which foreign-language studies are begun corresponds to the 7th year of - 1:6 - schooling). Plans call for subsequent improvement of the vocabulary and rules so that the machine "could be promoted from class to class" (Ba- bintsev). * Even greater significance is assumed by the mutual integration of the ideas from the theory of translation and the methodology of teach – ing foreign languages in the development of Self-instruction for people and the construction of self–teaching machines. In both cases the prob- lem revolves around the development of methods making it possible to derive new information in connection with some clearly limited subject on the basis of reworking information earlier introduced into the mem-- Ory. In this connection interest has been expressed in experiments to design machines to teach foreign languages (Morton). The teaching is accomplished by the establishment of a strict sequence of questions and answers leading the student to the disclosure of some new statement, i.e. , to the recognition of some language phenomenon and to the develop- ment of automated Speech. It goes without saying that such a sequence can be worked out only on the basis of a linguistic description of the structures of one 's native language and of the language being Studied, or more exactly, on the basis of the the ory of translation from the na- tive language to the foreign. Manu - A script . " * . Page . [Footnotes] . No. . . 7 F. Engels. Dialektika prinody [The Dialectic of Nature | Gos- politizdat [State Press for Political Literature, 1918, page 8 Translation by N. Lyubimov. See Miguel de Cervantes Saavedra, Don Quixote, Vol. II, Gospolitizdat, 1952, page 358. 1 O It should be noted that it was precisely during the course of literary translations that it developed, theoretical princ 1– ple S and the continuous disclaimers of the theoreticians not – 17 - 13 13 15 16 16 21 22 23 Withstanding, that there exists an interesting division of . labor into "mechanical," i.e. , the preparation of the literal "word-for-word" translation or, in other words, the most ex- act possible copy of the original, and into "creative," i.e. , the literary formulation of the text. What is of interest here is the fact that the "translator" has no knowledge of the language from which he is translating (compare $38 on the lack of a relationship between analysis and synthesis). It must be said that the process of literary translation has not yet been adequately described – we do not yet know suf – ficiently well the elements of which this process consists — and it is therefore difficult to say now useful, or converse – ly, how detrimental such a division of labor. However, its very existence is symptomatic. Index Translationum, Paris, Vol. 12, 1961. In this context let us present the following: "The modern World has the appearance of a tremendous translation machine working at ever-increasing speed. Each day the printing pres– ses produce some 60 translations of books, 3 to l; interna- tional conferences are convened and several dubbed motion pictures are shown in theaters; typewriters are pounding out copy and translators are dictating uncountable numbers of translated pages ; an infinite number of documents in a vari- ety of forms are being translated from one language to an — other" (Cary, page 62). A detailed account of the development of the principle of machine translation in the west is presented in the article by A. D. Booth and W. H. Locke entitled "Historical Introduc – tion" (Machine Translation, page S 15–32; the Weaver memoran- dum also appears here). For more on the development of machine translation in our Country, See Andreyev, Ivanov and Mel 'chuk, pages 3–25, as well as Mel 'chuk, 1961, pages l;0–61. Occasionally this distinct lon is qualified as being the dif- ference between the 100-percent approach and the 95-percent approach (see Mel 'chuk, 1961, pages 42–43). There was a time that the authors of this book were exces- Sively categorical in Setting off a literary translation ("translation"2) against a nonliterary translation ("trans- lation", ) ("Theses," 1958, page 26). As will become evident from the following, this distinction continues to prevail in this book (see Šiš). However, there is an indication of the relative nature of this comparison. Semeiotics refers to the Science of arbitrary sign systems; language represents a particular case of semeiot lcs. The contrast between the normative and the theoretical ap- proaches to the the ory of translation has also been noted in - 118 - 30 31 31 32 3!! 35 36 39 39 the work by Z. Klemensiewicz ("On the Art of Translation," page 86). - Compare the following sentence by A. N. Kolmogorov (from the foreword to the Lebesgue book): "According to Lebesgue , those definitions are excellent which properly reflect an extensive wealth of experience with respect to the real material world" (ibid, page 9). In another article , formulating the problems of modern Seman- tics, Firth writes: "The problem of translation also applies to semantics" (Firth, page 96). - There are significant differences distinguishing interlingual and intralingual translations. The most basic of these dif- ferences has been stated well by L. S. Barkhudarov : "With in- tralingual translation related units may enter into Syntagma- tic associations arising within the limits of a given text , whereas in the case of interlingual translation, that is , the actual translation, this is impossible (Barkhudarov, 1962, page ll). In this respect translation is of great significance not only in connection with linguistics, but with logic as well. For example, Quine examines the problem of meaning in logic in the example of a translation from an unfamiliar language ("On Translation," pages ll:8–172); translation also plays a signi- ficant role in the logical conception of the Polish logician K. Ajdukiewicz, 1959. It should be mentioned that even before De Saussure analogous ideas had been put forth by Baudouin de Courtenay whose opin- ions, together with those of De Saussure, defined the activ- ity of the most productive trend in structuralism, and name – ly, the Prague Linguistics Circle (the theses of this trend aré presented in the book: Zvengintsev, 1960, pages 69–85). * * For more on the new approaches to these problems see Ivan OV, 1961. There exists yet another concept of style and namely, the science of the facilities of expression in a language, pri- marily the facilities of figurative expression ("tropes"). In this sense the science of style continues the tradition of ancient rhetoric , i.e. , the art of speaking beautifully , and how the latter gravitates more to poetry than to linguistics. The intelligent combination of the science of Style in the old sense of the word with synonimic is apparently possible within the framework of general semeiotics which studies lit - erature in particular. For the concept of architectonic , see Rizel'. Let us note that M.V. Lomonosov himself was an outstanding expert in a number of languages and he was also an Outstand- ing translator (for example, he translated into Russian the physics text of H. Wolf, his teacher at the University of – 119 - 43 l;3 Marburg). one of these took place in 1953-1954 and resulted in the com- pletely unfounded rejection of the significant distinction ..... between "receptive," i.e., passive, and "reproductive," i.e. , "active" teaching programs. The entire issue of the journal "Inostrannyye yazyki v Shkole" [Foreign Languages in the School] (1954, No. 3) was devoted to this discussion. Lately one frequently encounters the advice in our méthodo- logical literature to the effect that the teaching of foreign languages must be based on a method involving the description of "models" of these languages, the very concept of model, however, remaining undefined. We regard as productive the ap- proach of L.V. Shcherba to the linguistic justification of efficient methods of teaching foreign languages, in particu- lar, his teaching of syntagmas as more or less complete com— ponents making up a statement. We are dealing here with the fact that in speech the stop should come not at rigidly es— tablished points – models , but at points making possible a less restricted transition from a given component to one of the many possible for the given point in the speech sequence. ..The movement to bring up to date the methods of instruction in foreign languages is known as the period of reform in the history of the methods of teaching foreign languages (the end of the 19th Century to the beginning of the 20th). - 50 - Chapter 2 THE PROCESS OF TRANSLATION FROM THE LINGUISTIC POINT OF VIEW $9. LANGUAGE AND METALANGUAGE To describe the process of translation from the linguistic Stand- point it is necessary to introduce a number of linguistic concepts which are basic to the theory of translation (and within the framework of this theory may be regarded as indeterminate). One of the fundamen- tal difficulties of any linguistic description involves the circum- stance that the facts of a language must be described by means of that Same language. Such an approach carries the danger with it of contradiction and even of a vicious circle. It has long been known in logic that when treating a language in the terms of that given language we can easily arrive at a contradiction. This is the nature of the following paradox: Let us make the following two statements: "A donkey is an insect." "Each of these two statements is false." In this case the second statement will be true, and thus contradict it – Self'. These are known as "semantic" paradoxes (Hilbert and Ackermann, page 187). These "semantic" paradoxes arise out of the fact that in the formulation of such phrases there is a confusion between Statements treating real objects such as , for example, "a donkey," "an insect," etc., and statements about certain real objects, the result being that — 51 - statements are made about other statements. - In this connection we have introduced a distinction whose Signifi - cance goes well beyond the limits of the subject paradoxes: "If we study, analyze and describe Some language L1, we will require a lan– guage I-2 in order to formulate the results of our investigation of lan- guage I-1 for rules for the utilization of language L1. In this case we will refer to Pl as the object language and we will refer to P2 as the metalanguage" (Carnap, pages 3–11). The phrase "each of these two statements" is not an expression of the "object language" which we are studying, but an expression of the "metalanguage." The entire difficulty in the Semantic analysis of real languages lies in the fact that there exists a confusion between the element S of the cb ject language and the metalanguage , as for example: & Mosc cº, is the capital of the Soviet Union. Man is mortal , and such expressions of "poetic metalanguage" as : MoSC ow. . . what wealth this sound conveys TC tr.é fùs Sian heart. (Pushkin) Man. H ºf proud this word. (Gor'kiy) In esser. 2 e any linguistic term is an expression of the metalan- guage, Sir.ce each Such term represents a Statement about some prºcper- ties of the object language. This is even more to the point for the theory of translation which has to do with at least two languages. It is not out of the question in this case that one language be - comes the metalanguage with respect to the other in the case of tranS - lation. For example, when it is maintained that the expression W11 de Rosen in German reflects a single concept , we base this contention on - 52 – the fact that in Russian it is rendered by the single word 'manoshuk' or in French by the single word &lantis, 1.e., the meaning of the expres– sion in the object language is established in relation to the other language which functions in the role of the metalanguage. Of course, this principle of establishing meaning is , to a certain extent , rela- tive. If we were to take English as the metalanguage, we would have to admit that wilde Rosen reflects two concepts , i.e. , Wild roses. $10. THE INTERMEDIATE LANGUAGE Let us suppose that we are dealing not with two languages, but rather with three (Russian, German and French) and that we first inter- pret the German expression with the facilities of the Russian language , and then the Russian expression through French. Thus one of the languages will be , as is Said in the the Ory of translation, the intermediate language. The intermediate language need not necessarily be a language in the ordinary sense of the word , i.e., a natural language. Any Sign Sys- tem may be an intermediate language, 1.e., any System of Symbols, given . the condition that these symbols are positioned in accordance with the words of the translated text. The simplest form of an intermediate language is the language of things. Moreover , in certain situations this intermediate language is the only means of establishing contact, for example, in encounters be - tween speakers of various languages or during the initial Stage of teaching a foreign language. It is not necessary in this case to de Vise phrases in the language of things (as is done by J. Swift's Laputans); as a substitute for the things, as 1s well known, the indication of an object may suffice (compare the role of demonstrative pronouns and con- Structions of the type Das ist ein. . . , C 'est un. . . , That is a . . . , etc. , – 53 - in textbooks on foreign languages). - - . . The problem of the intermediate language in the theory of transla- tion was subjected to a detailed analysis for the first time with res – pect to problems of machine translation, and it developed that it is - possible to isolate l; types of intermediate languages (Mel 'chuk, 1961, pages 58–59): l) Any of the natural languages can be taken as the intermediate language. But this approach is not suitable from a practical Standpoint , since all natural languages are characterized by a high degree of mul— tiplicity of meanings and they are not of interest from the theoretical 'standpoint since translation through such an intermediate language C3I) hardly be regarded as a model of what actually occurs in the process of translation. It is of course true that in practice (for example, in S1- multaneous translation) there arise cases in which the translation of some statement is first accomplished into one of the working languages of the conference and then into Some other, more exotic language. How – ever, situations of this kind are at the periphery of the phenomena with which the theory of translation is concerned. 2) The intermediate language must represent a standardized and simplified natural language. This was Dodd's idea ("Machine Transla - tion," page 223). In practical terms this solution is somewhat more ap- plicable than the first , but from the theoretical standpoint it is equally unsuitable for purposes of modeling. $ 3) The intermediate language may be one of the artificial interna – tional languages such as Esperanto or Interlingua. In essence, this so- lution differs little from the former. 4) A language especially constructed for that purpose is selected as the intermediate language. In the construction of Such a language two approaches may be pro- - 54 - posed (Melchuk, Ibid.; Leykina, pages 1-4). a) This language is designed to have its own vocabulary and gran- mar , i. e. , it represents yet another artificial language. This is the point of view defended by Leykina. - From the standpoint of modeling human translation we find the oth- er approach proposed by Mel 'chuk to be of interest : b) An abstract grid of correspondences between elementary units of sense ("semantic multipliers") and a set of universal syntactic rela - tionships suitable for all languages are selected to form the intermed- iate language. Let us note that in any discussion of translation the elements Of two languages are compared explicitly or implicitly with some other third system, whether these be "thoughts expressed in the text" at one end , or the abstract grid of correspondences between the units of the two languages, as is set up in the case of machine translation, at the other end. Thus there is always some form of an intermediate language available and it is therefore extremely difficult to devise a theory in which this concept is not employed. On the other hand, the absence at the present time of a suitable formal intermediate language makes it Im- possible to refer to a specific form of a corresponding intermediate lan– guage. Bearing in mind that progress in the development of theory Of translation is governed by progress in the practical realization of the idea of an explicit Intermediate language and vice versa, we will sub- Sequently proceed in the following manner. We will assume that an inter- mediate language has been devised and simultaneously, determining the relationships between this concept and others, we will formulate cer – tain properties which this intermediate language must satisfy. Thus, the concept of the intermediate language will assume more clearly de- fined forms as we proceed with this discussion. - 55 - $11. CERTAIN FUNDAMENTAL concerts of STRUCTURAL LINGUIstros Let us recall certain of the simplest, &nd at the same time, card- inal concepts of linguistics and let us attempt to clarify the extent to which these concepts may be employed to describe the process of translation from the linguistic point of view. Let us begin with an examination of the simple St , and namely , with an examination of the act of speech or the process of Communication. As is well known, it involves the following. There is , on the one hand, a speaker (writer) — a sender or subject (Bühler) 1n the general form — and, on the other hand, a listener (receiver or addressee). The sender and addressee enter into some contact. What is involved is the trans- mission of a certain message presented in the form of some sequence of Signals : Sounds, letters, etc. In order for the message to be understood there must exist a fully defined System of correspondence between the elementary forms of com- munication and reality, known both to the sender and to the addressee (see the concept of "code" in §12). This System of correspondence be- tween message and fact is known as the system of language or simply as language, contrasting this system of correspondence to the multiplicity of forms of communication generally known as speech. Thus the process of intercourse or communication 1S composed of the following six components (Jakobson, 1961). } Contact ‘. Message Sender -> Addressee Code Fact In this connection it has been proposed that we d 1st inguish the - 56 - following six functions of language: 1) The sender function (designated by Bühler as Kundgabefunktion), i. e. , the transmission of the sender's state . (for example, emotion); 2) The addressee function (designated by Bühler as Appelfunktion), i.e. , the effort to arouse a certain state (for example, emotional) in the addressee; - 3) The expression function (designated by Bühler as Ausdruckfunk- tion), i.e., the function of the form in which the communication is transmitted; 4) The language-system function, i.e. , the specific features of the language in which the communication is transmitted; 5) The reality function, i.e., the actual event which gave rise to the given communication; 6) The contact function (referred to by the English anthropologist Malinovsky as the phatic function), i.e. , the actual execution of the intercourse (compare the function of words such as "Hello"). Below we will see that a specific function in translation corres- ponds to each of the above functions. - To accomplish any of these functions the elements of communication must be represented by signs, i.e., formations having two aspects: Something designated and something designating (in French in De Saus - sure 'S terminology signifié and signifiant). In this connection, we distinguish two levels in language: the level of expression, 1.e., a definite sequence of Signals, and a level of content, i.e. , a definite grouping of categories or concepts. It is important to note that the level of content is a pure lin- guistic element and two languages describing the identical segment of fact may exhibit two varying levels of content. The difference between the level of expression and the level of - 57 – * content intersects the difference between the system of language and speech. We obtain the following table (Trubetskoy, pages 13-17). Level of Expression Level of Content Language Grouping of phonolog- Grouping of grammatical System ical categories (pho- and lexical categorie 3 nemes) - System of Continuous flow (con- Continuous series (con- speech tinuum) of sound sig- tinuum) of real pheno- nals Iſlena The most important distinction between the system of language and speech, as can be seen from this table, lies in the fact that in speech we are always dealing with a continuous series (continuum), whereas in the System of language we are confronted with categories ; each phoneme identifies some segment of a vocal series and each lexical category identifies a specific "segment of reality." It is also important to note that the reality as well as the level of expression represents some continuous series, or as is occasionally Said, a continuum, i.e. , in reality we have no clearly defined transi– tions, in reality everything is interrelated and it is difficult to Say where one phenomenon ends and another begins. Thus reality is a contin- uous series and speech is a continuous series. In contrast , the system of language consists of a finite number of elements. Thus correspondence is established between the continuous series of reality and some finite sequence of rules or , as is sometimes said, categories. In this way reality in the act of speech is represented as divided into a finite number of segments , i.e. , as some discrete series. As a matter of fact, Since a segment of reality such as , for example, 'window' is made to correspond with a Single Sign, 1.e., the word 'window," and another seg- ment 'wall' is placed in correspondence with another sign "wall , " the reality is represented as separated into several segments. For the theory of translation the fact that the "articulation of - 58 - reality" may vary in different languages is of fundamental significance. Let us take such an objective phenomenon of reality as time. The Segment of time identified in Russian by the word “cytkh', [sutki = the entire 24-hour day J in reality represents a continuous series – it is difficult precisely to define where day ends and night begins, where noon ends and where the time different from noon begins - none of these transitions is clearly defined in reality. But a language, let us Say Russian, divides this continuous Series into several segments. * For ex- ample, in Russian we have: day, night, morning, evening. Within a day We have noon, within night we have midnight. In the case of German the articulation of this same continuous series is different. This can be Seen from the diagram below (Zvegintsev, 1958). Tag l l) Day; 2) afternoon; 3} evening; 4) night; (5) morning; 6) forenoon; 7) day; 8) evening; 9) night; lo) ºne: il) the entire 21-hour } ay. , * . . . g Another, more interesting example is the real continuous Series of the light spectrum: red, orange, yellow, green, blue, indigo, violet. It would seem that this articulation of the Spectrum is the only One that is possible. [Translator's Note: It is obvious from the above that the discus – Sion in the following paragraph is based on the fallacy that the word 'indigo" does not exist in English. ) – 59 - As a matter of fact, the western European languages have Only a single word for the designation of the color blue: blue in English; blau in German or bleu in French. We kno' that in Latin there is no word for the color green. In the Uzbek language there exists the single Word "kuk' which covers the range of colors 'indigo, green, blue , " i.e. , each of these three words are combined into a single word (Zvegintsev, 1958). And, finally, We have the languages of the American Indians, among which the Shawnee language identifies only 4 colors and the Bas language [sic] distinguishes only two colors (Glisson). Let us present yet another characteristic example (Hjelmslev, 1961, 127-128). We can set up a table for the concepts of "brother" and "sis- ter." (see page 61). The problem of articulating reality can be approached from another point of view. Reality can be imagined in the form of a series of sit- uations, either physical, i.e., discrete placements of objects in the physical world or actions upon these Objects by man, and social situa- tions, i.e. , interactions between people. The C Oncept of Situation re- mains one of the basic fundamental concepts in the theory of transla - tion. * Unfortunately, the analysis of situations from the linguistic Standpoint (or more broadly, from the Standpoint of semeiotics) has not yet been carried out. ** We know only that a situation — part icularly a Social situation - must be described at the "level of perception or collective evaluation." "Identical physical things' may correspond to completely different semantic descriptions — depending on the framework of which civil 1zation this thing is being considered. This is valid not only for the terms of the immediate evaluation (good or bad, handsome or featureless), not only for Objects directly associated with the civ- ilization (house, chair, king), but for phenomena of nature. Hor Se , dog, mountain, Spruce , etc., will be defined in an entirely different manner – 60 - in a society which knows (and identifies) these objects as being indi- genous than in a society for which these objects are exotic phenomena. A dog is the subject of various semantic definitions, i.e. , the Eskimos treat a dog as a beast of burden, the Parsi regard a dog as a holy be - ing and the Hindu society despises dogs as a pariah. . . " (Hjelmslev, 1958, pages 133-134). Sense Hungarian RuSSian Indonesian Older brother bátya brat younger brother öccs [brother - Sudara Older Sister nerle SeStra younger Sister Ilug [sister] The theory of translation proceeds from the existence of Some Set of Situations characteristic of a form of life and the culture of peo- ples Speaking the given languages. It should also be borne in mind that within the limitations of a Single culture various linguistic means (and occasionally even differ- ent languages) may be used for the various sets of situations. In this connection, the concept of "sublanguage" takes on great significance for the theory of translation. - - The sublanguage of this language will be that minimum collection of lexical and grammatical categories included in the system of the given language, and needed for the description of the given objective area, i.e., a spectric sphere of activity. Strictly speaking, in anal- ogy with the Subdivision of language and speech we should introduce the Subdivisions sublanguage and assessen, but this is generally not done and the word sublanguage denotes a particular collection of texts in a given language as well. The division into sublanguages achieves the generalization of that which in stylistics in certain cases is known as division by functional styles and in other cases & J Jivision by termin- - 61 - i. . . . . . . . . .” - º, Jºs .. ... ological areas (the language of chemistry, the language of medicine, ` etc.). - : $12. THE BASIC PRocess of communica RT TION THEORY OF INFORMATION AND CERTAIN concepts FROM THE * ** *** with the following simple examples (Goldman, 33-39). *- t } º Students are being weighed in the infirmary and the resulting data | are transmitted to the dean's office by means of telegraph equipment . built by the students. The nurse is dictating the results of the Weigh- l ing to a typist who is typing them on a machine; these data are then t C On Verted into telegraphic Symbols by means of Special equipment and - * finally into telegraphic" signals which are decoded in the dean is office. The elementary communications which are transmitted at each Stage can be presented in the formſ of the following table: Russiºn Word *** * the Telč. h El IV A. . €gºrºa e (spoken) printed text . O strica.ºlegraph • elements elements l One l • (dot) trans - mit (+ two § :::::) rest (–) (+) I ten 5 eleven 6 tº º gº 7 8 twenty 9 O f tº e ge * forty forty one fifty ninetyeight ninetynine” *For convenience we have written the letters rather than a phone- – 62 - & - - \ * : * ! Before we take a Iook at the general process of communication, let - * → • • ? , -’ tic transliteration, but the words are written as one, as they are pro- nounced. For simplicity the process of converting nerve impulses into sound oscillations and vice versa is omitted. . Any set of elementary communications (I or II, or III, or IV) in the theory of information is known as an alphabet. With this definition for the term "alphabet" the number of alphabets in a given language is rather great. Let us present a number of these (the elementary communi- cations comprising the alphabet are indicated in each column): I II III IV Alphabet of Alphabet of Alphabet of Alphabet of German German German words German phonemes Letters (after Rudasch) phrases p, b, t, d, a q Aal \ º k, g b r aalartig C S Aalbaum . f, V, s, z, g, d t Aalbeere J, X, h, ſ e Ul É - f V # pf, ts, g W 3 All German m, n, l, r h x * * * * * . . phrases which i y #3 i.º. 1:, y:, u: j 2. | 5 cordance with . k - | * the rules of . 1, y, v l Zvnismus | < ... ºr e., O., o In Space ź. words shown * * Il º:” Zytologie in IV. Sae, 3, a;, &MO O -- * . - p An analogous table can be prepared for any other language. Thus at any level a communication consists of some sequence of symbols included in a given alphabet. It is clear that there are rules governing the construction of such series, with certain series permis- Sible and other's not. These rules are known as the code. "Any code represents some alpha- bet and a system of fixed limitations" Goldman, page 30). The relation- Ships between a code and the communication in a language are rather - 63 - complex (Ivanov, 1957, page 18). Let UlS note only that the point of - view or a linguist analyzing a language coincides with the point of view of an individual who seeks, on the basis of these communications, to restore the code or vice versa, having the given code , to devise some communication. Let us now examine the general circuit diagram of any communica – tions link: |axº~~ |: 6 *::::::::: /asyva ºp 3:#| > || || -- – - 4 | . - ſ - Tºn #. ſºft" | . Žiže Diagram No. 1. 1) Communication; 2) sender; 3) coding device; l;) decod- ing device; 5) receiver; 6) communications link; 7) code rule. The abstract nature of the diagram makes it possible to use it not only , for example, for the description of the transmission of ordinary telegrams, but in general for any transmission of communication by means of some system of signs (of language in the broad sense of this word). Let us examine the convent ional process of speech. How does com- munication take place between A (the brain of the sender) and B (the brain. Of the receiver). For a description of this process it 1s enough to use the special terms generally employed by linguists instead of the general terminology employed in Diagram 1 (see Diagram No. 2 on page 65). Thus we see that the process of intercourse considered in Šll is a Special case of this general diagram. Let us now turn to the information transmitted in accordance with - 64 - gºod º Hi tº ... " º, ommanen, 6 momyſare/ºr 7 É. Diagram No. 2. 1) Nerve impulse S ; 2) sound oscillations; 3) speech ap- paratus; 1) air medium; 5) ear; 6) sender; 7) receiver. this diagram. We know that as the measure of the information. On SOme experiment a in the theory of information the quantity H (2) * = —P, log, P, -Palog,P,... –Palogspa (l) is received, where Pi is the probability of the ith outcome of experi- ment o and 1 bit is chosen in this case as a unit of information, i. e. , the quantity of information obtained in an experiment with two equi- probable results. The formula yields the maximum value only under the condition that all P, are equiprobable, i.e., if each P1 = l/n. In this case Formula (l) assumes the simpler form: "... --É loc, ** (2) The ratio H/"max is sometimes known as the coefficient of compres- sion, while the quantity D = 1 - (H/Hmax) i (3) / *. . . . . is known as the exceS 8. It should be noted immediately that the direct application of For- mulas 1-3 is hardly feasible in the the ory of translation, primarily because of the fact that it is impossible for us directly to transfer from information measured in this manner to 1nformation of meaning. How- ever, there does exist an indirect relationship between these concepts. And it is to be hoped that translation will make it possible more ex- – 65 — - actly to resolve this problem. - - - - Let us imagine that translation rally retains the sense of the original and let us examine its value for the theory of information (Yagion, Dobrushin, Yaglom, page 109). If we use our formulas to compute the statistical information con– tained in this or in some other text, it will turn out that the p bits of text information in one language are equivalent to q bits of text information in the other language , thus making it possible to compare the "semantic content" of 1 bit of information in either language. It is of course necessary for the resulting conclusions to be checked against rather extensive statistical material. Investigations of this type were carried out by the Indian Scien- tists Ramakrishna and Subramanian. They established that in translation from English into German 1 bit of information contained in the English text is equivalent to l. 22 bits of information in the German text ; this difference is defined as the "economy" of English in comparison with German, so that an attempt to transmit the content of any segment with complete precision in any language results in an expansion in compari- son with the original because of the difficulties associated with the need to retain all shades of meaning in the translation that are ren- dered by means that are unique to the given language. On the other hand , in the translation from German into English 1 bit of information Con- tained in the original German text will be rendered by 1. O6 bits of in- formation in the English text. Hence Ramakrishna and Subramanian con- cluded that l bit of information contained in the English text is equal approximately to l. 15 bits of information in the German text ; more over, in their opinion, the process of translation requires the additional use of information (which can also be evaluated) which is associated with the need to account for the specifics of the other language. - 66 - $13. TRANSLATION AND INTERPRETATION - Let uS now consider the process of translation. How are we to re- fine that which we intuitively define as translation, proceeding from that scheme of the communication act which we have analyzed? In other words, how can we simplify the concept of "translation" so that it could be discussed in the terms of linguistics? -7- Hiſ ſl T | -- | - 0nopadume/re | //epedodwarar i //onywame/re t | I --~~~~! I r | F-1 FT º Curmeswar ! Cooſaemuel il Cucareer Cuc | | i ! Cucº" | | - Mſ7 wa //7 //7 /7 an ſº? /772 || || | l; % || T] . 6 |------- +--- th- - - - * Al º ... Diagram No. 3. 1) Sender; 2) translator; 3) receiver; l!) source lan- guage System; 5) communication in source language; j target language system; 7) communication in target language; 8) reality. II = P = perevodchik = translator Logically two basically different processes are possible and these can be described by means of this diagram. The first process can be described in the following mannelſ" : There is a sender A, and addressee B, but since A and B use dif' – ferent systems, a translator P is incorporated into the communications act and this individual acts simultaneously as the addressee with res- pect to A and as the sender with respect to B. Using,some system (which se will refer to as the source language [Iya]") A transmits communica- tion C, about some real situation D. The translator , using the source- l language system, establishes correspondence between Cl and D and then, using the new system (we will refer to this as the target language), he builds a new communication C2 about the same situation and this communi- cation C2 is received by the addressee B who, in turn, using the target – language system establishes correspondence between C2 and D (see Dia- – 67 – gram No. 3). - In Other words, the procedure is as follows: the translator re- ceives some spoken Sequence and from this sequence he transfers to a situation, examines the Situation and then, moving completely away from the communication which has been sent to him, bearing in mind only the given situatten, the translator communicates about the situation to an - other person. - In clearer, although paradoxic form, this principle has been ex- pressed by the Englishman Belloc who suggested that translations be ac- complished in the following manner: 1) read the foreign book in its en- tirety; 2) render in one's own language the effect which the book had on him; and 3) compare the translation against the original in order to achieve a closer approximation of the original, without losing natural freedom of expression ("On Translation," page 282). - º F-'. Diagram No. 14. Denotations the same as for Diagram No. 3. | * The second process may be described in the following manner" : A transr; its a certain communication Cl about a real Situation D, as in the first case, but P, receiving the communication C1, does not transfer from this situation to the Situation in reality, but to the intermediate language, i.e. , to a system of correspondence between the source and target languages, and then transfers from the intermediate language to the target-language system. By means of the target language - 68 - E] … - 1) Translator; 2) source-language system; 3) previous experience; li) target-language system; 5) re- ality. he formulates communication C2 which provides information to the ad- dressee about the situation in reality. This is shown in Diagram No. 4. It is indicative that Diagram No. 4 is a duplication of Diagram No. 1, with the decoding device of one combined with the coding device of the other. Thus Diagrams Nos. 3 and l, show two different processes. As is evident, the second process is a variant of the first ; however, it dif- fers significantly. First of all, it involves no direct referral to the situation taking place in reality. The transition from one language system to the other is accomplished directly in accordance with an ear- lier established system of correspondence. It goes without Saying what when this system of correspondence was established consideration was given to the reality and situations which reflect the corresponding categories in both of the languages. It is extremely important, however, that this is something that has taken place earlier and is not a part of the actual translation process. - From the above-considered differences between tºº two procedures – 69 – we find that there exists yet another difference which is significant for the comprehension of what follows. Having examined a certain seg- ment of reality in accordance with the circuit shown in Diagram No. 3, the translator makes use of his entire previous experience and all of his knowledge regarding this given segment of reality and its relation- ships to other segments of reality. In other words, that portion of the circuit in Diagram No. 3 which describes reference to reality may be expanded in the manner shown in the figure above. In the procedure indicated in Diagram No. 4, where referral to re- ality tº excluded, the possibility of referring to the earlier experi- ence of the translator is also excluded if , of course, this experience has not already been generalized in the rules of agreement, a S indica- ted earlier. We will refer to the procedure demonstrated in Diagram No. 3 as interpretation, while the procedure shown in Diagram No. l, will be known as translation. Both are part of a translator's activities; inter- pretation is more frequently encountered in the translation of literary material where the goal is to reproduce the reality expressed in the Original, while translation (in the above-defined sense) is noted clearly, for example, in the activities of simultaneous translators [interpreters J. However, translation is characteristic of all forms of translational activity in the ordinary sense of the word where the prob- lem involves the substitution of communication in One language by com- munication in the other, with retention of meaning. * Stressing the importance of the indicated distinction #3, we should point out that the concepts described here are not likely to be encoun- tered in their pure form in actual practice. Thus in any form of non- literary translation, including even simultaneous translation, there are possible moments in which interpretation is necessary, i.e. , refer- - 70 - ral to situations and the application of all previous experience to find correspondences not provided for in the rules. On the other hand, in literary translation it is quite possible to find qualitative COI* → respondences earlier established by the vocabulary and grammar , i.e. , translation in the sense described above. Let us now take note of the fact that in essence we have defined translation as the process which may be completely formalized, while interpretation has been defined as the process whose formalization at the contemporary level of our knowledge about language is regarded as impossible. With the initial appearance of machine translation (compare §3) two completely contradictory points of view with regard to this problem were expressed. One of these , most clearly expressed by L. I. Zhirkov, states that: "It is possible to maintain that the prosaic translation of any text by means of a machine is always possible, in- cluding a literary text" (Zhirkov, page 123). Another viewpoint was ex- pressed by L. S. Barkhudarov and G. W. Kolshanskiy who stated: "All that goes beyond the framework of elementary, formal and uniquely defined rules of linguistic functioning cannot be consolidated into a circuit and , consequently, cannot be the object of computer processing. It is not the development of technique that sets the limits on the possibil- ity of machine translation, but language itself. The computer can ac- complish only that which it is told to do by a human being, and in this case the human being is powerless to reproduce the entire program of the possible functioning of a language at the corresponding instant of a translation, because for this he would have to return each time to the creative activity of the mind or , in other words, to resort to as - sistance of his thought processes in seeking out the required results "on the basis of meaning" (Barkhudarov, Kolshanskiy, page 133). The renowned specialist in the theory of machine translation, - 71 - d Y. Bar-Hillel, has recently come around to this point of view and in a number or articles and statements has begun actively to promulgate the idea that machine translation is impossible. Bar-Hillel summarized his opinions with regard to the automation of translation in the following words: "Complete automation of this form of mental activity represents absolute Utopia for the reas on that books and articles are generally written for reader's possessed of a Certain amount of knowledge and cap — able of logical thought , including not only logical conclusions, but So-called probable reasoning. In no manner can this capability be re- placed, even given the most cleverly employed of all formal indicators of speech" (Bar-Hillel, page 205). We see that Bar-Hillel regards the given problem as one of general methodology: the machine in his point of view can never do that which a man can accomplish. At the same time he takes the position of those who hold that man cannot under any circumstances be regarded as an automa – ton (because in the opposite case man would be incapable of interpreta- tion) and that man possesses something that is unknowable and nonrepro- ducible. We cannot agree with Such a general methodological Weltanschau - ung approach. At any given instant of time the potentials of science and the potentials of exhaustive formal description by science of its objectives are limited ; however, as demonstrated by all experience in the development of science, all such limitations are temporary in na– ture. In particular, it Seems to uS that in resolving the problem of the possibility or 1mpossibility of automating translat 1 on the differ — ence between translation and interpretation should be borne in mind. The significance of interpretation varies from text to text — we have already spoken of the fact that it is important in literary text; and insignificant in texts whose primary function is communicative , i.e. , the reality function (Kuznetsov, Lyapunov and Reformatskiy, page lo8). – 72 - one cannot rail but notice in this connection that the language mater- 1a1 used to illustrate the impossibility of automation is remarkable in its artificiality. Thus, Bar-Hillel in one of his articles refers to the impossibility of removing lexical polysemy without resorting to knowledge of the real world and he cites the following example: "The box was in the pen (Bar-Hillel). Of the two possible translations of the Word pen in the given phrase — "quill (fountain pen)' and 'playpen – the human being would select the second choice because he is fully aware of the fact that a 'pencil case [box] cannot "be in" the ºpen (fountain pen). It must be taken into consideration, although Bar-Hillel himself does not regard this argument as convincing, that the algorithm of mach– ine translation may be set up in such a manner that specific operations Will be accomplished with consideration of information obtained from data -processing Systems. The entire matter rests on the extent to which the previous experience and accumulated knowledge on the part of the translator with regard to reality – which, as we have seen, are a char – acteristic feature of interpretation as distinct from translation – can be formalized. - As an example of the foregoing we can cite the solution to the problem of Selecting the article in the translation from Russian into western European language S. Let us note that unlike examples of the type discussed above, this problems arises in any context. if we are confronted with the problem of translating the phrase : "Tsar Fedor's son," we find that both forms of the article are possible in German, French, English, Rumanian, etc. : 1) ein Sohn, un fils, a son, un frate, 2) der Sohn, le fils, the son, fratele. In the case that the word had not been used before, the first of these translations indicates that the Tsar had several sons, while the — 73 - i. ...' r º … . . * tº second translation indicates that the Tsar had only one son. In the - ºf selecting the proper version. Let us consider yet another example. We are confronted with the problem of translating the sentence: "Factories and plants in the USSR are socialist property. TWO versions of this translation are possible: . - l) for example, German: Die Fabriken und Werke Sind in der UdSSR das sozialistische Eigentum. *- 2) German: Die Fabriken und Werke sind in der UdSSR sozialistisch- eS Eigentum. The use of either the first or second version depends on the ex- tent to which the translator is familiar with the social and governmen- tal structure of the USSR. It is clear that the first version of the translation — signifying that "The factories and plants in the USSR are the only form of socialist property' - must be reconstructed. Let us now select an analogous example involving the use of a noun in the plural. In translating the phrase : "representatives of the fif– teen Union Republics took part in the competition " the German will say: Vertreter der 15 Unionsrepubliken and in French it will be des repre- sentants des 15 republiques fédérées, while the phrase 'representatives of lo Union Republics' 1s. translated as: Vertreter von lo Unions repub – liken, des representants de 10 republiques fédérées. Today solutions of this type can be achieved only as a result of interpretation. However, in principle it is conceivable to describe these cases as translation in the form presented in Diagram No. 14. As a matter of fact, let us imagine that there exists some device which translates the above phrases in accordance with the rules of gram- matical correspondence. Let this device exhibit the property of being - 74 - specific case knowledge of history alone might assist the translator in able to indicate why it is unable to work out a translation in cert. 3 in specific cases. For example, in the case "Tsar Fedor's son" it will in- dicate that it cannot accomplish the translation, since it does not know how many sons Tsar Fedor had. Let us imagine that this device is connected to a large data-processing system containing the most impor- tant historical facts and which is capable of answering precisely for’ſ – ulated questions of the type indicated by the question of how many & Cº.: Tsar Fedor has. In this case all of the information contained in the data-processing system will be regarded as being included in the previ – ously established rules of correspondence and the problem of whether i : was translation or interpretation that was involved in this Case will 3: - pend on whether or not this historical fact had been recorded in the memory of the data-processing computer. Let us take note of the fact that the situation described here does not basically differ from the activities of a translator who under similar circumstances turns to ºn- cyclopedic references and historical Sources, etc. As demonstrated by the cited examples, the demarcation between translation and interpretation can shift and depends on the level or " linguistic formalization. We the refºre regard as unproductive the iX - tion of those who, like Bar-Hillel, categorically 5tate that Linc tºe i : ; no hope whatever of achieving success in the automation of translat, 3 ºr . At the very least , such a convention is premature : the work be in 3 dº on the simulation cf the Speech patterns of human beings, particula Ply in the formalization of semantics, is in its initial stašeš and expº 1 : — ing rapidly; for this reas on there is no basis to maintain in advah (; & that it will not achieve any positive results. To announce that any further research along this line will not be productive suggest S iſ ct only the blocking of the path to the solut icn cf practical problems , but to retard the development of linguistics - - 75 - polemic 1zing with advocates of the idea that automatic translation is impossible, in no way do we negate the advantages to be gained from - human participation in the preparation of the text , its simplification prior to its input into the computer , i.e. , preediting, and in improv – ing the quality of the translation printed out by the machine , i.e. , postediting. There is no basis for the assumption that ". . . this solu- tion of the problem. . . not only limits the significance of machine translation, but places in doubt its actual feasibility on a large scale" (Barkhudarov and Kolshanskiy, page 131. See also the similar op- inion of Bar-Hillel, 1962, page 205). Evidently another approach would be sensible , and namely : permit the preediting and postediting to that extent to which these operations are unavoidable, but use the results a chieved in the reworking of the text by a human translator to improve the translation algorithms, i.e., to enrich "the experience of the 11 ºf 3 machine," "teach" translation to the machine (Hays, pages 6, 13–25). Thus the share of the burden borne by interpretation will diminish, while that of translation Will increase. Finally, let us take note of the fact that posted iting is also ne– " and by no means does this ce 33ary in the case of "human translation, indicate that man is incapable of translation. $111. MEANING AS AN INVARIANT OF TRANSLATION Thus we have defined translation as a specific transformat Lon U 1 a communication. In any transformation there always arises the problem of . . . what remains unchanged during the transformation process or , as i:3 gen- e.’ally said, the problem of the invariant of transformation. It has always been recognized intuitively that meaning must 1’ema iri the invariant in translation (compare in this connect lon the usual re- quirement "to translate not words, but meaning" etc.). However, we are confronted here with the fact that the concept of "meaning" is preci., C – – 76 – ly defined only with great difficulty. * . . . . . . The initial difficulty involves the fact that meaning is not always uniquely defined with respect to the segment of reality indi- acetaw - C&A fºſſ 3 cated by the given communication (or, as we 1) Referent ; 2) com- munication; 3) mean- ing. Will Speak of it , with respect to the refer – ent. Occasionally the term "denotation" is used in the same sense. In other words, it is necessary to distinguish between communication, referent of communica– tion and meaning of communication. * For example, the communications: "Walter Scott" and "the author of Beverley" have the same referent, but differ in meaning (Church, page 18). Speaking of the differentiation between the concepts of referent , meaning and communication (symbol), Firth presents the following indica- tive example: "Generally various newspapers print articles about the Same event under different headlines. Let us imagine an event Such as the Sentence of Lord X. The headline in the "Times" will speak of the R. M. S. P. Case; the "News chronicle" will announce : Sentence Passed Upon Lord X; the headline in the "Daily Herald" will read: Lord X Sentenced to Jail for One Year; the "Daily Mirror" will have : Lord X Sentenced to 12 Months in Jail; the "Daily Mail" will state: The Sentence of Lord X Amazes London. And, Finally, the headline in the "Daily Worker," whose tone will imply 'it serves him right, ' will announce : Lord X gets 12 Months . . . The referent here is the same in each case , i.e. , the Sen- tencing of Lord X. It is identified by various symbols in the various headlines , but in various relationships — that is , the relationships between two sides : the headline and the event" (Firth, page 76). ** Let us note that in the case of interpretation (Tiºgram No. 3) - 77 – correspondence iS established through the referent (the segment of re- ality). Identity of meaning, generally speaking, is not required. As a matter of fact, in actual practice we are familiar with cases in which the referent remains the same on interpretation, whereas the meaning changes. Phenomena of this type arise in the simplest of translational situations Such as , for example , in the use of so-called periphrasis. Thus meaning is not retained in the translation of the German Messestadt [Fair City, i.e. , the city in which fairs are held | by 'Leipzig, " or the French ville-lumiere by 'Paris, ' etc. When rendering technical terms interpretation is frequently em- ployed. Let us assume that the term Bodenfreiheit [ground clearancel is to be translated in a technical description of an automobile. The term- inological equivalent is encountered with regard to tanks where it is expressed by the word 'clearance borrowed from the English. This term denotes the distance from the underside to the ground. With respect to automobiles, however, we more frequently encounter the term 'road clear- ance. " In this case we can resort to the method of descriptive transla - tion and translate : 'The distance from the lowest points of the automo- bile to the road. ' This descriptive expression is used in the descrip- tions of automobiles in Russian. Let us cite several other examples of descriptive terminological translation: - Unterflurmotor — the engine housed beneath the floor of the body; Solokraft rad – a motorcycle without sidecar; Strich platte - a graduated dial ; Abwalzverfahren hob method [generating process of gear cutting]. (Shvanebakh, Revzin, pages 80–81). With this approach to the original, i.e. , correlating the text with the referent, it is not translation that is taking place in the - strict sense of this term, but rather interpretation, it being diff i- cult to make the distinction here. Indeed, we have the obviously valid translation, or more exactly, the interpretation, of the word bonhome as ' the old man' in the context of the Short Story by Daudet "Un Teneur de livres": Brr. . . quel brouillard J. ... dit le bonhomme en mettant le pied dans la rue. ['Brr... what a fog - says the old man as he goes cut onto the street (Fedorov, page 1311). Particularly fraught with the danger of distorting the original is the interpretation of documentary, classical and similar texts. Instruc- tive in this connection is V. I. Lenin's experience as a translator. In the book "Materialism and Empiriocriticism" v. I. Lenin translates Marx is second thesis On Feuerbach: In der Praxis muss der Mensch die Wahrheit, d. h. die Wirklichkeit und Macht , die Diesseltigkeit seines Denkens be - weisen in the following manner: "In actual practice man must prove the truth, i.e. , the reality and power, the this-sidedness of his thinking. In this connection Lenin states: 'Plekhanov instead of "to prove the this-sidedness of thinking" (literal translation) writes: to prove that thinking "does not stop on this side of a phenomenon"... Plekhanov pr’O- Vided a paraphrasing rather than a translation... Plekhanov's paraphras– ing is not necessary for those who are interested in knowing precisely what Marx had written. . . "% Let us cite a Second example. In his "Who are these ‘Friends of ! I the People'. . . ." Lenin translates a section from the journal "Sozial- } politisches Centralblatt," where it was said of Populism that it ". . . wird herabs inken ... zu einer ziemlich blassen kompromiss fähigen und kompromiss Süchtigen Reformrichtung, in the following manner: "it will either degenerate into a rather pale reform movement Capable of comprom- 1Sing and seeking to compromise . . . " In a footnote to this translation Lenin states: "Ziemlich blasse kompromissfähige und kompromiss stichtige Reformrichtung - can apparently be rendered in Russian as: cultural Op- — 79 - portunism." As we can see Lenin clearly astinguished between transla - tion in which the meaning of the original remains unchanged and the in- terpretation of the author's intention by the translator. “ The problem of translation and interpretation becomes particularly acute in the translation of ancient texts when the likelihood of dis - tortion in the meaning of the original as a result of interpretation increases because of the difficulty involved in the determination of the referent (Shevoroshkin, compare, however, Markish, page 155). With regard to the process of translation (Diagram No. 11) the sit- uation is different. The referent is completely out of the picture and the problem consists in establishing such correspondence as ensures in- variance of meaning. This is particularly important because of the fol- icwing consideration. The intuitive concept of meaning is best defined in terms of translation. It is interesting to note that this is the ap- proach used in the scientific investigations which seek to attain maxi- mum accuracy. Compare the following definition: "The meaning of a sen- tence can be described as that which is mastered once the sentence has been understood, or as that which two sentences have in common in var- ious language S if they are properly translated with respect to one an– other" (Church, pages 31–32). - How is the invariance of , meaning to be reflected in the theory of translation? We have already, stated ($10) that each expression is made * to correspond with some grouping of units from the intermediate language. Hence the basic requirement imposed on the intermediate language: it S units must correspond to the semantic units 13 olated in both of the languages (both the source language and the target language). Thus the intermediate language and the invariance of meaning must be defined S1- multaneously and mutually refine one another as new linguistic material is brought into play. On the basis of the foregoing we will define the - 80 - meaning of some expression as the grouping of elementary intermediate - language Semantic units which correspond with the given expression. Let us take note of the fact that this concept of meaning corresponds fully with intuition, and the invariance of meaning in our concept corresponds with that which is meant when we say that "the theory of translation seeks to establish a correspondence between units of two (at least) different languages on the basis of common expression or semantic content" (Barkh- udarov, page ll). - Perhaps the most significant difference between that which has been discussed here and traditional theory lies in the fact that the in- Variance of meaning is not treated by us as an absolute category but as invariance with respect to the constructed intermediate language. $15. THE PROBLEM OF TRANSLATABILITY In §13 we considered the problem of whether automated translation was possible. However , of interest is the fact that the problem of translatability, i.e. , of the possibility of establishing correspond- ence between lexical and grammatical categories of two languages , as assumed by the translation diagram, has repeatedly been the subject of doubt. This is associated with the fact that reality is variously ar- ticulated in various languages (see Şll). Proceeding precisely from this fact , many linguists express doubt as to the possibility of trans- lating texts from one language into another. The linguist W. Humboldt most clearly expressed the idea of the impossibility of reconciling two views of the world described in various languages. The following is one of his characteristic statements: "Each language inscribes a circle about the people to which it belongs and it is possible to Step beyond this circle only if one steps into another circle. The study of a foreign language might therefore be compared with the acquisition of a new point of view in the previous conception of the ... . . . . . Tust because – 81 - -- -' • * - - - - - - - .* - - --- - - * - - .* us transfer our own ideas about the world and our own semantic views into a foreign language, we are not raily aware of the results of this process" (Humboldt, page 81). Humboldt's ideas have recently been restated in another form in Europe by Cassierer and Weisgerber, and by Sapir in the United States of America. The problem of language's influence on thought and the formation of concepts through a given language , where the expression of these concepts is apossible in another language , is considered in detail in the works of B. Whorf. Expanding on Sapir's idea that mean- ings "... are not ascertained exclusively through experience , as much as they are imposed by the tyrannical power which semantic form holds over our view of the world". (Sapir, page 78, quotation from "The Latest in Linguistics," Issue 1, page ll'7), Whorf draws the conclusion that the unique Structure of each language governs the culture and world outlook of its speakers. As did Sapir, Whorf derives his conception of the re- lationship between language and thought from observations of the cul- ture and languages of the American Indians. The exotic nature of these languages led Whorf to the following hypotheses: 1) our conceptions of (for example, time and space) are not identical in all people, but gov- erned by the categories of the given language and 2) there exists a re- lationship between the standards of culture and the structure of lan- guage. Whorf finds confirmation for these hypotheses in the languages of western Europe. He writes: "The Newtonian concepts of space , time and matter are not a result of intuition. They are set by the culture ard the language. It was precisely from these sources that Newton took these conceptions" (Whorf, "The Latest in Linguistics," Issue 1, page 168). The objectivized time that is, according to Whorf, characteris– tic of the western European languages corresponds to such features of western European culture as bookkeeping, mathematics, concern with pre- - 82 - Clse sequence, historic character and the like, such features being to- tally absent in the cultures of the American inaians. "we encounts a , " wrote Whorf in another work, "new principles of relativity which state that similar physical phenomena make it possible to create a simila tº picture of the universe only 1f the language systems are similar or at least related" (Ibid, page 175). At another point directly related to the problem of translatability considered here, Whorf formulates his idea as follows: "The formation of thoughts is not an independent pro- Cess, strictly rational in the old Sense of this word, but a port iOn Of the grammar of a given language differs insignificantly among various peoples in certain Cases , whereas in other cases the difference is sub- Stantial, as in the case of the grammatical structure of corresponding languages" (Ibid, page 174). Whorf does not reject the obvious uniformity of contemporary sci- entific thought. However, it seems to him that the Scientists who do not speak the European languages in which the basic concepts of science have been formulated may become accustomed to that science and may de – Velop its achievements exclusively through mastery of its thought pl’O- ce SSes and the cultures of Corresponding peoples, rather than on the basis of ideas formulated under the influence of language. At the same time Whorf's conception approaches the idea of the impenetrability of languages put forth by a number of IndoPEuropean linguists. Just as these linguists, Whorf draws attention to that which is original and unique in language, to the phenomena which have not been subject to the Pressures of other languages and cultures. To al considerable extent this approach was imposed on Whorf by the very subject, i.e., the lan- guages of the indigenous population of America, which during the course of many centuries developed independently of one another and of the Old World. — 83 – ence has not came aeross cases which do not lend themselves to transla- tion. And it is of eburge theoretically eleae that thers exist such categories of language among which it is impossibie to establish ºr- respondefice and, consequently, it is impossible to retain invariance of meaning. It is, however, important to define precisely the categories of language that are meant when we speak of nontranslatability.” since we are eanfronted with the questian of *lanes of meaning, we require a semantic classification of linguistic categories, Fºr than a formal classification. In first approximation we suggest the following classification. First of ºil, all categories , i.e. , all methods of articulating reality with the facilities of grammar or by means of the lexical ap- paratus, can, by agreement, be divides into semantically empty and se – mantically full. We will use the term semantically full for those categories which contain extralinguistie internation, with this sategory easable or be- ing operated upon by means of a given intermediate language. * Any gate- gory which aces not satisfy this requirement will be known as semantic- ally empty. In particular, we will regard as semantically empty those categories which contain purely linguistic information, i.e. , are used / * * * * ...” exclusively for the internal needs of a language (e.g., the syntactic categories). As examples of empty categories we have : the gender of nouns, all categories of adjectives with the exception of degrees of comparison, the gender, number and person of verbs, etc. As examples of full categories we have: number, 'definite and 'indefinite' nouns, mood and modality, verb tense, apparently all lexical categories (it is assumed that the intermediate language is a natural language or an ap- - 81; -- - - - -— Semantically empty categories are not significant in translation. The fact that the masculine noºn (for example, croſſ' ) will be transla- tea into another language as a feminine noun (for example, la table) in no way affects the semantic conversion of the text in question. The theory of translation is interested in semantically full cate - gories. These cana in turn, be subdivided into three groups* depending on the relationships which they enter into with other categories in the description of a situation. A. There exist categories whose use is dictated by a corresponding situation. For example, in sºammar we have the concepts of definite and indefinite, absolute and relative time: in the vocabulary we have all kinds of pleonasºns. Although these categories present difficulties for a beginner translator, they do not present any basic difficulties. In- deed, if a given category exists in the target language, an individual who knows that language will use that category (for example, the trans- *atiº gº the derinate-noun category from Russian into French). However, if there is no corresponding category, none is needed, since its use is defined by the situation. It makes no sense to seek a grammatical equiv- alent for the category of the mood of a Russian verb in translation-in- to a language which does not have this category: with the exception of the very rarest of eases (compare *columbus was happy not when he had discovered America, but as he was discovering it" (Dostoyevskiy), the use of the category of mood follows directly from the situation (and moreover, from the surrounding text).” There are categories in the vocabulary that are also completely predicted by a situation. We can easily notice, for example, that in the translation of Russian adjectives of the type ‘IpenBB16opºliſ , - 85 - --- - - - - - a . - . . . . ." . . - .' ' ' ' ', a - - - - - - *... * * * - - 1. * *** fall-union] into English, French or German, fºr the prefixes pred' and ºvse' cannot be translated. (compare, preelec- tion campaign is rendered in French as campagne &lectorale, and into German as Wahikampagne). [Translator's Note: the authors are again pro- ceeding on a fallacy that certain words do not exist in English, where- as "preelection" and words of the type "all-union" appear and are de – . fined in Web Stef's oneerinees Dictionary]. B. There are categories, whose use is compatible with some situa- tion, although not sover-a by that situation. This means that a given situation can be described through the use of various categories. For - example, this is the case in grammar with the categories of verb modal- ity and tense and the category of noun number; in vocabulary this is the case with the so-called ideographic synonymic s. (It may be main- tained that stylistic synonyms are part of one and the same lexical category. Compare ‘ecre' [to eat 1, ‘kyuare' [to take food] and ºxpate [to gorge] where the differences between these are not associated with var- ied articulation of reality). It is precisely the categories of this group that led to the idea of nontranslatability. As a matter of fact, if there exists in some language an articulation of time other than the articulation of time in the other language , to the extent to which the unique nature of this articulation is not governed by the situation, theoretically speak- ing, it is impossible to establish correspondence between these two categories. Observations of this kind, carried out with the American- Indian languages in comparison with those of western Europe, led to the formulation, as we have seen, of the Sapir–Whorf hypothesis of the non- translatability of text into various languages. In the area of vocabu- lary this type of nontranslatability was noted a long time ago. Compare the lack of correspondence between the lexical units ‘asſa' [cottage ) and - 86 - and xprs ſhut] and chaumiere; maison; or, on the other hand, the lack of correspondence between yº' (comfort], confort, Gemütlichkeit. (com- pare also the various problems which arise between vocabulary and gram- mar , as in the various forms of address — ‘Thi’ and “BB” in the Russian, you in English; the 'Bº' in the Russian and the noun that is used in this function in Polish, not to speak of the analogous differences between European and Oriental languages). C. There are categories whose Li Se logically contradicts situations. In grammar such use of categories occurs rarely. Compare, however, the phrase : "It was raining several times yesterday, ' where the form 'it was raining," denoting a continuous action, contradicts the Situation. Here we must include such cases of neutralization as the use of forms of the present and past tense to express categories of assumption in the present and future tenses. (Compare the French Si j'ai de l'argent, j'acheterai ce livre; si j'avais de l'argent, j'acheterais ce livre). In vocabulary the use of semantic differential indicators contra- dicting a situation is encountered quite frequently. Compare , for ex- ample, the use of the word emprunter in the phrase Pour pénétrer dans la maison le voleur a empruntē une fenêtre donnant sur la cour, literal- ly translated: "To break into the house, the thief borrowed the window facing the yard. ' It is obvious that the meaning of emprunter is "to make use of ; ' ' to borrow' contradicts the described situation. The presence of similar lexical categories does not result in non- translatability, because the given differential indicator is neutral- 1zed by the surrounding text: the disappearance of this indicator in the translation does not result in a loss of informat lon. This situation is confirmed by experience in the translation of idioms. There is the widely held opinion that the occurrence of idioms leads to nontranslatability. At the same time, idioms break down into – 87 - two groups from the standpoint of translatability: a) idioms including words whose semantic indicators do not contradict a situation, and b) idioms based on words whose meanings contradict the situation. As an example let us compare the following two idiomatic expressions: "Hey, look at what Fido can do, he 's barking at the elephant ' and "Hello, I am your aunt. ' The first of these expressions can be translated literal- ly, maintaining meter and rhythm (compare the German: El Schau, was das Möpschen kann, es bellt den Elephanten an , Since the original structure of this expression is sufficiently transparent and the comparis on there- fore does not contradict the situation. A literal translation of the second expression is impossible , s 1nce it results in nonsense. Thus we see that generally speaking only the Second group of full categories results in nontranslatability, and namely, when two differ- ent categories are compatible with a Single situation. This alone in- dicates the limited value Of the Sapir–Whorf hypothesis. This is not intended to say that the cited hypothesis can be rejected outright as baseless. Quite the opposite , 1t is important to note the importance of the facts which it established. The complete verification of this hypo- thesis is a matter for the future when, as a result of Semantic des – cription of the vocabulary, it will be possible to compare the vocabu- laries of various languages. At the same time, we immediately see two objections which, if they do not refute the Sapir–Whorf hypothesis, at least limit its significance. f As we have seen, the Sapir–Whorf hypothesis is a result of the fact that the categories of two languages are not comparable. This is true only in the absence of contact between the given languages which, generally Speaking, is possible , althCugh not characteri:3t 1C. Regular linguistic contacts and numerous translations of texts, as a rule , re- sult in the unification of semantic categories. It is indicative that - 88 - Whorf holds it possible to speak of the existence of a standard western European language, 1.e., about a uniform semantic system of the English, French and German languages. The appearance of this uniform system is , of course, explained by the disappearance of, or the substantial weaken — ing of , unique categorization of reality in languages included in this so-called standard European language as a result of linguistic contacts. In addition, let us take note of the fact that this standardization pertains exclusively to the full semantic categories, although within these limits certain phenomena differ substantially (for example , the forms of tense and mood). With regard to the nonfull semantic categor'- ies, they are , according to Whorf, also uniquely expressed in the lan- guages included in the uniform European standard. However, Whorf does not distinguish between semantically full and semantically empty cate – gories. It is therefore impossible not to agree with the idea that if one is to be consistent with Whorf's hypothesis, it must be admitted that each of the western European languages governs the unique nature of the thinking of that people and, consequently, complete mutuality of ... understanding between the Speakers of these languages is impossible. Thus, the Russian phrase “Manehbkhā Malibuhk Kataetch Ha romax ['The small sº is skating] in its complete rendering of all categories must be "translated" into English by the phrase He–Small he-boy drives himself on little he-horses (Barkhudarov, 1962, page 12). Generally, speaking, if we derine translation as the transformation in which there OCC Ullſ’S Il O loss, while all of the unique content and form of the original is re- tained (see Şll), we will have to admit that this transformation is ba- sically impossible, i.e. , to admit completely the validity of Whorf 'S hypothesis. If, however, we proceed from our definition (see $13), one will have to admit the possibility of the selection of such an inter- mediate language with which, as a rule , the semantically full Cat GEOr- – 89 – ies find expression. Of course, certain categories of languages cannot be reduced into a uniform semantic system and this occurs not only in such exotic languages as the language of the Hopi Indians of which Whorf Speaks, but in languages regularly in contact with each other. These categories, as a rule, pertain to the ordinary sphere of human activity and are situated at the periphery of the given language system. Given the above stipulations, having admitted nontranslatability, i. e. , the impossibility in certain cases to establish the correspondence called for by Diagram No. 4, we reject in principle the concept of non- interpretability which, apparently, is also assumed by the Sapir–Whorf hypothesis. By turning to the referent it is possible to establish cor- respondence for various articulations of reality and this is demonstra - ted not only by experience in translation, but by experience in lin- guistic contacts. Of course, there is one case widely known in the practice of trans- lation which, apparently, provides the basis not only of Statements on nontranslatability, but also of statements on "noninterpretability. "* We are dealing here with cases in which the form of the linguistic expres- sion becomes a significant element of the situation, i.e. , when we are dealing with a language-system function (a code function). Here non- translatable (and perhaps also noninterpretable) situations arise even on a base of empty categories. (A familiar example in the theory of translation is the example of Heine's poem "Ein Fichtenbaum, " Shcherba, 1923), not to speak of semantically full categories, whether grammat 1– cal or lexical. Apparently the Situations arising as a result of an idiom are also noninterpretable ; compare the well-known pranks of Till Eulenspiegel (known in German as Eulenspiegeleien). The foregoing per- tains to an even greater extent to poetic language (examples are well known). – 90 - . !! $16. ON THE PROBLEM OF "INTERPRETABILITY • \t, With a number of examples let us now demonstrate how in the 9V9% - the re — of various categorization of reality in two languages and with Ce SS sulting nontranslatability in the strict Sense of the word the pro tent b of interpretation is accomplished, i.e. > the rendering of Con y means of referring to reality. iX- A. The most obvious example involves words with subjective suff ond- es. The difficulty here lies not only in the fact that the corresp - tered ing suffixes in certain languages (Russian, Italian) are encoun n the very much more frequently than in other's (English, French), but i S a a various manners with Var*9% fact that these suffixes are comb ified iń Vari Cº. fixeS roots. For example, the Russian language has many enlarging suf 3. - c t , for exam - compare nomiſine. [huge housel, cunnina' [great forcel, etc. , but , e o a V c iO pl f'Or such Italian WordS a S 11 donn One ; la donnaccia; l'avarºa C. J J © re the there exist no direct correspondences retaining meaning; ComPº º woman]; Russian interpretations: 'Myxenonocha’ ” [a manlike 3. [a woman of the town l; ‘ckynell' [a miser). ith the same roots, ‘snaa xehnusha" - Even when corresponding suff”* * combined W uages pronounced divergences in the use of the Se words in the two languag, - e ar'e lein and others, in French -et , -ette. But, first of all, ther , rman Häus - considerably narrower than in Russian- Such words as the Ge et C . , a Fe chen, Tischlein, Tellerchen, the French maisonette , fillette , 5 3. ionally used, as a rule, in childrens literature, in fairytales, OC Caº 2 - nder the in familiar speech, not *9 mention the fact that they cannot re wealth of shadings as the various Russian SuffixeS. rd S Thus on the stylistic level the relationship between the WO espond even to Haus and Häuschen, maison and maisonette does not *** D - 91 - such a Simple relationship as the relation between the words ‘cron' [ta- ble 1 and 'erouak" [little table], and the word “Howar' [housel, as a rule, is translated simply by the German Haus, and by the French word maison, particularly in cases of the so-called "expressive" correspondence in Russian: ‘Manehb-khâ nomhk' [the little house ]; German das kleine Haus; French la petite maison. In addition, such series in the western languages, 3 S a rule, do not involve more than two or three words: Tisch, Tischlein, etc. It is impossible to render in German or French and, to an even greater extent, in English, the difference between ‘aomhk, Homok, Romhuiko' I'little house, a particularly little house, a tiny house ' ), etc. After all , the primary effect of these words is not the diminution itself (this can frequently be rendered through the use of adjectives), but the differing relationship of the speaker to the object. It is the translator's responsibility to render the relationship of the author or the hero to a given object as expressed by the suffix in the phrase, and perhaps even as expressed through the course of an extensive text. For example, let us take the translation of the K. Fedin novel "First Happiness," translated by G. Angarova. The heroine of the novel is called Anochka, but with a simple transliteration of the word (Anotschka) the German reader does not sense the particular relationship between the author and the heroine, a relationship which extends throughout the entire novel. And the translator seeks out the necessary words and techniques in order to express this relationship between the author and the heroine from the very first page. The novel begins with the Words: 'Ilebolka-Socohokka cuneia’ [A barefoot girl was sitting] , etc., which is ren- dered by the words: Ein kleine S barfüssiges Ding von etwa Zehn Jahren, etc. The expressively shaded word Ding is an excellent rendition of the relationship between the author and the heroine. B. The problem of the translation of Russian abstract nouns formed - 92 – by means of Suffixes ~ * , etc. , is close to the problem consid- ered above. We are dealing here with the fact that although al number of general comments may be made with regard to the method of rendering such words, each separate case requires reference to reality. However, it is important that this reference to reality - in certain cases - al- ways results in the possibility of finding a solution. The general prin- ciple, however, is based on the fact that correspondence in the target language is achieved through the introduction of certain new element S. * As an example let us take a Word having the suffix -ocmb. While the word 'sakonaccre' [legality] is translated into French as légalité, and into German as Gesetzlichkeit, such correspondences are rather rare Ca- ses, compare , for example, the word 'nºrmancı [party spirit ), which in German may be translated as Partellichkeit, while in French it is ren- dered as esprit de parti. It goes without saying that if such correspondences were firmly fixed in the vocabulary once and for all, it would not be necessary to speak of interpretation. The fact of the matter is , however, that this model is extremely productive in Russian and for this reason there f're - quently arise cases not covered by a dictionary. Let us take the word weaux, [1aeological content , i.e. , the ideological content of a work. Therefore, in certain contexts the Ger’— man translation Ideengehalt and the French conten; ideologique are pos- sible. In other contexts this Word is interpreted in German as Ideen- reichtum, in French as richesse d'idées ( 'wealth of ideas ' ) and inter- preted in German as Ideenre inheit [ 'purity of ideas ' ). In certain cases the word wieńhocts [moral fiber) can be translated by its Synonym ‘mphhuhnhanbhocte" [principled). However, there arises the question as to how to translate the Word ‘mphhuanhanbhocte" [principled 1. In the majority of cases it is interpreted in German as Prinziple: festigkeit or Prinzi- – 93 – pientreue, or in French as fidélité aux principes. The same approach could be used for the translation of words like “xobahnumha’ [the time of Chowanskil. In German this may be rendered as Chowanskimeuterei or -aufstand, or Chowanskizelt ['the Chowanski mutiny, 'the Chowanski period' ], etc. 'Kophanoammaa' can be translated as Korni- lowleute, Kornilowzeit, Kornilowmeuterei. ‘Kepenuka' can be translated as Kerenskizeit or Kerenskischmach (if the translator wishes to convey the emotional shading of the word). As can be seen from the cited examples a translator has no ready- made patterns for the translation of a given word of this type. Even when the correspondence seems to be uniquely defined, deviations are possible. Earlier We saw that the term snaprhāhocrbº is rendered in German by the word Parteilichkeit. It is also possible, however, to accomplish the translation on the principle of explanatory word formation: Partei- prinzip, Parteistandpunkt , Parteigeist. In one of Lenin's works the term ‘mapTuñhoctb' is used in an entirely special sense : "The system of parties in Russia is widespread and each party pre- sents a particular political front to the people." (V. I. Lenin, lith Edi- tion, Vol. 26, page 304). It is clear that none of the above correspondences satisfies the given case. The translator in this case found an extremely Interesting solution — the first clause was translated as follows: die Parteiungen in Russland sind sehr weit gediehen. The word die Parteiung is used very rarely and has the approximate meaning of 'division into parties , ' 'ad- herence by everyone to some party. ' Let us present yet another example. In Lenin's work "The Agrarian Problem in Russia" (Collected Works, 4th Edition, Vol. 15, pages 5–6) we find the following sentence : – 94 – "The ownership of land is spreading at an extraodrinarily rapid pace on a groundless basis. ' - The word ceceochoenocre ſtranslated above as 'groundless basis'] can be rendered only descriptively. Compare the interpretation of this Sen- tence : Der Bodenbesitz auf nichtständischer Grundlage nimmt ausserodent- lich rasch zu. C. Greater complexity is encountered in cases of finding lexical correspondence through interpretation by equating various cultural Sit- uations to one another (the authors of the series Stylistique comparee use the term equivalence in this case). * Russian candidate of science aspirant certificate of maturity French licencie bour'sier? d'études f baccalauréat English licentiate [M. A. postgraduate Student baccalaureate We frequently employ a similar procedure in the translation of id— iOmS ; German Ich will einen Besen fres- Sen, daf ... Wer das Wasser fürchtet, muſ nicht baden. Wenn meine Tante Răder hätte, wäre sie langst ein Omnibus.” Der Bauch läflt sich nicht mit Worten abspei- Sell. Wie man sich bettet, so liegt man. Unter den Blinden ist der Einăugige König. Blinder Eifer Schadet nur. Viele Köche verderben den Brei. Russian Tonoby maid Ha orceuehhe, * TO. . . - s Bonkob 6oatbcR – B mec he XO1HTB. Ecnh 6b, Ila kačb, Jia Bo pty pocnh rphôn, to Sto öhin 6b, he pot, a uenbuſh Oropoli. Coſioabă 6achām H. He Kop- M%T. - *H to moceeub, to H nox- HelLIb. ^ Ha 6espb.6bH H pak phlóa. * ..., ycnyx.nhabih nypak onac- hee Bpara. y cemn hähek Ahra 6es TJ1883. – 95 - English I'd give my right arm to. . . If you can't Stand the heat, get out of the kitchen. If you had Wings, "- you could fly. You can't pay the rent with promise S. As ye Sow , so shall ye reap. Among the blind, the one-eyed man is king. The road to Hell is paved with good in- tentions. Too many cooks Spoil the broth. • Liigen, daf sich die Bal- ken biegen. - * Wie der Ochs worm Berge stehen. Es wird nichts so heiß ge- gessen, wie es gekocht wird. Man Soll den Tag nicht vor dem Abend loben. - Wer A sagt, muß auch B Sagen. Mein Name ist Hase, ich weiß von nichts. Bpare, sax casus Mapua. WCrashtbºx, Haz 6apah Ha Hobbie popota. He tak crpauleh wept, Kak tro Mayº OIOT. - : 1) He roboph wrons, moka he nepecKouhuib. . . 2) LIHmnat no oceah cºn- Taior. . . . Bañnch saryx, he robops, HTO He AIOX. 3 he h, JIoIIlamb he moa, a He HSBO3HHK. Moń xata c Kpaio, a hauero He 3HäIO. To lie through one is teeth. To be as stubborn as a mule. The Devil is not as black as he is paint- ed. Don't yell "giddyap" before you're in the Saddle. Don't count your Chickens before they're hatched. You can't back out , Once you've begun. There 's nobody here but us chickens. Don't air your dirty linen in public. An interesting classification of pertinent cases was proposed by the American linguist Nida ("On Translation," pages 29–31). He des— cribes the following cases of interpretation: a ) the absence in the target language of a sign and corresponding referent, but the presence of an equivalent function accomplished by another referent . Example : '6eſocheikhhlā' [white snow ) is translated into a language in which the concept of 'Snow' is lacking (because the referent 'Snow' 1s lacking) as white as egret feathers. } b) the presence in the target language of a referent, but With an- Cther function. * EXample: The word “cepnue' [heart 1 is translated into the Kabbalaka lan — guage (Equatorial Africa) as 'liver, and into the Conob lan- guage (a form of the Mayan language in Guatemala) as 'abdom- inal cavity. ' c) the absence in the target language of an equivalent word and a – 96 - corresponding Situation. In this case it becomes necessary to resort to borrowing, in which case the borrowed word, as a rule , is accompan- ied by an explanation. The examples given by Nida for this case are not very illustrative. For this reason we refer the reader to the examples employed for this category of cases by Fedorov (page llllff). Manu- script page No. 59 6O 6 5 67 7 O [Footnotes Whorf discusses this problem in detail, page ll:5ff. Compare the definition of this concept in the series stylis- tique comparée : "A situation is a specific or an abstract re- ality described by a statement. In certain cases it is pre- cisely the situation which governs the translation by answer- ing the question "What is said in similar cases in the lan- guage into which the translation is being accomplished?" The situation also produces equivalance , e.g. , aller aux nues : Einen Bombenerfolg haben ºi. 1961, page ll). The story so far: Résumé des chapitres précèdents" (Vinay, Dar- belnet). Firth made an attempt in this direction: "We can cite such ordinary situations as : a) the address: Simpson ( ; Look here, Jones ; My dear boy ; Now, my man; Excuse me , madam. - $ Greetings, farewells, mutual recognition of social position and rank and entry into contact, determination of relationships after contact, breaking up of relationships, renewing relationships, changes in relationships. c) Situations in which words, frequently established by law or custom, are used to obligate a pers on to perform a specific act or to free him from certain unnecessary - respons i- bilities in order to impose others" (Firth, 93). - The quantity. H (a) has been designated as statistical or selec- tive information; a number of authors have also introduced the concept of "semantic information," but this last has not been developed, unfortunately, so that it could be applied to the theory of translation. IYa [MH) is the source language; PYa [IIH ) is the target lan- guage. Compare the comparable terms in English - Source lan- guage, target language; and the French – langue de départ, langue d'arrivée. Here, we cannot agree with the opinion of Cassierer (see "On Translation," page 291) who held that any translat lon is ºn interpretation (understanding the word interpretation to mean - 97 – To 77 77 79 8O 8O 8l4 85 85 90 95 95 precisely the reproduction of reality). The difference between a true translation and the reproduc- tion of the original is stressed by Malblanc who speaks di- rectly of two forms of translation (il y a deux especes de traductions, Malblanc, page 29). See A. A. Reformatskiy (1959) on the relationships bet concept and Subject. Firth has taken his example from Straumann, page 28. V.I. Lenin, Collected Works, 11th Edition, Vol. 14, page 92. V.I. Lenin, Collected Works, 4th Edition, Vol. 1, page 255. G. V. Shnittke drew the author's attention to this example. The relationship between the property of semantic completeness and the properties of the intermediate language is defined as follows: the intermediate language may be built So as to in- clude a greater or smaller number of concepts. The example of the translation of article titles dealing with chemistry into the information language is instructive. In Such a transla - tion the words describing the titled work: "Remarks on. . . ," "Investigation," "On the Problem of . . . " etc., as well as Words expressinº an evaluation of the covered material: "A New (Method)," "Simple," "Difficult," etc., are not transla - ted at all (Lakhuti, Stokova, pages 3-5). This classification in certain respects is close to the Clas – sification suggested in the work: Gougenheim, 1960, pages 3– l O. Compare the possibility of the mechanical selection of mood form in machine translation into Russian. (Nikolayeva, 1959). The term "interpretability" is introduced in analogy with the term "translatability" and has the same relation to the form- er as the term "interpretation" has to the term "translation." As far as it is possible to judge, Ya. I. Ret Sker use S the term "adequate replacement" in similar circumstances (compare, Retsker, 1962, page 47). is, Compare the French: Avec un "Si", on met trait Paris dans une b oute ille. This same case has been analyzed in detail by Fedorov (page l61) where there is an analysis of the translation of 'fia- cre as 'driver , " 'concierge ' as porter, ' and 'police commis- sioner' as 'policeman on duty, ' etc. Chapter 3 Analysis AND SYNTHESIS §17. BREAKING DOWN THE PRocess OF TRANSLATION INTO ITS TWO BASIC STAGES Diagram No. 4 shown in § 13 describes the process of translation in a slightly general form. For this reason we are reproducing its basic portion at this time and highlighting two stages which are extremely important for a study of translation. Let us now examine this process from the Standpoint of a transia — 6 Cºcºa Coadweave //7 a /27 | |nocaedraſar --T. 8 9 - *::::...” - A/A/3 | yearanoámezaw C////753 Coomdemºcadºw . Diagram No. 4a. 1) Communication in source language; 2) translator; 3) source-language system; l;) intermediate language; 5) target –language system; 6) communication in target language; 7) analysis; 8) transition in accordance with earlier established agreements; 9) synthesis. } *is º t or making the transition from the source-language system to the target – language System. His work breaks down into two parts: 1) he must 1so- late certain elements of communication in the target language and Set them into correspondence with the categories of the Source-language Sys – tem and then with the categories of the intermediate language (this pro- cess corresponds to the understanding of a certain text in a certain – 99 — language) and 2) he must make a transition from the given set of cate – gories in the intermediate language to the target language and construct some communication in the target language (this process corresponds, for example, to speech in some language). The distinction between analysis and synthesis is of great signi- ficance from the Standpoint of general linguistics. It develops that the problems and difficult leS which arise in the analysis of a text are rad- ically different from the problems of synthesis. Theoretical linguists generally fail to separate the process of translation into its compon- ent parts and therefore overlook these problems. The basic difference between hearing (analysis) and speaking (synthesis) has been noted only in recent times (Jakobson, 1956). At the same time, our remarkable lin- guist L. V. Shcherba (compare $8), granted in an entirely different context (in connection with the method of teaching foreign languages), ncted the basic difference in the structure of grammar as depending on whether or not it is directed toward the active mastery of a language ("synthesis" in our terminology) or the passive knowledge of a language ("analysis"). An explanation is called for here. We know that the term "grammar" is used in two senses: 1) the grammatical structure of a language and 2) the description of the grammatical structure of a language. However, we are not always aware that the resulting use of the word leads to the absolutization of the concepts of "grammar." From this it is but a Sin- gle step to the widely held conviction that only one "grammar," in the sense of describing the grammatical structure, can be regarded as S C i- entific. This explains the claim by various schools and trends in lin- guistics of having formulated the only true and uniformly scient iſ lo methods of describing a language. At the same time, the description of a language System wholly de- - 100 - pends on the practical purposes with which a researcher seeks to deal, and the scientific nature of the promulgated principles may be evalua- ted only from the standpoint of the validity of the posed problem and the degree of approximation to the intended goal. A serious argument against the breakdown of analysis and synthesis , the passive and active aspect, might be the following. Passive grammar presupposes a transition "from form to pure meaning," while active gram- mar presupposes the transition "from pure meaning to the form of expres- sion." Since it is impossible "to strip of meaning" or "for thought to exist without language," it is impossible for such a transition to take place. Here we must bear in mind the following. Language can be under- stood to refer to any system of signs which serves to establish thought , whereas "existence of thought without language" is , of course, impos- sible. However, it is the general practice to use the term "language" to refer to "a grouping of specific means of communication among people in a given country," i.e., a specific language (Russian, English, Ger- man, etc.). The existence of meaning (or of an elementary semantic unit) outside of such a specific language is a theoretically fully acceptable assumption. As a matter of fact, it may be assumed that meaning is eS - tablished in some artificial language in which mutually uniquely de- fined correspondence is established between that which is designatiné and that which is designated and where, in view of this last cºnsidera- ºn, it, is possible to speak Of "pure meaning." As an examble of this type of artificial language we can cite, for example, the intermediate language of which we spoke in §lo. Machine translation shows that the intermediate language is something more than 3 useful abstraction. Machine translation assumes the existence of some abstract lan- cºnse which is treated in the machine program. Such an "intermediate language" (independent of the number of specific layºuage; which can be - 101 - <- translated into this intermediate language) may be built as a system Of uniquely derined established abstract meanings or, if convenient, pure meanings (both grammatical and lexical). If we regard machine transla- tion as the modeling of Some process occurring in the teaching of a language to a human being and if there is general agreement as to the possibility of cybernetic analogies, our assumption as to the exis – tence of an analog of such an intermediate language in any translation is completely valid. In machine translation the presumption of the pre- sence of a system of "pure meanings" is now being subjected to experi- mental verification. Let us also note that it is precisely machine translation that has demonstrated the need for a divided approach to the problem of analyz— ing a text in which we actually proceed from form to the intermediate language , i.e. , to the function or the pure meaning (the passive grammar of L. V. Shcherba) and to the problem of text synthesis where we actual- ly proceed from the intermediate language, i.e. , from the function or pure meaning, to form (the active grammar of L. V. Shcherba). When we say that the process of communication breaks down into two stages, i.e. , analysis and synthesis, this does not indicate that there is no interrelationship and interaction between these , or what is more , that they represent equal elements in the act of communication. In the description of the act of communication, particularly when this COiſmun- 1éation is being accomplished by means of translation, the first step must be the examination of the synthesis (the outcome of the communica – tion). This is all the more useful, because analysis (recognition of the units of communication) can be described as a process which is the opposite of synthesis, i.e., as the process of reestablishing the meth- ods of its creation. – lo.2 - $18. MODELS FOR THE CREATION OF A TEXT In this section we would like to discuss briefly the basic ideas associated with the construction of models of the creation of a text. * The basic concept in the construction of such models is the concept of a sensible phrase or, as is said, one that is "marked," it being as - Sumed that this concept is understood intuitively and therefore not de - fined. It is also assumed that the set of intelligent phrases in the given language is given (Revzin, 1962, pages 60–61). For the theory of translation, sensible communications are those which are to be transla - ted. Thus, let the set of sensible phrases be given. The following im– portant question was then raised (Chomsky, page 422): what kind of sys- tem of rules must we have in order to: 1) obtain the largest number of Sensible phrases through the use of this system; 2) be incapable of producing a single nonsensible phrase with this system; 3) keep the sys – tem comparatively simple. In this connection there may arise the prob- lem of a comparison of various arbitrarily constructed systems with each other, and namely, if one system produces a greater number of Sensible phrases than another, it may be regarded as "stronger;" on the other hand, if one system contains fewer rules than another, it may be regard- ed as simpler. The construction of Such formal grammatical systems or grammatical models and their mutual comparis on is the subject of the remarkable W: Ork by N. Chomsky entitled "Syntactic Structures." Briefly, the idea of Chomsky's first model reduces to the following: ^. / t C & (, —-0 —- 0, Diagram No. 5 - 103 - Diagram No. 6 Let us have a certain mechanical device which can assume a finite number of states: 91, C2, . . . °n. One of these states, for example 90, is the initial State, and one of these States, for example C is the In 2 final state. Let the transition from SOIſle State Ci to some state 9k be 39°ompanied by the output of some word. Let all such pairs of states be fixed and known as connected States. Now, let the device pa SS through a number of connected states from CO through 0n. Any sequence of words ob- tained through such an Operation will be referred to a S 3. phrase. Each Set of phrases produced by a device of the type described here will be known as a language with a finite number of states, e.g. , the set of phrases the man 9. Ones and the men come can be achieved by the device, assuming the five states shown in Diagram No. 5. In the transition from Co (the initial state) to C1 an "empty word" is produced; in the transition from Cl to C2 the word the is produced; from C2 the device may make the transition to two different States 03 and C13 in the first case the word ſman is produced; in the second case the word men is produced; the transition from C3 to C5 results in the Word comes and the transition from Cl to 05 results in the word come. C5 is the terminal state. From this example we can draw the conclusion that a language with a finite number of states must Consist only of a finite number of phra - Ses. As a matter of fact this is not the case. After all, the device "lay return from several states to the same State, producing some word; - 104 – for example , let the device be capable of making the transition from C2 not only to C3 and Cl; but also to C2, producing the word old in this case. Let us represent this transition by an arrow in the diagram, pro- ceeding from C2 and again returning to C2 [see Diagram No. 6]. The language will then consist of an infinite set of phrases having the form : the man come S, the old man comes , the old. . . old. . . old. . . man COIſle S . . . The diagram seems rather simple and convenient , since the final set of States makes it possible to produce an infinite multiplicity of phrases. Nevertheless, it turned out that this diagram, on the one hand, exhibits an extremely limited explanatory capacity , i.e. , it does lit – tle to make comprehensible the process of phrase formation by a human being, while on the other hand, it can be replaced by a more general model of which a language with a finite number of States is a particu- lar case (Chomsky, pages 1122–430; Revzin, l962, page 139). Below we will consider a stronger formation process. However, we should first note the following important advantage of the subject diagram: it makes it possible to apply the apparatus of the theory of information to the analysis of language ($12). The theory of information has already considered diagrams similar to the diagram of the formation of a language with a finite number of states, with the only modification that, a certain probability *ik iS assigned to each transition C1 - C, (Ashby, Chapter 9). Let us now return to the process of formation referred to at the beginning of this section; we will add only the following transitions : - 105 — C1 - C6 yielding the word this Cl - or iſ these C6 - Clo | WOIſla (l 11 C7 - Co women Clo - Cº It COIſle S C9 - Ca . !! º C OIſle yielding the empty Word II - C3 - Co The over-all diagram will have the form: ,- ~"Nº- _2&^*. *ss e vº. Diagram No. 7. 1) Empty word. For the cases in which from the given state Ci we make the transi - tion' to the One and only One state 9k, the corresponding probability *ik = 1. For the remaining cases we can define the frequency by investi- gating in a rather lengthy text the number of times we encounter : after this the words woman and man / after these the words Women and ser: after the the words woman, women, man and men at the Tbeginning (after the T"empty word") the word this. : This analysis (granted, on limited material) demonstrated that ap- parently : P., (the). P., (this) > P., (these) ..ie (woman) = Pºs (man) - P.4 (men) ~ P.s (women) -- These results are obviously not accidental. We can state the hy– - 106 - pothesis that it in a language with a finite number of states two words belong to a Single grammatical class and to a single Semantic class , the conditional probability of appearance (after any given word) for cºe of these is approximately equal to the conditional probability of the appearance of the other. It is possible that in languages with a finite number of States the opposite is true , i.e. , from the fact that the conditional probabil - ity for the appearance of one word (after any given word) is approxi- mately equal to the conditional probability for the appearance of an — other word follows their membership in one grammatical class and one semantic class. This hypothesis is not refuted by the fact that the appearance probability for the word man after the word the is equal to the appear – ance probability for the word men, although they belong to various gram- matical classes. After all, after this the probability of appearance for the word man is some positive number, whereas the appearance probability for the word men is virtually equal to zero. The fact that there apparently exists a definite relationship be - tween the statistical and grammatical-semantic structure Of the text is very likely because of the following considerations. The theory of information demonstrates the following fact : the greater the freedom to connect individual elements of a code, the Smal- ler the excess (see §12). If a language were to be built So that it S eX – cess were equal to zero, each phrase in that language would be Sen Sible , and the diagram for the formation of that language should be construct - ed so that the transition from any state would be possible with ident 1– cal probability to all of the remaining states. Conversely, the greater the limitations imposed on the sequence of the elements, the greater the code excess. - 107 - The presence in a language of many various classes of words with the resulting complex structural formation of the language and the pre- sence of specifically linguistic grammar are apparently explained by the need for extensive excessive information. From the very conditions of speech intercourse language is inconceivable as an optimum code, Since it must exhibit extremely great resistance to interference. In this connection the following remark by Yngve is of interest : "Order and disorder complement each other to some extent. Statistical in- dependence assumes the absence of structure, and any deviation of ran- domness may be regarded as an indicator of structure" (Yngve, 1956, page 106). - . For similar investigations the diagram for the formation of a lan- guage with a finite number of States is extremely useful, Since it re- presents the simplest model of grammar. On the other hand, it is inter- esting to analyze the more complex formation models which better reflect the properties of real languages. Employing the method of analysis based on direct components (NP), * N. Chomsky constructed a formational process which corresponds to a much stronger than the process considered earlier (Revzin, 1962, page 139). Let us outline the basic ideas of this formational process, em- ploying the concept of configuration used in our literature. We will de – note as configuration: a) the sequence of grammatical classes corresponding to the skele- ton of the phrase , i.e. , the classes corresponding to the subject and the predicate (this configuration is known as the base configuration); b) the sequence of grammatical classes playing the same role as a single class (this configuration is known as the nonbase configuration). - 108 - The fact that we are not considering word sequences such as, for example, 'the man came, a sick man, very sick, came home,' etc., but, Sequences of corresponding classes, 1s associated with the need to re- duce the number of derivation rules. Thus, the sequence adverb + ad- jective masculine, singular, nominative can Subsequently be replaced by a very large number of combinations such as , for example: very good very bad extremely mediocre rather primitive , etc. The formation of a sensible phrase can be imagined in the follow — ing manner. We have the initial symbol Zo- This symbol is replaced by some base configuration. Then each of the elements of the configuration is in turn replaced by a configuration, etc. Then the rules replacing the symbols of classes by individual words take effect. Let us explain the above in the following example: ...Let Sl be a class corresponding to a plural noun in the nominative - case ; Vl be a class corresponding to an intransitive verb , plural, past tense ; Vo be a class corresponding to a transitive plural verb in the past : tense ; - : be a class corresponding to a plural adjective in the nominative Case ; - be a class corresponding to a feminine Singular adjective in the accusat, lve case ; be a class corresponding to a feminine Singular noun in the ac- - cusative case. Thus the formation of the phrase: "Soviet chess players won a great victory' can be represented by the - loS - following sequence of substitutions: 1) zo- SiVi 2) Si - A1S, 3) V1 → V2s2 l!) S 5) S 6) A 2 - A2S2 l -* ches splayers l - Soviet 7) V2 – won 8) S2 - victory 9) A2 - great - Let us note that in all of the Substitutions there is a Single Sym- bol in the left-hand part, While there may be several symbols in the right-hand part. Each series of Operations of the type X → y leads to a transformation of the original symbol, and this can be Written in the following manner: Zo SiVl (according to 1) A1S1 V1 - (according to 2) Al SiV2S2 (according to 3) Al ShV2A2S2 (according to 11) Soviet chess players V2A2S2 (according to 6) p - e e º e g © • © e e º º e g e e º º º e º e º e ! Soviet chess players won a great victory (according to 9). Chomsky referred to this series as derivation. We will refer to the symbols 40, A, S and V, etc., as metadenotations. The grouping of °oncluding lines given by rules of this type will be treated as the set of sensible phrases formulated by the given grammar. The set of sensible phrases, 1.e. , the concluding lines in the der- ivation was designated by Chomsky as the terminal language. - 110 - tolencrue alºner, *-aw wº- ** - . .. 3 l; ..., 5 . 1) Soviet; 2) chessplayers; 3) won; 1) a great; 5) victory, Formation in the form of a derivation sequence is equivalent to the process which is presented in the form of the so-called formation tree. Each substitution of a single element by several elements corres – ponds to a node from which several branches emanate. Thus, the cited sentence can be presented as being formulated in the manner shown in the diagram. The subject model is also dº interest from the cybernetic stand- point. On the basis of this model Yngve formulated the following inter- esting hypothesis (Yngve , 1960). pr - //oc/nog///ag . o .* Adaſame * * Memoodendo 3- fº/cm/ſočºczópouqaw | 66oãa. - ſazzº f a devāoãa - i. - ; : “a - * - t Aewaaouee ! . Woodcardo - b- - 1) Permanent memory; 2) input and output device; 3) high-speed memory; 11) solution device. Let us assume that man actually synthesizes phrases in the manner considered above. It is clear that he does not immediately produce a - lll - finished phrase, but constructs words in a sequence, one after the other, and he is not always certain how the phrase started with certain words will finish. Generally speaking, the phrase may be as long as desired (as an example we can cite the humorous English poem 'The House that Jack Built, with which we are familiar from the S. Marshak translation). On the other hand , the volume of man's high-speed memory, as demonstra- ted experimentally by the psychological research done by Miller (Miller, lg56) is extremely limited. Yngve proposes a model which takes this sit- uation into account, and namely, a machine whose circuit is presented on page lll. ..", In this case the permanent memory contains rules of the form Xi → Y }, the substitution operation is performed in the solution device , and the intermediate information which must be memorized to carry out certain operations is stored in the high-Speed memory. For example, to derive the combination : "won a great victory" on the basis dº the rules l) V1 → V2S2 2) we – won 3) S2 - A2S2 1) S2 – victory 5) A2 - a great and until the fourth, rule takes effect, the high-Speed memory must con- tinue the word victory which can be released (and erased from the high-Speed memory) only when the words 'a great ' have been derived. From this example it is clear that the high-speed memory must Store at least two words, Since otherwise it is impossible to accomplish the sequence of words that we require. Yngve demonstrates that the minimum capacity of the high-Speed mem- ory needed for the formation by the described device of any phrase in - 112 - the given language depends on whether we are dealing with so-called pro- gressive or regressive” constructions in the language. Yngve identifies progressive constructions as trees or subtrees (parts of trees) in which the branching proceeds toward the right, i.e º • 5 having the form: Yngve demonstrates that the indicated device, with a limited cap- acity for the high-speed memory, may produce phrases of any length if the trees of these phrases are progressive constructions. Yngve identifies regressive constructions as trees or subtrees in which the branching proceeds to the left , i.e. , having the form: It is precisely as a result of these constructions that greater requirements are imposed on the capacity of the high-speed memory. After these preliminary comments let us turn to a discussion of Yngve 's basic hypothesis. Yngve holds that : a) phrases actually used in the spoken language exhibit a regres- – 113 – sive structure of limited length; - b) in all languages there exist methods of limiting the regressive constructions, and in general for phrases with branched regressive structure Synonymic constructions of progressive structure are employed. Yngve 's hypothesis is presently being subjected to experimental verification. There are a number of serious objections to this hypothe- sis (in particular, those expressed by Chomsky and Liz, see Symposi- . um. . . ). Nevertheless, Yngve 's basic errort to relate the mechanism of text formation to the features of the high-speed memory are de Serving of the Eost serious attention. In particular, we shall see in §31 that hypothesis b) is of important significance for the understanding of phenomera encountered in oral translation. However, no matter how important the model of the direct components for the comprehension of linguistic phenomena, there exists, as has been develcpe i, a number of phrases which cannot be explained on the basis of this model (Chomsky, Revzin, 1962). As an example let us present a number of German phrases : l. 3estern baute er da S Haus. Yesterday he built the house. 2. Wie hat er das Haus gebaut'? How did he build the house? 3. Das Haus wird von ihm gebaut. The house is being built by him. Tries e phrases cannot be produced by the method of direct C Cmponent S for the following reasons: a) With formation by this method any phrase must be divided into two continuous part S (corresponding to the subject and the predicate), and this does not occur in our phrase S ; b) Intuitively it may be said that all three phrases are derived from the phrase er baut das Haus, which, in turn, can be derived, for example, from the configuration: subject-predicate (er baut). However , it is easy to prove that there is no such transformation for which phrases 1–3 are derived from the phrase er baut das Haus by - lll: - *g means of replacing only a single element of the configuration (without changing the sequence of the other elements). c) In the formation of phrases by the configuration method each Substitution Xi - Y, was independent of the other substitutions carried out earlier. However, in this case the substitution of baut by wird ge- baut is associated with the Substitution of er by von ihm. In this connection, Chomsky develops a certain generalization of the theory of formation on the basis of direct components, and namely , he assumes that in addition with the phrases obtained from the base con- figurations (these might be referred to as the nucleus of language), there exists an additional set of phrases obtained from the phrases of the language by applying certain transformations. The formati on process can now be represented as the expansion of configurations and the substitut lon of abstract symbols by the words of a language, or as in the previous model, with the subsequent transforma – tion of the derived phrases. Thus, there exists a grouping of phrases of the nucleus, obtained by the configuration method, and a grouping of phrases obtained from the phrases of the nucleus (or some other phrases), obtained by trans – formation. We will demonstrate, for example, that phrases l–3 may be obtained from the phrase er baute das Haus by a transformation of the form des– p cribed above. Let us introduce the following transformations: Transformation l. er baute gestern — gestern baute er (this type of transformation is the simplest, since here all elements remain the same , i.e. , this might be referred to as a "transposition transformation"). Transformation 2. er hat — wie hat er (this type of "transposition transformation" may be referred to as "in- - 115 - terrogative transformation"). Transformation 3. baute das Haus — hat das Haus gebaut (this is not the simple substitution of the element baut by hat gebaut, of the type which were carried out in the previous section, but the original transformation, Since here , first of all, a certain transposi- tion takes place and , what is the most important , it is not some arbi- trary element of Some class Such as , for example , geschrieben, gemacht , etc. , which appears , but that element which render's the same meaning. Such transformations may be referred to as "morphological"). Transformation H. er baut das Haus -- das Haus wird von ihm gebaut. Analogously, in Russian : oh ctport HOM->nom crportcº Hū ["he is building the house — the house is be – ing built by him' This is "passive transformation." Let us now demonstrate how by combining the methods of configura - tive formation and transformational modifications we can obtain our phrases. From the configurational modifications we will need the fol- lowing three : Ph l . . Pha ging – baute das Hause [went — built the house ] Ph3 baute — baute gestern [built — built yesterday ). These modifications form the following phrases Of the nucleus : / ! Z — er ging er ging [he went •. er baute das Haus [he built the house ] er baute gestern das Haus [he built the house yesterday ). Paraphrases 1–3 are obtained by the application of the follow ing modifications : lst phrase: Phl, Ph2, Ph3, Tl 2nd phrase: Phl, Ph2, T3, T2 3rd phrase: Phl, Ph2, Tl The transformations may be extremely varied in form. These may be – 116 - the transformations changing: a) a portion of a phrase into a portion of a phrase, as for exam- ple, T2, T3 b) a phrase into a phrase, as for example, Tl| c) a phrase into a part of a phrase. Until now We did not consider any transformations of this last kind. At the same time, they play an extremely important role in lan- guage . l (a 4our croounce w == mocmooeve &aar (*) 2. - - - cmpouwerd war &&. 3 . > l; o, compoum &ow * l) Compare: 'the house is being built by him; 2) the building of the house % him); 3) the house being built by him; l;) he is building the building the house; 6) the house being built; 7) to build a he who is building the house , etc. house ; 5 house; 8 The majority of the sentences in our language contain Certain transformations such as , for example, the sentence taken at random (page 615) from the "Brief Course in Mathematical Analysis" by A. Ya. Khinchin : f - "In the development of a rigorous the ory of infinite series Abel } achieved fundamental results simultaneously with Cauchy. ' This sentence has been obtained from the following: l) Some one devised a rigorous theory of infinite series 2) Cauchy obtained fundamental results 3) Abel obtained fundamental results, i.e., in the construction of this sentence at least two transformations were nvolved: — ll'7 – a) someone devised a theory — formulating a theory b) Abel obtained results – results were obtained by Abel and two rules of conjunction which must also be treated as a transforma- tion: - - c) when combining 1), 2) and 3) we have the rule: formulation – -> in the formulation d) with coincident predicates in 2) and 3) we have the rule: Cau- chy did something -- simultaneously with Cauchy. The introduction into the sentence of homogeneous terms my be ac- complished by means of the algorithm proposed by Ye.V. Paducheva (Padu- cheva, pages 17–18). Such an algorithm can be constructed in the follº lowing manner. The algorithm must compare synthetic trees of two com- pleted sentences and if it turns out that certain portions of these trees are occupied by identical words, (i.e. , corresponding sequences coincide in all respects with the exception of the Sequential number of the words in the sentence), these two trees can be "grafted together," i. e. , reject that portion of the second sentence which here is identical With the first sentence, and introduce the remaining portion into the first sentence by means of a conjunction. As a result of such an Opera - tion it 1s possible, for example, from two sentences such as : The Straight line A intersects the straight line B, and the Straight line A * intersects the straight line Y To obtain a sentence having the uniform terms: The Straight line A intersects the straight line B and the Straight line Y. The trees may be grafted "from above," as in the previous case, where two composed sentences have identical subjects and predicates, as well as from "below," as in the sentence: - ll& – The straight line A and the straight line B intersect the straight line Y, where the tops of the trees are different and the lower por- tions of the trees, beginning with the predicate , coincide. $19. TRANSFORMATION AS INTRALINGUAL TRANSLATION The abstract definition of transformation given in §18 reduces to the fact that with transformation a number of rules is given for the replacement of certain phrases by others, in which case it is assumed that meaning remains unchanged. For us it is precisely this last fact that is significant. For this reason we will define transformation as the translation within a language or , more exactly, that particular case of translation in accordance with Diagram No. 14 in which the source lan- guage coincides with the target language (see the concept of intralin- gual translation in §6). The difference between transformation and in – terlingual translation, by the way, is found in the fact that in the case of transformation we do not encounter the various categorizations of reality. Of particular interest is the classification of transforma- tions proposed by L. Tesniere. * Tesniere based his classification on that part of speech which he denotes with symbols corresponding to the endings of the parts of speech in Esperanto: O (noun), A (adjective), I (verb), and E (adverb). Here he distinguishes two basic types of transformations: the first includes the transformations of words With full meanings – of a term in the sentence into the word of another CatC – gory (transformation of the first degree); the second transformation in – volves the entire sentence (transformation of the second degree). With- in each of these types we distinguish transformations that are simple, dual, triple , etc. , depending on the number of transformations to which the base construction is subjected. Transformations are further clas Si- fied on the basis of the nature of the resulting (rather than the ori- ginal) category. Consequently, transformations are examined in a noun , - 119 — an adjective, an adverb and a Verb . Tesniere includes among the simple transformations of the first issree : the transformation of an adjective, adverb or verb into a noun; the transformation of a houn, adverb and verb into an adjective; the transformation of a houn, adjective and verb into an adverb; the trans – f Jr'ſ at ion of a noun, adjective and adverb into a verb. Dual, triple 2 - - C - , transformations of the first degree are described in a Similar £ Farrier. Thus, according to Tesniere, the dual transformation of the -º- first degree in accordance W1th the formula I S. O > E represents the tºansformation of the personal Verb form into a verbal noun (I X O) and fººther into the adverb (O > E) as, for example: Compare : Je vous défends de Sortir, Bernard cherche a COmprendre. Here, in accordance with the theory of Tesniere, the indefinite ** 1s the verb changed into a noun (first transformation), and this - = 5: by means of the preposition gie is converted into a functional equiv – a-erº of an adverb, playing the same role as , for example, rigoureuse- :=:: (compare je vous défends rigoureusement). The same applies to the fºr a comprendre. The facts of translation, apparently, C Onfirm the † = - i. 3 it y of this approach, since, for example, in the translation into ****r of similar constructions there appears the Original verbal noun - > 2), with the formation of a prepositional group frequently play 1ng *::= adverbial role (O > E). Compare : Il se décida a traiter. Er entschloss Sich zur Unterhand- - +5. [He decided on the deal]. The transformation (I S o S E) is characteristic Of the Russian is guage. Compare: "from Something to do," obtained by the transforma – - * ºr of the infinitive into an adverb by the preposition from (analo- £2 25 to the French preposition 'de 1). The transformation of the second degree (denoted >>) describes the - 120 – modification of an independent clause into a subordinate clause. This includes the transformations I 2 o (Je dis + Alfred a raison > je dis qu' Alfred a raison, where the subordinate clause functions as a noun). I >> A (les livres que vous avez, where the subordinate clause func- tions as an adjective), I XX E (je vous ai recu + vous étes arrives —- → je vous ai regu quand vous étes arrivés, where the subordinate clause functions as an adverb. It is easy to see that these transformations describe an auxiliary, attributive and adverbial clause. A Second-degree transformation, just as a first-degree transforma – tion, may be of the multiple kind , i.e. , consist of several verbs. Thus, the compound Subordinate clause J 'approuve ce que vous faites is des — cribed as a dual transformation: vous faites — que vous faites (I X A)+ + que vous faites >> ce que vous faites (A >> 0). Tesniere refers to certain transformations as elliptical. Thus, the dual transformation of the second degree in accordance with the formula I XX E >> O in German may correspond in French to the elliptical transformation I XX O. Com- pare, Lehren, wie man worsichtig Sein muss with enseigner la prudence (compare the concept of modulation in $20). The theory of transformations makes it possible to resolve the problem of constructive homonymy. If one and the same sequence of sym- bols is obtained in two different transformations , we will say that this sequence represents a constructive homonym. AS an example let uS * present a case of coincidence between the so-called subjective and ob- jective genitive case with an abstract verbal noun. If we take the grouping of words die Entdekkung dieses Laboratoriums [the discovery of this laboratoryl, it can be explained as arising as a result of the transformation of the phrase: Das Laboratorium ent deckt et was — die Entdeckung des Laboratoriums | the laboratory discovers something' – 'the 'i'i scovery of the lab - - 12l - oratory' or as a result of the transformation of the phrase: Man entoleckt das Laboratorium - die Entdeckung des Laboratoriums ['Someone discovers the laboratory — the discovery of the laboratory). In the terminology of transformational formation it is possible for us to discuss a problem which has a long history and plays an im- portant role in the theory of translation and grammatical stylistics. We are speaking of the case in which two various grammatical construc – tions such as, f Or example, in the German the participial and preposi- tional constructions with verbal nouns having the same meaning. 1) Näher tretend, hêrte er Klänge einer Blechmusik (Bredel", "Rel- atives and Acquaintances"). 2) "Aha," sagte er im Nähertreten (Mann, "The Loyal subject"). By the way, this is borne out by the translation: 1) Stepping closer, he heard the sounds of a brass band. 2) "Aha!," he said as he moved closer. Various definitions were provided for the phenomenon of grammati- cal synonymy; apparently, the most suitable of these is the definition provided by Bally who regarded grammatical synonymy as a special case of substitution (supplétion): "The substitution signs have precisely the same meaning, but various indicators" (Bally, pages 196–198). Cit- ing examples of substitutional signs, Bally, by the way, says: "Let us, by the way, take note of the fact that the gerund changes the verb into a noun: Il lit en Se promenant into Il lit pendant sa promenade ['He is reading as he walks" into "He is reading during a walk' ]. However , all attempts to define this phenomenon encountered insur - mountable difficulties associated with the lack of clarity in the ex- pression "having the same meaning" or "having similar meaning." We will refer to two constructions as grammatically synonymous if they : - l?2 - a) evolve from one and the same initial node on the tree of phrase formation or, in other words, if they function in the same grammatical role ; - - b) are included in phrases derived, one from the other, by trans- formation. Let us now turn to the definition of the concept of transformation in §18. On the basis ar this definition we can make the one significant statement : - In the phrases of the nucleus of some language it is impossible to find two phrases containing grammatical synonyms. This is obvious , Since the phrases of the nucleus are formed without resort to any transforma- tion. Thus there are at least two sublanguages, one of which, and namely the nucleus, is an idealization of the concept "neutral style," consid- ered in stylistics. With regard to the second sublanguage, this may be regarded as an entire grouping of sublanguages , with each of the Sub- languages capable of characterization by those transformations by means of which the phrases of this sublanguage have been obtained. Thus it be - comes possible for us to obtain objective characteristics of that which is generally referred to as "functional style." Indeed, transformations of the type : Er legt etwas zugrunde—-Zugrundelegung. | "He is laying the foundation for something' – 'laying the founda- tion' ] - Erhält die Prinzipien ein – Einhaltung der Prinzipien: [ He observes the principles' → 'observation of principles' ] are characteristic of commercial style, while transformations of the type : Er legt etwas zugrunde — etwas zugrundelegend [ He is laying the foundation for something' — laying the founda- tion for something') - 123 - a hit die primipien en-die Principien “inhº" I He iS observing the principles' — observing the principles' ] are used for literary descriptions • At the same time, in ordinary speech neither form of transforma- tion is suitable. Of course, the separation of the sublanguages by types of suitable - 3. * t º methods discussed in §18 may serve at least as an auxiliary tool of Sty listics. § 20 SEMANTICS IN THE FORMATION MODEL. THE CONCEPT OF MODULATION The formation models examined in §18 exhibit a single significant e l drawback from the standpoint of the theory of translation : they fal eS either entirely to consider the lexical content of the formed phras Ullſlo - or these are considered only indirectly. This corresponds to the a58 p - O - tion that unlike grammatical structures which are developed in the prº cess of formation from the simpler structures , the lexical content 15 always given in some manner and is not subject to modifications during • * I} - the course of which units that are simpler from the standpoint Of mea the ing are replaced by more complex units. It is of course clear that - a"; rºſna - process of translation can be described in accordance with this f'O l tion, model (in the majority of textS On translation treating grammatica vº “, type 1S modifications and lexical content separately , something of this by P assumed in essence). In this case the intermediate language must be rºe S - built as a system of correspondences between entire words and eXp * S 3 S - sions (analogous to a two-language dictionary 9* * phrase book, a sumed in the texts of the above type. his The heart of the matter's however, lies in the fact that in t f'in- case the number of correspondences become 3 extremely great (even in - 124 - ite, if it is assumed that the wocabulary can always be enlarged with new words). If it were possible to isolate a small number of elementary Semantic units which could be combined to give the meaning of each word, the problem of establishing correspondences would be significantly simp- ler. Apparently, it is precisely such a description which best of all Simulates the activity of a human being. during the course of transla - tion. This consideration makes it possible more clearly to formulate the problem of which units are best regarded as the initial units and how to construct the intermediate language in which these units are Written. Certain thoughts in this connection were expressed by V. V. Ivanov (Ivanov, 1957, page 56). These remarks reduce to the following. The en- tire set of words is divided into a series of "semantic fields" such as, for example, a field of words denoting the doer of deeds , a field of words denoting the mental capacities of people, etc. "The meaning of each lexical unit may be regarded as a Set consisting of a countable number "of sems" (or "semantic differential indicators"). The selection of the "sems" of which the meaning of the given lexical unit consists iS defined by the number of semantic fields in which this unit is found" (Ibid). As an example let us take a group of Russian words describing the relationships of possession (compare: 'to give , ' ' to present, ' - 'to will, ' ' to exchange , ' 'to sell, 'to acquire, ' ' to inherit, ' , "to receive, to take away, ' ' to deprive, ' ' to appropriate, etc.). Each Of these words can be represented as a complex syntactical formation de S - cribing a situation in which two parties (A, B) interact and in which we have a certain object (O). The meaning of the word 'to have ' is ele- mentary for all of the words in this group. This is not intended to Say that the word ' to have cannot be broken down into more elementary Se- mantic units: let it suffice to say that it describes the situation in – 125 - which the subject (A) and the object (O) are participating and between whom there exists a certain relationship of being (compare : "the city has a library with 'in the city there is a library " or with 'the li- brary of the city," etc.). However, it is precisely that fact that the verb 'to have describes the situation in which only 2 elements are participating provides the basis for the application of the meaning of this word in its elementary form for the ssRantic description of these last words. Thus the word 'to give ' will be represented as consisting of the following semantic units: 1) to cause ('to do so that. . . ; French, faire ; English, make; German, lassen); 2) to have , i.e., by means of this word we describe a situation that is characterized by the fact that. A causes the possession of o f'Orº B (compare: A gives the book to B). The word 'to present ' will represented as follows: 1) to cause; 2) to have; 3) gratis. Since to cause + to have — to give, the word "to present " can be represented as to give + gratis. We obtain the follow – ing approximate table of formation for certain words denoting posses – sive relationships: - l) to cause + to have — to give 2) to give + gratis — to present 3) to give + in exchange for money — to sell It is easy to see that the words cited here are defined from the point of view of A causing the posse S Sion : (compare the French faire avoir I to give and faire don [to present ' ' ). However, since we are speaking of a two-sided relationship , we can also present the COI"I'e S - ponding situation from the point of view of B. We have only to Sub St 1– tute the positions of the acting Sides , i.e. , in the place of A to cause + to have O for B (i.e. , 'to give ') to write : B to cause + to have O for A, thus denoting the word 'to receive. " For Relationship (2) we obtain the word 'to receive as a present ' and for Relationship (3) we, obtain the word 'to buy. ' It is easy to see that the relationships - 126 - - expressed in this manner are similar to the verbal relationships: from the semantic point of view (it is precisely this approach that is of primary interest for the theory dº translation) there is no basic dif- ference between the expression of sense by means of lexical or nonlexi- cal means. Compare the English B is given a book by A with B receives a book from A i.e., in Russian, B receives a book from A or A gives a book to B). The transitive-reflexive relationships differ from these relation — Ships semantically: 'to take ' and 'to give ' mean to cause possession, but to give means that A causes possession of O for B and 'to take ' means for B to cause possession of O for B. A fully possible situation 1S One in which A gives B a certain object, but that B does not take this object. Let us now illustrate how we can describe the meanings Cºf the Russian words: 'to have, ' ' to give , ' ' to deprive, ' ' to take , ' ' to lose, ' ' to pick up, ' ' to buy, ' ' to appropriate, ' ' to sell, ' ' to pre- sent " by means of four elementary indicators: l) possession, 2) CauSe , . 3) reflexivity (for oneself), l!) payment and by means of a single logical operation - negation; (each of the ci- ted indicators can be negated). We can imagine how the meanings of the corresponding words 3. I’é formed from simpler words. We illustrate this process by means of trees, agreeing to show the positive meanings of the corresponding indicator (+) on the left side, and to show the negative meanings (–) on the right Side. In this diagram the two different meanings are neutralized : to cause possession (independent of the object) equals 'to given' and to cause possession not for oneself = 'to give2, since in Russian the same - 127 – 3 e . 22 & - cºda Ayma/me 'ſº mºſºme anywdawe 20 ñº 23- * * * * * * = ame * = <= * * =ms º ºsmº, smºs, cº, 3* savvenue wº tº º * - •º * sº º sº 44 savémae - º aº l) to exist; }} 1st meaning; 3) words; 4) to have; 5) not to have; 6) 2nd meaning; 7) to give ; s not to give; 9) to deprive; 10) not to de'. prive; ll) 3rd meaning; 12) to take; i5) to lose ; 14) to pick up; 15) 4th meaning;T6) to sell; iT) to present; 18) to redéem; 19) to expro- priate; 29) to buy ; 21) to appropriate; 22) to give way" (tô'seil); 23) to alienate. Word 'to give ' is used for both cases. In French faire avoir is used in the meaning of to give1, While donner is used in the sense of to give2. ' In our diagram we used only a single logical Operation, and namely , negation. It is of course understood that other operations may be used, e.g., conjunction which forms more Complex constructions from the indi- cators defined above, these being connected by the conjunction (and. t * Thus We Can define : to exchange means for A to cause possession for B, and for B to cause possession for A ( 'to give " + "to take 1); to Steal means for A to cause nonpossession for B and for A to cause possession for himself without payment (to pick up + to appropriate); to borrow means for A to cause possession for B+ in order for B to Cause possession of the same thing for A (to give in order to receive - 128 – in return). An analogous process of formation might be constructed for the case in which the initial meaning indicated the possession not of a ma – terial object but of certain bits of knowledge (in this case "to givea' would correspond to 'to teach, ' which in German is rendered by lehren, in French by enseigner, in Rumanian by a preda; " to take " would cor- respond to 'to learn, ' and would be rendered in German by lernen, in Rumanian by Tnvaya, etc. , while 'to steal ' would correspond to 'to pla– giarize '). - It is clear that each of the described meanings can be modified by the addition of new indicators which are characteristic not for the en- tire group, but only of a single word. Thus through the addition of the indicator 'posthumously " to the group of indicators for the word 'to pre- sent it would be possible to obtain the word 'to will, ' etc. This process of formation might be referred to as lexical trans- formation. However, from the standpoint of terminology it would be more convenient to separate semantic concepts from the grammatical. We Will refer to the process of formation in which the transition from one word to another is accomplished with retention of the basic meaning as modu - lation. * Thus, the transition from the word 'to deprive ' to the word 'to take away ' will be modulation. A particular case of modulation is identical modulation in which the entire grouping of elementary meanings is retained in the transition from Word to Word. Thus, from the word ' to present " it is possible to form not only the word "to will, but the lexeme 'to leave in a will which is identical in meaning. Thus that which is known as Synonymy is a particular case of modulation, or more exactly, Stylistic Synonymy. This circumstance impose S yet another requirement on the nature of the Intermediate language. In the formation of a word not only the elemen- - 129 - tary meanings of the word must be given, but the set of characteristics indicating its utilization in the given functional style, i.e., the set of Stylistic characteristics. This process can be presented as the addi- ... #ion of the characteristic (familiar, colloquial), etc., to the set of slementary meanings. Thus, from the grouping of indicators "to take away' + 'to appropriate, ' we can obtain not only "to steal, but to #ull down, ' 'to carry away, ' 'to take away (a machine) etc. º We have demonstrated how words close to one another in meaning are analyzed (or, as we will say later on, relate to a single "semantic series"). Let us note that such groups are easily isolated if we use the follow ing system of definitions: A) Let us refer to the words x- and y as semisynonymous, if there exist contexts (in some target language) in which they are translated by a single word (homonyms are treated as different words). B) Let us refer to the words x and y as semiantonymous, if x and "y plus negation" or y and "x plus negation" are semisynonymous. C) We will say that x and y belong to one semantic series if it is that l) possible to construct a sequence of words Xo, x1, x • , X 2 ” tº º *i-l and xi are semisynonymous or semiantonymous (O < 1 < n) and 2) * n° xO = x and xn = y. It is easy to prove that the semantic series do not intersect. The concept Of a Semantic Series is interesting because of the fact that it always contains the possibility of modulation (it is possible to refine the concept of modulation so that modulation in turn not exceed the lim- its of the semantic series). The importance of this concept has alway S been recognized intuitively in the theory of translation. Unfortunately , the principles of separating that which is known as a "synonymic ser– ies," "translational synonyms" and "antonymous translation" have not been sufficiently clarified. - - 130 - We can see from the definition of modulation that this concept al- SO includes the formation of Words generally referred to as ideographic synonyms, i.e. , words which differ not only in their suitability in a given sublanguage, but in the number of elementary meanings. Thus, for example, 'to appropriate' and 'to buy' differ only in a single elemen- tary meaning (the fact of payment). As in the case of phrase formation, the base constructions making up the nucleus of a language played a basic role, but it is also possi- ble to regard the words which serve as the base for the formation of an entire group or words as the lexical nucleus of a language. Thus , for a group of words denoting 3. possessive relationship , the word 'to have ' will be the nucleus. - Thus the term modulation is called upon to denote the phenomenon long since intuitively recognized in the theory of translation and fre- quently referred to as translational Synonymy. In works on the the ory of literary translation such cases are gen- erally described in an examination of methods for the attainment of an integral translation. In particular, it is indicated that it is possi- ble to concretize undifferentiated and abstract concepts. For example, the sentence: His Lordship jumps into a cab , and goes to the railroad is translated as follows: Jopi Kºło opkhyn B H3B03HWHElo kapety H Hpakasai beath ceda ha xeneshyo Iſopory’ ['Lord Q hopped into a cab and Ordered that he be taken to the railroad ' ) (Retsker, 1950, cases re-17). In connection with the problem of breaking down words into their semantic multipliers we must take note of the following. In translation, in addition to general words of the type considered above , we always are dealing with words of terminological significance to Some extent. It is important to distinguish the terminological meaning of a term, and, in the words of A. M. Terpigorev, its "literal meaning." – 131 – The terminological meaning of the term is its relation to defini- tion. However, the "literal" meaning of a term corresponds to the set of Semantic multipliers isolated for the given word in the given lan- guage . As Terpigorev indicates, the literal meaning of a term may cor- respond, not correspond or contradict its terminological meaning (Terpi- gorev, 1953, page 73). For example, the Russian term "liquid corrosion' is used for the concept defined as the "corrosion of metals in a liquid medium" (Terpigorev, 1953, page 74). The literal meaning of the Russian term corresponds to its terminological meaning. In examining the analo- gous German term feuchte Korrosion we find only partial correspondence with the literal meaning of the word and its terminological meaning: the definition of feuchte in this combination means 'moist, ' 'raw, ' and not "liquid. ' Let us present another example : The German word Feldstärke has the component part denoting the literal concept of 'force of field. ' However, since the terminological meaning of this word is defined by the content of its denoted technical concept , a specialist in electronics will translate the word Feldstärke into Russian by the term 'field intensity. ' Consequently, in evaluating and translating a term it is extremely important to take into considera– t 1 on the relationship between its literal meaning and its terminolog1 – cal meaning (Bozhno, page, 11). Fr Om the standpoint of the theory of translations terminology can be defined in one of the following three way S : a ) as words which cannot be modulated. b) as words whose only transformation is the substitution of the term by its definition or by replacement with an absolute §ynor lym. c) as words whose translation is accomplished not through an or— dinary intermediate language, but through an intermediate language - l32 – which is the language of the given scientific discipline (independent of whether or not this language has or has not been formalized). The third definition is apparently the most important, since it is from this definition that the first two follow. It is also important for the determination of the specifics of that which is generally under- stood when referring to scientific-technical translation. The specifics of Scientific-technical translation are found precisely in reference to a particular intermediate language. It is not in vain, therefore , that all writers on the subject of scientific-technical translation speak of the need to be competent in the branch of knowledge in question. * The cited considerations pertain in part to words of the general language which denote specific objects (as we stated earlier, the lan- guage of things in the given case becomes the intermediate language). However, here the situation is facilitated by the fact that in the des — ignation of specific objects there occurs generally an identical cate- gorization of reality. Of course, in cases of the type presented in §ll it is a good idea to have smaller units in the intermediate language, e.g. , individual units corresponding in the French to main, in German to Hand, in English to hand and in French to bras, in German to Arm, in English to arm, etc. However, it is necessary to bear in mind that the same set of in- dicators (in the particular case of specific objects – a single indica- tor) can be used both for the denotation of an abstract concept as well as for the denotation of a specific object. Therefore they must be a1- vided in the intermediate language, e.g. , the elementary meaning Cor- responding to the word 'heart should be represented as having two ele- mentary meanings: "heart 1" (specific object) and 'heart 2' (abstract in- dicator). This is necessary, in particular, for the proper translation of Russian adjectives. - 133 – In examining the problems as Sociated with the translation of Rus- Sian adjectives, we find that the very nature Of the adjective is of decisive importance, Since depending on whether it is qualitative, i.e. , it corresponds in the intermediate language to the abstract sense, or relative , i.e. , it corresponds in the intermediate language to the spe – cific sense, the nature of the correspondence in the target language may also change. - - 1 As an example let us consider the adjective 'cardiac " or "heart - felt " [Translator's Note: given by the same word in Russian]. Let us take a combination Such as 'cardiac activity, 'cardiac [heart l disease on the one hand, and heartfelt reception " on the oth - er hand. If this word in the analysis of both combination groups were set into correspondence with only a Single unit of the intermediate language, incorrect translations would result from the synthesis. For example, if this unit in the intermediate language were to correspond to the French word cordial, to the English word cordial, or to the German word herz- lich, in addition to the meaningful combinations in German of herzlicher Empfang, in French of accueil cordial, and in English of cordial recep- tion, there would be formed the senseless combinations: herzliche Krank- helt, maladie cordiale, etc. One might say that sense here depends simply on the capacity for - the adjective herzlich, cordial, etc. , to combine with other words. This approach is possible, but hardly feasible , since there prevails the general qualitative relationship here that is suitable not only to this example, but to the translation of an entire series of similar ad- jectives. Let us take another example — the word ‘uapckhā' ['the Tsar's, tsar- ist ' ) and let us examine such combinations as 'uapekaa moub’, ‘uapckan ceMba', - 134 - on the one hand and ‘uapckań Pocchā’, ‘uapckhâ manau', ‘Hapckuh pexhm' on the other hand. If we do not isolate the elementary meanings of "tsara.' and "tsara, we will again obtain a number of senseless combinations. Cor- rect translation is achieved only by the isolation of the two different elementary meanings, because ‘uapckan moub’ is translated into French as la fille du tsar, into English as the tsar is daughter, and into German as die Zarentochter; on the other hand, the combination ‘uapckaa Poccua' (tsar- ist Russia] is translated into French as la Russie tsariste, and into German as das zaristische Russland. Let us now consider an adjective which is frequently encountered in translation. We have reference to the adjective ‘Maphuā' ('peaceful Orº universal ' ). For example, we frequently encounter such combinations as etc. Peace treaty' refers tº O an peace treaty, ' peaceful labor, agreement on peace. ' Here the adjective has relative meaning. We will in translate it as Friedens vertrag. It is clear that it is impossible this case to say friedlicher Vertrag [a peaceful agreement). 'Peace- ful labor has the qualitative meaning of 'quiet labor, labor not inter- rupted by catastrophies , ' and we therefore translate this phrase as : friedliche Arbeit [peaceful labor ). However, in this case the Corres - pondence is not uniquely defined. The problem here is that ‘Maphblå tpyl' can occasionally be translated as Friedensarbeit. In this Case , th; translation my have the meaning 'lab or during peace, " or 'labor for peace. ' The combinations 'childrens home, ' 1 childrens literature ' have relative meaning (Vinogradov, pages 2011-205) and we therefore translate : Kinderheim, Kinderliteratur. If , on the other hand, we say: 'childrens babbling (in the sense of idle chatter'), this has qualitative In earn - ing; compare the translation kindliches Gerede • - 135 — - We find an interesting example of a change in the meaning of an ad- jective in the writings or M. Gorºkiy: 'She smiled, her face was arrogant and immobile , only her gray oceanue eyes were smiling' ("The Artamonov Affair"). Generally the adjective oceans has the relative meaning of a 'fall leaf ; compare : Herbstlaub, with 'autumnal days ' Herbsttage. Here, how - ever, the meaning is clearly qualitative. The translator therefore should have translated this point as follows (Kuznetsova, page l87): Sie iáchelte : ihr Gesicht war hochmütig und unbeweglich, nur inre grau- en herbstlichen Augen làchelten. -- § 21. UNITS OF TRANSLATION. CONTEXT Let us now turn to the work of the translator. He is presented with a communication which is to be translated. First of all, the trans- lator must "understand" this communication. But what precise meaning can we impart to the words "understanding the text?" It is obvious that in the the ory of translation these words will vary in meaning, depend- ing on whether or not it is interpretation (Diagram No. 3) or actual translation (Diagram No. 4) that is taking place. In the case of inter- pretation the understanding of the text reduces to the establishment of correspondence between text and reality. But what is the meaning of these words "understanding of text" in translation? Apparently, it is expedient to treat the understanding of text in translation as refer- ring to : a) the establishment of correspondence between certain Seg- ments of the text and the elements of the intermediate language and b) the establishment of syntactic relationships between these Segment S. The text segments with which correspondence 1s established through the intermediate language will be known as the units of translation. * The isolation of the units into which the analyzed communication breaks down from the standpoint of translation is one of the basic pur- - 136 – poses of text analysis. It would be possible to adopt a word (the interval between SpaceS in Written text and the interval between pauses, i.e., "a phonetic word," in spoken text) as such a unit. However, serious objections have been raised to the using of words as units of translation. Of these apparently the most interesting ob - jection is the following. "The main reason why we cannot use a word 3. S a unit (we are referring here to "a unit of translation") lies in the fact that the dual nature of the sign does not show up with sufficient Clarity in a word, with the denoting sign moving to the forefront at the expense of the denoted sign "(Vinay, Darbelnet, page 37). The authors of the cited book resolve the problem of translation units by the isolation of the so-called units of meaning (unités de pensée). However, in practice it developed that the authors reduced the problem of units of translation to yet another classification of phrase - ological units. The problem of units of translation was treated in considerably greater depth in books dealing with machine translation. # , -, ** In machine translation the isolation of the units of translation Corresponds to the operation of reducing specific segments of a text to the information recorded in the memory of the computer. It was proposed that we distinguish between the following forms of words ("Machine f & Translation," 127): a) The input word is a unit of text defined in the following man- Ine I" . - A Word is any sequence of letters which precedes and which 1S fol- lowed by a Space or punctuation. - Two Words are identical if the letters in the corresponding posi- tions Coincide. - 137 – b) The output word is a group of words in the target language which are in correspondence with the given input word. c) The stored word is the collection of information pertaining to the given word and stored in the computer memory. The first problem associated with the isolation of text unit S in- volves the following: should the "input word" coincide with the "stored word?" For example, if we are confronted with the phrase: "The magni- tude of the angle is defined by the ratio of the arc to the radius, "* is it necessary for the stored words to have the form: magnitude angle , etc. In conventional dictionaries this is obviously not the case; here the so-called "dictionary word" (German Stichwort ſcuel, Polish haslo) is given in its original form (the noun in the nominative singular, the verb in the infinitive form, etc.). It is assumed here that a person will easily find the original form for the given word from the text. However, in the case of a computer precise rules of identification for the input and stored words are necessary. In essence the following two cases are possible : 1) The input and stored words coincide. This indicates that the computer memory must be provided with Such varied words as , for example : "magnitude, ' ' of the magnitude, ' ' to the magnitude, ' 'by the magnitude, ' "about the magnitude, ' ' of the magnitude, 'to the magnitudes , ' ' in the magnitudes. ' Generally speaking, this case may be applicable (for exam— ple, for languages such as English or German). However, for Russian this approach would increase the capacity of the computer memory tenfold, which is technically disadvantageous. 2) Formal rules are set down to reduce a series of varied input words such as , for example, "magnitude, ' 'in the magnitude, ' ' of the - 138 - magnitudes," etc., to a single Stored word. In this case the stored word may be in 16s original form, as it would appear in dictionaries (e.g. , 'magnitude '), or it may appear in some other form (e.g. , of the magnitudes "). Since the criteria of convenience are decisive, generally Speaking, the form of the stored word may be set arbitrarily. It is convenient to use the stem of the word, as shown by the example with 'magnitude; ' however, this term must also be precisely defined with ºcc. pect to the computer. The reduction of the input word to the stored word can be accomplished by the addition or (what is more convenient) the elimination of certain letters. For this reas on the term "stem" is understood to refer to that portion of a word which remains graphically unchanged in all of 1ts forms (Kulagina, Mel 'chuk, 1956, page lll). Let us take note of the fact that this is not what is generally meant by the term "stem" in conventional grammar books. Let us take the input words 'ocen', 'ocha', 'ocnob’, ‘oclamh','ociax' [ 'donkey, ' 'of the donkey, ' ' of the donkeys, ' 'by the donkeys, ' 'about the donkeys ' ), etc. We can set a rule by which the machine will eliminate the endings -a, -os, -axia, -ax , etc. , and the input words ‘ocha', ‘ociob', 'oclamh', 'ocnax' Will then be reduced to the stem: ‘ocin' . It is more difficult to reduce the word ‘ocen' to this stem. The computer may be called upon to eliminate the letter e in the next -to-last position, but this will cause all of the rules to be – come more complicated. Apparently, it is more advantageous to have two stems: ‘ocn' and ' 'ocen". It Sometimes becomes necessary to Store three Stems such as , for example, 'wot', 'uox' and “Mou', or four Stems such as , for example, ‘ºn', 'uin', ‘ hiº','ulenui' for the verb ‘Hath' [ to go ' ) and, per- haps , even more. Thus in machine translation, as a rule, we isolate those units which are shorter than the word, and namely, the units corresponding to morphemes. On the other hand, experience in machine translation has de- - 139 – monstrated that it is advantageous to introduce entire terms of speech is into the computer dictionary in the form of stored words (for which purpose it is convenient to have even a special dictionary of phrases). Combinations of words or "phrases" in machine translation repre- sent a more extensive. concept than, let us say, "idioms" or "phraseo- logical combinations," "we will understand the word phrase to refer to coºbinations of severai Words whose literal translation in sequential order (in ace ordance with the data contained in our dictionary), with- out consideration of the form of their combination, would result in distortion of meaning" (Kulagina, Mel'chuk, page 113). If we proceed from this definition, it becomes obvious that many of the combinations of words which are treated in linguistics as talo- matic are not regarded as phrases in machine translation, e.g. , the German den Kopf verlieren ('to lose one 's head') or Koloss auf tonernen Füssen ('Colossus on clay feet '). It has been shown in machine translation that any algorithm of analysis has as its final goal the comparison of segments from a target - language text with some fact existing outside of the text. Such a fact might be represented, for example, by the target –1anguage text for the target –language System, the Source-language System and , f' inally, the system of the intermediate language. -- In this connection, T. M. Nikolayeva proposed the following class 1-, fication of translation units (Nikolayeva, 1961): - . . . l) A unit of translation is identified as a segment of source-lan- guage text in relation to a segment of target –language text. This type of unit has been identified in a number of works on Iſla Chine translation , e.g., in the Dostert experiment of which we will speak later on ($38). Here there is direct equating to one another of parts of words , of words and of individual word combinations (theoretically the length of - 140 - a segment can be increased as much as desired, selecting even entire sentences and paragraphs, but in practical terms, of course, one should not go beyond the framework of word combination — this is covered in greater detail in §32). As we will see later on, the role of logical and grammatical rules in such an algorithm is reduced to the minimum; 2) The unit of translation is identified as a segment of the source-language text in relation to the target –language System. With this approach, having analyzed, for example, an English noun with the preposition by , we immediately assign the Russian indicator "instrumen- tal case," etc. This type of analysis is extremely widespread in mach– ine translations ; - - 3) The unit of translation is identified as a segment of source- language text in relation to the Source-language system, independent of the language into which the translation is being accomplished (see $38, independent analysis). 3) The unit of translation is identified as a segment of source – language text in relation to a universal intermediate language. It is obvious that in the general theory of translation it is a good idea to have such units of translation as are identified as func - tions of the intermediate language whose properties, generally Speaking, in turn are functions of that pair of languages participating in the translation: that which is a unit in the translation from Russian into French need not be a unit in translation into German. Strictly speaking, this does not contradict the idea put forth by the authors of the Stylistique comparée, since phraseologisms and idioms are particular cases of translation units * in any of the four considered approaches. The basic difference between the approach under considera - tion, however, 11es in the following: here it is tº k . . . int. J C on Side. I'l- tion that the text does not consist directly of semā...tic ur, it , i.e. , – 141 – units on the level of content, but that the semantic units are isolated precisely OYl the basis of comparis on with something outside of the text and for this reas on Something on the level of expression is taken as the unit of translation, i.e. , in the immediate reality which is the text. Proceeding from all of these considerations, we propose the follow- ing definition for the unit of translation: . A unit of translation is that minimum Segment of Source-language text which corresponds to Such a set of elementary meanings in the in- termediate language as can in turn be made to correspond with some text segment in the target language. * * Thus, units of translation here are isolated relative to some tar – get language. Below we shall see that this definition can be simplified so that the unit of translation is independent of the target language. In the identification of the units of translation it is important to bear in mind that by our definition a unit of translation may be an individual morpheme. This occurs , for example, in the case in which a morpheme expresses the meaning of modality and is made to correspond with a separate set of elementary senses in the intermediate language , and then with an individual segment of text in the target language. For example , let us examine the French phrase Deux ouvriers aur- aient été tués. - - t * In translation into English (Vinay, Darbelnet, page 142) we must identify a single unit which corresponds in English to the independent verb ; compare the English: Two workers are reported killed. The same in the Russian : ‘Coobulaetch, TroythTo usoe padouhz'. The individual unit may be a morpheme of the form: compare the - ll;2 - Russian on rosapusan'. [He used to say 'l; German: er pflegte Zu Sagen, er sagte oft French: Il aimait répéter. Il répétait souvent. We have defined the unit of translation with respect to transla - tion into some other language. For this reason, as we have stated, Vaſ'-- ious units of translation may be identified in a given text (depending on the language into which the translation is being accomplished). At the same time, we sense the need for such units as would be independent of the language into which the translation is being accomplished. Since transformation and modulation represent a form of translation (compare the third approach to the identification of units, page 140), a unit of translation can be defined in terms of transformation and modulation, i.e., in terms of translation into the same language. The units of translation obtained in this manner will be referred to as the funda- . . .e. *- mental units of translation. Let us give an example: "Better operas and ballets appeared as a result of the Creative cooperation between composers and poet S. ' Here we can identify the fundamental units appeared as a result ' (compare the transformation, ' were the result ) and 'creative coopera- tion ' (compare the modulation joint creativity'). º Thus, at the analysis stage we first of all separate the text into units of translation (with respect to the target language) or into the fundamental units. The presentation of the text in the form of a suc- cession of units will be known as the segmentation” of the source-lan- guage text. Let us take note of the fact that, generally speaking, Several – 143 – various sets of elementary meanings in the intermediate language, satis - fying the definition of a unit of translation (we will refer to such units as homonymic or not uniquely defined) may correspond to one and the same Source-language text segment , i.e. , to an identical unit of translation; in addition, several target-language segments (we will re- fer to such segments as synonymic) may correspond to a single set of elementary meanings in the intermediate language. If we can now proceed without great expansion of the unit of trans- lation from a unit of translation not uniquely defined to one that is uniquely defined (or at least sharply to reduce the number of corres- ponding sets in the intermediate language), we will refer to this new expanded text as the microcontext of the given unit. However, if it is possible to proceed from the text unit that is not uniquely defined to one that is uniquely defined only through a considerable expansion of the text , we will refer to this new expanded text as the macrocontext of the g1ven unit of translation. The macrocontext cannot be taken into consideration in formal the ory. Therefore, when we again have occasion to speak of context we will, as a rule (if it is not so stipulated) have the microcontext in mind. The following classification for the types of translation as a function of the length of the Syubject context has been proposed (Yngve, lS55): : a) word-for-word translation; b) phrase-for-phrase translation; c) sentence-for-sentence translation. Continuing this classification, we might identify : d) paragraph-by-paragraph translation, etc. It is necessary to have a clear 1dea of the fact that with this – 144 - - setting of the limits of context We need not necessarily resort each time to the process of running through the entire context that we have selected, but in any event, we should not go beyond the limits of this context. This means that if, for example, a sentence-for-sentence trans- lation is in process, We Can imagine that the text is being separated into phrases, each being handed to an individual translator, and the resulting translations are subsequently combined in proper sequence. It is clear that in practice such a limitation results in reduced transla - tion quality. However, it is extremely important to define precisely that average length of context in order to keep error's comparatively infrequent (certain attempts in this direction have been made, see the work by Kaplan, 1955). The identification of minimum context takes place in actual pra C – tice, and namely , in the case of so-called consecutive translation. Here there naturally arises the problem of the length of a speech Seg- ment. It is governed by the conditions of verbal intercourse In tranS - lating dialogue (discussions, negotiations, interviews, etc.), small segments are translated consecutively , and as a rule each of the Se rep- resent a self-contained unit. In the consecutive translation of speeches by a single individual (appearances, announcement S > etc.), the length of segment S is defined, on the one hand , by the nature Of the speech (logical relationships , for more on the nature of these relationships see $29) and, on the 9th- er hand, by the need on the part of the audience to understand the com- munication. If the segments are lengthened significantly or , what is more, if the speech were to be translated in its entirety after the speaker had concluded, the primary characteristic of spoken inter’00”5° would be lost, and namely, the simultaneity of communication. However , if the speech were to be translated sentence by sentence , 9” by indivi- - lºt — dual syntagmas, the characteristic features of monologue speech will be distorted. Consequently, it may be said that the length of a trans- lated segment will be defined by these opposing requirements. As a rule, it varies from 1 to 5 minutes of Speaking, rarely extending to 10 min- utes. Manu- script Page No. 103 108 ll3 ll.9 128 129 133 [Footnotes According to the program in the course of linguistics for the appropriate divisions, this section assumed familiarity with the basic concepts of formation grammar. Readers not familiar with this course are advised to skip $18 and to proceed to $19 immediately. A discussion on the analysis method involving direct compon- ents can be found in the Glisson book (Glisson, 1959). See the concept of progressive and regressive sequence in the book by Bally (Bally, pages 218–312). Tesniere 's transformational syntax is not as rigorously for- malized as Syntax in the Chomsky theory. Following Bally, Tesniere introduces the concept of translation, denoting ". . . the translation of a word of full meaning from one gram- matical category into another..." (Tesniere, page 364). He does not provide the rules for phrase formation by means of translation, but instead provides a detailed description of translation. Briefly discussing this classification, we will use the term "transformation." In Russian reflexive is not compared with nonreflexive , un- like the French préter, German leihen [ 'to lend ' ) and the French emprunter, the German borgen [to borrow. " ) J This term has been taken by us from the authors of the series "Stylistique comparée" (see Vinay, Darbelnet, page 51, Mal- blanc, page 28). These also cite many examples illustrating the significance of this concept for the theory of transla - tion. At first glance this requirement seems trivial (as if in the case of a nontechnical translation there is no need to refer to reality, i.e., to be concerned with interpretation) and even erroneous (it is precisely in a technical translation that the reference to reality is at a minimum; as proof of this we can cite the machine translation of technical texts), but it reflects the valid idea that in a technical transla - tion the process is accomplished through another intermediate language. - 146 – 136 138 ll l 1113 For interpretation certain text segments set into correspond - ence with elementary situations will serve as analogs of the units of translation. The given phrase has been taken from those translated ăs part of the Georgetown experiment (see "Machine Translation," page l73ff). Compare the definition of idiom with respect to a Specific language, as proposed by Bar-Hillel and Mel 'chuk (Bar-Hillel. "Machine Translation," page 251, Mel 'chuk, 1960 b). The authors of the series "Stylistique comparée" use the term découpage in this sense (Malblanc, page 23, Vinay, Darbelnet , pages 161 and others) – 147 – Chapter l; WAYS OF ACCOMPLISHING THE TRANSLATION PROCESS $22. CERTAIN IMPORTANT conceFTS IN THE TRADITIONAL THEORY OF TRANSLATION For a long time the theory of translation has given consideration ' "adequate translation," to such concepts as "literal translation,' "free translation," etc. These concepts are not based on a uniform cri- terion: on the one hand, they indicate linguistic phenomena (corres- pondence or noncorrespondence of a given target –language element with an element of the source language), while on the other hand they indi- cate artistic-esthetic phenomena , i.e. , correspondence or noncorres — pondence between image , genre or individual features, etc. In the for- mulation of the scientific theory which seeks to provide a constructive description of the process these concepts must be replaced by ones that are more precise. - . Bef cre we undertake the aſseussion of these new concepts , it would be a good idea, however, to explain the meaning traditionally imparted to such terms as "literal," "adequate," "free" translation 1n order to make an attempt at retaining that which is of value from the linguistic * . . . . * standpoint in this terminology. We will proceed from an analysis of the ' since it is in this term, as a rule, that term "literal translation," linguistic significance ls found. The term "literal translation" is used in many ways: l) it indicates "a sense of Strangeness" in the translation, i.e. , the rendering of poetic , generic and individual features of the original literary text , contradicting the esthetic norms of the literature of – lll:8 - which the translated version is to become part. In this sense, the term "literal translation" is set against the so-called "forcing the translation to good usage," i.e., to a transla- tion in which the original is freely paraphrased. This comparison evolved into a comparison of literal translation with that of a free translation, without losing any of its theoretical-literary essence ; 2) it indicates a translation broken down by elements (word-for- word, phrase-for-phrase, etc.). The term "word-for-word translation" is occasionally employed in this meaning. . In this sense the term "literal translation" is apparently used only for the case in which the element-by-element correspondence has not been established in conformity with the specific standards of lar;-- guage, style, genre, etc. This term is contrasted with the term "ade- ' where the term adequate is understood to mean trans- quate translation," lation with consideration of extensive context (and generally not ex- clusively from the linguistic standpoint). With regard to cases in which the element-by-element translation does not contradict any standards, the term "literal" will either be avoided entirely, or used in conjunction with the term "word-for-word" only for the case in which the unit of translation is represented by a word; - 3) it indicates a word-for-word translation for the case in which the lexical-grammatical Standards of one language are rendered in the other language , apparently contrasted in this other language to the correct translation accomplished with consideration of the standards and norms of the target language. - Of course , all of these are not precise definitions, but only ap- proximate descriptions representing an attempt somehow to refine the Current usage of the word. - llig - Let us now proceed to the introduction of certain terms refining the above -enumerated words. - These terms will describe the various means of accomplishing the fundamental translation procedures, depending on the nature of the pre- determined correspondences which are assumed. $23. FUNDAMENTAL TYPES OF CORRESPONDENCES BETWEEN UNITS OF THE SOURCE LANGUAGE AND UNITS OF THE TARGET LANGUAGE - We have seen that it is advisable to construct the intermediate language so that for each unit of translation in the target language and its complement in the target language there exists in the intermed- iate language : a) a set of elementary semantic units, b) the formational history of the corresponding text segment (both from the standpoint of incorporation in specific configurations, as well as from the standpoint of transformational modifications), c) stylistic characteristics. We will refer to the complement as full if the source-language un- it corresponds to the same Set of characteristics in the intermediate language as the target –language unit into which it is being translated. The word terms whose meanings are limited by their subject-matter relationships are included in the class of words among, which it is pos - sible to establish complete correspondence. •, This class includes not only scientific terminology, but all the units of a language which denote categories common to two languages. For example, words denoting numbers, Spatial relationships , the funda– mental concepts of economic and political life of European peoples , etc. , are common to Russian and the Romance and Germanic languages. Preliminary calculations show that about 10% of the full-meaning words (nouns, adjectives, verbs and adverbs) in sociopolitical texts on inter- - 150 - national Subjects are always translated into French, English, German etc. , with the same words. In connection with the phenomena considered here, compare the fol- lowing definition: "Let us identify as congruent such two elements of compared dictionaries which, occupying identical positions in the para - digmatic system, exhibit identical syntagmatic functions. As an example of complete congruence we can cite the term "static sensitivity," ren- dered in French as sensibilité, in German by statische Empfindlichkeit * in Spanish by sensibilidad estatica, in Italian by sensibilita Statica , in Dutch by statische gevoeligheid, in Polish by Wyda Jnosc statyczna and in Swedish by statisk känslighet" (Andreyev, Zambrzhitskiy, Pazukhin, Pak, page 1). Let us now turn to incomplete correspondences. First of all , here we can identify the following three groups, depending on which chara C – teristics of units are not coincident : a) correspondences with semantic or stylistic invariants , i.e. , those in which only the characteristic of formation fails to coincide. AS an example let us examine the translation of the follºwing RuS - sian sentence : * = 1Kerzon died in 1925, only barely outliving his political demise. In the German and French translations of this sentence the Sen Se of consequence can be rendered by relating two basic clauses without A A ** subsequent transformation: - t German: Er tiberlebte seinen politischen Tod nicht lange , (denn) er starb l925. French: Il ne survécut pas longtemps a sa mort politique : il mour” – ut en l925. When we speak of retaining stylistic characteristics . We have in mind not only the marks appropriate to individual words, but the general nature of the structure, i.e., that which is sometimes referred to as - 151 – the "stylistic figure." It should be pointed out that this characteris- tic is not as diffuse as it might seem at first giance. Since the basic figures can be enumerated (this, by the way, had been done as far back as the ancient courses in rhetoric ), it is quite possible to imagine an addition to the algorithm of syntactic analysis which checks On the Structure of a phrase from the standpoint of determining whether or not it contains one of the given figures ; * b) correspondences with a semantic invariant , i.e., those in which not only the characteristics of formation, but those of stylistics, can be distinguished. Divergence in Stylistic norms is characteristic of the translation of certain newspaper texts such as , for example , a Sports report. Ger’— man newspapers contain many Special expressions, generally quite unique, but quite meaningless to the "uninitiated." The majority of these words must be marked "familiar" or "natural." Let us present an example (Sh- vanebakh, Revzin, page 83): Für die eindrucksvollste Worstellung sorgte Halbweltergewichtler Niegsch, als er favorislerten Strauss mit einem schweren Recht-Haken abfing und zu Boden zwang. - The retention in the translation of the stylistic characteristics for individual words results, in the best of cases, in a specimen of poor "sports jargon." ['The welterweight Niegsch put on a good show when he caught the favorite Strauss with a right hook and forced nim to the ground. ' ) Compare this "less exciting" translation: ['The welterweight Niegsch fought more effectively by knocking out his opponent, the favorite Strauss, with a hard right hand. ' ] Failure to retain the stylistic characteristic frequently leads to phenomena which in translation pract ice have come to be known as 'Smooth ling over. ' An example of this is the translation into Russian of the German sentence : In Chemnitz war der Boden für ihn zu heiss geWorden - 152 - ['It became too dangerous for him to remain in Chemnitz. " ) Here, how- ever, it is possible to retain the stylistic characteristic and to leave the figurative expression in the translation as , for example, "It became too hot for him in Chemnitz. Another example of 'smoothing over 1 : X. n'y va pas par quatre che- mins [Y writes rather openly about this; 1. c) correspondences which do not retain meaning. In connection with this third case we must mention the following. It is clear that the complete nonretention of meaning contradicts the Very definition of translation. However, the situation here involves the fact that in the set of semantic units corresponding to the text Segment there may be some which are more or less important , i.e. , a certain hierarchy of components of meaning may be introduced. As an ex- ample, let us consider the sentence: "The African people won their in- dependence with their blood. ' This was translated into French in the following manner: Les peuples d'Afrique ont acheté l'indépendance au prix de leur sang. The Word 'won' is rendered in the translation by the word acheter ., * , (to buy), which means "A causes to himself the possession of a certain object in exchange for money." At first glance it may seem that the meaning of the Russian word has been entirely lost. In actual fact, however, this is not the case, because in both the words to win and in to buy we retain the ele- mentary meaning: "A causes to himself the possession of a certain ob- ject." This general meaning in the given context is relative, while the meaning indicating the means by which possession is caused (force, in exchange for money, etc. ) is not relevant in this case. In other Words , the correspondence under consideration here is established so that the basic elementary Semantic differential indicator of a certain group is - 153 – , * retained, while the indicators distinguishing these words from one an- other are neutralized. Thus the words to acquire,' 'to buy,' 'to win, 'to steal," etc. , having the general meaning "to cause possession for oneself," can replace One another. - Thus, in the set of elementary meanings compared to the word it is possible to identify several of the fundamental and most important words which relate to the situation. This subset of identified most important meanings we will refer to as the primary meaning of a word. The primary meaning of a word will be a Semantic invariant which is retained in the translation. Here it should be borne in mind that the primary meaning cf a word may be identified from the very structure of the intermediate ianguage , i.e. , predetermined. This then accomplishes the translation procedure in accordance with Diagram No. 14. An analogous phenomenon takes place when it becomes necessary to identify the primary meaning of a word, but when this identification cannot be given, but is derived from the situation. Let us examine the following expression: 'Two long rows of dense trees are growing virtually at the edge. " In translation into French there is no direct correspondence for 'to rest against . . . . [Transla– tor's Note: translated in the sentence above as 'growing at ' ). This is understandable, in view of the complexity of the semantic structure of the given word. This word has three definitions in the dictionary: l) "2 oming into close contact with a part of its body or with the end of some object;" 2) "while walking, encountering something, detecting Some cbstacle" (simple); 3) (figurative) "stubbornly not to agree with something, on something. . . " These meanings are illustrated, respective – iy , by the following examples. "To take a firm stand; he walked and he walked and he ran into a fence ; he turned everything down and didn't go anywhere ' (see the "Dictionary of the Russian Language" S. I. Ozhegov, - 154 – 1953 edition.) Strictly Speaking, none of these definitions Corresponds to the meaning of the Sought word. Intuitively , we understand that from the cited illustrations the meaning of ‘ynepetbca' is more precisely ren- dered by the phrase 'he walked and he walked and he ran into a fence. This definition is & overned by the fact that a row of trees may be looked upon as a line of equally spaced uniform objects in a specific Sequence, one after the other. This elementary meaning of emanating **on one point (the origin) in the direction of another (the end) is included in the meaning of the word 'ynepetsca', both in the context of 'he walked and he walked and he ran into a fence, ' and in the context 'a row of trees £rowing at the edge. In the Synthesis of the French expression it would be possible from the numerous French equivalents of the word “ynepetbca’ to select the word buter, i.e., "while walking, te encounter an obstacle." With this approach, however, we assume that the French reader has had the same experiences in life as the Russian auth- Or and that he will °onsequently be in a position to see in the word buter the elementary meaning "to emanate from one point (the origin) in the direction of another (the end)" and to reject the attresential in- dicator "while walking, to encounter an obstacle." But the translator can replace the word "they are resting..." by another word including the elementary meaning "to emanate from one point (the origin) to an– other (the end), " for example, by the word are going toward. . . . [vont jusqu'al. "Two rows of trees arºe growing at the edge' will be transla - **d as follows: Deux rangées d'arbres vont jusqu'au qual. We have thus described incomplete correspondences from the Struc – tural-semantic or systemic point of view (and namely, from the point of View of that set of characteristics in the intermediate language through which the correspondence is established). Let us now turn to another aspect of the problem. Let us C On Sider" - 155 - Here we can identify two forms of incomplete c crrespondences : a ) conditional and b) probable. A conditional correspondence is one in which the translation unit of some source -language sublanguage corresponds to units of the target language, given the presence of specific surroundings for the given un- it. This class includes words with diverse categorization in two (or more) languages, i.e., a substantial majority of words. Compare foot with Jambe/pied and hand with main/bras. In order to establish corres- pondence with these words in the target language, we must know their surroundings, their context (compare "The child was not Sitting on his legs' and 'He kicked him with his foot ' ). Thus, the concept of conditional correspondence reflects the idea familiar to the theory of translation that translation is a function of context (see $21). Thus in the translation of the phrase 'A large square Spread out in front of the palace, ' by the way, we encounter the fol.- lowing difficulty: the word 'wide " [Translator's Note: translated as 'large ' in the sentence above ) has no complete correspondence in French: it is contrasted with the word 'narrow, ' whereas in French large is contrasted with the basic word long. However, the word large suggests the elementary meaning 'extension in space, ' which is implied in the W Crd Vaste. If we take into consideration the microcontext of the word 'wide ' in the cited phrase (large square), in the synthesis of the French text we will select as conditional correspondence for the word wide the French va Ste. Complete and conditional correspondences have in common the fact - 156 - that for a certain unit of translation a direct correspondence (in the given sublanguage) can be found. Let us now consider a case in which such uniquely defined corres – pondence is impossible , even for Some sublanguage. We will designate as probable that correspondence in which the translation unit in the Source language corresponds to several transla - tions in the target language, basically not reducible to one another, and for each of these translations it being impossible to render a con- text that is restricted from the Standpoint of sense and which uniquely defines that translation, it being possible only to indicate some like – lihood [probability that a given translation will be selected. For ex- ample, for the word "discussion" it is impossible to indicate an intel- ligently restricted context which would determine the selection of the French causerie, the German Unterhaltung or the French entretien, or the German Gespräch, Unterredung. However, it is possible to determine the probability of the appearance of one of these correspondences for the given Sublanguage. Let us examine the following sentence: "This institution's sources of income are nebulous." There is no direct correspondence for the word "institution" in English, French, German and other languages [sic]. As a rule this word is encountered in combination with words with which it forms a single translation unit (compare, educational institution, in- st it ution d'enseignement, Mittel- und Hochschule , etc.). However, in this text we are speaking of a corporation Shrouding Some improper po- litical activity in the guise of a scientific research institute. Only through the knowledge of this fact, derived from the entire article, can correspondence with the word institution be accomplished with a high degree of probability: Cette officine dispose de res sources d'ori- gine d oute use. - 157 - It must be pointed out that if complete and conditional correspon- dences describe a situation for which reference to reality in accord – ance with Diagram No. 3 is not necessary, the probable equivalance des— cribes the case in which a uniquely defined solution is possible only in the event of reference to reality, while for Diagram No. 4 only an indication of the probable correspondence is possible. Note. As a rule , direct correspondences coincide with mutually un- iquely defined correspondences. Let us recall that a correspondence satisfying the following requirements is designated as mutually unique - ly defined: A corresponds to B and only to B; B corresponds to A and cnly to A. As an example of such correspondence we can cite the words: Januar, Janvier, January ; Sieben, Sept, Seven (grammatical mutually uniquely defined correspondences are also possible : Der Himmel ist blau, le ciel est bleu, the sky is blue. Or : Er hatte gesehen, il ava it yu, he had seen. * However, let us note that the concept of direct correspondence is more extensive than the concept of mutually uniquely defined corres- pondence. Correspondence may be direct, but need not be mutually unique- ly defined; for example, Russian hauhharb, German beginnen, anſangen, En- glish to start, to begin, to commence. In this connection, the procedure analyzed in the following is a generalized version of the one proposed earlier by the authors (Revzin, Rozents veyz, pages 56–57). § 24. FUNDAMENTAL DEFINITIONS FOR THE THEORY OF TRANSLATION Let us introduce an auxiliary concept , and namely the concept of the "supercategory." We will say that two units of translation, of which one is identi- fied in the source language with respect to the target language, and the other is identified in the target language with respect to the s curce language, belong to the same supercategory, if it is possible to establish a one-to-one correspondence between them; they belong to dif- ferent supercategories if it is impossible to establish such a one-to- one correspondence. For example, Wandzeitung and 'wall newspaper' be - – 158 - long to one supercategory, while ‘yot' [comfort, coziness j and confort belong to different supercategories. Let us now take two languages Ll and P2 and let us identify all units of translation in the Se languages. Each of the units of translation be- longs to some supercategory based on an analysis of the correspondence between these two languages. Some Supercategories ë include units of translation from two Diagram No. 8. 1) Common languages, while others include units of part. translation only of one of the two lan- guages. We have obtained two sets of Su- percategories which, generally speaking, intersect ; see Diagram No. 8. In the cross-hatched portion we find the supercategories including the units of Ll and I-2 or, in other words, the units between which it is possible to establish a one-to-one correspondence. Let us now examine the procedure involved in the translation from Ll into I-2 Two radically different types of translation are possible : or from I-2 into Ill. 1) the element a falls within the common part of the two languages I-1 and L2. Since the supercategory of element o includes some element É such that between a and B a one-to-one correspondence is established, the transition from a to B is accomplished by simple recoding. We will refer to this type of translation as interlinear. A particular case of interlinear translation — one in which the unit of translation is a word — is represented by a word-for-word translation. We will consider translation as a recoding process in greater de- tail in $25. - 2) element a does not fall within the common portion of the two – lº.9 - languages Ll and L2. - Here the situation is more complex and several various Subtypes - are possible. a) The process of translation is accomplished as if the element falls within the common portion of the two languages, i.e. , it is held to be possible to establish a one-to-one correspondence between o and some 5 in L2. In other words, the P2 System is expanded here and it is complemented by a number of supercategories from L1. We will refer to such a translation as a literal translation. b) The process of translation is accomplished so that element a iS replaced within Ll by some element Gl which falls within the cross- hatched portion, with element al being chosen in Such a way as to re- tain the Sense of the Statement. A one-to-one correspondence is then es – tablished between Gl and 5, where f3 is included in the same supercate- gory as al. We will refer to a translation of this type as a simplify – ing translation. * The difference between this diagram and the previous lies in the fact that here the original element does not necessarily fall within the C omiſſion portion of the two languages. There is also the poss 1b 111ty of the case in which the translating element is not necessarily chosen from the common portion. c) The process of translation iS accomplished in its initial Stage in the Same manner" as indicated in 1 or 2b, 1.e. , a correspondence 1s established directly between O. and 5 if both fall within the common portion, or as given in b, a correspondence is established between O. and dºl and then between Gl and 5. Now, within I-2 We have either ſmodula - tion or transformation taking place, i.e., those segments of text are being selected as have the same meaning as B (1.e. , Pl, P2. . . . Pn), and a correspondence is established between the initial unit of translation - 160 — and any of the examined B, (Bl Orº B2. . . or Pn). we will refer to a translation of this type as an exact translation. d) A particular case of an exact translation is a translation in which the selection of the Correspondence is accomplished with reten- tion of the stylistic characteristic and with consideration of the rules of conjunction in language I-2 between the given Segment and it. Surrounding segments, obtained as a result of the translation of other units (1.e., with consideration of the context in the target language). This translation will be known as an adequate translation. It is easy to see that on the basis of this definition any ade- quate translation is simultaneously an exact translation. The reverse is not true. With regard to the relationship between the interlinear and exact translations, we might take note of the fact that each interlinear translation is simultaneously an exact translation. An adequate inter- linear translation, including a word-for-word translation, is possible Only under the condition that the Stylistic characteristic is retained and that the rules of conjunction are observed, i.e., when any exact translation becomes adequate. d) It is also possible to have a translation in which the selec – tion of the translation element is not accomplished On the basis of the rules of correspondence, but exclusively on the basis of the rules of . C On Junction. We will refer to a translation of this type as a free translation. Since a free translation in the sense of this definition may se described only by resorting to the concept of poetry, we will not deal with this type in the following. We have devoted §§25–29 to a detailed analysis of the various means of accomplishing the translation pr’OC eSS. - - 161 – $25. The Portion common - - - T9 TWO LANGUAGES AND ITs BORROWING OF WORDS AND #Rāś". #oº" IN TRANSLATION: In any procedure for the teaching of another language and the re- S . . . - ulting establishment of correspondences between two languages, we must 2 ide -- • . ntify a certain number of one-to-one ° Orrespondences achieved by di- rect indication of objects. It is precisely the area of objects in which we find the "intersection of two cultures" (Quine, 1953, page 62) that serves as the basis for bilingual intercourse, and thus as the ba- S is for translation, this area gradually expanding with increasing con- tacts (compare $6) and finally leads to the formation of that which we have referred to as the common portion between two languages. In order more easily to understand the process by which there arises a portion common to two language Systems, let us recall how the process of studying a foreign language by students in primary or second- ary schools takes place in its initial Stage. With a textbook and a teacher the student I'e Cognizes the meanings of words and grammatical forms in the language, i. e. , their translation into the native language. As a rule, this translation is Word-for-word, since in this stage the foreign language consists of elements each having only a Single corres – poisease to the native language. As a result of repeated C Omparis Ons - these correspondences are memorized and form, in Shcherba is expression a lar-3-age with two terms. In this stage of the teaching pr’OCe SS this J) 1s the portion common to the native and the ** ***. - The portion common to two languages stands out with particular Clarity in Vocabulary. This is understandable: the word units of a lan- Buage are less interrelated than the phonological, morphological or syntactic elements. However, the effect of language contacts is notice – able both in morphology and in Syntax. From the Standpoint of the theory and practice of translation 2 - 162 - particular attention should be devoted to the phenomena making it pos- sible to establish one-to-one correspondences and the formation of a common portion for the two systems of languages: cognates and transla - tion loans of identical thought processes. Let us turn to the cognates. We know that languages borrow words from each other, as well as combinations of words. Some of these words, the so-called internation- alisms, are found in many languages (compare, for example , 'democra - cy" – democracy, democratie, Demokratie; "socialism' – socialism, so- cialisme, Sozialismus; 'revolution' – revolution, revolution, Revolu- tion; 'party' – party, parti, Partei ; 'university " — university, uni- versite, Universität ; "tragedy." — tragedy, tragédie, Tragödie; 'basket – ball' – basket-ball, basket-ball, Basketball). Words of this type are found in many languages, including those that are not related; however , this is not true for all languages. It was pointed out in this connec – tion that the term "internationalisms" is not precise. It has been sug- gested that this term be replaced by "regionalism." However, even this term is not convenient, since words belonging to that group of words may be found in languages geographically remote from each other". We will remain with the term "internationalism," considering the inclusion of the Se WOrdS into the language of international, political and Scien- t if iC intercourse, as well as their inclusion in the international ar- tificial languages. It is obvious at first since that despite the phon- ological and morphological relationships, these internationalisms be – come subordinate to the System of the given language , although they continue to retain certain similarities (they stand out rather clearly in written form). From the standpoint of the theory of translation it is important to point out that these words are included in the common portion of corresponding languages and , consequently, are translated – 163 – interlinearly. Internationalisms are frequently encountered in the translation of political text. However, basically they belong to the language of sci- ence which is international in content." Formed in a given language, primarily from ancient Greek or Latin morphemes, international scienti- fic words and word combinations spread into other languages (compare the recently evolved terms: 'cybernetics, ' 'entropy'). Primarily this category pertains to nouns (including compound nouns: 'dialectology, ' 'spectrometer," etc.). The uniquely defined nature of scientific terminology does not , in principle , contradict the fact that in individual scientific disciplines such as , for example, in linguistics, inconsistencies are noted in the use of terminology (compare 'language, ' 'speech," langue, langage, par- cle, 'phoneme, ' 'morpheme, ' 'syntagma, ' etc.). With the development and refinement of the basic concepts, Scientific terminology becomes more aniquely defined. The more abstract a scientific discipline , i.e. , the ºcre it is based on the oretically defined concepts , the more exact its terminology. (Compare, for example, the uniquely defined terminology of mathematics). The development of international trade , communications and Scien- : ific-technical contacts has resulted, as is well known, in interna – tlonal standardization of technical terminology , i.e. , 10 the expansion of the common portion of incºnses. In addition to internationalisms, the common portion of two given languages includes cognates, i.e. , translated with the use of the for – eign word. (Compare, for example, such Russian words taken from the French as : "pomah’, ‘Bhsht’, ‘Mahebpbu', ‘napthaah' [' novel, ' 'visit, ' 'maneuver 5, 'partisan' ] and, on the other hand, such French words taken from Russ lan as: Soviet ). - 164 – - However, it is important to bear in mind that not all words taken by language A from language B fall within the common portion. Thus, for example, the word “kohdepahche' [master of ceremonies ) corresponds in the French not to conferencier, but to présentateur. The French word con- ferencier corresponds in Russian not to 'master of ceremonies, " but to lecturer. ' The Russian word 'demonstration' (compare, 'street demon- stration'), taken from the German Demonstration, corresponds in the French in this meaning not to démonstration, but to manifestation. In addition, there are no one-to-one correspondences between such words borrowed from the German as 'sanatorium' ('a therapeutic institution equipped for therapy and rest ' ) and the French sanatorium. Words such as 'demonstration' – Demonstration, demonstration at first glance seem to be international, whereas in fact they are not. These are the so-called faux amis of the translator. Many of these "faux amis" can be traced etymologically to the same source, thus making it more difficult to distinguish them properly. Compare the word 'passenger' (denoting a general concept) and the French passager ("a passenger aboard a vessel'); the French voyage ('a trip") and the English voyage ('a trip at sea"); the Russian 'trivial, the En- glish trivial and the French trivial ('vulgar'); the French salaire - ['the wages of a working man'] and the English salary ['the earnings of a professional man' ]; the Russian loyal, ',the French loyal and the En- glish loyal ("true'); the Russian figure," and the French figure ('face '). - - Portions of word combinations may also be 'faux amis. ' Compare , for example, the English fresh water and the French eau fraſche ('cool water'). Words of this type nevertheless appear in translations. Thus, for example, the Word practically (in the sense of actually ) is frequent- - 165 – ly translated as ‘mpartitueckh' which in Russian 1s compared with the word theoretically." The word paghetts (in the sense of touching,' 'pºtt- ful, ' 'pathetic ') is always being translated as 'pathetic. ' Compare , also, according to liberal estimates with 'according to broad calcula - tions' (in the place of 'according to modest calculations'); comforta- ble income should be reºdered not as 'comfortable income , " but by 'good income, ' etc. / f Some of these 'faux amis' are such only in certain Sublanguages. Compare , for example , the French attaque, the English attack. in the following context: les attacues contre les droits démocratiques (the at- tacks on democratic rights) – with 'mayueme (a he ataka) ha memokpathueckhe npasa...' ; compare also in the German gegen die demokratischen Rechte At- tacke reiten. Let us note that as a result of linguistic contacts a bor- rowed word may affect the meaning of its corresponding word in the lan- guage from which it had been taken. Thus, the Russian word 'brigade ' was in one-to-one correspondence with the French brigade only in mili- tary texts. In combinations of the type "brigade of workers' this word was translated as équipe. Recently, however, it has been translated in these combinations as brigade. (compare, 'communist brigades' with les brigades communistes). Such means of expanding the common portion of languages and , consequently, the potentials of interlinear translation call for a Special Study. This we will undertake immediately in connec – tion with the problem of rendering identical thought processes. As a result of regular language contacts it is not so much borrow- ing that takes place as the rendering of identical thought processes , i. e. , Changing the structural meanings of words in accordance with the specimen of structural meaning in words of the other language to pro- duce new words (and new word combinations) rendering the structure of the words from the foreign language. (Compare the Russian 'touching' - 166 - with the French touchant; the Russian 'prejudice with the French pré- Jugé; the Russian 'influence with the French influence; the ºwner emulation socialiste with the Russian 'socialist competition; the French sans-parti with the Russian without party; ' the Russian ‘ww.pobosspeline' with the German weitanschauung 3 the German Planwirtschaft With the Russian 'planned e Conomy; ' the Russian ‘Hapcya' with the English Peoples' court; the Russian ‘patºpax' With the English workers' faculty; the Russian “Mom othbxa' with the English Rest House; and finally, the * 'labor book' with the English Labor Book. Having borrowed a Word, a language takes over its phonetic sub- Stance; in the case of transferrence of 1dentical thought processes this is * the case: the foreign word (or word combination) is translated with the structure of the model being retained in the translation (com- pare , grandi cantieri del 90mmunismo, grands chantiers du communisme, die grossen Bauten des Kommunismus With the Russian the great con– Structions of Communism; ' critica e autocritica, Critique et autocri- tique , Kritik und Selbstritik with the Russian 'criticism and self-cri- ticism; omnipotens, tout -puissant, allmächtig with the Russian 'Omni- potent '). We can therefore understand Why newly intrºduced identical thought patterns are frequently treated as Something Strange, and some are never accepted. * In order for a thought pattern to be accepted, the language must experience a need to express new concepts ſor Which words have", been formed in another language or, in other words, the need for uniform categorization of reality and the need for a uniform semantic System resulting from the process of contact and, in turn, encouraging C Ontact. When the Russian combination 'shock brigade' is rendered 1nt O French as brigade de choc this is translation — the rendering of thought processes today is no longer regarded as literal translation, - 167 – but is treated as interlinear translation, because that segment of re- ality, that content which is denoted by the words 'shock brigade" did not exist in French and when the need arose for such an expression a combination of words was devised to reproduce the Semantic Structure of the Russian word combination. - It was pointed out in $16 that, as a rule, in the Romance languages the words corresponding to Russian nouns ending in -ocmo are not formed by means of a suffix added to the corresponding adjective (this is a Russian word-formation model not characteristic of these languages), but formed by a combination of words (compare the Russian ‘map.ThäHocrb' with esprit de parti; Haeflhocte: ['wealth of ideas ' ) with richesse d'idées, richesse idéologique, haute tenue ideologique, etc., and the same in Rumanian). At the same time , with Stronger contact between languages, the more noticeable the trend toward duplicating the thought processes of such words. * (Compare the following words which have appeared in Ru- manian: partinitate [' party spirit ' ), ideinitate ['wealth of ideas']). Thus, regular language contacts result in a narrowing of the gap be - tween word-forming models: that which was known as literal translation may become acceptable interlinear translation. The boundary between lit - eral and interlinear translation depends on the boundaries between the common portion of the two languages, and this boundary 1S variable. The rendering of identical thought processes is an extremely wide- spread phenomenon in international organizations. The activities of the United Nations are instructive in this connection. The spoken and written translation of all speeches and texts mak- ing up the documentation of the UN and its various organizations into the official languages (or into some of them, i.e., into the working languages adopted by a given organization) call for rigorous termino- logical standardization and unification of concepts. This is associated - 168 - -- with the publication of a great number of terminological references and dictionaries of great interest for the study of linguistic contact and means of forming the common portion. As an illustration let us present an abbreviated list of terms in English, French and Russian pertaining to the Structure of the United Nations in connection with problems of education, science and culture (UNESCO). The rendition of identical thought patterns stands out clearly and is easily seen in a comparison of the following terms : General Conférence Tehepawibhan Kohſpepehutha Conférence Générale • Président of the General TIpemcenatejib Tehepawibhoà Illerence Kohqepehu HH Président de la Conférence générale - Programme Commission Komhcchā no mporpamme Commission du programme - Administrative Commission AIIMHHHCtpath Bhaā KoMhc- Commission administrative CH; - - Division of Regional Acti- Otmen perhohajibhoh Hex- vities Tejib HOCTH Division des activités ré- gionales - - - - Department of Social Sci- Ilenaptameht couhajibhbºx eIICeS - - Hayk Département des Sciences sociales Division of Applied Social Otmen mphkian HBix - colla- Sciences anbHEIX hayk • * * Division des sciences socia- - . les appliquées - Statistical Division Crathcthueckhā otgen Division de statistique An interesting and as yet unresolved problem in the theory Of translation is the question of the cases in which a translator borrows WordS are tº which cases he renders identical thought patters? The an -Tº swer to this question might be sought, for example, in the nature of the semantic element expressed in the language contact. It has been noted that when the need arises to transmit a semantic element which is foreign to a given reality, as a rule, borrowing takes place (French borrowing from the Russian: samovar, cosaque, steppe). Studies of bor- rowed words have led to the conclusion that they denote specific items - 169 – t .- - •.- ſ• * of custom, trade, etc. (Compare 'suit, " 'coat, 'goblet, ' 'powder, "soup, 's 'ple, 'electrician, etc.). - Another interesting feature is the introduction into a text of a borrowed word as the basic means of translating "multilevel" words, i.e., having many meanings, in the translatiºn of Oriental scriptures. Thus, W. N. Toporov, the translator of the shuddíst doctrine "Dhammapada" ['Words of the Doctrine' ] from Pali consistently follows this principle. Unlike the numerous other translators of the doctrine, attempting to render the meaning of the basic term of the book "Dhamma" (the word "dhamma" denotes "virtue," law, 'teaching, 'religion, ' 'element, 'quality, ' ' thing, ' 'phenomenon, ' etc.), V. N. Toporov introduces this word into the Russian text (Toporov, 1960, pages 32 and 133) [Trans- lator's Note: The Encyclopaedia Britannica defines Dhammapada as "words of the Doctrine"). It is possible that the problem of rendering identical thought patterns or the borrowing of words may be resolved by resorting to cer– tain theoretical-informational concepts. In borrowing the morpheme al- phabet of the target language is expanded; in the rendering of identi- cal thought patterns, however, the morpheme alphabet remains unchanged. Let us now take a look at some communication (a word which we want to translate). If this communication is frequently employed (for exam– ple, if it identifies a customary object), it makes sense to gode this word with a single morpheme (as a rule, a borrowed word cannot be bro- ken down), even at the expense of a slight expansion of the alphabet. However, if the word is rarely used, it would be more advantageous to have a longer communication (in terms of the number of morphemes), but in this case the alphabet need not be expanded. Let us note that the solution to this problem is of great significance from the standpoint of general linguistics (compare with the question of whether it is ad- - l'70 - vantageous to treat diphthongs and affricates as single phonemes). Among renditions of identical thought pattern the so-called popu- lar expressions occupy an important position, i.e. , turns of Words re- suiting from the translation of popular literary sources. These include expressions of biblical origin, words and phrases from ancient mytholo- gy, as well as the statements of leading people. - In view of the importance of popular expressions both from the practical standpoint (these frequently confront an inexperienced trans- lator with difficult problems) and from the standpoint of theory (these clearly show the importance of contact between the European languages), French Que la lumière Soit. - ILe Fruit dé- fendu. Le doigt de Dieu. CEil pour ocil, dent pour ~dent, Je m'en lave les mains. Aller aux på- tres. Qui sãme le vent récolte- rala tempête. Aller A contre courant. Jeter des per- les devant les pourceaux De l'abondan- ce du coeur la bouche parle. English Forbidden fruit. Fig-leaf The finger of God. Eye for eye, tooth for tooth. I wash my hands. Begathered to ones fathers. Sow the wind and reap the Russian IIa 6ynet cBet. 3anpethbuſh IIJIOI. Qbhrobbiſ, Jih- CTOK. Tlepct 6oxhā. Oko 3a oko, 3yö 3a syó. 9 yMEIBalopy- KH. Otripabhtbca K npaoTuam. KTo ceet Be- Tep, noxhet wirlwind. 6ypio. Swim against TInbutb mpothb the stream. Teueh H5. Cast pearlsbe- Metath, 6acep fore swines. nepen chhhh. $4MH. . Out of the a- Or Ha6Bitka bundance of ceptua rvia- the heart the romat yeta. mouth speaks. - 171 - we present at this point a number of illustrative examples : German Es werde Licht. Verbotene Frucht. Feigenblatt. Der Finger Gottes. Auge um Au- ge, Zahn um Zahn. Ich wasche meine Hände in Unschuld. Zu seinen Vä- tern Versan- melt werden. Wer Wind sā- et,wird Sturm ernten. - Wider den Stromschwim- Inen. Die Perlen Wor die Säue wer- ſen. Wes das Herz voll ist, des gehet der Mund über. French Personne n’est prophète dans sa patrie. Rendre à Cé- sar ce qui ap- partient à Cé- Saſ. L'argent ne sent pas mau- vais. tº Et pourtant elle tourne (se meut). (Cp. Htanbāh- cKhā ophth- HàJI: Eppur simuo- vel) Sans égard pour person- Ile. Pomine de dis- corde. L'enferest pa- vé de bonnes intentions. On ne parle pas de corde dans la mai- Son du pendu. Tout est bien, qui finit bien. Errer est hu- main. La place au Soleil. English No man is a prophèt in his own country. Render onto Caesar, what Caesar's is, and onto God, what God's is. Money has no smell. [Compare Without res- pect of per- SOITS. The apple of discord. The (road to) Hell is paved with good in- tentions. All's well that ends well. To err is hu- II].811. - Russian Her npopoka TB CBOem Ore- Ornati, 6ory 6oxbe, a Lie- capio IIecape- BO. Jehbrh He naxhyt. A Bce-Takh oHa Bepthtch. the original He B3hpaa Ha JIHLI3. - $16moko pas- Ilopa. AR BEIMouleh RoºpblMH Ha- Mepehhāmhs B Home nobe- lileHHOTO He roBopat o Be- pebke. Bce xopolilo, *To xopoulo KOHHaetCŞı. Henobeky cBoſłctBehho Ouihôathca. Mecto IIOI! •COJIHLeM. - 172 - German Der Prophet gilt nichts in Seinem Va- terlande. Gebet – dem Kaiser, was des Kaisers ist, und Gott, was Gottes ist. Geld stinkt - nicht. Sie dreht sich doch. Italian] Ohne Ansehen der Person. Der Zank- apfel. Mit guten Ab- sichten ist die Hölle ge- pſlastert. Im Hause des Gehängten spricht man nicht vom Strang. Ende gut — alles gut. Irren ist menschlich. Der Platz an der Sonne. *. § 26. LITERAL TRANSLATION As was stated earlier (see $24), a literal translation presupposes the establishment of complete correspondence between units of the source language and the target language, where no such correspondence exists or where such correspondence is present only in incomplete form. For example, let us consider translation from a native language into a for- eign language. Essential to literal translation in this case is the fact that the arbitrary correspondence is established on the basis of cate – gories in the native language , i.e. , the unit of translation is synthe- sized not in correspondence with an intermediate language, but on the basis of the level of content in the native tongue. In other words: a text produced by , literal translation belongs in terms of level of con- tent to the source language (the native language), while in terms of the level of expression it belongs to the target language (the foreign language). * Let us describe certain of the most common cases of literal trans – lation demonstrating the mechanism of this type of translation process (Rozents veyg, Uman, 1962). These examples will be considered in detail, with an explanation of the actual mechanism involved in literal translation, the need for a detailed analysis being associated with the fact that in the normative approach to translatiºn the Sections on literal translation are reases to . purely negative 'evaluations and to advice not to translate literally, occasionally simply being restricted to the presentation of anecdotes and incidents from translation practice. We have not even mentioned the fact that frequently the words "literal translation" a rºe used Simply as a synonym for "unreliable translation." In the theoretical approach to translation adopted in this book, cases of literal translation provide the material for a more thorough – l'73 - penetration into the mechanism of language contact which takes place in translation. Let us take note of the ract that "errors" are occasional- ly as regular as correspondences, and the analysis of these "errors" is of interest from the point of view of general linguistics (see Frei). In the initial stage of translations of texts from Russian into French, German, Hºglish and other languages, a tendency is noted to con- trast the forms of the personal possessive with some form that is inde – . pendent of person and would express possession on the part of the Sub- ject of the action, as is the case in Russian : compare "the parents loved their son' with the French les parents aimaient son fils (instead of leur), with the German die Eltern liebten seinen Sohn (instead of ihren), etc. In order to explain this phenomenon it is obviously neces- sary to provide a precise description dº the structure Of content and form of expression of the possessive in Russian and French (or German). In Russian the System of the possessive is based, as is Well known, on the contrasting of the forms of the personal possessive pronoun with those of the reflexive possessive prº Onoun. The personal and reflexive possessive pronouns represent parts of a single opposition. The comparison is based on an indicator common to both — the expression of the relationship of ownership. The indicator distinguishing these is the presence or absence of the meaning of Own- ership on the part of the subject of, the action: a reflexive posse S Sive pronoun expresses the ownership ſºossession) of an object by the subject of the action — 'I am taking my pencil, you are taking yours. ' The per- sonal possessive pronoun does not express the ownership of an object by the subject of the action (the object belongs to some other person): ! I am taking your pencil, you are taking mine. ' In this case the reflexive possessive pronoun, always expressing the ownership of an object by the subject of the action, requires no – 174 – special forms to indicate person or the number of owners, since the owner is the subject of the action as expressed in the words of the pers on doing the Speaking. The French has no such opposition as is inherent in Russian. French does not distinguish whether an object belongs to the subject of the action or not. It is in this that we must seek the cause for literal translations of corresponding forms. - Let us now examine the errors which arise from the translation Of Russian conditional Subordinate clauses: Si j'aurai le temps, je viendrai vous voir (instead of si j'ai). [[If I will have time, I will come to you'. Si j'aurals le temps, je viendrais vous voir (instead or si j'avais). - ['If I had had time, I would have come to you. " ) Analogous phenomena are encountered in translation into German and English. The nature of this erroneous translation makes it possible to as - Sume that various differential indicators are characteristic of Russian and French. In Russian, 1n the conditional subordinate clause there is a con- trasting of the real and Supposed act , which is expressed by: a ) the contrasting of the forms of the indicative and subjunctive moods ; b) the contrasting of the conjunctions "if and "if [subjunctive). In French the real and the supposed action is also distinguished. However, here we encounter a number of distinct features: - l) Despite the fact that the action of the main and subordinate clauses in the cited examples pertains logically to the identical in- stant of time, French contrasts the verbal forms of these clauses (pré- sent with futur; imparfait, plus-que-parfait with conditionnel). – 175 - 2) In expressing a presumed action French distinguishes its time relation (Si j'avais le temps aujourd'hui ou demain with si j'avais eu le temps hier). Also of interest is the literal translation of particles of affir - mation and negation. Studies of the translation of Russian particles of affirmation and negation into French show that in many cases the translation is correct. Thus, for example, the answer to a question of the type: 'have you read this book?' is always translated correctly. Yes, I have read. Oui, je l'ai lu. No, I have not read. Non , je ne l'ai pas lu. - on the other hand, in a number of other cases a literal transla - tion is a frequent occurrence. Thus the affirmative answer to a ques – tion of the type : "You have not read this book?' in the majority of ca- ses is wrong (compare: 'No, I have read ') : with Non, je l'ai lu. with Cui, je l'ai lu instead of S1, je l'ai lu. A negative answer is occasionally also incorrectly translated (compare: 'No, I have not read'): with Cui, je ne l'ai pas lu instead of Non, je ne l'ai pas lu. The absence of a literal translation in the first group of cases (the question put in affirmative form) makes it possible to assume, the presence of the identical differential indicat or in both languages. The consistent error in the second group of cases (the question put in nega- tive form) indicates that in one of these languages there is no differ- ential indicator inherent in the other language , and namely: in Russian there is no indicator (inherent in French) by means of which it is pos- sible to contrast the particles 'yes' and 'no.' Let us examine corresponding situations in greater detail. - 176 - 1) A does not know whether B has read the book, and he turns to B with a question. . Affirmative and negative answers are equally probable for him, and each of these answers Will be uniquely defined, since the answer Will either affirm or negate the statement made by A. A: "You have read this book?" B: "Yes, I have read" (Qui, je l'ai iu). "No, "I have not read" (Non, je ne l'ai pas lu). In this case we have a clear contrast between the affirmative and negative particles in both languages. As a result they can be used on an isolated basis in the elliptical clause : "You have read this book?" "Yes" (Oui). "No" (Non). - 2) A does not know whether B has read the book, but makes an as- sumption which B can either affirm or deny. In either case A expects a uniquely defined answer. However, the problem is made more complicated by the fact that A's statement contains a negation to which B must re- fer. The elementary meanings forming the meaning of B's statement can easily be traced in the Russian examples. "You have not read this book?" "No, I have read it." It is impossible simply to answer 'no', ' since the isolated use of no would alter the meaning of the reply. Apparently, there must be - two components in an arrirmaviv- answer to a question containing a nega — tion : - : a refutation of the negation; b ) affirmation. Let us examine a negative reply to the same question : "You have not read this book?" "Yes, I have not read." The use of a Single 'yes' is inadequate for a uniquely defined - 177 – * - ^. statement. Here We also find two *** ***. - 8. arrirmation of negat, • § a negative rjijº ion; Conse quently, B's statement in reply to a Question containi n negation consists of two COIIlp Onent S : 9, a a) affirmation or **, *enent; refutation of the he&ation contained in the pre ) an affirmative * negative reply to the Question In Frenc - º ** Particles by means of which a negation is refuted confirmed are clearly contrasted : Orº Vous n 'avez pas lu ce 1 - i: 2 Si, (§ l'ai lu). vre": Non, (Je ne l'ai pas lu). The p possibility of the isolated *ść of 'Si! and 'non in ll tical clau - - an ellip- Se is borne out by the fact that each of these particl €S pos - Sesses the two necessary components of meaning The particle 1 si : a) o refutes the negatio In ; b i tive reply. ) gives an affirma – The particle 'non f : : a) confirms the ne ation : reply. 8ation; b) gives a negative Thus, wit • . ºf • 2 h respect to the negation contained in the previous Stat ment, th - - - 6 - ° particles "sit and 'non' are Contrasted with respect to th indicator, refutation — affirmation € In Russi ... • - Sian there is no contrasting of the particles 'yes' and "no Or, the ba . O * * * **ator. Compare you have not gene this book? with N - OOK'. o, I have read (affirmative *Ply) and 'No, I have not d : y - 2 I’ea **h 'Yes, I have not read (negative reply) The pa ! particle 'no' was used in the affirmative reply refuting th negation - 2 e 8 On contained in the previous Statement, but it does not t yet pro- the two Components of meaning required for a uniquely defined repl - y e – 178 - Therefore, in this case the use of the single particle 'no' is not enough. A complete answer is necessary: "No, I have read it. ' The particles 'no' and 'yes' are possible in the negative reply ('No, I have not read. Yes, I have not read'). Here the particle 'no' contains : a) a confirmation of the negation contained in the previous state- ment ; - b) gives a negative reply. The use of 'no' is quite sufficient for a uniquely defined reply: "You have not read this book?" "No. 11 The particle 'yes' contains confirmation of the negation, but does not give a negative reply. For a uniquely defined reply the use of 'yes' is not enough. A complete answer is necessary: "Yes, I have not read. ' The lack of such a differential indicator in Russian is the main reason for the literal translation in the categories being considered here. § 27. SIMPLIFYING TRANSLATION A significant difference between a simplifying (as well as between an exact and "an adequate) translation and those considered earlier lies in the fact that here the translation process is broken down int O Sev- eral stages, and namely, the actual translation and the various (both preliminary and subsequent) modulations and transformations. * Of significant importance for the the ory of translation 1; the fact that at least for any pair of European languages the nucleus of each of the languages (from the standpoint of transformation) and the common portion of the systems of these languages (from the standpoint of correspondence) are sufficiently close to make it possible in the majority of cases to achieve a simplifying translation through a number of modulations and simple transformations. - l'79 - Let us now examine examples of how the actual translation is com- bined with transformations and modulations in a simplifying translation. Let the following word combination be translated into German : ! At the height of the foreign military intervention. . . ' The difficulty in finding correspondence arises in the translation Of the word at the height. ' The dictionary has a translation for the group 'to be at the height' – in vollem Gange sein - but the combina - tion in vollem Gange der Intervention is not usable in German • The cited phrase can be translated as follows. Let 95 replace the words at the very height of the foreign intervention' with the words 1 when the foreign intervention was at its very height. ' As We have Seen , for the group to be at the height' is a one-to-one correspondence, i. e. , we again fall within the common portion of the two languages. In Our €28- ample we have carried out a transformation prior to the translation. We have identified ($21) a translation a S simplifying if the ele – ment of the language through Some transformation or modulation is brought to a form falling within the common portion of the two language systems and is then translated in accordance with the rules of COr’re S - pondence. The term simplifying translation is employed because the phrase in the target language is constructed in the manne” of a certain phrase in the source language 5 which, of course, cannot stylistically impoverish .* that phrase . Nevertheless, the simplified translation is a higher de- gree of translation than the interlinear translation. A simplifying translation is encountered with particular frequency in the translation of verbal formations, i.e., precisely those construc - tions obtained by transformations from verbal phrases. Of interest from this standpoint is the translation of the following sentence containing a participial construction which can be rendered only after simplification: - 180 - Doch nach vollzogener Aurt . . . -- Zeilung der Neuaufteilung mit noch heftigerem §:#":...:” der Kampf um die Of course, it is impossible to translate: 'after the concluded division of the world , " but the translation: 'after the conclusion of the division of the world' does not satisfy the stylistic requirements because of the accumulation of nouns disrupting the style of the ori- ginal. This sentence is best translated by the wordinate clause : I strue...": tºº...:*:::"...º.º...".º.º..." f e ! tº Orº Ce (Or : "However, after the division of the world had been com— pleted, the Struggle intensity º. 39, f'Or its reorganization flared with even greater Another example: Einst; #. ############ Relativitätstheorie durch Winkelsumme in einem Dreieck ..º.º.º.º.º. The interlinear translation of the underscored words: 'long before Einstein's formulation of the theory of relativity is stylistically not suitable. - In such cases the prepositional group inalesting the actor is fre- quently converted into the subject , while the verbal noun is changed into the predicate of the Subordinate clause: . ... betteº Einstein formulated the theory of relativity (or eve - German nºta: : theory of relativity), the grº" angles in a triangle need not º the sum of the It might seem that the translation of § tions through nontransformeč *stan eX transformed German construc – pressions is a result of the features encountered in the relationships between the Se two languages. However, this is not the case. It is indicative of this rest that in the translation from Russian into German (and into a number of other western European languages), vice versa, the nontransformed construc- tions in the target language correspond to the transformed RuS Sian Con- Structions. - 181 - Let us take the phrase The Soviet Union is ready to Cooperate with the western states, given adherence to the principles of recipro- city and implementation of assumed responsibilities. ' The following translation does not correspond to the stylistic norm: German unter der Bedingung der Einhaltung der Prinzipien der Ge – gense it igkeit und der trfüllung übernommener Verpflichtungen; nor does the French Sous la condition du respect des principes de réciprocité et de l'execution des engagements contractes. . . - The problem here lies in the fact that the genitive case in German is used in written form, and the abundance of words in the genitive case enhances the impression of booklshness. The repetition of the same article in der ... der . . . der . . . in German or the repetition of the preposition with the article du . . . de la in French causes the phrase to become stylistically cumbersome. The basic method of translating such groups of words, as a rule , is based on the fact that Russian gen- itive constructions are formed on the Stems of verbal nouns. In this case we have the words "adherence " and 'implementation. ' The principle of translation reduces to simplification, i.e. , the construction is transformed into a phrase of the nucleus so that the verbal noun could be translated by a verb. In our phrase, the group : 'given the implemen- tation is replaced by "given the condition that . . . will be implemented, " and then the corresponding subordinate clause is constructed in the tar- get language : 4 - German: Die Sowjetunion ist bereit mit den westlichen Staaten zu- Sammenzuarbeiten unter der Bedingung, daSS die Prinzipien der Gegen- seitigkeit eingehalten und übernommenne Verpflichtungen erfüllt werden. French: L'Union Sovićtique est préte a collaborer avec les pays occidentaux a condition que solent respectes les principes de récipro- cité et executes les engagements contractes. Simplifying translation is extremely important in teaching trans- lation, since the knowledge to employ this form of translation protects - 182 – the trainee from error. In this connection we find the following exam- ple of considerable interest. As is well known, the dative case Cannot be attributed to a noun group. This occurs rather frequently in Russian: ! the report to the XXIInd Party Congress, ' 'the award of the interna- tional Lenin prize to an outstanding worker in the field of social aſ — fairs, ' 'a monument to Mayakovskiy, ' 'resistance on the part of certain governments to others. " In translation we frequently employ prepos l— tions. One cannot Say Der Rechenschaftsbericht dem Parteitag; it must } necessarily be rendered as an den Parteitag: 'resistance to the decree ' must be rendered as der Widerstand gegen das Diktat, rather than as Wi- , derstand dem Diktat . - - Occasionally it is difficult here to find an appropriate preposi- tion, and the best means in this case is reconstruction of the type that we examined. It is impossible to say Politik der Gegentiberstellung der einen Staaten den anderen. In the translation it is necessary here to use a more expanded form, and namely: Politik, die darin be Steht , die. einen Staaten den anderen gegeliberzustellen [A policy which consists in the confrontation of certain states with others l. Yet another example: "The actors of the Little. Theatre after ex- tensive discuss 1 on at a general meeting and at a session of the artists ' council on the result of activities in the past season decided to. ' The underscored words cannot be translated as nach der Erörterung in der allgemeinen Versammlung und 1n der sitzung des Künstler rates der Ergeb — nis Se . . . The word order must be changed, but in this Specific case it 15 simplest of all to use the verbal form: Die Schauspieler des Kleinen Theaters zogen auf der Versammlung und in der Sitzung des Künstler rates nach einer eingehenden Erörterung das Fazit der verlaufenen Spiel Zeit und be Schlos Sen. . . - 183 – Frequently a simplifying translation is used in the translation of Russian sentence S involving abstract nouns ending with -ocine. Compare "Here are a number of facts indicating the need to apply Soviet exper- ience." We can transform this sentence in order to achieve a subordin- ate construction (compare "Here are several facts which indicate that . . . ") and translate, respectively, into French (compare Wolla quel- ques faits qui prouvent que l'expérience sovićtique est d'une grande portee. ... the German Hier einige Tatsachen, die davon zeugen, wie wert- voll (von welchem grossen Wert) die Sowjetischen Erfahrungen sind. Compare also the following expression: "We must take into consider- ation the experience of recent years, indicating the unilateral applica – tion cf the UN charter' (the same as '... indicating that the UN charter has been applied unilaterally'), which in simplified translation into French yields: I: faut tenir compte de l'expérience de ces dernieres années, at- testant aue l' on a fait un usage unilatéral de la charte de l'ONU; into German: Man muss die Erfahrungen der letzten Jahre bertick- sichtiger, die davon zeugen, dass die UN-Statuten einseitig angewandt wurden. A simplifying translation of nouns ending in -ocme is particular- ly necessary when they are formed by means of prefixes (compare 'devoid [without ) of ideas, ' 'without promise, ' 'antiparty spirit,' etc.). Com- pare the sentence: "The tribunal admitted the charges to be without ba- sia' which is the same as 'admitted that the charges were baseless " is translated into French as Le tribunal a reconnu que l'accusation était mal fondée. Compare also: 'The statements to the effect that the African states were nonviable were refuted ' and the French translation: Les affirma- tions prétendant que les Etats africains n'étaient pas viables ont été refutees. - 184 - Another construction frequently calling for transformation in or— der to achieve a simplifying translation is the passive construction. - We know that if not out of grammatical considerations, then at least because of stylistic reasons, we are frequently compelled to give pre- ference to active constructions in translation. Compare the following example: Der Strohertrag wird durch die späte Stickstoffgabe weniger bee influsst. 'The yield of hay is less af- fected by the delayed application of nitrogen fertilizers. ' In English texts we frequently encounter passive constructions, whereas in Russian the passive voice is used much less frequently. In translation the passive constructions are frequently replaced by active constructions (Aristov, pages 21–22): He was given a book by the TIpobeccop man emy Khary. professor. - The room was entered by Heckonibro heſiobek Bouvià several men. B KoMHaty. The book was much spoken Cty Mehth Mhoro robophnh about by the students. o6 Stoà KHHrē. Our delegation was wel- JIronh mphberctBoBaſih ha- comed by the people. luy Henerallhko. The passive construction is frequently replaced by the active in translation from Russian into French: "In the newspaper was printed the Statement of the Council of Minister's " — Le Journal publie un arrêt du Conseil des Ministres. . . ; "To German militarism was assigned a leading role in the plans of worldwide reaction' - Les projets de la réaction mondiale réservent au militarisme allemand un rôle de premiere impor- tance. f . - - tº “. simplifying translation in vocabulary is expressed in the subst1- tution of words which do not have a One-to-One correspondence in the target language by a word which has the same basic meaning, but lacks one or more of the elementary senses contained in the word of the Source language. This simplification is characteristic of beginner translators - 185 - and, in general, of the initial stage in the study of a foreign language. In this stage the speaker may make proper use of the vocabulary (and the grammatical structures) of the language being studied, but he lacks synonymic variation. Thus, as an equivalent for the word to end he Uls e S the French finir, se terminer, but he makes absolutely no use of such words as about ir a , Se Solder par, etc. This correst, although by no means fluent speech, is a positive phenomenon corresponding to sim- plifying translation in terms of the means employed to accomplish the translation. It has been noted in the theory of machine translation that the application of the principles of simplifying translation can reduce the Volume of the dictionary in synthesis by replacing groups. Of . Synonyms with a single "word of basic meaning," known as the cover word in the theory of machine translation. Let us present the following example (Pimsleur, pages 12–13): the German word Schwer has a number of correspondences in English, and namely: ein schwerer Stein – a das Dach ist schwer von heavy stone Schnee — the roof is ein Schwerer Wein — a laden with snow strong wine . - er hat ein schweres Schick- g sal — he has an unfor- . tunate fate - The American linguist Pimsleur holds that in all of these meanings the word heavy ban be employed, since it exhibits the most general mean- ing. In a number of cases the use of this word causes the phrase to be - come stylistically rough, as for example: the roof is heavy With Snow ; however, the phrase remains completely understandable and its communica- tive value is not reduced. A particular form of simplifying translation is the translation from a natural language into some artificial language such as , for ex- - 186 - ample, the language of symbolic logic. Let uS examine the following statement: l. Every Dutchman knows how to ride a bicycle. 2. Some Dutchmen know how to ride a bicycle. Phrases l and 2 can be transformed as follows: #. Every object that is a Dutchman knows how to ride a bicycle. . Everyone who is a Dutchman knows how to ride a bicycle. 5. Some object which is a Dutchman knows how to ride a bicycle. 6. Someone who is a Dutchman knows how to ride a bicycle. Similar versions are sometimes referred to as "stylistic" (Suszko, 18, Quine, 1950, pages 83–86, 127-128), underscoring that the sense re- mains unchanged. In our terminology we used transformation, which in general form may be reflected as follows: l) Each X exhibits a given property, *- everyone who is X exhibits the given property. * +. - 2) Some X exhibits a given property, - someone who is X exhibits the given property. These transformations are particularly important because the phra- ses for which phrases of the type given in 3 through 6 frequently en- countered in real languages may be regarded as transformations, are con– veniently denoted in symbolic logic as follows: 1) (v.) P (x) iS read as : - * * "For any x—P (x)', '': 'W is known as the quantor of community 2) (3.) P(x) is read as: "There exists an x such that P(x)': 3 is known as the quant or of existence. It must be pointed out that reduction to a form convenient from the standpoint of symbolization is not always accomplished as easily as in the previous example, since the structures of the real language are frequently structural homonyms, i.e. , the identical construction is de- rived as a result of the application of various transformations. Let us introduce the following denotations: bl is the property to be light blue Bu is the property to be a book Lö is the property to be a lion Rt 1s the property to be a beast of prey St is the property of being *;a - 187 - -kº. * : In addition, we use the losis symbols & and D. The sign & is a Sign or conjunction; A & B denotes A and B; D is a sign of implication. A D B denotes: from A follows B. - - In this case the German sentences: 1. Alles ist blau. - 2. Es gibt ein blaues Buch. 3. Der Löwe ist ein Raubtier.. 4. Der Löwe ist satt. 5. Alle Löwen sind Raubtiere, must be written in the artificial language of symbolic logic in the following manner: . (Wx) (bl (x)) ... • . (3-) (bl (x) & Bu (x) ) . (v.) (Lö (x) → R* (x)) . (ax) (Lö (x)) & St (x) ) . (W.) (Lö, (x) D Rt (x)) i We see that the form-varying statements (3 and 5) Correspond to a Single formula in the artificial language. This circumstance, however, does not disturb the Symbolization, because the various linguistic forms of Statements 3 and 5 may be regarded as the result of the application of various transformations to the same phrase : 'Everyone who is a lion is a beast of prey. ' Jeder, der ein löwe ist, er ist ein Raubtler. Considerably more important iS the circumstance that the same arti- cle in 3 and l; corresponds to Various quantors, and it is impossible to say, on the basis of the form of the statement which transformation of the phrase it represents. However, it cannot be sala that this diffi- culty is insurmountable (see $13 an analysis in connection with the article). $28. EXACT TRANSLATION Thus we have considered cases in which we changed over to an ele- ment Gl of the common portion for the translation of some element O by - applying some transformation or modulation, finding for this new ele- - 188 - ment a correspondence with B. It goes without saying that the element f8 was Situated in the common portion. In addition, by means Of some trans- formation or modulation it is possible from here to proceed to an ele – ment Pl lying beyond the limits of the common portion. Let us examine the following phrase: "Despite the assurances of , official circles in the USA that they are allegedly concerned primarily with the economic wellbeing of the countries of the East , the allocations of military "aid" to these countries continued to increase. ' Let us translate this phrase in the form of a Simplifying translation: LeS milieux officiels des USA prétendent qu'ils sont préoccupés tout d'abord par la prospér- ité Économique des pays orientaux. Cependant, les crédits militaires accordés a ces pays augmentent sans cesse. This translation is the re- sult of the transformation of a Russian phrase which, for the sake of brevity, is being omitted here. It corresponds to the norm of the French language ; however, it may be subjected to further transformations: LeS gour vernants amércains voudraient faire accroire qu'ils ont avant tout le Souci du bien-ètre des pays orientaux. Cependant, les crédits mili- taires alloués a ces pays he cessent d'augmenter; A entendre leS mil- ieux officiels américains, ils ne font que veiller S Ull? le blen-etre des pays orientaux. Or, les credits qu'on accorde a ceux-ci a titre a'assistance militaire ne cessent d'augmenter. . . Of course, other transformations a rºe possible. We are dealing here With the fact that With transformations leading to the common portion of two systems, many elements correspond to a single element, whereas in the transition from the common portion to the entire language the op- posite is found to be the case : an entire class of units corresponds to a single unit. This is fully understandable. In developed languages a given thought can be variously expressed and for this reas on a number of transformations and modulations may correspond to any . . . º "a se. — l89 – Thus, some set of translation phrases may correspond to each phrase of the original. Each such phrase in the language of the translation will be referred to as a variant of the translation. - - The average number of translation variants for a single phrase of text may , by the way, serve as a measure of the difficulty involved in the translation of texts from the given sublanguage. It is interesting to note that from this standpoint the Sublanguages are treated precise - ly in that sequence in which they are generally considered in the theory Of translation in connection with the problem of evaluating accuracy (for example, see L. N. Sobolev). The least number of variants per phrase of text is found in commercial texts; the number is considerably great- er in sociopolitical texts and, finally, the maximum number of variants per phrase is found in literary text. Difficulty in translation here 1s associated with the fact that to obtain an adequate translation (in the sense in which this was defined in $211) the translation must be famili- ar with the greatest possible number of potential variant S. Let us note that , as a rule , the textbooks on translation fail to devote sufficient attention to the problem of accuracy as we have de- fined it, i.e., they do not devote sufficient attention to the exist – ence of a set of variants rendering the same meaning. Generally, having completed a certain phrase the authors are satisfied with a single var– iant which they regard as the best. However, if the author tries sever- al variants, as a rule, he does so only to reject all save one. An out - standing exception in this regard is the textbook by Aristov (see "Bib- liography") in which for virtually every phrase several equally prob- able variants are presented, 1.e. , identically exact in our definition of the word. Here are several examples. In speaking of the translation of passive forms, the author indicates that the phrase This question was discussed at the conference can be translated in the following ways : - 190 - 3rot Bompoc 6bin oocyxneh Ha Kohgºepehuhh. Stor Bonpoc occyxmanca Ha Kohgºepehuhh. 3rot Bompoc oºcykmann ha Kohipepehu Hà. Kehºepehuna oºcyſtina stor Bonpoc. Particularly many variants are encountered in an examination of correspondences to English verbal formations such as , for example, in the section on passive participles: When heated, mercuric oxide decomposes rather readily. rt is pointed out that the translation of such phrases can be ac- complished in three ways, e.g. : - When heated, solids expand little as compared with liquids. verbal Adverb: ‘Harpesaacs, thepuble Bellectea paciuhphotch He3H8th H- Tembho no cpabhehhio c XHAKHMH . - Subordinate Clause: "Korna Teepage beinectBa Harpebaiotch, ohā pºcºP" ca. Heahauhtembho, no cpabhehkiko c XHIlkHM* . Noun with Preposition : Tipu Harpenahan thepmae Tena pacuapāiorch Heaha” Tembho, no cpabhehhio C XHAKHMH'. The same applies to other forms such as , for example : Having successfully solved ycnelliho peuwb 3a/lauy pac- the problem of splitting uſenſiehha atomhoro RJ1pa, the atomic nucleus, the yuehbie cMorniſ paspado- scientists could work out Tath cnocoöbi hcmojib30- the methods of utilizing Ba HH% aroº 3Heprlih. O atomic energy. * - Tocine toro, Kak yuehble f ycneuho peulhih 3alaqy pacuennehha atomhoro ampa, ohh cMornh paspa- 6otatb criocoöhl McIIojib- 3obahha atomhoſt 9Hep- - THH. After closing the circuit 3amkhya uenb, MB, H3Meph- we measured the resist- Jih compothalehhe. Orº 2 IICe. Iloc.ne 3ambi Ka HHR Heſi H Mb, H3Mephinh compotlib.ne- HHe. Orº - 191 – *. Splitting, the atom is a difficult task. - f Transformers are used for changing d.c. into a.c. © \, Iſocyte Toro, kak Mbi 3am- Khyna IIemb, MH HaMephinh compothamehne. Pacuſemineh He atoma HB918- etch Tpythoſ. 3alaueh. Orº Pacinemath atom - TPYA". Hayſ 38Ja‘īā. MeRºi TpahcºpopMatoph mph - P. juia, mpeodpasoba- HHA noctonhhoro Toka B nepemehhhh. Or Tpancpopuatops npº. HaroTczi AJIH TOTO, mpeo6pasobHeath nocrº #Huárok a nepemehhhh. The translation of the infinitive in the subject is in Russian by means of the infinitive or a noun : To transmit the image by radio has become a pract tical possibility not so long ago. Being placed in the open * #. rusts and deterio- rates. IIepenaeate H306pakehºe ño pañño crano npakº accomplished ueckh BosmoxHBlm Cpab", HHTenbHo Hejſa BHO- IIepenaua H306parkehāº no pańwo crana mpaktºeckh Boamoxhoºl cpabhºtelib" Ho Heſla BHO- Bynyun momeuſehºbº a orkplatom Bosſlyxe, kelet so pxabeet H paapyuaet- CŞ. Korna (eciſii) xeneao Home" uſaerca ha orºplutoM Boº Ayxe, oho pikabeer M Pas- pyutaetch. The problem of exact translation become S particularly urgent in the synthesis of a phrase from the set of units in th 4 * - - - language. This problem was studied in great detail by N. N. Leont 'yeva e intermediate in connection with computer synthesis (Leont 'yeva, pages 1-9). Given the requirement to synthesize a Russian phrase who can be rendered by the words: "Our dissatisfaction with this was brought about by the hostility of Zurbagan. In the intermediate language this phase corresponds to some S*- - 192 - se meaning quence of Symbols ABCDEF with the following rules of word synthesis A B C - D We are dissatisfied this because OUII? dissatisfied Such because SO brought about . Zurbagan Zurbagan's to bring about - having been brought abou E F hostile hostility negative relationship behaves negatively The rules of synthesis state that from each set of correspondences one is selected and then the corresponding sequence of words is checked to make certain that it does not contradict the limitations imposed on Russian. It is easy to of variants in the 1. We are not 2. We are not tively. 3. We aré not H. We are not tively. 5. We are not 6. We are not attitude. see that of 600 theoretically possible combinations given sequence (223.525) only ll, will be correct : Satisfied with this, because Zurbagan is hostile. satisfied with this, because Zurbagan behaves nega – Satisfied with this because Zurbagan is hostile. satisfied with this , because Zurbagan behaves nega- satisfied with this because of Zurbagan's hostility. satisfied with this, because of Zurbagan's negative 7. Our dissatisfaction with this is brought about by the fact that Zurbagan is hostile. w -" 8. Our dissatisfaction with this is brought about by the fact that Zurbagan is behaving negatively. 9. Our dissatisfaction with this is brought about by Zurbagan's hostility. 10. Our dissatisfaction with this is brought about by the negative attitude of Zurbagan. ll. Our dissatisfaction with this is brought about by Zurbagan's hostility. 12. Our dissatisfaction with this is brought about by Zurbagan's negative attitude. - 193 – 13. We are dissatisfied with this, because of the hostility of Zurbagan. ... . . . . . 14. We are dissatisfied with this because of the negative attitude of Zurbagan. - It is clear that with a larger segment of text the number of such variants increases and the problem of Studying these variants becomes e Ver” more complex. . e Since, generally speaking, there may be many exact translations, we have the ract that phrases in the Source language and phrases in that same language, obtained through reverse translation, do not coin- cide. (Reverse translation, as is well known, involves the translation of some text into another language, the resulting text subsequently be - ing translated again into the source language). - Considerable interest has been shown in experiments involving the repeated process of reverse translation. The most precise procedure was employed in the experiment described by van der Pol. In this experiment a text excerpt was translated from English into French and then by an– other translator from French into English; a third translator then trans – lated the English into French and, finally, a fourth translator trans- lated the French into English. It turned out that the general meaning of the text was retained in its entirety, but the style was altered be - yond recognition (van der Pol, 1956). Note should be taken of the fact that in the event that not two but several languages are involved in the experiment , retention of meaning cannot be assured. Thus, we know of the experiment conducted by the Dane Jensen (see "On Translation, page 282); a short story (700 words) in Danish was selected and translated into Swedish and then into German; it was then translated into French and , finally, retrans- lated into Danish. Although each of the translators exhibited brilliant command of the corresponding pairs of languages, the final result coin- - 1914 – cided with that which is observed in the familiar childrens game "Ru- mor." This result is explained by the fact that in view of the diverse categorization the successive modulations and transformations alter meaning slightly, the shift in meaning, undetected in each pair of lan– guages, accumulates and becomes noticeable on transition to new systems of categorization. - It should also be pointed out that the study of the variants of exact translation apparently makes it possible to approach the problem of excessive text material from another side (see §12), and thus to ap- proach the problem of text flexibility as recently posed by A. N. Kolmo – gor OV. The information measure of which we spoke in $12 (Formula 1) is in essence the measure of the number of various texts of a given length N that can be constructed in a given language. Thus, in an alphabet consisting of 32 letters it is possible, if there are no restrictions, to construct 232N texts. The limitations imposed on a language consid— erably reduce the degree of diversity. However, it is significant that by no means all of the diversity permitted in a language is used to transmit sense. Investigating poetic speech, Kolmogorov drew our atten- tion to the fact that the requirements of rhyme and rhythm significant – ly restrict the degree of diversity. If a poet is able to express a GG2 – quired thought within the limits of the given restrictions , this can be explained by the fact that a great portion of the diversity is employed in a language not to transmit some content , but to devise various forms of expressing a given content. In this connection Kolmogorov proposed the decomposition of the entropy of a language into two components: a) the measure of diversity employed to render extralinguistic (semantic) information and b) linguistic entropy. As we know, entropy has been defined on the basis of experiments - 195 - involving the guessing of text continuation. This method, proposed by K. Shannon, the founder of the theory of information, is based on the hypothesis that each individual fluent in a given language possesses the knowledge to determine at any given point in a text the likelihood of the āppearance at this point of the following text elements. This method was improved by Kolmogorov (see Rychkova, 1961) who proposed the use of the guessing method in connection with translations: the indiv- idual being tested, knowing the original and thus the sense of the text, must guess the continuation of the translated text. Apparently, these experiments make possible precise evaluations of the measure of flexi- bility in the target language. § 29. ADEQUATE TRANSLATION - Let us now consider a "text" in which in the place of each phrase of the original we have written the entire set of possible translation variants. What principle is to be applied in the construction of the actual translation text? Is it or is it not possible arbitrarily to make a selection on the basis of a single variant? Experience in trans - lation produces a negative reply to this question. It is clear that the translation variants are not indifferent with respect to each other and the selection of one of the variants limits the selection of the others (the selection of words of different sty- listic shading, word order, the Selection of conjunctions, articles, \ etc. , all that which we referred to in $24 as the rules of conjunction). The definition of adequate translation, as given in $24, reduces to the following: An adequate translation is that selection of variants in which : a) each variant is obtained on the basis of the given rules of correspondence, with these rules retaining at least a common meaning, as well as the set of stylistic characteristics; - 196 - contradict the rules of conjunc- - b) the selected variants do not tion. - . . . . - We have already stated that an interlinear translation (a word- for-word translation as a particular case) may be treated as an ade- quate translation. As a matter of fact, item a) of our definition here is always satisfied. However, it is not out of the question that a word- by-word translation will contradict the rules of conjunction and thus not be adequate. - Let us take note of the fact that our definition of adequacy gen- eralizes the rules of translation for words of multiple meaning — rules known to everyone – and namely that it is necessary in the selection of a translation variant to know not only the semantic correspondence but its capacity to enter into combination s with others (both lexical and grammatical). - Thus for each phrase in a translation we have two significant meas — urements: its prototype in the source language and the form of the Sur- rounding phrases in the target language. A significant deviation of the proposed definition is the fact that it allows a multiplicity of solutions within the limit S of ade- quate translation, which is in accord with the current practice of translation. As a matter of fact, the multiplicity Of adequate trans – lations for a Single text provides the flexibility that is required for those various rules of conjunctiºn (the organization of the complex linguistic entity) which have existed at various times. Let us also take note of the fact that the set of solutions may be an empty set, i.e., there will be no adequate correspondence. After these general remarks, let us undertake the Specific analy- Let us begin with several of the simplest examples. — l97 - As is well known, Russian participial phrases may be translated into German in various ways , and namely, by subordinate clauses, pre- positional phrases, participial phrases (less frequently), etc. The most frequent correspondence in German is the definitive sub- ordinate clause. However, context may require Other translations. Let us analyge the translation of the following sentence : "The foreign policy of the African countries, based on the princi- ples of peaceful coexistence, pronounced by the Bandung Conference of Asian and African Nations, is supported by the Soviet Union and peace- loving States. " - º . Let us examine the translation: Die Aussenpolitik afrikanischer Staaten, die auf den Prinzipien der friedlichen Koexistenz beruht, die von der Bandung Konferenz der Länder Asiens und Afrikas verkündet worden sind, findet bei der Sowjet- union und allen friedliebenden Staaten Unterstützung. This is an exact translation; however, we can see that the repeti- tion of the conjunction die causes the Style to become cumbersome. It might be possible to use another form of the relative pronoun by re- placing the verb , e.g. : der die Prinzipien der Bandungkonferenz Zugrun- deliegen. This translation is also exact ; however, even this transla - , tion contradicts the laws of conjunction which state that it is not permissible to make double use of a Single conjunction in varying func– tionS. • Frequently the following method is employed : in the place of a second relative clause, extensive use is made in German of a compara- tº enrage; Prinzipien der Koexistenz, wie sie von !... verkündet wor- den Sind. However, the following difficulty arises here. The subject die Aussenpolitik is somewhat too far removed from the predicate findet Sep- arated from the subject by two subordinate clauses. Let us now take note of the fact that the Russian participial phrase frequently assumes a causal definition. This is what happens in the given sentence. Employ - - 198 - ing this circumstance, we make use of the subordinate Clause with da which can be placed after the predicate. Thus We found the following variant: Die Aussenpolitik afrikanischer Staaten findet bei der Sow jetun - ion und allen friedliebenden Staaten Unterstützung, da dieser Politik die Prinzipien der friedlichen Koexistenz Zugrunde liegen, wie sie von lºanaung-Konferenz der Länder Asiens und Afrikas verkündet worden This variant , first of all, is exact and, secondly, corresponds to the laws of conjunction in the target language. It may therefore be regarded as adequate. : We should take note of the fact that in many cases the participle is dropped entirely. Let us take the combination 'the amount of mach- inery used per meter of mine face. Let us assume that this combination of words is contained in a sentence which, without this last phrase , contains many relative subordinate clauses or participles, the use of yet another contradicting the laws of conjunction. Although the trans – lation into German: Die Anzahl von Mechanismen die auf ein Meter Abbau- Strecke ent fallen is an exact translation, it is replaced by another : Die Anzahr der Mechanismen je (pro) Meter Abbaustrecke. ** This pertains to such other participles as being, 'representing, etc. For example: the Soviet delegation arrived in Delhi, a major in- dustrial center' is translated into German as follows: Die SCW jet is che Delegation kam nach Delhi, einem grossen Industriezentrum e e º h ~, Compare arso the French translation: La delegation sovićtique est arrivée a Delhi, important centre industriel... It frequently is the case that it is precisely the laws of con- junction that govern the use of Russian verbal adverbs. Compare the French translation of the following text: 'They hold. . . that in these countries it is necessary to establish Strong prowestern governments whose military leaders can be regarded as obedient. — lS9 – "Having attained this goal, the military establishment, as before, casts its lot with. . . " Ils pensent qu'il faudrait installer dans ces pays de S gouverne - ments prooccidentaux dont les chefs militaires seraient obedients. Aus- Si bien les militaristes misent-ils sur. . . The conjunction of the two cited phrases, expressed in Russian by a Verbal adverb is expressed in the French text through aussi bien. A translator concerned with translation into German employs the rules of conjunction when he is confronted with the problem of using a Subjunctive of indirect speech during the synthesis. We have seen in the section on simplifying translation that in many . cases the noun group is changed into a subordinate Clause , and that in the majority of cases such a subordinate clause is introduced by a con- Junction (for example, when using dać in a translation into German). Let us now take note of the fact that the conjunction (daſ) is one of the most commonly used words, even without the fact of the previous Case , and if a translator makes extensive use of the transformation Of a noun group into a subordinate clause, he will again encounter diffi- culties with the laws of conjunction: having eliminated an accumulation of nouns, he will produce a text where the conjunction (daſ) is used S Omewhat too often. The problem therefore arises as to the manner of avoiding repetition of the conjunction. The Sub junctive of indirect Speech is one' such solution. / * ... " Let us take the following phrase: When the rebellion WaS Sup- Pressed, the Minister announced to the House of Commons that the very thought of war on the part of one great state against another, because of the small country which mounted the rebellion, would be insanity and that only the nearsightedness of the rebels can be held responsible if any among them intertained this thought. The group 'the thought of war on the part of one great State - 200 - against another, because of a small country' in translation into German is transformed as follows: der blosse Gedanke daran, dass eine Gros S - macht wegen eines kleinen Staates, der den Aufstand begann, gegen eine andere Grossmacht Krieg führen wirde . . . The words 'responsible if in German correspond to: daran schuld Sel, das S. Now the translation contains four dad (two for the transla– tion of the existing conjunctions and two as a result of the reconstruc- tion) which pertain to various words, thus contradicting the laws of conjunction. Let us now consider a variant using the subjunctive of in- direct speech: Als der Aufstand unterdrückt worden war, erklärte der Minister im Unterhaus, der blosse Gedanke da ran 2 dass eine Grossmacht wegen eines kleinen Staates, der den Aufstand begann, gegen eine andere Grossmaciſt Krieg füren wirde, wäre wahns innig, gewesen, und nur die Kurzsichtig - keit der Aufständischen allein sei daran schuld, das . . . This variant no longer contradicts the laws of conjunction and may be regarded as adequate. Let us now consider another example: "The newspaper carried an ar- ticle on the policy of the imperialists in the Near East. The article says that this policy will lead to intensification of tension. Accord- ing to the words of the author of this article, the maneuvers of the English fleet in the vicinity of the Suez Canal are not accidental. The article points out that all Arab patriots are incensed by the behavior of the imperialists, etc. ' The laws of conjunction in German require that sº intracter, Words. Such as " in the Words of the correspondent Or 'as Stated in the newspaper, or 'the author shows that..." must be eliminated in the translation; it is enough in the first sentence to Say: Eine Zeitung. . . hat einen Artikel úber die englische Politik im Nahen Osten veröffentlicht, and then to use the subjunctive of indirect Speech. It is interesting to note that in translation from German into Rus — - 201 - sian the opposite is found to be the case. It is precisely the laws of conjunction that compel us to use various introductory words in Russian in translation of sentences using the Sub junctive of indirect speech. Let uS analyze the following excerpt from this point of view: Sehr charakteristisch daftir 1st die Meinung der Kollegen im Büro für Erfindungswesen. Sie sind der Auffas sung, bei ihnen wäre die Tech- nik so weit entwickelt, dass man jetzt, keine bedeutenden Werbes serungs- vorschläge, mehr erwarten kann. Zum anderen wären nicht immer die Mit- tel für die Einführung und Prämierung der Worschläge vorhanden. Folg- lich werden auch nicht regelmässig Themen über Schwerpunkte in einzel- nen Abteilungen popularisiert, um die Initiative aller Werktätigen auf die die Planerfullung hindernden Probleme zu lenken. In the second sentence the indication of indirect speech is con- tained not only in the form of the subjunctive , but in the words Sie sind der Auffassung: "They believe that their technology has developed to such a point, that it is unlikely that any Significant efficiency suggestions will be forthcoming. ' In the following sentence it is only the subjunctive that lndi- cates indirect speech. However, in translation into Russian the 1ndica – tive mood is employed. If there are no additional indicators, the read- er will not recognize whose opinion being expressed in the sentence : 'Moreover, facilities are not always available for the introduc – tion and rewarding of efficiency proposals. ' It is most likely , he will even think, that this is the opinion of the author of the article , since this is followed by a phrase in which the author's own thoughts are expressed (note that the verb here ap- pears in the indicative ioca). Therefore, in order to avoid duality of meaning, we introduce the words in their opinion, ' 'they say' or, if we want to indicate irony, "do you see , ' 'he says, ' etc. For example: 'Moreover, they say, 'facilities are not always available for the introduction and rewarding of efficiency proposals. ' Let us now deal with the problem of the so-called 'actual articu- - 202 - lation of a phrase' which is so important for translation. In translation from German the relationship between the actual ar- ticulation of a phrase and the use of the passive (three-term) construc- tion is very important. This is explained by the fact that in German not only the position of the declined form of the verb is fixed, but also by the fact that the direct object is very rarely placed in the first position. A normal word order is preferred here. Therefore, if for the Sake of continuity with the previous (or, in general, to iden- tify the logical predicate) it becomes necessary to place at the begin- ning of a sentence a group of words which in active construction would be the direct object, use is made of a passive construction in which this group becomes the subject. - Let us analyze the following excerpt from this point of view : Die Gültigkeit dieses Gesetzes, das Schon früher Cavendish auf Grund anderer Beobachtungen aufgestellt hatte , wurde von Coulomb durch genaue Versuche mittels einer elektrischen Drehwaage festgestellt (West – phal, Physik, S, 231). - Would an active construction be possible here? No, for the reason that the word order: Die Gültigkeit dieses Gesetzes hat Coulomb . . . restgestellt is somewhat too emotional for the given discussion and , in general, is not characteristic of the German language. However, with ordinary word order in active construction: Coulomb hat die Gültigkeit dieses Gesetzes fest gestellt the relationship to the previous is en- tirely lost (in the previous phrase we were speaking of the law , rather than of Coulomb) and the general lege of the discussion is disrupted, thus violating the laws of conjunction. After all the words die Gültig- keit des Gesetzes, make up the logical subject, while Coulomb is here the logical predicate. The passive construction is used in order to overcome the contra — diction between the logical and the grammatical structure of the Sen- – 203 – tence. - - - .- With regard to Russian, the direct object can very properly stand in first position and it is therefore entirely unnecessary to use the passive voice. We can translate the second sentence as follows: 'This law (better than: the validity of this law) was established by Coulomb. . . " - - Let us now analyze an analogous example in connection with trans- lation from Russian into German. - - ** us take the sentence : "In the Hall of columns of the House of Unions was convened a conference devoted to the International Day of Women. The conference was opened by the chairman of the Antifascist Com- mittee of Soviet Women 1. The translation Of the first clause presents In O articulty. However, how are we to translate the second clause? First of all, let us examine the interlinear translation: Die Versammlung eröffnete die Vorsitzende... This is a poor translation primarily because the Subject-object relationships are not clear in this case nor is it clear Which is the subject and which the object. These coincide in form and for this real- S On the clause is felt to be Stylistically poor and contradicting the laws of conjunction. If the order is changed in the words Die Vors it - Zende eróffnete die Versammlung, the translation Would be quite correct if the second Clause were translated Separately, out of context, but in this case the translation violates the laws Of conjunctiºn between the second clause and the first. - In our example it is the word 'conference which is the logical Subject - the basis of the communication which sets everything into mo- tion - that relates this clause with the previous , and for this reason this word must be in first position. The Second portion of the clause – - 204 - "The chairman of the Antifascist committee of Soviet Women..." repre- sents a new element and therefore, if this new element assumed the in- itial position, the logical relationship would be disrupted. . - To satisfy both requirements, i.e. , to ensure clarity of the log1- cal relationship between the given clause and the previous clause and to ensure the clarity of the subject–object relationships, it is necessary in this case to use the passive construction: Die Versammlung wurde von derº versitzenden eröffnet. Compare also the French: La réunion a été ouverte par... ... - - With this word order we obtain complete correspondence between the logical structure of the two clauses , although grammatically they are different. Logically the structure here is the same : the given is in first position, the new is in second position. * Of greatest importance for the comprehension of the essence of ade- quate translation as we have defined it thus becomes the problem of architectonics” or the Syntax of a complex linguistic entity. It is not only the relationships within a clause that are impor- tant here, but primarily the relationships between the clauses and even larger units. --- The transition to a new paragraph, for example, is a positive means of articulating a text. It is indicative that articulation into paragraphs in each language occurs uniquely. For example, let us €2Caſſ - ; ine the text "Appear to the Peoples of the Entire World" adopted by bpre- conference Of Representatives of communist and Labor Parties (Moscow, November, 1960). Our eye is immediately caught in a comparison of the Russian text of this Appeal with its French translation by the ract that the number of paragraphs is different : in the Russian text there are 69, while in the French text there are 55. The conjunction of the French text, as evident from the following excerpts where corresponding - 205 - means dº expression are italicized, is accomplished by means of con- Let us take the excerpt: junctions and the italicized constructions, unlike the Russian text in "In 1935, fascist Italy attacked Abyssinia and enslaved it. She - - [Italy] attacked without any justification or cause. ' Russian text% 1) Whence comes the threat to peace in the entir World? . - All governments Speak of peace, but it is not words that count, but deeds. - As before, the reaction- ary, monopolistic and milli- tary circles of the imperial- • ist countries are the Organ- izers, and initiators of ag- gressive wars at the present time. sº 2) Is a strong peace in the entire world possible? We, the Communists, re- ply: war is not unavoidable . . . 3) The possibility of averting war follows from the real facts of the new State. of affairs in the world. The worldwide socialist system is becoming an increas- ingly decisive factor in the current state of affairs. We have already said that the rendition of stylistic characteris- #. which word order serves as an important means of conjunction. French text ++ D'où vient la menace qui pèse sur la paix dans le Inonde 2 • * Tous les gouvernements parlent, certes, de la paix ; mais ce ne sont pas les pa- roles qui comptent, ce sont les ac Gr, aujourd'hui comme hier, ee sont les milieux réactionnaires mo- nopolistes et militaristes des pays impérialistes qui sontles organisateurs et les initia- teurs des guerres d'agression. Une paix durable est-elle possible dans le monde P Nous, communistes, ré- pondons : Oui 1 La guerre .n'est pas fatale... - La possibilité de conjurer la guerre découle des donnés de la nouvelle conjoncture mondiale. c Le système socialiste mondial devient, en effet, la force de plus en plus dé. e cisive des temps modernes. The rules of analysis prevailing within the limits of this excerpt established that here a) we have repetition and b) two clauses within the excerpt , joined to each other. In synthesis it is by no means necessary for the repetition to occur at the same position; it is important that it be reproduced with- in the limits of the excerpt, and the conjunction that is characteris- tic of the target language must be accomplished. These requirements are satisfied, for example, by the translation into French: En 1935 l'Italie fasciste attagua et Soumit l'Ethiopie. Cela sans aucune raison, sans aucun motif. $30. MEANS OF ACCOMPLISHING THE TRANSLATION PROCESS AND THE FUNCTIONS OF LANGUAGE IN THE ACT OF COMMUNICATION Having defined the various means of accomplishing the translation process, we assumed that the basic Significance of communication re- mained invariant. This contention requires refinement. When we spoke of the invariance of the primary meaning, we had in mind the correlation of the sign with the reality which is accomplished by this primary meaning. This selection of the invariant emanates from the basic establishment of the communication act, i.e., from the commun- 1catiye function of language. However, as we have seen §11), other func- tions of language are carried out in the communication act and these , respectively, modify the accomplishment of the translation process. Thus, in the practice of translation it becomes necessary to take into consideration the unique features concerning the Speaker in commercial tics required for inadequate translation assumes the transmission Of the principles of construction or "stylistic figures" of the original speech (juridical text, scientific literature, official documentation, (see $23). etc.), in publicity, in literary speech. * Commercial texts involve pri- marily an intellectual function of language – the referrent function; - 206 - - - 207 – in publicity texts, we are dealing primarily with the addressee func- tion (the vocative case, the imperative function); in literary texts we are dealing with the referent function, the contact function, the sender function and the addressee function (with various forms and gen- res within literary Speech are identified in dependence on some func - tion). The need to take into consideration the multilateral act of speech intercourse in the practice of translation has long been recog- nized. Moreover, it was proposed to formulate the very definition of the process of translation as a function of the goals sought to be ac- complished by the translation ("On Translation," page 19). Of interest is the characteristic of the basic historically governed types of trans- lation proposed by Thieme. For translations of official and sacred texts (pertaining to reli- gion) the following is characteristic : ". . . scrupulous accuracy. This requirement relegates to a secondary level the respect for standards in the language into which the translation is being accomplished: language must be subordinate" (Thieme, page 58). Literary translations are divided by the author into 3 subtype S : a) the rendition of the content of the original (for example, the translation of a poem through the use of prose, as is frequently done in France); b) the creation of a self-contained work in the native language by the free reformulation of the thoughts and contents of the original ; c) equivalent reproduction of the original in the native language. With regard to translations of commercial text, they are charac- terized primarily by the referent function. With this classification we can compare the types of translation identified by Savory : 1) translations of informational text. - 208 - 2) translation of text in which "... that which is described 1s more important than the manner in which it is described" ("...the matter is more important than the manner" savory, page 21). 3), translations of text in which the manner of expressing the con- tent is as important or even more important than that which it des – cribes. - 4) translations of scientific-technical literature. Finally, we should make mention of Sobolev's attempt to establish Specifics in the accomplishment of the translation process as a function of the purpose sought by the author of the text. Proceeding from the con- tention that thoughts, feelings and tendencies expressed in a given lan- guage can be expressed in other languages, Sobolev maintained that there is no uniform criterion of translation accuracy: "... the measure of ac- curacy varies as a function of translation purpose , the nature of the text being translated and as a function of the reader for whom the text is intended. Therefore, specific criteria of accuracy are more conven- iently considered separately , as a function of the nature of the texts: . literary, publicity and commercial." (Sobolev, 1959, page ll;3). A trans- lation will be accurate from the literary standpoint if 1t retains the poetic idea of the original; it will be accurate from the standpoint of publicity if correspondence is achieved for each detail of the desired political effect ; and it will be considered accurate from the standpoint of documentation if foreign terms are replaced by tersºn the native language and if the standards of the native language are maintained. One cannot help but recognize the validity dº the statements cited here — they are based on a rich international tradition of translation in various forms of literature. However, one would like to refine these thoughts. As a matter of fact, how are we to understand the statement that the measure of accuracy varies as a function of tº Carlslation pur’— - 209 - w pose? It seems to us that the means of accomplishing the translation process, including an accurate translation, are uniform for all forms of text, S. Let us recall that the units of translation in the source and target languages are compared on the basis of the set of semantic, gram- matical and stylistic characteristics. In this case it is by no means necessary for each of these units always to carry the same weight. We can assume the opposite to be the case: the weight of a single Semantic, syntactic Orº stylistic indicator will vary as a function of the func – tion of the language in the communication act. Consequently, it is pos- sible to achieve any type of translation (i.e., literal, interlinear, exact and adequate) and at the same time to embody a given function of language. As an illustration of the foregoing let us use the example pre- sented in the forementioned work by Sobolev, and namely, the last lines of the tragedy "Othello" in the scientific-philological prosaic trans- lation by M. Morozov and the poetic translation by B. Pasternak. K.' Lodovico (to Iago): O Spartan dog, more fell than anguish, hunger or the sea. Look on the tragic loading of this bed; this is thy work. The object poisons sight, let it be hid. Gratiano, keep the house and seize upon the fortunes of the Moor, for they succeed to you. (To Cas- sig). To you, lord governor, remains the censure of this hellish villain, the time, the place, the torture. O, enforce it., Myself Will straight aboard to Venice and with heavy heart this heavy act relate to the Doge and to the Senate. Lodovico (to Iago) Cnaptanckas codaka, uto 6ypa, mop a romon mpen todoñº B3rj1HhH. Ha ºp. rpys noctenh stoh. Teoh padora. ChnB. Her cworpeth. . . yxpoſite ux. 3añMHre now, Tpanbaho, – 210 - Bctynhre Bo Bianehbe scem Hoºpom, Octabulimch ot Mabpa. Bhi – Hacienhhk. Bam, rocnothh ſipabhrenh, othaio CyAhth anomen. Błudephte Kapy, Ha3Haubre Rehb h cohepuihre kashb. A 8 mpo 3ty ropekylo y ſpary C taxenbum cepanew monoxy Cehary. Lodovico O Spartan dog, More fell than anguish, hunger, or the sea 1 Look on the tragic loading of this bed; [To Iago. This is thy work: the object poisons sight:— Let it be hid. Gratiano, keep the house, And seize upon the fortunes of the Moor, For they succeed to you. — To you, lord Governor, Remains the censure of , this hellish villain; The time, the place, the torture, — O, enforce it ! Myself will Straight aboard; and, to the state, This heavy act with heavy heart relate. [Exeunt. In commenting on these two translations, Sobolev noted that both translations are exact , differing only in the criteria of accuracy: ". . . one translator, Morozov, set himself the goal of providing a sci- entifically validated interpretation of "Othello" in the translation into Russian, while the other translator, Pasternak, presented a liter- ary translation of "othello" (Ibid, page ll!!!). - This conception, it seems to us, is inaccurate and contradictory. The author regards it possible to speak of Morozov's translation as a precise scientific translation precisely for the reason that it pre- Sent S literal lexical-equivalents dº the English text , which seem strans- in the Russian. With this approach the very concept of accuracy loses its meaning. It seems to us that the cited segment of the Morozov translation Should be defined as an adequate documentary translation on the basis of one evaluation and as a literal-literary translation on the basis of the other evaluation, 1.e., having been obtained as a re- Sult of carrying out the translation process as a literal translation, – 211 - T-m- identifying and retaining the text function as the primary function in the documentary ("scripture") approach. It goes without saying that this segment of the tragedy could have been translated in a Simplified manner (as is the case in modified editions), retaining the speaker function, i.e., the expressive function of language in the text. In that case, we would define this translation as a simplified-literary translation. However, we regard the Poetic translation by Pasternak as a translation arrived at by the adequate approach, with identification and retention of the Speaker function, i.e., we define this translation as an adequate-literary translation. It is obvious that Within the lim- its of the requirements that the identification and retention of the eXpre S Sive function be retained, the cited Segment might have been translated not adequately, but freely. Thus it is proposed that the classification based on the means Of a CC Omplishment be combined with the Classification pertaining to lan- £uage ſanction. (See the table on page *). It is clear that this is only an approximate approach which can be improved in many respects : it is possible to refine the forms of the 4- *- text 3 in dependence on the functions, more Completely to take into con- S ideration the interaction of the functions, particularly in literary S Cee ch, etc. 5: However, it is important for us only to note the basic trends in the interaction of the functions and the means of a C C Cmplishment. It is Indicative that if in the reality function (the intellectual function), all types of translation with the exception of free translation are possible, in the case of the text function only interlinear translation is possible or , if the Categorization does not Coincide , literal trans- lation (translation of thought processes). * In the case of other func- tions interesting limitations also arise, but we Will not dwell on these – 212 - this point. | l; 5 lvemo- Verãonka Ha TeKCT Ha Heſſ- || Wotahobra . otnpabh- cTBhtenb- || Ha Tekct | ****P** Teng Han- iina | Hocre (scarpan. (ºr pecarºy- & JIH- (ºneſioban has peubs) º JIoxect- 33 lili H peubs) ckah pe Behhaji peub>) 6 -l sº hhtepjihhe- hitepnhhe- apho- apho-cak- *- *º- 7. IIenoboſ 8 pajibhbh 9 hhtepuihhe- aphbū 6ykBanbhbiſh GykBanbho-6ykBanbho- 6ykBanbho- 6yKBanbho- Heſioboſ, cakpanb- nyCºnh- xyno- lO kk hbiſh IIHCTH- XKecTBeh- l2 ueckhā l3 hbih 11, ſipoulakouth? - - ympo- ympo- ymp "ºome. Lilakollie- IIIa KOLIle- l; JIenoboſ; myönh- xyllo- LIHCTH- XKeCT- 16. ueckhā Behhbiſh (ranan- lT Tall H%2 anº, 2O Touhbiſh TOUIHO- - - *º- 2O HeſioBoš anekBathbā lanekBatho- - a Hekbatho- aſleRBaTHO- JIenobok myönh- xyllo- 2l - LIHCTH- XKeCT-. 22 ueckwä24 Behhuà BombH buſ -> - BOJIbHO- BOJIb HO- - - myóJih- xyno- 25 26 incth- 27 ×ect- . ueckhā Beh 14 blkſ omplishment; 2) function; 3) reality, function ("commer- *:::::A; l; ) text #ºi speech"); 5) addressee nction ("publicity, advertising speech"); 6) sender and addressee II] xt function ("literary jºš. 7) interlinear; 8) #:::::::::::1 rcial; 9) interlinear-liturgical; iO) literal; ll) *:::::::::::: 3. literal-liturgical; 13) literal-publicity; ll;) literal-li ºit e simplifying; 16) simplifying-cºmmercial: l'7) :*:::::p. gº Simplifying-literary ("adaptation ); 19) exact ; 20) ::::::::::::- al.; 21) adequate; 22) adequate-commercial; 23) adequate-pu iñºr .) adequate-literary; 25) free; 26) free-publicity; 27) free- ary. we was point out only that the combination of the classification means of accomplement with the classification by function of lan- lage makes it possible theoretically to formulate the problem posed as – 213 - early as 1796 by wilhelm von Humboldt : how to combine in a translation the integrity of the original and the unique nature of the translation language? (Humboldt, Brief). In the terminology of the table that we have presented here, this problem, extensively considered in connection with literary translation (Mounin), reduces to the selection of a type of translation which takes into consideration the function of language and the means of accomplishment . $31. THE TIME FACTOR IN THE PROCESS OF TRANSLATION (WRITTEN AND ORAL TRANSLATION) * - . It is possible to distinguish translations not only on the basis of the table shown in $30 (for example, "literary," "publicity," "docu- mentary-commercial, "technical" and others), but also as a function of whether the translation is being accomplished orally or in writing. Al- though intuitively the need for this last division is sensed rather strongly, the formal criteria on the basis of which this division is accomplished remain unclear. The fact that a certain statement is ren- dered during the course of the translation into the target language in graphical or accoustical form makes essentially no difference from the standpoint of the translation procedure. With regard to the nature of the translated material and the function of language , it is highly un- likely that in a modern language it will be possible to delineate a clear boundary between the sublanguage of oral speech and the sublan- guage of written speech. The difference between written and oral translation is governed by something else. It seems to us that this difference should be sought in the fact that the written communication is ordered in space (from left to right), while the oral communication is ordered in time. The time factor in general plays a Significant role in any translation, Since it is not only the "orderedness" of the signals that is important, but the – 214 - time during which a certain communication must be retained. An artistic-literary work and its translation retain their value and currency for a long time. An "academic" scientific translation re- tains its lifespan, although to a smaller extent than works of art, particularly in recent times. How ever, the reprocessing of scientific information, including translation, is transient in nature: after a cer – tain period of time – in certain branches of science this involves months - the value of the scientific information and, consequently, of the translation diminishes sharply and then drops to zero. - However, in this case we are not interested in the time of informa – tion Storage , i.e. , the time factor in the broad sense of the word, but we are interested in the fact that an Oral communication involves 3. Span of time, and in principle time is irreversible. We have seen that the units of translation vary as to volume and Sometimes require rather lengthy text to determine correspondence. We also know that a unit of translation cannot be rendered in parts. Hence it follows that during the translation process the intermediate informa – tion must be accumulated somewhere for use in the transmission of speci- fic units. This always takes place , and up to this point we assumed that this point was self-evident. Thus, in the translation of a segment of a literary text or article from a newspaper, the translator carries out a preliminary analysis of the text and – consciously or subcon- ./ y sciously - isolates the units, writes out words, aid combinations, their translations, and also drafts preliminary versions of translations for more complex excerpts, etc., frequently turning back. In the transla - tion of an oral communication something else takes place , since other – wise the translation would be entirely impossible. However, in view of the irreversibility of time, the load on memory is increased, and we have in mind here the so-called "high-speed memory" (see $18). – 215 - When we Speak of a written translation we will be referring to a process in which the volume of the memory is virtually unrestricted, i.e. , it will be possible "to memorize" (record or retain in some Other manner, for example, to remain in the text to which we repeatedly re- turn) a segment of text of any length. When we speak of oral transla - tion we will refer to that process in which only a limited number of units (not exceeding the volume of the high-speed memory of a human be — ing) can be simultaneously memorized. We have seen that if a device exhibits a memory of limited speed ('bral speech"), it cannot synthesize the series of phrases which a de – vice with a high-speed memory ("written speech") can synthesize. We al- so know that if the second device (simulating written speech) can ac- complish both regressive and progressive constructions, the first can accomplish any progressive constructions and regressive constructions limited in length. Thus, for example, a typical regressive construction, not characteristic of Spoken language , is a widespread attribute of the German language (das erweiterte Attribut). In Russian such are clauses with a participial attribute , preceding that which is being modified , as for example : . Of glass , green and dull Glowing at his bench, Stood the lamp, that dear present, That the bride had given to the groom. , ("The Young Guard," No. 12, 1960 page 9) In connection with the above-considered features of memory, the spoken language – and consequently oral translation – presupposes a cer – tain automatic procedure in speech (synthesis) and perception (analy- sis). In speech the automatic approach involves the fact that each sub- sequent word is a function of the series of preceding words (the choice of words is limited), this selection having to be accomplished in the shortest period of time. The selection is accomplished from among the – 216 - most generally used, shortest and most frequent words (and grammatical structures). It is clear that this dictionary is limited in volume only by a part of speech. Translation in this case, as a rule , is interlin- ear" Olſº simplifying, although it is not out of the question that the rules of conjunction are considered (adequate translation). - With perception (analysis) the automatic approach consists in the fact that each speech element is recognized and perceived on the basis Of the preceding. Here consideration is given to extralinguistic in- formation perceived visually (gesticulation) and to the situation in which the entire spoken intercourse is taking place. Manu- script [Footnotes Page NO. 152 The author's would like to take this opportunity to express their gratitude to I.Yu. Shekhter for pointing out the impor- tance of the given point in the formalization of the concept of "adequate translation." 158 In this connection, the concept of isogrammatism is of inter- est (Golāb, 1960). 16O The term "simplifying translation," as well as other terms used here, are not intended for purposes of evaluation. Simplification should be understood only as the Selection of a form more convenient with respect to the target language. l6|| "... the exclusivity of meanings for words is characteristic of the scientific method. Since similar definitions have been introduced by scientists in all countries, the translation is facilitated by the uniquely defined correspondence of scien- tific dictionaries" (Brillouin, page 14). - 167 In this connection, let us cite the following statement by I. R. Gal perin: "Indeed, what is this "rendition of precise thought patterns?" Slavish copying of the foreign structure of a phrase into one 's native language? If "rendering thought patterns" is understood to mean destruction of the standards for the language into which the translation is being accom- plished for the sake of retaining the syntax of the language from which the translation is being accomplished, then why is it necessary to refer to this as the "rendering of thought patterns," and not simply as a grammatical error? The trans- iation by the girl student in Chekov's story "An Expensive - 217 — w Lesson"; 'He was walking on the street and met the master of - - 212 - Let us recall that the term "literal" in our approach is not his friend, and said: where are you going, seeing your face - used as an evaluation term (in the sense of "poor"), but so pale, it pains me" must also be recognized as "rendition * . • points to the objective need for expanding the common portion. of thought patterns," if "rendition of thought patterns" is - - With regard to the evaluations "good-poor," "successful—un- understood to refer to some degree of coercion on the struc - successful," they may be applied with regård to any type of ture of the language" (Gal 'perin, page 136). - - translation. - 168 Compare with the German, where, as a rule, suffix formations correspond to compound words such as Ideengehalt, Ideen- reichtum, Ideenreinheit, Partelstandpunkt, Parteigelst. The term Parteilichkeit "has come into use recently. 173 In the terminology of the theory of language contacts, liter- al translation is described as an interference phenomenon (Rozentsveyg, 1963). 179 It is therefore understandable why in the teaching of trans - lation we are required to use not only a two-language dic- tionary (which is sufficient only in the case of interlinear translation), but primarily the single-language dictionaries of the two languages , i.e. , dictionaries of definitions and dictionaries of synonyms. In this connection, let us recall Shcherba 's thought on the need to include definitions and syn- onyms in a two-language translation dictionary (Shcherba, 1959, pages 54–92). 2O5* The principles of replacing the active voice with a passive construction in the case of unclear subject–object relation- ships had been worked out by that school of latinist teachers of which we spoke in §8. Cauer, for example, cites the fol- lowing phrase from Horace : notum si callida verbum reddiderit iunctura novum and says that the literal translation of this phrase into German: wenn ein bekanntes Wort eine geschickte Verbindung neu gemacht , hat ('when a familiar word is cleverly combined in a new way") is incomprehensible, proposing the following alteration: wenn ein bekanntes Wort durch geschikte Verbindung neu geworden ist ('when a familiar word becomes comprehensible because of clever combination of the words'). 2O5** The indication of the importance of this problem for the the o – ry of translation is contained in the series "Stylistic com— parée," where the term charniere corresponds to our term "con– junction" (Vinay, Darbelnet, page 230). . * - - 2O6% "Programmnyye dokumenty bor by za mir, demokratiyu i Sotsia — ºr lizm" ["Program Documents from the Struggle for Peace, Demo- cracy and Socialism"], Moscow, Gospolitizdat [State Publish- ing House for Political Literature | 1961. - 2O6++ "L'Humanité," 10, XII, 1960. 2O7 This problem is covered in detail by Fedorov (Chapter 6 en- titled "Various Forms of Translation as a Function of the Genre Type of the Translated Material"). For this reason we restrict ourselves here only to a few comments. - 218 – – 219 — Chapter 5 CERTAIN SPECIAL PROBLEMS IN THE THEORY OF HUMAN AND MACHINE TRANSIATION §32. THE PROBLEMS OF THE DICTIONARY Let us now take a look at how translation units are identified. Since generally the intermediate language is not explicitly present, correspondence is simply established between a text segment from the source language and the text segment of the target language. It is pri- marily the dictionary that is used to identify the translation units. By the way, a somewhat neglectful attitude toward the two-language dictionary has arisen in the theory of translation and among Some pro- fessional translators. It is even maintained that the dictionary is more of a hindrance than an aid in translation. Neverthele SS , the dic – tionary is the primary list of given correspondences with which we have to deal in translation practice. The following Statement by a contem- porary theoretician of translation is worthy of note: "a dictionary may be defined as a tool which makes translation possible without knowledge of either of the languages" (Cary, 1956 c, page ll). The lack of faith in dictionaries , although a paradox, is brought: . . about by their advantages. A dictionary provides Somewhat too many translation variants, and translation requires that each word be trans- lated in only a single way. Generally speaking, this is to some extent conditional. It is quite easy to 1magine a Situation in which each word is rendered in a given text by a set of various correspondences. * Very instructive from the theoretical point of view is an analysis - 220 - of such a possibility. - Let us imagine a dictionary in which each set of variants is marked off by a vertical line, and that within these Set S the basic transla - tion is written in first position, the remaining dictionary variants written in parentheses. Let us take the following Russian sentence: 'Successful experiments in machine translation have now been car- ried out . " - - We obtain the following translation of this sentence into German: . In (nach, zu, um)||gegenwärtig (echt, wirklich, wahrhaft, richtig) || Zeit Lerfolgreichſ durchgeführt (geführt, verwirk- licht, gelegt)|| Erfahrung (Experiment, Versuch)|| iber (ent- lang, durch, O, nach)|| Maschinen — ||Ubersetzung (Ober- führung, Versetzung, Anweisung)||. - With a translation into French, we would obtain the following: A (dans, en)|présent (véritable, vrai, authentigue)| temps (heure, saison)|| avec succès (heureusement)|| entrepris (con- duits, menés, accompagnés, passés, construits, installés, posés)|expériences (essais)|| Sur (par, selon, suivant, d'après, à raison de) ||mécanique (à la machine) traduction (version, thème, transfert, mutation, aiguillage, mandat postal)||. Even such a translation makes it possible for a person knowing Ger– man. Or French to understand the eS Sence of the matter and to translate this phrase without knowledge of languages , however, it is clear that this is a poor translation, if for no other reason than that it gives too much , it contains much that is useless and interfering (see Mickles – sen). •: $t .. Let us now See in what way Š this useless information can be elim- inated. Any adequate dictionary establishes not only the words, put the frequent combinations of words , i.e. , the microcontext of the corres- ponding Words. For example, for the word 'present we may find the com- bination 'at the present time rendered as gegenwärtig, actuellement ; - 221 - for the word "experiment' we may find the combination: to carry out an experiment [Experiment durchröhren) and the combination experiment On [Experiment mit (experiences sur, essais de)]. - At the same time, a good dictionary lists the combination 'postal transfer, ['money order") rendered in German as Postanweisung, and in French as mandat poste, and since we did not encounter this combination, the corresponding translations will be eliminated. Thus, we will give only a single translation, and this will be the most probable transla- tion, if the word was not encountered in the specific combinations, and especially if the word WaS encountered in a specific atotionary-estab- lished combination with the given word or with the word of the given gr’Oup . -- Another means of reducing unneeded information involves reference to the macrocontext. However, this is done only in those cases in which it is possible to establish the corresponding sublanguage for the in- atviauai translations. Thus, with the German word Kolben we can, for example, present the following given in the Great German Dictionary edited by E. A. Meyer (Moscow, 1934): (military); rifle butt {###".:*: º ear (of corn) (zoology) the buzzing mechanism (in insects) Let uS note that with this solution to the problem it is advisable to establish a specific hief archy between the criteria of combinability and the area of application. Preference should be given to the former. As we stated in $20, one and the same word may have a common ("literal") and terminological meaning, i.e. , it may be used in one sense as a term, and in another Serºl Se a S ar) ordinary word. Let us take the German word Lösung for the French solution. In chemical terminology these words have the meaning of solution. However, we may find the combination Lö – – 222 - Sung der Aufgabe , la solution du probleme ('solution of the problem ) used in a chemical text as well. By referring to the indication (chem- 1stry), a translatºr may employ the translation the solution [dissolv. ing] of the problem." In order to prevent this, he must first of all examine all of the possible phrases in which the given word might be employed and only then should he refer to the indication (chemistry). We see that the basic problem in text analysis is the elimination of homonymy or the multiplicity of word meanings. In the examples present - ed here, homonymy is reduced through consideration of the context in- volving the hononymous or variously defined word and on the basis of its inclusion in some sublanguage. Occasionally, however, the elimina- tion of homonymy calls for interpretation of the text, i.e. , reference to reality and determination of what is meant. Thus, for example, the dual meaning of the phrase : "The tests of this weapon threatened peace [world]," resulting from the homonymic word "Mup' (English peace and world; French, paix and monde; German, Frieden and Welt) cannot always be eliminated through context analysis. This may also be said of the phrases : 'These student groups held two general [mutual meetings to discuss problems of scientific work" (English, Eeneral and mutual; French, général and commun; German, allgemein and gemeinsam). Note Should be taken of the fact that the complete definition of all terms of phrases is virtually impossible in a dictionary. It is therefore frequently advisable to forloº another course, and namely, to provide in the very structure of the intermediate language an ident if i– cation of elementary meanings which would make it possible to ascertain the uniquely defined units without reference to the microcontext. The following example (Tesniere, page 2911) is instructive from this stand- point. Let us examine the French translation of the German phrase: – 223 – German: . Er tieff' die Zeitung holen. French: Il encoya churcher le journal. . Russian: He sent for the paper. It is of course possible to establish the following correspondence from the dictionary: Holen lassen – aller chercher. However, if the three elementary definitions are identified in the intermediate language, We will have the following: ! A aux. an. * an. ā- 3JI. *. * | * * * * º . - - envoyer derde c new agus lasen, - holen ID pyck asuk | E * sa A) Intermediate language; 1) Elementary causal meaning; 2) elementary meaning of walking; 3) Elementary meaning of finding something; B) French; C) German; D) Russian; E) send for. Here we can establish the individual correspondences between en- voyer and meanings l and 2; between chercher and meaning 3; between lassen and meaning l; between holen and meanings 2 and 3. It is clear that this method of establishing correspondences is more economical. One of the most frequent situations in analysis involves the fact that for a certain segment of text no correspondence whatsoever, can be *A* found in a 41ctionary. This can be explained by Such various ractors a S , f'Orº example: 3. In inadequate dictionary, or it can be explained by the fact that the given word has not yet come into any dictionary (the so- called neologism), * or by the fact that the given text segment does not represent a word (formulas, proper names, diagrams, etc.). In this case it is useful to distinguish two cases from the Stand- point of the theory of translation: – 224 – A. The word is reproduced from context (compare the phrase "trans- lator's surmise" encountered frequently in literature). This is possi- ble in those cases in which a phrase contains excess information (com- pare $l.2). Let us analyze the following example (Zvegintsev, pages 161-162). Let us take a series of phrases in which a single word has been left out , e.g. : - We went to various. . . This . . . has been covered thoroughly by me. Let us discuss this problem from various. . . . Proceeding from the context of the phrases, selecting words, as Zvegintsev says, "on the principle of subject relationship," it is enough simply to fill in the corresponding gaps. * Let us note, however, that this restoration assumes that the remaining words have been placed in correspondence with the groupings of elementary meanings which thus make it possible more or less uniquely to define the omitted words. This restoration is particularly important in the case of simultaneous translation (see $31). B. The word is not restored from context. The following cases are . then possible. - :* a) The state of analysis leaves the word unchanged and transfers it in this form into the intermediate language and then, in the synthe- sis stage, to the target language. We will refer to all such text ele- 'A * ments as formulas. f b) In the analysis stage correspondence is established on a letter- for-letter basis between the letter's or groups of letters of the trans- lated word and the corresponding phonemes. Thus, the intermediate lan- guage includes correspondences between letters and phonemes, in addi- tion to correspondences between words. This principle is used primarily to analyze proper names, geo- - 225 - graphic names, In 81mmeS of institutions, newspapers, Steamships, etc. , words denoting so-called social realities, as well as neologisme, 1.e., words pertaining to the so-called vocabulary without equivalents. For additional information on the words of this group see $37. $33. CONFIGURATIONAL ANALYSIS We have already seen that a dictionary may be able to define not only individual words, but combinations of words as well. Let us have a language containing exactly 1002 words (the so-called Basic-English is an example bf a language much like this) and let the length of a phrase in this language always equal 5. Now, let the dictionary contain all possible combinations containing five various words. The number Of Such combinations in principle is equal to: Cº., -998x999×1000×1001x1002=1000000000000000–10” Even if it is assumed that from 100 possible combinations of five words only one makes sense, we obtain an astronomical figure of lo!3. However, the following is interesting. If we were able to prepare a dic – tionary containing lo!3 various combinations, translation from this hy- pothetical language might be accomplished without any grammar: each phrase would be looked up in the dictionary and its translation found there. This example demonstrates the position of grammar and its role in translation. Grammatical rules make it possible to translate a combina - tion of Several words in the case in which this combination is not six- en in the dictionary. In other words, it might be said that grammar makes it possible to reduce the number and length of translation units. In this case we are of course interested in Such rules as would be applicable to the largest possible number of combinations. It is clear that if the rule is applicable only to a single word , it will differ 1n no way from the simple definition of the given combination in the dic – * ... - - -e 226 º tionary. Hence it follows that , unlike a conventional grammar which provides several exceptions for virtually every rule and thus becomes difficult for anyone using its facilities, in the theory of translation we are interested in a grammar which provides rules which know of no exceptions, all of the exceptions being defined in connection with the corresponding Words in the dictionary, i.e. , the translation unit S are expanded. By the way, this is the point of view of theoretical linguistics. Thus, Shcherba held that all mass phenomena should be reflected in the grammar, whereas everything solitary is a fact for the dictionary. The basic distinction between grammar and dictionary lies in the fact that the grammar pertains not to individual words but tº U entire classes of Words. Let us now consider some hypothetical language. Let this language contain a very large (but finite) number of words, but let all of the Se be divided into 32 grammatical classes (this is a completely realistic - figure, the number of classes in certain languages being even smaller). Now let us look at all of the possible combinations of five words in - cluded in the various classes. Of these we will have : Cº-28×29×30×31 ×32 = 30% = 20 000 000 - If we assume that of every 20 combinations only one has (grammati- cal) meaning, we come up with one million possible combinations. Let us assume that a dictionary has been compiled with this grammar. (It is considerably easier to accomplish this than in the previous case). In this case the translation would be accomplished in the following In 3 rºllſle ſº for each word an index would be looked up in the dictionary to denote membership in a specific class and for a resulting sequence of indices there would be a table to indicate the grammatical structure in the target language and the corresponding Words would be included into this – 227 – structure in accordance with spectric rules, thus providing us with a translation. We see that a reduction in the volume of the dictionary would result in more complicated rules; however, these rules are appli- cable to a great number of phrases and their entry into the dictionary occupies considerably less space than the entries of lo!3 combinatious. However, it is possible to take yet another important Step. It was demonstrated in §18 that every phrase can be formed by expanding con- figurations corresponding to word combinations or phrases [syntagmas 1. Without resort to any formulas it is possible to calculate that the number of configurations in the language varies from 50 to 200. It is very easy to memorize these configurations. The idea of machine translation by means of configurations was first employed by T. N. Moloshnaya in Setting up rules for translation from English (Moloshnaya, 1957, page 92ff). Let us first illustrate this method in simplified form by means of an example involving the translation of a German Sentence : 1 2 3 4 5 : 6 7 8 9 Ein sehr junger Arbeiter erzählte uns von Seinen Arbeits- mºn - - - Let a correspondence be found for each word in the dictionary, and let there be information on the membership in a corresponding clas S. In addition, we have a list of configurations such as, for example: a) adverb + altered form of adjective; b) altered form of adjective + noun; c) possessive adjective + noun; d) noun + s + absence of space + noun written with lower-case let — ter; . e) verb + pronoun + noun in indirect case ; f) verb + preposition + noun; – 228 – g) noun in nominative case + verb. Having C Ompared the information for words 2 and 3 with the list of configurations, we establish that they form configuration a. In this configuration it is only the basic word (3) that is taken into consid— eration. Words 3 and l; will be combined in precisely the same manner (4 remains); with 9 and 10 combined (10 remains); with 8 and 10 C OIII – bined (10 remains); with 5 and 6 combined (5 remains); with 5, 7 and 10 combined (5 remains); and finally, with l, l; and 5 combined. This method makes it possible to establish all of the relationships between the Words. Once all of these relationships have been established, the RuS Sian correspondences can be found. It goes without saying that this must take place in reverse order, at least for the reas on that we do not know , for example, in what form the word junger is to be translated until we have translated the word Arbeiter. Thus, first correspondence is found for the configuration g, then for the configuration f, etc. This, in general, is the method of translation by individual C OIl - figurations. Its fundamental features reduce to the following: 1) the minimum combinations of words are identified (configura- tions); - - . . . - 2) the dependent member of the configuration is not taken into consideration in the subsequent analysis; - - 3) the identification of the configurations is accomplished in a', '' rigorously defined sequence; 4) for each source-language configuration a corresponding config- uration is sought in the target language; - 5) the formulation of the configurations in the target language is accomplished in reverse sequence (in relation to the identification of the configurations in the source language). – 229 º This method of translation exhibits the advantage that it is suit- able for any language. - . . • • , - Let us now examine the application of the configurational method of analysis in the formalized form in which it was proposed by Molosh- naya for machine translation from English (Moloshnaya, 1957, 1960). Moloshnaya proceeds from the structural classification of word Classes developed in descriptive linguistics (Fries, * 1953) and expand- ed somewhat by her with respect to the goals of machine translation. In order to enable the feader to use the primary source, we will retain the system or indices adopted by Moloshnaya. In particular, Moloshnaya identifies the following classes of Words : . - noun; - - personal pronoun in nominative case ; - personal pronoun in objective case ; In intransitive verbs in personal form; - transitive verbs in personal form; o *- - the verb after which the so-called accusative case is used with an infinitive (for example, I made the boy run; – the verb be in personal form; * * – the verb have in personal form; 2T2kh2V – the verbs shall and will ; I I – the verb let; participle I (active); participle II (passive) of transitive verb; g - – adjective ; ! . possessive and demonstrative prº On Oun ; - articles; - - preposition; — subordinate conjunction; : D: {-> — interrogative or relative pronoun. Moloshnaya then presents a number of formulas for the expansion of the configurations, i.e. , the replacement of these configurations by the index from which the given configuration might be developed as part of — 230 - • the process of formation, e.g. : *) 31 *2 = 12 (the subscript refers to the number of the word in the phrase) 2) I 1 12 = 12 4) 2' 12 = 2i, etc. Here we must take into consideration that on the basis of these rules a single sequence of indices can be expanded in various Ways. For example, the phrase the old man Corresponds to the sequence of indices Ill 32 13 which can be grouped in two ways: 1) II 3 1 Hill A 3. 1 It l l 2) 3 It is clear that the first method of expansion results in an unre- liable analysis of the phrase? Perhaps we should entirely reject the II 3 = 1. However, this would be valid for a sublanguage which does not contain phrases of the type : The rich enjoy life, which is analyzed Specifically in the following manner: II 3 ''2' 1 —T- -ā- where the configuration l 2T is regarded as the base configuration. Moloshnaya therefore selected another course of action. She first as Sumes the sequence in which the rules are to be applied; thus, the modification 3 l = 1 is carried out before 3 = 1; this last modifica- tion is then accomplished only in the case that the index 1 does not follow immediately after index 3. Now, in addition to the above given four rules of expansion, let the following rules also apply: — 231 – 5) 2° l l =2- . 5) 21, Pa 2-—2- . Let us examine how the mechanism of expansion operates in the translation of the following English sentence : Let us call the above-mentioned lemma the .2”. Pinz 2°, H, 3, 1, II, fundamental lemma C ls The expansion can be seen from the following table: 1 Tipmennewaa - AhavihahpyeMax nocnemoBatembHOcth qopMyna - 2 MHAekcos - 3s lºwl, - 3s 19-1, 2!' Prº 2: II, 1, II, 1, II. ls=ls º 2|| P 29 l. 1 Il, lo-l, 2. p". 2: 6 * 0. 2; 1, 1,–25 -, -ma -s 2!' Pn, 2, P.2, P.P. 1) Formula employed; 2) analyzed sequence of indices. The configurational analysis facilitates solution of the problem of grammatical homonymy, e.g. , conversion in English. As is well known, there are very few words in this language which are exclusively nouns Such as , for example, development, kingdom, nor are there many words which are exclusively verbs such as , for example, develop , protect ; the overwhelming majority of nouns can simultaneously act as verbs; (by the way, it can also be said that the overwhelming majority of verbs can simultaneously be nouns) as for example, turn, rule, and talk. In addi- tion, almost any noun may be used as an adjective. - All of the foregoing represents one of the basic difficulties in translation from English (Moloshnaya, l958, page 215). The difficulty in analysis here lies in the fact that the form of the word provides inadequate information regarding its function in the sentence. It is precisely for such languages that configurational analysis is particu- - 232 – larly advantageous. - Grammatical homonymy results when several indices are Set into Correspondence with a Single word or, as Moloshnaya Says, 'a single COImpound index. I For example, let us take the Sentence : . No prior knowledge of vector theory will be assumed. The compound index (l, 3) Corresponds to the word vector, i.e. , this word may function both as a noun and as an adjective. The follow- ing rule is therefore given: "If, beginning with the end of the sen- tence, we note a compound index (l, 3), we check to see Whether the in- dex 1 corresponding to a noun Contained in the dictionary follows the Compound index. If the answer is yes, the subject someone index is re- placed by the index 3" (Ibid, page 240). • Let us analyze yet another example offered by Moloshnaya. Given the Sentence: dy All the solutions of # = Ay approach O as t → co The following sequence of indices Corresponds to this sentence: NII F 1 (2+) 1 Y1, Where the compound index (2* 1) corresponds to the Word approach be- Cause of the fact that the word approach operates as a verb in certain Contexts, and as a noun in others. - - - . . " Here the following rule prevails: "Having (1, 2*) or (l, 27), we check to see whether the unanalyzed index 1S or is not preceded by 2 F, 22'm, F, 2°2'... F, 2× 2.d F. If one of these trai- CeS does precede the unanalyzed index, for the compound index we Sub Sti- tute the index 2+ or 2T. If none of these indices precedes the unanal- yzed index, we check to see Whether there are any of the indices 2*, 2*, 2* 2',4, 29 2 3, 2* 2°.4, 2b 2 ing: 2° 2 d. can be found to the right and left be - - 233 – s fore the period or can of J or 1. It yes, We replace the compound in- dex with the index I: if no, the compound index is replaced by 2” or 2-1. (page 241). - Similar cases of grammatical homonymy are analyzed in like fashion. It is interesting to compare the methods of configurational lan- guage analysis with an analysis method frequently employed in foreign machine-translation literature. This method was used in this connection by Bar-Hillel (Bar-Hillel, 1953), and it is based on the idea of Synta C- 'tic connectivity put forth by the Polish logician K. Aidukiewicz (Aldu- - kiewicz, pages l-27; Suszko, page 9). The Aidukiewicz theory is based on the concept of a semantic cate - gory (Bedeutungskategorie), i.e., it originates in the sense of the ex- pression. More precisely, two words or expressions A and B belong to the same semantic category if A, taken in sense 2: and used in Some statement Sa can be replaced by B in sense y so that the meaning of the remaining parts and the articulation of the statement retain their mean- ing and so that Sb is also a statement. Aidukiewicz notes that this semantic category includes not only individual words, but entire £roºps of words (i.e., that which Molosh- naya referred to as "configurations"). He has constructed his conne C- tivity theory in such a manner as to assign to a group of words ("con- figuration") an index of the same category as an individual word belong- ing to that category. This is accomplished by the method of "cancella- tion indices." To each word is assigned the index : a) n, if it is a name; b) s/n, if it is an intransitive verb; s/(n)[n], if * * * tran- - - ſº If it is contained in the sitive verb ; and in general tº H.I.T., herºe W base configuration with the words having the indices di and P3, o, and P; are indices composed of the elements n and [n]; i - - - 234 – c) (s/n)/(s/n), if the word is an adverb in the verb group: n/a, if the word is a noun attribute; (n/n)/(n/n), 1s the word is an adverb modifying the attribute , and in general a/a, if the Word is a cancelled term in the configuration whose main word has been assigned the index Cl3 d) certain elements may be regarded as operators converting the element of a given characteristic into the element of another. Thus, prepositions convert the element with the index n into the element with the index (s/n)/(s/n) or n/[n], i.e., the attribute of a noun or verb; auxiliary verbs (of the type of the German sein) change the noun or the attribute of the noun into an element with the index s/n. Accordingly, a preposition may be assigned the index (s/n)/ſ (s/n)/ /(n)} or (n/[n])/(n) and the auxiliary verb may be assigned the index (s/n)/[n], etc. Now, let the corresponding indices for each of the elements of the language be obtained by application of the indicated procedure. In this case Some sequence of indices: 11 12- . . *n will correspond to each phrase. Let us now agree that if in the Sequence of indices two Such indices are next to each other that the entire right-hand index co- incides with the denominator in the parentheses of the left-hand index or if the entire left-hand index corresponds with the denominator in the square brackets of the right-hand index, the coincident indices cancel each other out. *, For example, let us be given phrase : "A very small girl was walking there. ' The following sequence of indices corresponds to this phrase : Il $ (n) ſº n, n >, F. (*) 0 " : - 235 - By the method of successive cancellations We obtain : is n – s 1) ś, ni 2) n # 3) s Analogously, for the German example given by Aidukiewicz : Der Flieder duftet stark (n) Il T - Il T Tº T We finally obtain S. T The following sequence of symbols will correspond to the phrase : 'He climbed the high hill 1 : -7-d (n ) Successive cancellation will again give us S. The Significance of Aidukiewicz's index notation lies in the fol- lowing. As we have already seen, a Single index does not correspond to every word, Since a single word may enter into various Configurations. A word such as vor in German must obtain indices both as a preposition and as an element pertaining to a verb. Comparing corresponding sequences of indices, We Can determine in a purely mechanical manner which index must be selected in the analysis of the given sentence; compare: ./ er schlägt etwas vor Hir Peter With the two possible indices for vor: 3. $ . II In -- and wr--> º $ |#| n) # i. n In - - - - - n) | If on cancellation the second sequence yields S, the first , as is and 2) Il ( ( n ) #| easily proved, on cancellation yields the sequence # S ( n) # which cannot be reduced any further. This is then an indication that one of the indices was improperly chosen. Thus the validity of grammati- cal text analysis can be verified using Only a computer. - $31. ANALYSIS THROUGH SYNTHESIS - - Removal of grammatical homonymy, as we have seen, is accomplished by restoration of the process involved in the formation of the homony- mic construction. Similar methods of analysis through synthesis are needed not only for the solution to the problem of hononymy - in impli– cit form they are present in any analysis. The restoration of the con- struction-synthesis process on the basis of direct components or in accordance with the rules of transformation is particularly effective each time the relationships between the elements of some grammatical structure are not clear. As an example, let us consider the analysis of the German defini- } - “. . ) ,' tive combound Il OUIIlS • - * ... Such words, for example, as Produktionsmittel [production facill- ties l are easily broken down into the component parts: Produktion and Mittel; the translation of such words has been standardized; as a rule, this is accomplished in the manner shown in the model: - 237 – **s-ºut, 1 There are many such words and for these *> - it is not necessary to memorize the entire cºeñº accusiºcamás (asºn, efs): - . r- 1) Facility of produc Word, e.g. , a Word Such as Produktionssteige- tion (genitive case, singular). lºung . It is easy to formulate even a machine (automatic) rule for the breaking down of such words into their compon- ent parts. The computer must collate the input word with the stored word, find from left to right the largest of the stored words entirely contained in the given input word, a stored word of this type in the case of Produktionsmittel being the word Produktion; with this stored word there must be an indication of the conjunctional element S which this word employs in word formation (in our example, -s-); this con- junctional element must be set aside and the remaining portion must be checked cut against the dictionary in the same manner. This procedure can be repeated several times, and it is thus possible to obtain an element-by-element translation of words of the type : Bleidioxydschicht ['lead peroxide layer' ], Gefrierpunktserniedri- gung I'lowering of the freezing point ' ), etc. It is clear that it is not advantageous to enter all such words in the dictionary, all the more sc, because each author forms new words of this type. It is enough for the dictionary to contain in their entirety only words having a non- standard translation as, f'Orº example, Stickst Off Initrogen"), etc. . . . . Of course, the analysis of compound Words in individual cases may present significant difficulties; for example, E. Reifler points to the possibility of a dual breakdown of the compound word ("Machine Transla- tion," page 198). As an example we can cite the word Arbeitsamt which can be broken down into two Words : Arbeit-s-Amt Arbeit-Samt. - 238 - However, Some of these problems arºe resolved by the introduction or in- dications as to possible conjunctional elements (see above); with re- gard to the remaining problems it may be maintained that such Cal Se S 3 rºe quite rare (this is tie case with virtually all of the examples pre- sented by Reifler). - - - However, it is interesting to note that in principle in any analy- sis of the so-ealled definitive compound nouns in German it is neces – sary to refer to the history of the formation. Indeed, in order to translate the word Unions-Landwirtschaftsausstellung, we must know the relationship between the components, and for this it is necessary to understand the history of the formation of this word. However, if we Set up the formati on diagram for this Word, having the ſolºſh J (hions-Laha; actºrs jawsstellong and if we know that at each branch point the left-hand portion is a function of the right, the translation becomes easy: 'All-Union Agricultural Fair. Here we have examined questions related to the history of config- urational formation. Let us now turn to the transformational history of the translated text. * / *3 S. .. J In order to place the article in the case of uniform terms, it is necessary to know whether these uniform terms are a result of eXpansion from a single element – in which case the Corresponding combination de- notes unity and the article pertains to the entire combination – or they were formed through a merger of two clauses — in which case the corresponding combination denotes two various concepts and the article must be placed in front of each word. Suppose we are given the combina - — 239 - tion workers and peasants' in the sense of all workers and peasants. we would translate this by means of the combination; die Arbeiter und Bauern, since it may be maintained that this configuration corresponds, for example, to the word die Werktätigen. However, when translating the C Ombination 'Workers and capitalists ' the article must be repeated: die Arbeiter und die Kapitalisten, since generally these combinations arise out of the merger of two or more Clauses. Let us cite yet another example in which knowledge of the forma- tional process is necessary for the proper placement of the article. Let us take the combination: 'The 5th London Congress of the Party. ' If the formation of this combination can be represented by the scheme: 5- if nondoncrud cread magnuu - - - - - - - 3 #} 5th; 2) London; 3) congress; l!) of the party. Only a single article is needed, but the numeral 5th defines the en- tire combination 'The London Congress of the Party , " and the impress lon then is that all five party congresses were held in London. However, if the formation is presented as a transformation accom- - § - plished through the consolidation of two formational diagrams : 54 cºod modman | | l 2 3 moniºcaud g à nazmuu f l) 5th; 2) congress; 3) of the party; l;) London; 5) C Ongre SS ; 6) of the party. – 240 - . It becomes necessary to repeat the article twice (or the article + noun); compare the German der rúnrte, der Londoner Parteitag, with the French le se Congres du Part 1, le Congres de Londres. An analogous example. If for the translation of the combination "The IIIrd Communist International' we use in German the combination die dritte Kommunistische Internationale, or in French la troisieme In – ternationale communiste , this would indicate that all 3 Internationals We Iº e communist; it is well known, however, that this cannot be said by any means of the IInd International. Thus, it is necessary here to use the form die dritte, die Kommunist is che Internationale, for the German and la Troisieme Internationale, l'Internationale communiste, for the French. The knowledge of transformational history facilitates delineation in Russian of the two forms of multiple negation. 1) The repetition of negation for purposes of intensification; this multiple negation is rendered in the target language by a Single negative. 'I cannot say anything about this " can be translated into Ger’— man as: Ich kan darüber niemand etwas erzählen or Ich kann darüber kein- em et was erzählen. . . . - Negation of this kind, no matter how many times it appears, does not affect Sense and therefore phrases with a variety of negatives can be treated as variants of a single phrase. - 2) The double negative arters tº a transformation retaining its positive Sense as , for example: "I could not not laugh' ['I could not help but laugh' ] transformed from 'I burst out laughing' or I should have laughed. " In the translation of such phrases we may make use of constructions such as the German nicht umhim können which compare with the French ne pouvoir s 'empécher. Transformational history must also be known to translate Russian - 241 - phrases involving verbal adverbs. The situation here revolves around the fact that in Russian a phrase involving a verbal adverb may exhibit a great variety of adverb- ial meanings arising out of the transformation of various subordinate clauses. This may involve the sense of time. For example, 'he treated - everything with disinterest and disdain, he held everything in Contempt , and even upon finishing his tasty meal, he sniffed disgustedly (Chekh- ov, "Kashtanka"). "Having eaten his tasty dinner, ºr when he had eaten his tasty dinner. " The Sense or cause: 'Sensing an unavoidable separation, he desired . at least to remain her friend (Turgenev, "Fathers and Sons"). "Sensing Separation * since he sensed a separation. ' The sense of condition. "However, chattering with you, no mush- rooms will be picked " (Pushkin, "The Young Peasant Girl"). chattering with you + if I chatter with you.' The sense of concession. 'Although a heavy hammer Smashes glass, it can forge a sword (Pushkin). 'Smashing glass - although it smashes glass' etc. - The restoration of the transformational history is needed in the analysis of relationships between combinations and the abstract verbal nouns. Alone the transformational analysis of these combinations Will Show whether the second term here is the object or the logical subject of the verbal noun. compare, for example, the German die Regierung Frankreichs ['the government of France' ] and zur Zeit der Regierung des Königs Heinrich IV [' during the reign of King Henry the IV' ]. compare the analogous phenomenon in the translation of the combina – tion: 'strengthening of the State. "This combination may correspond to two clauses: - 242 – a) '(workers and peasants) strengthen the state' and b) "the state was strengthened. ' In translation into German the meaning of a) will most readily be rendered in translation by die Festigung des Staates, while the meaning of b) will most readily be rendered in translation by das Erstarken des Staates. Compare also 'Each of the two congresses played a significant role in the unification of the masses," i.e., "... was of great significance from the Standpoint of unifying the masses; ' compare the French Chacun des deux congres a été tres important dans l'oeuvre de cohesion des . Iſla S Se S. $35. GENERAL PROBLEMS OF SYNTACTIC TEXT ANALYSIS In §§33 and 34 we paid particular attention to the meaning ac- quired with respect to the theory of translation by the concepts worked out during work on machine translation. However, we made no mention of the manner in which in the course of a machine translation the syntac- tical relationships between words are as certained for the general Ca Se , i.e. , when the terms of the configurations are not necessarily in di- rect contact. This process is important for the theory of translation because it represents an apparent simulation of that method of trial and error which frequently (at least in the initial stages) is what takes place in the "human" analysis of a text in an unfamiliar language. Initially, it would not be a bad face to take a close look at the man- ner in which this process is accomplished by a human being. Let uS take the sentence : Oil immersed transformer windings are now usually made with paper covered conductors. Analysis ("passive grammar") assumes, as stated in $17, that for each word form a complete list of its functions is given, i.e. , its – 243 – gramatical meanings. Let us take the first word oil; it may be the subject, translated by the word vacao, in the nominative case. The se- cond word immersed must then be treated as the predº.cate (the given form of the word can easily be a verb in the past tense). We get 'Macno morpysano' ('oil loaded"); the third word is then translated as the object : "oil loaded transformer. ' The following word is also a noun and for this reas on the preceding words cannot be translated as a noun. Let UlS rework the translation : "the oil loaded transformer winding. HOW - ever, the following word are shows that the earlier assumptions regard- ing the functions of all of the words are not accurate, since are is necessarily the predicate or a part of the predicate ; in addition, there is no Subject for this predicate. Only the word windings can be the sub- ject. Thus we have 'transformer winding is made. ' This compels us to assume that immer Sed 1S not a verb, but a participle (this function is also consistent with the initial enumeration of meanings) and should be translated by the word “morpyxehhaa’ ['immersed ' ). Hence it follows that the word oil is a function of this participle and must be translated in the indirect case. We thus obtain a final translation for this portion : "Immersed in oil transformer winding is made . . . " This is treated as a final translation because we encounter no additional Words which Would indicate that our analysis has not been properly carried out. The second portion of the sentence 15 analyzed in the Same manner. Let us now an- alyze an analogous German example : Die zusátzliche Produktion kann nicht losgelöst von den Planauſgaben in Angriff genommen werden, wie es in der Vergangenheit vielfach geschah. - The key to the translation is the correct identification of the subject and the predicate. The subject of the main clause is easily found (Die Produktion). The predicate is clearly compound. The first portion of the predicate (kann) stands, as is proper, in second posi- - 24h - tion. But which is the Second portion? If we assume that this is gen Om- men werden, the translation, as is easily proved, is senseless. In this case it is necessary to know that in Angriff nehmen represents a unity having the meaning of 'undertaking something, ' 'beginning Something. ' However, let us make certain that there is not some other form of the verb in the clause , which might also serve as a portion of the predi- cate. From this standpoint let us C OnSider the participle loºselëst. On the basis of formal indicators this participle might be included in the predicate (die zusátzliche Produktion kann nicht losgelöst werden). However, in this case, first of all, the indication of the object (von den Planaufgaben) could not stand after losgelöst, and secondly, the portion of the sentence in Angriff genommen werden would be hanging in the air. Thus we have established that losgelöst does not form part of the predicate. In other words, it forms an independent group losgelöst von den Planaufgaben I'separated from the planned assignments') (Shvane- bakh, Revzin, pages 101-102). - Translations of this type , as our initial variant , inevitably arise with each individual studying a translation; however, they are general- ly simply qualified as errors, which in no way guarantees that such er- rors will not appear in the future. If these completely proper "er- rors" are used as a stage along the way toward determining a transla - tion, effective mastery of translation habits will only increase. There is the widely held logical principle which states that recognition Cf contradictions always leads to increased knowledge. Below we will discuss one of the methods of general syntactical text analysis. This method has been proposed by Yu. S. Martemyanov. The method involves the following. The syntactic analysis of a phrase refers to the establishment of the syntactic relat lonships between words with indi: ...tions of the – 245 – & maïa. . . . . ii. º. & ºs.º. tº fº -zº ----- . . . . . . . - . - s: - . - • ‘ läärönships as the "dominating" or subject relationship. and dependent terms. In the following we will refer to these re- -º-, * . **i. *:: *; To determine the maximum number of dominating relationships encom- passing the words of a phrase, we have found it convenient to have three main classes of words and two rules relating these classes. This classification is based on special syntactic indicators: 1) the active capacity to be dependent (abbreviated Az); 2) the active capacity to be the main (abbreviated AG); 3) the passive capacity to be main (abbreviated PG). For indicators i and 3 an additional distinction has been intro- duced to take into consideration the relative position of the words in speech: sequence in word order – from left to right (p) and reverse or- der – from right to left (1). Correspondingly, the indicators AZ and PG arise in two variants: *. and AZ1; PG, and PGl, i.e., "the active capacity to be dependent on the left," "the passive capacity to be main for the right" and the same "for the left." - We are given the following two basic rules: I. "If for a word x from the class of elements with the indicator AZ (short for – from the class AZ) a word on a given side of a clause is found from among the class of elements with the corresponding indicator PG, (in the absence of certain restraints – we will have more to say about this later on) a dominating relationship should be established be- tween X and y, with the word y being the main." II. "If after the word x from the class Ag the clause contains some additional words y, t and z not from the class AZ, each of the subse- quent words should be combined with the word x in a dominating rela- tionship, with x as the main term." $ These two rules, with their variants, make up the basis of syntac- - 246 - - - tic analysis. Let all syntactic indicators (with the exception of the three main indicators, Several auxiliary indicators have been introduced ; see be - low, page 21.9ff) be attributed to words in the dictionary. Then, let the required words be copied from the dictionary into a summary table of clauses and let each of the words be assigned sequen- tlal (absolute) numbers. In this case the establishment of the dominat- ing relationships will obviously involve the successive seeking out of the active term (AZ or AG) covered in Rules I-II in the seeking out of a second term (from the corresponding side and with consideration of specific conditions) and in the accomplishment of the corresponding Se- quence : "assign to the dependent term of the relationship in the col- umn 'relative numbers' the absolute number of the main term." Thus, the relative number indicates the word to which the given word is subject. At the beginning we encounter a group of operations establishing the dominating relationships with respect to Rule I; Rule Ia for AZ left and Ib for AZ left. A special group of operations establishing the dominating relationships with respect to Rule II is used later on. The obvious "outcome" of each group of operations is the appearance of relative numbers for the corresponding words. Thus, for the words of the English sentence (P) : The (1) chief (2) - members (3) of (!) the (5) common (6) Market (7) seem (8) too (9) little (10) perturbed (11) by (12) Maudling's (13) proposition (14) of (15) the (16) free (17) trade (18) area (19) including (20) only (21) seven (22) nations (23) (the absolute numbers of the words are shown in the parentheses) we obtain the following results. – 247 – Results of the Syntactic Analysis of the sentence (P) by Groups of Qperations tº Tpynna onepauwä Ia) l 2 adoon. Hoºtepa & -- 1 2 3 4 5 6 78 9 10 ll 12 13 14 15 16 17 18 1920 21 22 28. of Hoc. Homepa 2 7 7 10 1 1 14 18 18 22 23 . . l, Tpynna onepauhh 16) aôcon. Homepa | 1 2 3 4 5 6 78 9 10 11 12 13 14 15 16 17 18 19 20-21 22 23 * * othoc. tºoltepa - 2 3 7 7 10. 11 8 13 14 14 18 18 7 ſpynna onepanah II . - 8 accom. Homepa 1 2 3 4 5 67 s 9 to 11 12 13 14 15 16 17 18 192021 22 * Q orhoc. Homepa - 2 3 7 74 10 l l 8 (3 14 12 14 1818 15 1922 23 1) Group of operations Ia; 2) absolute number: 3) relative number; 4) group of operations Ib 3 3 :::::::::: number ; 6) relative number ; 7) group of operations II; 8) absolute number ; 9) relative number. Let us now consider the auxiliary syntactical indicators of the words. To evaluate the words in accordance with Rule I it is necessary to introduce an auxiliary indicator of "orientation with respect to the main or nearest indefinite word." Immediately to the right of the AZp - word (or to the left of the Azi word) there are in the sentence two corresponding PG Words, and they are associated by a dominating rela - tionship; The (AGE) Common (PGA) Market (PGA). According to the general rule (Rule I) any AZ word must be associated with the nearest PG Word, as is the case, for example, for words 9-10 (too little) * *** (on- ly seven). However, for the word the the general rule would not produce a desirable relationship with the more distant of the two PG Words - Market. In such cases the AZ indicator must be refined through the in- troduction into the syntactic coding of the words of an auxiliary indi- cator "of orientation to the main." It developed that a word with the AZ indicator may be a function of the more distant of the two PG Words - 248 - - only in the case that the farther word is associated with the acter in a dominating relationship and is the main word at the same *. To use this auxiliary indicator a determination must have been made at the instant of evaluation of the AZ word with regard to which of the two PG words 1s the main. This is accomplished by the fact that the Selection of the words in the analysis follows a specific sequence. In evaluating the Azp WOI’ds the selection proceeds from the end of the Sentence t e o the beginning; in the case of the AZ1 Words, in the oppo -- site direction. - It should be pointed out that the indicator of "orientation to the main" does not pertain to any other data with the exception of those . attained as a result of the preceding syntactical evaluation. From the standpoint of grammar, the desire to associate the with Market is gov- erned by the fact that in English the "article" element of a part of speech is generally combined, i.e., forms a "configuration," with the "noun" elements of a part of speech. But at the same time, In a proper- ly formulated English sentence , the article is nevertheless not separa- ted from its noun by any word which might be the main word for that noun (for example, a verb cannot stand between an article and its noun, but only an adjective and an attribute of the adjective). It is pre- cisely for this IP6a S CIT that in the analysis of a complete and proper English sentence the knowledge of the specific part of speech with which a given article must be associated, “. . not mandatory. The AG class evaluated in accordance with Rule II include S Words such as of and by (prepositions) as well as words of the type seem and perturb (verbs). This consolidation leads to the need for the introduc- tion of a certain additional distinction; in our sentence the word of, if we apply the second rule, would be the main word for the Subsequent word seem, while the word by would be the main word 'or the subsequent tºº 249 º :i i • 'º' • * * * word of, which is undesirable. Using the terminology or conventional - parts of speech, We should limit their "capacity to be main" by the •. nearest "noun" on the right. However, instead of referring to a Speci - fic part of speech, we will refer to the results of the previous evalua- tion, introducing the refining indicator of "limitation by the neare. . . main word of the following words." This is all the more convenient be- cause, for example, in English the subordinate effect of the preposi- tion is limited not only by the noun, but primarily by the sequence of the subsequent nouns (of the free trade area, by the chief members, [see below]). • - Upon ecºpietion of these operations, We find that the words of the Sentence are carriers of specific syntactical relationships associating these words with others of the same sentence. The Syntactic relation- ships noted in a given word may differ from the syntactic relationships of another a cri in terms of quality and quantity (see the example shown On page 245). The unique nature of the syntactical functions of the various words, established in the corresponding stage of the analysis, is entirely brought ab cit by the initial distinction in the syntactic coding of the words (see the table on page 255). The Syntactical indicators seem to show in expanded form all possible variations of linguistic word rela- tionships * r any sentence. These indjeators make it possible to break down into corponent elements those inseparable complexes of syntactic properties which arºe generally ascribed to S Ome traditional part of speech. The latter appear before as sets of "differential" indicators. This is clearly borne out by the table of sentences at the end of this section. At first glance, in the syntactic coding of the various words cer- tain stable sets of indicators are easily recognized and are regularly - 250 - ** * - repeated for many words. FOrº example, the set of units in columns l'7-18- 19-20 corresponds to a "preposition;" the units in the columns 10-12-16 denote an "adjective;" the units in the columns 10-12-19–20 denote an "article;" the units in columns lll-l'7-18–19–20 denote a "subordinate conjunction." The description of the enumerated parts of speech by means of the airferential indicators seems quite simple and at the same time 1s. rather complete, thus pointing up the extremely syntactic nature of the corresponding parts of speech. The remaining traditional parts of speech in English (verb, noun) are syntactically not uniform: if the verb appears in the form of the infinitive (or in the present tense), it is described by the set 15-16-7-19-20 (seem). The verbal form end- ing in -ing (including corresponds to more extensive syntactical infor- mation: 10–11-12-15-16–17–19–20, providing the verb with additional in- dicators which bring it closer to the adjective (see the table). The noun in the direct case (Maudling, Market, area) is coded by two indicators 15–16; in the indirect case – Maudling's – it is more informative: 10-11-12-15-16. Consideration of the indicators ascribed to the affixes may elim- inate certain syntactic indicators : the verbal Stem having the suffix —tion (or -ment, -ence, -er) uses indicators 17–19–20 in the manner of a noun (proposition). - The eleven indicators used in coding are adequate to ensure dif- ferent evaluations not only for the various parts of Speech such as , for example, "preposition" and "adjective," but for individual words within the same part of speech: thus the preposition of , generally de- pendent on the preceding noun, is assigned the indicator AZ1, whereas for by which is found after a verb exhibiting its own subordinating properties (the AG indicator), the AZ1 indicator would be unnecessary. After the description of the parts of speech in differential indi- - 251 - * cators , the syntactic evaluation Of each part of speech breaks down into al number Of successive cycles replacing but a specific portion of the syntactic properties. The advantage of this separate evaluation of the individual syn- tactic indicators is obvious : there are considerably fewer elementary syntactic indicators than there are parts of speech to describe them. "The syntactic significance" of these indicators was revealed in the restricted number of analysis rules. Within these limitations, so long as an individual part of Speech may be described in the proposed indi- cators, the feature of its evaluation will lie exclusively in that par- ticular combination of rules corresponding to these indicators. In § 23 we had occasion to speak of syntactic homonymy. Let us now consider the discussion of this problem in the terminology of the Mar- tem'yanov theory. A word which under two various conditions satisfies different indicators will be known as a homonym. Thus an adjective (10– 12–15–16) for which indicator's lo–l2 are not achieved under certain specific conditions is syntactically similar to a "noun" (15–16) under these conditions. The preposition (18-19-20) for which no subsequent noun was found in the sentence, thus leaving the indicator 18 (AG) un- used, turns into an "adverb" (19–20). Even greater possibilities for such homonymy prevail in the forms ending in —ing, whose Syntactic code contains the sets of indicators characteristic of a "noun," "verb" and f "adjective." - In all these cases the syntactic homonymy of the word may be re- garded as the result of diverse contextual satisfaction of a given set of syntactic indicators. This homonymy may be lacking in its entirety in the permanent syntactic code of the word. Words of the type common , any, bad should always be assigned indicators 10-12; words of the type with and by should be assigned indicators l8; words ending 1n —ing - 252 - should be assigned the maximum set of possible indicators Thi - - - e Various means of attaining these Sets... are left entirely to the syntactic C Ontext of the °01’responding sentence Le t us now return to our Sentence. At the conclusion of the text evalua: - - **on in accordance with the rules of procedure (see page 248) in th ° analyzed sentence there remained several non interrelated s egments : Segment (1–2); - - Segment # Segment (8-l j; Segment ig–33). : The described method is 1nadequate for the establishment of domi- nating relationships between these segments. The syntactic analysis based on the differential indicators Of Words is possible only to the extent to which the indicators are per Iſlan ently assigned to the W Ord ; it is understandable that a Dermanent C Ode may be useful only in the event that it leads to the establishment exclusively of desirable relationships for any Sentences in which this W Ord might be en COuntered. The desired subordination of the word trade to the WOrd area (the Same as the Word chief to the word members) Cannot be provided With any indicators which Would not run the risk of leading to an undesirable relationship in some other sentence. he limitation on the use of syntactic Coding of words cannot be regarded as a Shortcoming of the method being analyzed here. The de- e . . . . *, *. pendence of the Preceding noun on the following noun may be regarded a S a rºl effect of Speech based on the fundamental potentials of the mutual. C ---- Ontact positions of the words. These potentials are realized in vari- OUIS V - languages in arious Ways : English prefers the preposition and f'Or, J € Xa - cº mple, French prefers a postposition. In this Case the described meth- Od O. - Of analysis, based On permanent syntactic pr’Operties of WCrds, in - 253 - ... * its very essence cannot be extended to the case of the type trade area. Cases of association"explained by the fact of a characteristic mu- tual positioning of adjacent words should be identified in a particular syntax; we will refer to this as syntax II. Rules Ia, Ib and II serve exclusively for the description of syn- tax I. They do not lay claim to the identification of all relationships existing in a sentence. Their role is to expand the sentence to the minimum of segments whose further association is either totally unneces- sary or represents no difficulty whatever. To describe syntax II (in our case, for the establishment of dom- 1nating relationships chief members and trade area) a special group of rules III has been prepared, but of which we have not spoken here. Finally, an entirely particular case is the relationship between the two remaining segments : the chief members . . . and seem perturbed. . . This is no longer syntax II which has to do with the mutual posi- tioning of adjacent words: nor is this syntax I, since it is entirely impossible in the permanent syntactic code of the word to provide for a nonpermanent "subject" and "predicate" function. But this is not ne– cessary : for the majority of sentences the Subject and predicate appear automatically during the course of the expansion as the two main inde – pendent "vertices" of the sentence, whose recognition presents no dif- ficulty. - }. # To the syntactic indicators of a word which are assigned to its stem in the dictionary (or in the "table of stems") it is possible to add the affix indicators noted in a corresponding morphological table. Therefore, syntactically the algorithm is predicted by a single stage - the rules of morphological analysis (the group of rules A) thanks to which the identification of all morphemes of the text words with the table morphemes is accomplished. This stage concludes with the Selec – - 251 - tion of information from the general tables of the language and its com- pilation into a summary table of sentances (see table) subsequent to which syntactic analysis described above is carried out . OPHEH- T | A T H II A 3 3TAP ha T 6 p l sº - i t : Ta&nhua g . -- 3. # | # 3 a 3 | 8 || 3 || 3. … g. = | 5 | #.| 3 || 3 | # 5 3. t; ; R ſº :: E * & B- > § 3 ; s | 3 || 3 | #| 5 || 3 || 3 & 8 S 5 3 : 5 tº : a l ; ºr K = 10] 11 12 13 14 | 15 16|17|18|19|20/21 the | ! - 1 1 ; chief - - 1 1 - 3 member 1 1 3 & 1 1 1 | 1 4 the 1 l | 1 : common l 1 . l 7 market ... , 1 1 8 Seeml 1 1 1 1 1 9 too l l 10 little l l 1 - 1 perturb 1 1 1 1 1 11 -ed 1 1 1 12 by - | 1 | ! 3 Maulding l 1 | . 13. -> 1 - - more. 1 1 1 1 1 14 ition l 1 1 14 of l l 1 1 1 # the l 1 1 1 # free l 1 18 trade | ; 19 a ſea - include - 1 1 | 1 1 ; -ing 1 1 1 - 21 only 1 l 22 seven l l l | 23 ºtion 23 f * .. e Sr. ; 6) NP; 10) from the 1) Table; 2) AZ; 3) orientation to; 4), PG; 5) AG; 3. e Hº ii) łº Ieft; 12) main; 13) near; 14), indefinité; º the right; 16) for the left ; 17) unlimited; 18) limited; 19) to the ...ht: 30) to 'the left; 21) absolute number. §36. SEMANTIC PROBLEMS OF SYNTHESIS As a result of the lexical and grammatical analysis Of the S Our"Ce - language text some succession of translation units and -relationships - 255 - sº • * = - * l) First person, singular; 2) A causes possession for B and B CauSe S possession for A ; 3) possession, first person, singular; l!) the son of the same mother and father; 5) 500 rubles; 6) B causes possession for himself with payment; 7) automobile. between these units is revealed, as We have Seen, 1.e. , the plan of communication content is established in the form of a certain set Of elementary meanings and relationships in the intermediate language. Let us examine the communication analyzed from the point of view of translation, i.e. , having passed through the stage of analysis. Let us take the Sentence : - ! I lent my brother 500 rubles to buy an automobile. ' We can imagine the analyzed text to have been broken down in such a manner that the communication, as a result of the analysis, assumes the form shown in the figure. In order to accomplish the process Of translation that sequence of units from the content plan must be replaced by elements from the plan of expression, i.e., the text must be prepared in the target language. This is the role of synthesis. J In what manner" and in accordance with what rules will we select the target-language units required for the preparation of the text? Let us examine this question briefly. - AS a result of the analysis we have at our disposal a set of ele- mentary meanings. Some of these Will correspond uniquely to the Words in the target language. These include many words indicating Subject re- lationships: automobile,' tree, " 'flower, ' 'dog,' 1 table, ' 'pencil , ' - 256 dº 'to breathe, to read' and, for example, the French automobile, arbre, fleur, chien, table, Crayon, respirer, lire, etc., or such abstract concepts identically constructed in both the source and target languages (compare 'physiognomy, 'arteriosclerosis, ' 'balance of payments, ' 'ne- cessity, probability, etc. , and the French, physionomie, arterlo- sclerose, balance de paiements, necessité, probabilité). It is true that in some languages there are variations of words used in various ethnic communities. Thus, for example, in translatien into English the American versions of English words must be borne in mind (compare baggage with luggage; bill with note can With tin; car with coach; corn with tin [sic]; store with shop, etc. Nevertheless, in principle this does not alter the problem of the synthesis of such words: it is enough to give a corresponding indica- tion of this in the dictionary. - The number of words identically articulating reality in the source and target languages , i.e. , coinciding in the content plan, is particu- larly great in scientific texts, as well as in other branches in which regular contacts between the languages are maintained. As a rule, however, between the set of elementary units from the intermediate language and the words of the target language no uniquely defined correspondences exist. As an example, let us examine the English Sentence Bleriot flew acroSS the channel. The translation of this Sen- tence into French is : Blériot traversa la Manche en äyton (Vinay, Dar- beinet, page 105) where the English words flew across correspond to traversa en avion. What is the mechanism of this translation? At the stage of the analysis of the English sentence the following sequence Of semantic units was identified: moved + by air (flew) + by crossing (across). Then the following French words were synthesized from this Set : traverser [to move by crossing overl + en avion (by air). Thus, - 257 – i & * * , * there has taken place some redistribution of the elementary semantic units pertaining to the words, but on the whole, the translation unit has been synthesized, i.e. , a communication has been constructed in the target language, and it corresponds in terms of meaning to the original text. - - It is essential to point out that in the synthesis of the cited French phrase the redistribution of the elementary semantic units by words was an absolute necessity. This redistribution will be needed in the translation of this phrase into Russian, and in this case three elementary meanings (to move oneself + by air + by crossing over) will be expressed by the single nepeneten [flew over]. The fact that the ex- pression of this set of meanings in Russian differs from the English or the French is associated with a different procedure for word formation and is important in its own right , but plays no decisive role from the standpoint of synthesis: the important fact here is that in the case of synthesis it is necessary to construct a communication that corresponds to the set of meanings in the content plan. However, the method of ex- pressing this meaning may vary: one and the same meaning may be ex- . . pressed lexically through a variety of words or it may be expressed morphologically (Mel 'chuk, 1960, 1961). Thus, for example, the word * [to give 1, containing two elementary semantic units (to cause + + to have ) may in its tyanslation into English or into French be syn- thesized not only through give and doner, but also through let have and faire avoir. This method of synthesis so popular in English and French, as well as in German, permits flexibility of expression in the subject - object relationships. For example, it is enough to add the auxiliary verb raire to the French intransitive verb in the infinitive in order for this verb to express the fact of action, causality. Thus the mean- ing of sending or signaling a communication can be expressed not only - 258 – by the corresponding verbs (dire, montrer, etc.), but by verbs denoting reception of the communications (entendre, voir, etc.). Compare the following texts : "A sort of rumbling noise was heard' – un bruit se fit entendre; the bell rang hollow " - la sonnerie sourde venait de se faire entendre ; "he showed him the horse ' - il lui fit voir le cheval; the news which I am bringing" - les aventures qu'on va lire; 'the people were muttering '- on entendair murmurer; "he repeated the anecdote " – il se fit l'écho d'une anecdote; "he remained silent – il ne soufflait mot; 'he laughed self-assuredly" – il partit d'un éclat de rire assure. . In translation from German or from Russian into French it fre- quently becomes necessary to reconstruct the entire sentence, particu- larly in expressing cause-and-effect relationships. We are dealing here with the fact that French in expressing causality by means of the verb faire or with similar devices assumes a word order in which the Subject of the action is in first position, while the corresponding objects follow after the combination faire + the verb in the infinitive. Compare Robert regarde un paysage and Albert fait voir un paysage a Robert , whereas the German and Russian languages prefer to express the causal relationship by means of adverbial constructions. This is the cause of the corresponding reconstruction in the synthesis of the French sen- tence. Compare the following examples (Malblanc, pages 27–28): Davon Zittern die Fensterscheiben — Cela fait trembler le S Vitre S [This caused the windowpanes to rattle ]; Bei diesen Worten erbleichte Hans – Ces paroles firent pâlir Jean [At these words paled Hans (Jean )] (on a par with A ces mots Jean pâlit, [At these words Jean paled]); Vor dem nahen Feind ent flohen die Einwohner in die Wälder - L'approche de l'ennemi fit fuir les habitants dans les forêts [The inhabitants º fled to the forests before the approaching enemy J. Compare also the an- - 259 - alogous examples of translation from the Russian: "You are afraid of my words' — Mes paroles vous font peur. "He laughed and at times he knew not nor did he understand why he laughed Il riait et parfois ne savait ni ne se rappelait lui-même ce qui le faisait rire; "That which makes men alive (L. N. Tolstoy) – Ce qui fait vivre les hommes; "Through them she forgot her black misfortune' (Dostoyevskiy in the "Idiot") — Ils lui firent oublier sa misere (on a par with Par eux elle oublia sa misere noire (Tesniere, page 296). "The danger does not let me sleep' – L'orage m'empêche de dormir; "I love her nevertheless " — Ça ne m'empé- che pas de l'aimer. • , . . . The difference between the negative anticausality ('nonhindrance " ) expressed by the French ne' pas empêcher and the positive causality (per- mission) is noted in the synthesis of the French text in the introduc- tion of words having the elementary meaning of concession (si vous vou- lex, ge veux blen, j'y consens, S oit , d'accord. (Compare: 'Perhaps , clean caps might be put on ' (Gogol) and the French "On peut leur mettre des bonnets propres S1 vous voulez; Je veux bien qu'on leur mette des bonnet S propres, jºy consens; Soit, qu'on leur mette des bonnets pro-. . pre S ; D 'accord , qu'on leur mette des bonnets propres (Tesniere, page 299). - The semantic interrelationship of the subject, object and circum- stances of the action is expressed, as we can see, lexically, morpho- logically and syntactically, which is understandable , since these re- lationships are essential to virtually each of the forms of spoken com- munication. Certain syntactic expressions of this interrelationship were considered by the authors earlier ($27) (for example, the relation – ship of active and passive constructions). Here we should devote S Ome attention to the redistribution of the elementary Semantic units in the synthesis of phrases expressing voice - 260 - - at relationships. compare the following examples: Pour quelques personnes, la fortune du vieux était un objet d'orgueii. "There were people who. were proud of the Old man's Wealth; ' Le bonhomme Grandet devint maire. (Balzac). The honored and respected Grandet was made mayor (Dostoyev – skiy's translation). We frequently find redistribution of elementary meanings express – ing the circumstance of an action. The presence of auxiliary verbs ex- pressing causality makes it possible in the synthesis of text to alter a group of words expressing the circumstance of action into the subject group. Compare , for example: "In l962 the production of steel was increased by 8%. In French, L'année l962 a vu la production de l'acier augmenter de 8%; in English, 1962 saw the steel production rise by 8% [sic]. - The Russian and German adverbs, respectively , are replaced by the French verb, while the verbs having the meaning of a specific action are replaced by impersonal forms of the verb. Compare, der Fluss steigt unauf'hôrlich – le fleuve ne cesse de monter [the river is rising inces – santly l; lesen Sie weiter [ 'read on 'l — continuez a lire; er leugnet hartnäckig ["he stubbornly rejects ' ), il s 'obstine a nier; Man hat so- gar gesagt ['it has even been said " ), on est alle jusqu'a dire; the German gern and the Russian 'gladly' are given in French by aimer: ich lese gern - j'aime a lire ['I read gladly"). Compare also the transla– tion of the comparative degree of the dema: gern-lieber, as well as the Russian "better' by the French préférer, including, in comparison with aimer, the additional elementary meaning expressed in the C Ompara - tive degree (je préfere – j'aime mieux): Ich lese lieber ['I had better read'], je préfere (j'aime mieux) lire. Compare the opposite in French J'aime le vin with the German ich trinke gern Wein. In conclusion, let us consider the following problem. In §32 we – 26l - saw that words which have no correspondences in the intermediate lan- guage and are not formulas are placed in correspondence with the phon- emes by sequences from the intermediate language. With synthesis, in this case we must set the sequence of letters in the target language into correspondence-with the sequence of phon- emes in the intermediate language. - - - Here it is necessary to distinguish the following possibilities (Superanskaya, pages lik–81): - l) the intermediate-language phoneme corresponds to the phoneme in the target language; - 2) there is no such correspondence. In the is: case the phoneme of the intermediate language in synthe- S is receives the same literal fixation as the corresponding phoneme in the target language as , for example: {{...} From the French Chesne Russian Shen gaus) From the German Gauss Russian Gauss In the second case the intermediate-language phoneme is rendered by a letter corresponding to one of the target-language phonemes which are from the articulator -accoustic Standpoint is closer to the inter- mediate-language phoneme. 's- This situation is encountered with particular frequency in the synthesis of words from the phonemes of the intermediate language cor- responding to the English phonemes (compare Aristov, 39–40), but 1t iS also characteristic of translation into other languages such as , for example, German (shvanebakh, Revzin, 34–35). Experience shows that a certain lack of coordination arises in this case; frequently the surnames of an individual are written (or, a S is generally said, transliterated) in various ways in various publica- tions. Thus, for example, the German diphthongs ei and eu are not ren- dered as they are pronounced in accordance with the general German pro- – 262 - nounciation standards ("oy," "ay"), but the way they are pronounced in southern Germany ("ey"). Therefore - - Feuchtwanger is rendered as Feykhtvanger rather than Foykhtvanger Neues Deutschland is rendered as Neyes Deychland rather than Noyes Doychland; - - Leipzig is rendered as Leypt Sig rather than Laypt Sig. Until recently it has been traditional to render the sound h (spi- rant) through the Russian g (associated, apparently, with the south Rus- stan and Ukranian pronounciation of 'g' [khl which is close to the German h. Therefore : Heinrich Heine is rendered as Genrikh Geyne rather than Khaynrikh Khayne. Recently, the sound h has been rendered phonetically as kh. In all such cases an exception is made for those names and designa- tions for which a definite tradition in written Russian has been estab — lished as , for example : Humboldt — TyM60Jibnt Herbert — ſep6ept Reuter — Peñtep Wien — Beha Paris — Tlaphx Washington— Bauhhrton Sachsen — Cakcohha Mailand — Mhnah. The synthesis process described above - occasionally known as transliteration – is the final process only for proper names, Since the latter are not subject to any categorization in the target-language Sys- tem. The problem is somewhat more complicated in the case of neologisms or realities for whose synthesis it is necessary to provide the word with all categorie S of the corresponding part of speech in the target language. - 263 – As an example, let us take the German correspondence to the word . ‘kolkhoz' (see Revzin, 1959, pages 2011-205). Let us assume that we have to translate this word by the method of transliteration. It turns out that it is not enough Simply to introduce some word; this word must be – come part of the language System, i.e. , it must assume its position in the declension, and this means that it must be assigned some gender. Since the word has no particular indicator S ; proceeding from various considerations, various people assigned different genders - masculine (der Kolchos) was used out of consideration of the fact that in Russian the combination 'kolkhoz ' is masculine ; neuter gender (das Kolchos) is used out of consideration of the fact that "kollektivnoye khozyaystvo" [' collective economy' ] is neuter; the feminine gender (die Kolchose ) is used because Kollektivwirtschaft is feminine (a rendition of thought patterns, as used before). With regard to the form die Kolchose , it must be admitted that this is not a suitable approach because of its Sound, since it is homonymic with the nonsensical, and yet related, combinat 1 on : Kohlhose. The form das Kolchos is not suitable because this word gravitates to a Specific declension classification, and namely , to the class of neuter-gender words ending in -os as , for example, Epos; in the plural , the ending -os is replaced by -en, and there is a tendency to pronounce daS Kolchos as die Kolchen. Therefore, the form der Kölchos should be regarded as the most Suitable. However, as of the moment all three forms are encountered. - . We have considered the possibilities of synthesizing nonequivalent vocabulary in translation. It is clear that in the case of interpreta- tion, i.e. , referral to reality, other solutions are also possible (see §13). - 264 - $37. Syntactic PROBLEMs of SYNTHESIs Synthesis in general form is accomplished in accordance with the diagrams of formation which we analyzed in $19. In particular, configurational synthesis is accomplished as a pro- Ce SS which is the opposite of configurational analysis. In configurational synthesis (Moloshnaya, 1960, page 268) the tar- get-language configurations will be accomplished in a sequence opposite to that of the source language. On page 230, we find a list of word classes identified by Moloshnaya for English. Let us now present exam- ples of corresponding classes in Russian: P is a personal pronoun; l_is a noun; - - 2 is an intransitive verb in personal form (irrevocable) 3 is an adjective. Here we will agree to ascribe to each index some morphological in- formation (indication of case, number, type, etc.) and, moreover, through the use of the sign Y, written above the index, to indicate agreement, the determining Word being identified by means of the in- dex ~ as , for example , the expansion formula * T., 3 f Shows that the noun is expanded in the sequence noun + adjective, where the adjective agrees with the noun. In §23 we examined in detail the analysis of the English clause: / Let us call the above-mentioned lemma the fundamental lemma and we - * . . . . - • saw that it leads to the base configuration: Ps 2"s - We introduce the following rules of synthesis (corresponding to the expansion rules analyzed above : J) 2-2 to be 1 accusative l instrumental. 2) 1 =3 I - 265 – - * The process of synthesis can now be shown in a table: IIpHMehaeman ChhteahpyeMaa nocneñobarènehocre 1 popMyna . . . 2 HH ReRCOB - - - 3 - " -- l, 5 P, 2 s =2 s 6yń. 1 n., 2 a 6yn. 1 M., Mii. v. M.H. W.H. . 2,+2, * BñH. le #. 2Ts 6yſ. 1 M., Ms. v. 1, BRH. 1, #. 1.-3, !, -- X. ſº-3, i. 2 sºyā, l a. *s, *. 3. 1. Bhh. - -- 3, ls TB. 2,+=Hasbath 8 . • . . 3s-Bhuieynoºshytºtº 9 ls, le=Memma lo 3s=ochobhoi ll gasobem Bºuleynominhyrylo nexmy ochoahoº Jiemºmoi 1) Formula used; #} synthesized sequence of indices; 3) to be; l;) per- sonal;5) plural; 6) accusative; 7) instrumental; 8) to call; 9) above- mentioned; lo) lemma; ll) fundamental; lº) let us call the above-men- tioned lemma the fundamental lemma. Let us dwell in some detail on the problem of the limitations im– posed on the target language and thus affecting the synthesis of the phrase. The most general case of limitations imposed on language is the Word order necessary for whatever reas on in the given language. Let us consider an example of Synthesis for a German combination in which one of the words of the intermediate language corresponds to a Russian adjective. Let us take the combination "The Union of Soviet socialist Republic's ' , rendered in German as ate Union der Sozialis - t is chen Sowjetrepubliken. It is easy to see that in the synthesis some repositioning has taken place in the second and third places: compare: ' l 2 - 3 Compare : "The Union of Soviet Socialist Republics. ' l - 3 2 - Die Union der Sozialistischen Sowjetrepubliken. The rearrangement is a result of the fact that the words 'Soviet Republic are translated by the single word: die Sowjetrepublik, while - 266 - the adjective sozialistisch assumes the first position. Following this same model *A large socialist enterprise is translated as : ein Sozial- 1stischer Grossbetrieb. This is , of course, associated with the fact that the semantic set 'large enterprise on synthesis corresponds to the single word (Grossbetrieb). Let us now consider certain other problems of synthesis into Ger– man. Suppose we are given the combination 'servicing the process of merchandise flow ' and in the interest of eliminating a string of geni- tives we would like to form a compound word. Theoretically we have the following possibilities: 1. die Betreuung des Prozesses – die. Prozeßbetreuung 2. der Prozeß der Zirkulation=der Zirkulationsprozess 3. die Zirkulation der Waren =die Warenzirkulation. However, in this phrase it is possible to use only the third corn- pound word. Why? It turns out that in the formation of compound words in German the following rule applies: a noun Subordinate grammatically to another noun may be changed into the first component of a compound word only in the event that it is independent of all other words. Consideration of limitations is important also in the synthesis of correspond ing Russian double nouns, and namely, nouns connected with a hyphen, such as 'deputy-communist , ' 'poet-democrat," 'writer-fighter, ' 'artist—amateur, ' 'scientist—artist, etc.; [Translator's Note: in each of -the foregoing cases the second part of the hyphenated Word Serve S as an adjective describing the first part of the compound word, e.g. , 1n the first case the deputy is a member of the communist party; in the second case the poet is a democrat ; in the third case the writer is also a fighter, etc. l. - It is impossible to translate these combinations literally into German or French; for example, it is impossible in German to say Dich- º 267 º-º ter-Demokrat. We are dealing here with the fact that in Russian the se– quence is progressive, whereas in German the sequence is regressive, and in German, as a rule , such words, if they exist, are words of the definitive type in which the first component determines the second as , for example, in the combinations 'deputy-communist , ' 'poet-democrat where the main words are 'deputy and poet, ' and therefore in the trans- ition from the intermediate language to the German it is possible to change the second word into an adjective: der kommunistische Deputierte and ein demokratischer Dichter, etc. §38. THE INDEPENDENCE OF ANALYSIS AND SYNTHESIS In the familiarization with the examples used in the previous Sec – tions the impression might have arisen that translation is a function of the source language or , more accurately , that the synthesis Stage is a function of analysis. Generally speaking, this is frequently indeed the case with translation. This was the Situation in the construction of the so-called binary translation algorithms; Russian-English, English- Russian, French-Russian, etc. In this case the intermediate language was a system of correspondences between two given languages and the pro- perties of this intermediate language were a strong function of the pair of languages. Moreover, in certain respects it is convenient to construct precise- ly such binary intermediate languages whose rules of translation from language Ll to language La in a number of cases prove significantly air- ferent from the rules of translation from I-2 into Ill. In other words, these rules are irreversible. Thus, it was shown in machine translation that the relationships between Russian and English are such that in the simpler situations the rules of translation from Russian into English may be very simple, whereas the rules of translation for the correspond- ing English phrases into Russian are extremely complex. - 268 - As an example let us consider a phrase from the first experiment with translation into Russian ("Machine Translation, "page 171ſf). As we have seen, in this experiment the following phrase Wa S translated: Beauma yraa onpemenaetch othouſehnem-myrh kipanhycyſ "The magnitude of the angle is defined by the ratio of the arc to the radius' ]. The words ‘yrna', ‘othouſeilhem’, ‘Ilyth' and 'panaycy'ſ ' of the angle , ! 'by the ratio, ' ' of the arc ' and 'to the radius' ] on analysis are divided into stem and ending, where it is shown that the endings a and u express the genitive case; the ending em indicates the instrumental case and y indicates the dative case. On synthesis the meaning of the genitive case is set into correspondence with the element of ; the meaning of the dative case is set into correspondence with the element to; and the meaning Of the in - Strumental case is set into correspondence with the element by. Then a rule is given: the element corresponding to the meaning of the case is placed before the word to which it pertains. As a result we obtain the translation: Magnitude of angle is determined by the relation of arc to radius. Let us note that the form is determined and the article in front of the word relation was simply taken from the dictionary, i.e. , We arºe convinced On Ce again that the division into dictionary and gram- mar has by no means been finally established. The success in the described machine-translation experiment can be explained, among other reasons, by the absolutely unprejudiced approach to problems of grammar, a result of the over-all course Of development of American descriptivism. The problem also lies in the fact that herºe a particularly suitable pair of languages has been chosen for the trans- lation. The "passive grammar" for Russian is more or less simple: vir- tually every word indicates its function with its ending. The "passive grammar" of English is considerably more complex and on analysis it is necessary each time to refer to context and to carry out complex logical – 269 – operations. Conversely, the "active grammar" of Russian is extremely °omplex; each meaning corresponds to a number of forms (for example, de - - - 2 º pending on the type of declension or conjugation); the "active grammar" of English 1s simpler. However, these statements are valid only for specific Sublanguages. Also, in the analysis of a Russian phrase homonymy is frequently en- Countered (including grammatical hononymy - see Nikolayeva, 1962, page loſis). - On the other hand, in research on machine translation it was re- Cognized rather quickly that there was an advantage to be gained from the onstruction of rules such that the analysis of the Source-language text not depend on the properties of the target language, and vice ver- Sa. The gain here is primarily quantitative. Let us imagine that we are confronted with one hundred languages and we want to translate from each into each. In this case we will need 9900 various sets of rules, since we have precisely that number of paired combinations of languages. If we make use of independent analysis and Synthesis, i.e., if we have a separate set of rules for analysts and ror Synthesis for each lan- guage, it will obviously be enough to carry out 100 analyses and loo Syntheses (Uspenskiy, lg59, page l;0). Independent analysis, i. e. , transition from the Source-language text to the 'intermediate language, correspond; to the "understanding Of the text," i.e., to that general understanding which, in the unanimous Opinion of all theoreticians, is a necessary prerequisite of tranna- tion. In formalization, understanding (analysis) is reflected as the translation of vocabulary and syntactic relationships of the given text into some universal vocabulary and universal Syntactic relationships (Mel 'chuk, 1960). The system of notation for this universal vocabulary. - - 270 - and the universal Syntactic relationships is the intermediate language. Knowing the means of expressing the general concept and relationships with the facilities of the target language or, in other words, knowing the system of correspondences between the intermediate language and the given specific language, we can easily synthesize the text in that lan- guage. * * Independent synthesis, i.e. , the transition from the intermediate language to the text of the target language corresponds to the creative instant in translation, and namely, "to the expression of thought." The independence of analysis from synthesis is of interest from the standpoint of simulating this approach to translation. Note. The above-mentioned experiment with no distinction between analysis and synthesis, in general, is very characteristic of the ini- tial results of the so-called 95%—approach to machine translation (§3) and therefore , although its practical value today is in doubt, it re- tains an important instructional Significance, since it illustrates by means of a simple example the principles of the pure machine (that is to say, antilinguistic) approach to translation. Below we will present a detailed description of this approach, restricting ourselves to the phrase which we cited earlier, and namely: 'The magnitude of an angle is defined by the ratio of the length of arc to the radius. 'The dictionary in this case has the form Shown in the table. - 2 - l, Kakoe l pyccKne Ahraſińcklift 3xaneanehr | 3 Ko A M npashno C.T.O.33 P A I II I II. III eTCſ. * - A w” Bºiii!!!MH3 magnitude - -* - 6 * t yrn- Coal, angle 121 – 25 2 y -3 of 13| || 222 25 3 onpeneſiaetcais determined orhouehh- relation, the relation 151 *= — 5 -emſ by 13| | – || – 3 AJIH II- length - ess - 5 - -H! of 13| — 25 § myr- aſ C - *s- - -14 of . 13| --- 25 3 K to, '' for 212 || – 23 2 panhyc- radius - 221 - 6 - -y to 131 || – - 3 #} Russian words; 2) English equivalent ; 3) codes; 1) the rule employed. – 271 – - We present 6 syntactic rules: * Repositioning. * If the first code is llo, the third code of the Preceding complete word must be recognized. If the code is 21, the se - - Quence of these words must be changed (the word with the code 21 must follow the word having the word llo); in the Opposite case, the word Order is retained. In both Cases, for the word with the code ll O a a- Éle English equivalent is Selected. 2. Selection on the basis of subsequent text. If the first code of the given word is 121 it is necessary to recognize the second C Ode of the following complete word or part of a word (root or ending). If this Second code is 221, for the word having the code 121 the Ist English equivalent must be taken; if the second code is 222, the IInd English °quivalent must be selected. In both cases the word order is retained. 3. Selection With repositioning. If the first C Ode of the given Word is 131, and the third code of the preceding complete word or part of a word (the root" or ending) is 23, for the word having the code 131. We must take the TInd English equivalent and retain the word order. If the first code of the £iven word is 131, and the third Code of the pre- Ceding word is 25, then for the Word having the code 131 We must take the Ist English equivalent and reposition the words. 4. Selection ãccording to the Preceding text. If the first code Of the given word is 141, and the second C Ode of the preceding complete Woºd or any part (root or ending) of the preceding separated word is 24l, then for the given word we take the Ist English equivalent. If, however, the second code is 242, we take the IInd English equivalent. In both cases the word order is retained. - 5. Selection with omission. If the first code of the given word is 15l, and the third code of the preceding complete word or any part (root or ending) of the preceding separated word is 25, then for the word having the code 151 we take the IInd English equivalent ; in the opposite C a Se , We take the Ist English equivalent. In both cases the Word Order, is retained. - - 6. Division. If the first code of the £1ven word is "-" then for this word we take the English equivalent I and the Word order is re- tained with respect to the preceding word. These are the basic features of this machine-translation experi- ment . - The translation programs for Russian were considerably more com- - 272 - plex, first of all, because here we were dealing at once with the 100%- approach and , secondly, because of a different relationship between the languages. It is impossible for us at this point to present the entire program and dictionary required for the translation of even only a sin– gle sentence (see Kulagina, Mel 'chuk, 1956; Kulagina, 1960, pages 181- 208; Kulagina, 1962). - - - Below we will present only a general description of the stages in- volved in the analysis of the first French phrase translated by means of the "Strela" computer at the Steklov Mathematics Institute of the USSR Academy of Sciences in June 1956: Nous allons considerer d'abord deux systemes d'équations. - l. The first program selected the following stems and endings from the dictionary: l Ochoba 2 Okohuahue 3 *Iacrb peah IłOUIS l; Mectonmehhe all OITS ; º: - ef TJiaº O * mpenſior abord cyulectBhtenbHoe deux ułłCJIHTenbh O6 ºme - S : cyluectBhtenbHoe 3. mpeſinor équation S 7 cyldecreatenshoe 1) Stem; 2) ending; (3) part of speech; 4) #: § verb ; é) preposition; 7) noun; 8) number. *... " 2. The evaluation program found the expression d'ab ord in the phrase (which for the computer represents two independent words de and abord). Instead of a word-for-word translation the computer provided adverbial information ('from the beginning'). - 3. The following program of recoding Information produced a Scale for the parts of speech and a scale of words exhibiting individual fea- tures such as, for example: 1) a scale of verbs OlloCOOO; 2) a scale of nouns 00000101; - 3) a scale for the verb aller Oloooooo (a word with an individual feature). - for the l;) Since in this phrase there are no homonyms, the program for - analysis of homonymy is not functioning here. - 273 - . 5. As a result of the analysis program we obtain the following: a), the case of the first word was determined — the pronoun "we" (nous) – and it was given as nominative, and not as one of the indirect cases, because following the pronoun is the related verb ; b) the combination of words allons considérer was analyzed and it was translated by means of the verb 'to consider' in the first person, plural, future tense; c) the genitive singular was assigned to the noun 'system' (sys- temes), since it was preceded by the noun 'two' (deux); d) the number of the noun 'equation' (equation) was determined to be plural, on the basis of the ending -s to the noun and the case was determined to be genitive on the basis of the preceding preposition. 6. The synthesis program, employing the derived information for the Russian word selected from the Russian dictionary, and the data de- veloped by the analysis program, gave us the required forms of the Rus — Sian words. * . . .' As a result, the computer printed out the phrase: "We are first of all considering 2 systems of equations. ' Manu- script [Footnotes Page No. - 22O By the way, this was the initial idea of an automatic dic- tionary ("Machine Translation," page 17). 224 This includes the cases in which situations that are ordinary for the culture of a single people are not found in the other culture. The authors of "Stylistique comparée" make special mention of this case by referring to gaps ' (lacunes). Compare Malblanc, 31. / : - + -i-.' • . - 225 Let us note that these exercises are extremely important for the development of determinations from context , both on the part of translators, and in general for students learning a foreign language. Compare, for example, the following exer- cise in the filling in of gaps in the phrase: Notre repre- sentant . . . vous voir la Semaine prochaine (passera). Je Vous . . . que nous venons d'engager un nouveau représentant (informe). Vous constateréz que nos prix. . . toute concurrence (défient) (Mauger, page 82). - 23O Fries used this classification as the basis for the textbooks of the English language which had been developed under his guidance. The generalized discussion of the principles in- gºe 27: - 245 271 272 volved in the preparation of these textbooks are presented in Fries' report to the VIIIth International congress of fin- guists (Fries, 1957). Yu. S. Martem 'yanov graciously gave his permission to have his own discussion of this method published in our text (see Mar- tem'yanov, pages 2–20). The endings M and B are not distinguished here. The first rule in the translation of our sentence does not apply , but we presented here for the purposes of a complete description. We will explain its effect in the following ex- ample : The Russian expression "general-major must be trans- lated into English as major general. This is accomplished by assigning code 21 to the word "general and the code 110 to tº: Word 'major. ' Thus Rule I accomplishes a correct transla– OIl . REFERENCES Akulenko, V. V. , sushchestvuet li internatsional' naya leksika [Does an International Vocabulary Exist? ), "Voprosy yazykoz- naniya" ["Problems of Linguistics"), No. 3, 1961. Andreyev, N. D. , Mashinnyy perevod i problema yazykapos redni- ka [Machine Translation and the Problem of the Intermediate Language], "Woprosy yazykoznaniya," 1957, pages 117–121. Andreyev, N. D. , Meta-yazyk mashinnogo perevoda i ego primen- eniye [The Metalanguage of Machine Translation and Its Util- ization], "Materialy po mashinnomu perevodu" ["Materials on Machine Translation"], I, Leningrad, l958, pages liO-60. Andreyev, N. D., Ivanov, V. V. and Melchyk, I. A. , Nekotoryye Zamechaniya i predlozheniya otnositeino raboty po maShinnomu perevodu v SSSR [ Certain Remarks and Suggestions with Respect to Work on Machine Translation in the USSR), "Mashinnyy per- evod i prikladnaya lingvistika" ["Machine Translation and Applied Linguistics"], Moscow, 1960. Andreyev, N. D. , Zambrzhitskiy, W. L. , Pazukhin, R. V. and - 275 - S lC. ll. 12. :: * * * * Pak, G.A., Postroyeniye i funktsionirovaniye slovarya yazyka-posrednika v pole vkhodnykh yazykov [Preparatioh and Functioning of the Dictionary of the Intermediate Language in the Field of Input Languages], "Doklady na konferentsii p C corabotke informatsii, mashinnomu perevodu i avtomatiches- komu chteniyu teksta" ["Procedings of the Conference on the Processing of Information, Machine Translation and Automatic Reading of Text"), Moscow, 1961. Aristov, N. B. , Osnovy perevoda [Fundamentals of Translation], Mcs c cº, l959. Bat iſ...tsev, A. A. , Yaponskaya perevodcheskaya mashina "Yamato" Tºle Japanese "Yamato" Translation Machine], "Mashinnyy per- evci i prikladnaya lingvistika" ("Machine Translation and Ap- iieſ: Linguistics"], No. 3 (10), 1959. P all. O. , Obshchaya lingvistika i voprosy frantsuzskogo yazºra [General Linguistics and Problems of the French Lan- guage . , Translated from the French, Moscow, l055. Baily C. , Frantsuzskaya stilistika [French Stylistics J. raris lation from the French, Moscow, 1957. Bar-ºillel Y. , Idiomy [Idioms), "Mashinnyy perevod" ["Machine Translation"], Translated from the English, Moscow, 1957. Bar-Eillel Y. , Budushcheye mashinnogo perevoda [The Future of Machine Translation], "Filologicheskiye nauk" ("The Phil- ological Sciences"], No. 11, 1962. Barzºudarov, L. S. , Obshchelingvisticheskoye znacheniye teorii perevoda [The General Linguistic Significance of the Theory of Translation], "Teoriya i kritika perevoda" ["Theory and Critique of Translation"], Leningrad, 1962. Barzºudarov, L. S. and Kolshanskiy, G. V., K voprosu o vozmozh- - 276 - l!!. lº. l6. 17. 18. 19. 20. 2l. 22. 23. nostyakh mashinnogo perevoda (The Problem of Machine-Transia- tion Potentials], "voprosy yazykoznaniya" ("Problems of Lin- guistics"], No. 1, 1958. Bozhno, L. I. , Tekhnicheskiye terminy v nemetskom yazyke [Technical Terms in the German Language], Moscow, 1961. Booth A. D. and Locke W. H., Istoricheskoye vvedeniye [Histori- cal Introduction], Collection entitled "Mashinnyy perevod" ["Machine Translation"), Translated from the English, Moscow, 1957. - Brillouin L. , Nauka i teoriya informatsii [The Science and Theory of Information], Translation from English, Moscow, 1960. Vinogradov, V. V., Russkiy yazyk [The Russian Language], Mos- cow, 1947. Woprosy khudozhestvennogo perevoda [Problems of Literary Translation], Moscow, 1955. Gal" perin, I. R. , Perevod i stilistika [Translation and Styl- 1stics], Collection entitled "Woprosy teorii i metodiki wenenose perevoda" ["Problems in the Theory and Methods of Instructional Translation"], Moscow, 1950. Hilbert D. and Ackermann V., Osnovy teoreticheskoy logiki. [Fundamentals of Theoretical Logic], Translated from German, ..f ..? Moscow, 1917. . . .x. *, Glisson, Vvedeniye v deskriptivnuyu lingvistiku [Introduc- tion to Descriptive Linguistics], Translated from English, Moscow, 1959. - - - - • . Goldman, G. , Teoriya informatsii [The Theory of Information], Translated from English, Moscow, 1957. . - Griftsov, B.A., zametri potekhnike perevoda [Notes on the - - 277 - 211. 25. 26. 27. 29. 30. Technique of Translation], "Woprosy yazykoznaniya" ("Problems Of Linguistics"), No. 5, 1952. . - - - 31. Humboldt W. , , Q razlichii stroyeniya chelovecheskikh yazykov 32. i ego vliyanii na dukhovnoye razvitiye chelovecheskogo roda [Variations in the Structures of Human Languages and Its In- fluence on the Spiritual Development of Humanity), "Khresto- matiya po istorii yazykoznaniya XIX-XX vekov" ["Reference on 33. the History of Linguistics in the XIX-XXth Centuries"], Mos- cow, 1956. Hjelmslev L. , Metod Strukturnogo analiza v lingvistike [The Method of Structural Analysis in Linguistics J. 3/4. Hjelmslev L. , Mozhno li Schitat', chto znacheniya slov obraz- uyut strukturu? [Is it Possible to Maintain that the Meanings of Words Make up Structure? ), "Novoye v lingvistike" ["Lin– guistic News"), Issue 2, Moscow, l06l. Yespersen O. , Filosofiya grammatiki [The Philosophy of Gram- mar], Translated from English, Moscow, 1955. Khirkov, i. i., Granitsy primenimosti mashinnogo perevoda 35. [Limits of Applicability for Machine Translation], "Woprosy yazykoznaniya" ["Problems of Linguistics" j, No. 5, 1956. Zhirmunskiy, V. M., Natsional'nyy yazyk i sotsial'nyye dial- ekty [The National Language and Social Dialects], Leningrad, 1936. Zaliznyak, A. A. , Opyt obucheniya anglo-russkomu perevodu S - 36. pomoshch'yu algoritma [Experience in Teaching English-Russian Translation by Means of an Algorithm], "Pitannya Prikladnoi lingvistiki, Tezi dopovidey mizhvuzivskoi naukovoi konfer- entsii" ["Problems of Applied Linguistics, Topics of the Re- 3 7 © ports to the Scientific Conference of Institutes of Higher – 278 - Education"], Chernovets, l960. Zvegintsev, V.A., Semasiologiya [Semasiology], Moscow, 1958. Zvegintsev, V. A., (Sostavitel') Khrestomatiya po istorii yazykoznaniya XIX-XX vekov [ (Edited by) Reference on the His– tory of Linguistics in the XIXth-XXth Centuries J, Moscow, l956. Zvegintsev, V. A. , Istoriya yazykoznaniya XIX i XX vekov v ocherkakh i izvlecheniyakh [The History of Linguistics in the XIXth and XXth Centuries in Outline and Excerpts j, Moscow, Part II, 1960. - Ivanov, V. V. , Lingvistika kak teoriya othosheniy mezhdu yaZykovymi Sistemami i yeye Sovremennyye prakticheskiye p rilo- zheniya [Linguistics as a Theory of Relationships Between Language Systems and Its Contemporary Practical Application j, Collection entitled "Lingvisticheskiye issledovaniya po ma– Shinnomu perevodu" ["Linguistic Studies of Machine Transla– tion"], Issue 2, Moscow, l961 (photoprint edition). Ivanov, V. V. , Lingvisticheskiye voprosy Sozdaniya mash innogo yazyka dlya informatsionnoy mashiny [Linguistic Problems In- Volved in the C OnStruction of a Computer Language for a Data- Process System], "Materialy po mashinnomu perevodu" ("Mater- ials on Machine Translation"], I, Leningrad, 1958, pages lo- 39. x * Ivanov, V. V. , Ponyatiye neytralizatsii v morfologii i leksike [The Concept of Neutralization in Morphology and Vocabulary ), "Byulleten' ow"earneniya po problemam mashinnogo perevoda" ["The Bulletin of Problems in Machine Translation" j, No. 5, l957 (in mimeograph edition). Ivanov, V. V. , Yazyk v Sopostavlenii s drugimi Sred Stvami per- – 279 - 3 9 edachi i khraneniya informatsii [Language in Comparison with other Means of Transmission and Information Storage), "Dok- lady na konferentsii po obrabotke informa & Sii, mashninomu Perevodu i avtomaticheskomu chteniyu teksta" ["Proceedings at the Conference on Tnformation Processing, Machine Transla- tion and Automatic Text Reading" j, Issue 7, Moscow, l06l. Ivanov, V. V. , Kod i Soobshcheniye [Code and Communication], "Byulleten" ow"earneniya po problemam mashinnogo perevoda" ["Bulletin on Problems of Machine Translation" J, No. 5, Mos- cow, 1957 (in mimeographed edition). Isachenko, A. V. , Grammaticheskiy stroy Yºu SS Kogo yazyka V SOPostavlenii Go Slovatskim [The Grammatical Structure of the Russian Language in Comparison with the Slovak Language J, Morfologiya [Morphology j, Part I, Bratislava, lgºl. Isºchenko, A. and Pyurovich, L., O sopostavitel' nom izuche- nii Slavyanskikh yazykov [The Comparative Study of Slovak Languages], "Russkiy yazyk v. natsional'noy shkolye" ["The Russian Language in the National school" ), No. 3, 1958. Kulagina, O. S. , Operatornoye opisaniye algoritmov perevoda | Operational Description of Translation Algorithms j, "Mashin- *YY Perevod i prikladnaya lingvistika" ["Machine Translation and Applied Linguistics"), No. 2, (9), 1959, pages 6-22, 3 *. • ‘ (10), pages 3–34. Kulagina, O.S., O mashinnom perewode S frant Suzskogo yazyka na russkiy, I, Slovar' [Machine Translation from French into Russian, Dictionary I) "Problemy kibernetiki" ("Problems in Cybernetics" J, Issue 3, Moscow, 1960; II, Algoritm perevoda S frantsuzskogo yazyka na russkiy [The Algorithm for Transla– tion from French into Russian], "Problemy kibernetiki" – 280 - 113. |l. l6. li'7. 118. ["Problems in Cybernetics"), Issue li, Moscow, l062. Kulagina, O.S. and Bakulovskaya, G. W. , Opytnyye perevody s frantsuzskogo yazyka na russkiy na mashine "Strela" [Experi- mental Translations from French into Russian by Means of the "Strela" Computer], "Problemy kibernetiki" ["Problems in Cy- bernetics" j, Issue 2, 1959, pages 283–288. Kulagina, O. S. and Mel' chuk, I. A. , Mashinnyy perevod s frant Suzskogo yazyka na russkiy [Machine Translation from French into Russian], "Woprosy yazykoznaniya" ["Problems in Linguistics"), l956, No. 5, pages lll-l21. Kuznet SOV, P. S. , Lyapunov, A. A. and Reformatskiy, A. A. , Osnovnyye problemy mashinnogo perevoda [Fundamental Problems of Machine Translation], "Woprosy yazykoznaniya" ["Problems of Linguistics"), 1956, No. 5, pages 107–ll0. Kuznetsova, N. N. , Perevod na nemetskiy yazyk othositel'nykh prilagatel'nykh russkogo yazyka [T] anslation into German of Relative Russian Adjectives ), "Uchenyye zapiski LGPIIYA" ["Scientific Notes of the Leningrad State Pedagogical Insti- tute of Foreign Languages"), New Series, No. 2, 1955. Larin, B.A., Nashi zadachi [Our Tasks], "Teoriya i kritika perevoda" ["The Theory and Critique of Translation"), Lenin- grad, 1962. g Lakhuti, D. and Stokolova, N., O zadache polska khimicheskikh referatov po Zaglaviyam [The Problem of Seeking Out Chemical Abstracts on the Basis of Titles], "Doklady na konferentsii po Obrabotke informatsii, mashinnomu perevodu i avtomatiches- komu chteniyu teksta" ("Proceedings of the Conference on the Evaluation of Information, Machine Translation and Automatic Reading of Text"), Issue I, Moscow, 1961 (photoprºnt edition). - 281 - #9. 5 o 5 i e 5 3 Lebeg., G., Ob izmerenii velichin [Measurement of Magnitude], Translation from the French, Edited by A. N. KolmogoroV, MOS- cow, 1938. Leykina, B. M. , K voprosu o strukture yazyka—posrednika [The Problem of the Structure of the Intermediate Language], * Ecklady na konferentsii po obrabotke informatsii, mashinnomu rerevodu i avtomaticheskomu chteniyu teksta" ["Proceedings of ++e Conference on Evaluation of Information, Machine TranSla- tion and Automatic Reading of Text"), Issue 5, Moscow, 1961. Leont yeva, N. N. , Model" sinteza russkoy fražū na osnove S e Iſl— a-icheskoy zapisi I MGPITYA [A Model of the Synthesis of a 2 : -:ssian Phrase on the Basis of Semantic Notation I MOSCOW s—ate Pedagogical Institute of Foreign Languages], "Doklady na konferentsii po obrabotke informatsii, mashinnomu perevodu s artomaticheskomu chteniyu teksta" [" Proceedings of the Con- == rence on the Evaluation of Information, Machine TranSla- -i-n and Automatic Reading of Text"), Issue 9, Moscow, 1961. • *- * Icronosov, M. V. , Polnoye sobraniye Sochineniy [ Complete Col- - * * * ** Lyapunov, A. A. , O nekotorykh obshchikh voprosakh kibernetiki 2-2-ain General Problems of Cybernetics], "Problemy kiber- re-iki" ("Problems of cybernetics"), Issue 1, Mosco", 1958. Lyapunov, A. A. and Kulagina, O. S. , 9 rabotakh po mashinnomu = revodu Matematicheskogo instituta AN SSSR [Research on Ma- p crine Translation at the Mathematics Institute of the USSR Acziemy of Sciences], "Theses from the conference On Machine -: anslation"], Moscow, 1958. Mazowskiy, M. M. , K probleme tak nazyvay emoy "internatsional' - ncy" leksiki [The Problem of the so-called "International" - – 282 - 58. 59. e !! !! e tika ["Machine Translation and Applied Linguistics"), No. 4 1960. * - / t Vocabulary], "Woprosy yazykoznaniya" ["Problems of Linguis- tics"], No. 1, 1960. Mashinnyy perevod [Machine Translation], Translation from the English, Moscow, lS57. Markish, S. , Neskol" ko zametok o perevodakh s drevnikh yazykov [Some Remarks About Translations from the Ancient Languages], Collection entitled "Masterstvo perevoda" ["The Art of Translation"), Moscow, l059. Martem" yanov, Yu. S. , O kodirovke slov diya algoritma Sintak- sicheskogo analiza [The Coding of Words for the Algorithm of Syntactic Analysis ), "Doklady na konferentsii po obrabotke informatsii, mashinnomu perevodu i avtomaticheskomu chteniyu !! 11 - teksta" [" Proceedings at the Conference on the Evaluation of Information, Machine Translation and Automatic Reading of Text"), Issue 10, Moscow, 1961. Materialy po mashinnomu perevodu, I [Materials on Machine Translation, I.J., Izd-vo LGU [Leningrad State University Press], Leningrad, 1958; 2–1963. ! Mel' chuk, I.A., K voprosu o "grammaticheskom" v yazykepos red- nike [The Problem of "Grammar" in the Intermediate Language ) Col - - !! e * * lection entitled "Mashinnyy perevod i prikladnaya lingwig- 3 Mel' chuk, I. A. , Mashinnyy perevod i lingvistika [Machine Translation and Linguistics], "O tochnykh metodakh is sledo- vaniya yazyka" ["Exact Methods of Studying a Language"), COm- piled by O. S. Akhmanova, I. A. Mel' chuk, Ye. V. Paducheva, R. M. Frumkina, Moscow, 1961. Mel' chuk, I.A., Nekotoryye vyvody obshchego kharaktera V – 283 – 63. 611. 65. 69. svyazi s mashinnym perevodom S vengerskogo [ Certain Conclu- SiOnS of a General Nature in Connection w th Machine Transla- tion from the Hungarian], Collection entitled "Mashinnyy per- 70. evod i prikladnaya lingvistika" ["Machine Translation and Applied Linguistics"], No. 6, 1958. Mel' chuk, I. A. , O mashinnom perevode s vengerskogo yazyka na russkiy [Machine Translation from Hungarian Into Russian], iſ problemy kibernetiki" ("Problems in Cyberneties"], Issue l, 71. Moscow, 1958, pages 222–2611. Mel' chuk, I. A. , O nekotorykh tipakh yazykovykh znacheniy [ Certain Types of Linguistic Meansings], "O tochnykh me todakh is sledovaniya yazyka" ["Exact Methods of studying a Language"), Compiled by O. S. Akhmanova, I. A. Mel' chuk, Ye.V. Paduche Va, R.M. Frumkina, Moscow, 1961. Mel chuk, I. A. , O terminakh "ustoychivost' " i "idiomatich- 72 ost " [The Terms "stability" and "Idioms"), "Voprosy yazy- e koznaniya" ("Problems in Linguistics"], No. 4, 1960b. Mel' chuk, I. A. , Raboty po mashinnomu perevodu V SSSR [Work on Machine Translation in the USSR), "Vestnik AN SSSR" ["Her– ald of the USSR Academy of Sciences" ), No. 2, 1959, page:S -,-4 li3-li'7. Mendeleyev, D. I. , Osnovy khnmiki [Fundamentals of Chemistry], ºth Edition, St. Petersburg, 1906. Moloshnaya, T. N. , Algoritm perevoda S angliyskogo yazyka na russkiy [Algorithm of Translation from English into Russian ], iſ problemy kiberietiki" ("Problems in Cybernetics"), Issue 5, 1960, pages 209-272. Moloshnaya, T. N. , VOprosy razlicheniya Omonimov pri mashinnom 75. perevode s angliyskogo yazyka na russki [Problems in Distin- – 284 - 73. WH. guishing Hononyms in Machine Translation from English into Russian], "Problemy kibernetiki" ("Problems in Cybernetics"), Issue l, Moscow, 1958, pages 215–221. Moloshnaya, T. N. , Nekotoryye voprosy sintaksis a v svyazi s mashinnym perevodom s angliyskogo na russkiy [Certain Syn- tactical Problems in Association With Machine Translation from English into Russian], "Woprosy yazykoznaniya" ["Problems in Linguistics"), No. 4, 1957, pages 92–97. Nikolayeva, T. M., Algoritm nezavisimogo grammaticheskogo an– aliza russkikh tekstov [The Algorithm of Independent Gramma- tical Analysis of Russian Texts), "Doklady na konferentsii po ob rabotke informatsii, mashinnomu perevodu i avtomatiches- komu chteniyu teksta" ["Proceedings of the Conference on the Evaluation of Information, Machine Translation and Automatic Reading of Text"), Moscow, l06l. [Translator' s Note: This reference has been omitted in the Original]. Nikolayeva, T. M. , Opredeleniye Vida russkogo glagola S. pom- oshch' yu konteksta [ Determination of the Mood of a Russian Verb from Context], "Mashinnyy perevod i prikladnaya lin- gvistika" ("Machine Translation and Applied Linguistics"), No. 2 (9), 1959. f i Nikolayeva, T. M. , Struktura algoritma grammaticheskogo anal- iza (pri MP S russkogo yazyka) [The Structure of the Algor- ithm of Grammatical Analysis (in the Case of Machine Transla– tion from the Russian Language) ), "Mashinnyy perevod i pri- kladnaya lingvistika' ["Machine Translation and Applied Lin- guistics"), No. 5, 1961. Novoye v. lingvistike [Linguistic News), Issue l, 19: " : – 285 - 76. 77. 78. 79. 80. 8l. 82. 83. oischk I. --- --- ow, 1962, Issue 3, Moscow, i963. Issue 2, Mose - º º - , Istoriya nauchnoy literatūry na novykh yazykakh [The History of Scientific Literature in Modern Languages], Moscow, 1933. Paducheva, Ye.V., O pravilakh porozhdeniya predlozheniy stan- dartizovannogo yazyka geometrii [The Rules of Formation for Phrases in the Standardized Language of Geometry], "Doklady na konferentsii po obrabotke informatsii, mashinnomu perevodu i avtomaticheskomu chteniyu teksta" ["Proceedings of the Con- ference on the Evaluation of Information, Machine Translation and Automatic Reading of Text"], Issue 5, Moscow, 1961. Panov, D. Yu. , Avtomaticheskiy perevod [Automatic Translation], Moscow, 1958. Peshkovskiy, A.M., Ob" ektivnaya i normativnaya tochki zreniya na yazyk [Objective and Normative Points of View with Regard to English], Cited in the book by W. A. Zvegintsev, see page 33. Polivanov, Ye. D. , Opyt chastnoy metodiki prepodavaniya rus- skogo yazyka [Experience in the Method of Private Tutoring of Russian], Part I, Tashkent, 1961. Rakhmanov, I. V., Ocherk po istorii prepodavaniya novykh zapadnoyevropeyskikh inostrannykh yazykov [An Outline of the - - - History of Teaching Modern Western Western Europeans Foreign Languages], Moscow, 1947. Revzin, I. I. , Modeli yazyka [Models of a Language], Moscow, 1962. Revzin, I. I. , Po povodu retsenzii K. A. Levkovskoy na knigu M. D. Stepanovoy [In Connection with the Review by K. A. Lev- kovskaya of the Book by M.B. Stepanova.], "Woprosy yazykoz- – 286 - --- --- -º-º: maniya" ["Problems in Linguistics"], No. 5, 1955, - 85. 86. 87. 88. 89. 90. Revzin, I.T., Fºntsipy postroyeniya teoreticheskºgº ºursa perevoda na inostrannyy yazyk [Principles of Devising a Theor- etical Course in Translation into a Foreign Language], "Uchen- yye Zapiski I MGPriya" ("scientific Notes I, Moscow State Ped- agogical Institute of Foreign Languages"), Vol. XXII, 1959. Revzin, I. I. and Rozentsveyg, V. Yu. , K obosnovaniyu iºnºvº- ticheskoy teorii perevoda [The Basis of the Linguistic Theory of Translation], "Woprosy yazykoznaniya" ("Problems in Linguistics" j, No. 1, 1962. Reformatskiy, A. A. , Vvedeniye v yazykoznaniye [Introduction to Linguistics), Moscow, 1917. Reformatskiy, A. A. , Lingvisticheskiye voprosy perevoda [Linguistic Problems of Translation], "Inostrannyye yazyki v shkole" ["Foreign Languages in the School"], No. 6, 1952. Retsker, Ya. I. , O zakonomernykh sootvetstviyakh pri perevode na rodnoy yazyk [Regular Correspondences in Translation into the Native Language], Collection entitled "Teoriya i metodika uchebnogo perevoda" ["The Theory and Method of Instructional Translation"), Moscow, 1950. Retsker, Ya. I. , Zadachi sopostavitel'nogo analiza perevodov [Problems in the Comparative Analysis of Translations], "Teoriya i kritika perewoº ["Theory and Critique of Trans- lation"], Leningrad, 1962. Rizel', E. G. , 0 t. n. arkhitektonicheskoy funktsii yazykovykh stilisticheskikh sred.stv [The So-Called Architectonic Func- tions of Linguistic Stylistic Facilities], "Uchenyye zapiski MGPIIYa" ["Scientific Notes, Moscow State Pedagogical Insti- tute of Foreign Languages"], Vol. 10, Moscow, 1956. – 287 - 92. 93. 911. 97. 98. Bozemtsvey& V.Yaº, o yazykovykh kontaktakh (Linguistic con- tacts J, "voprosy yazykoznanaya tº Problems in Linguistics"), No. 1, 1963. Rozents veyg, V. Yu. [Raboty po mashinnomu perevodu s inostran– nykh yazykov na russkiy i s russkogo na inostrannyye v Sovet- Skom Soyuze [Research on Machine Translation from Foreign Languages into Russian and from Russian into Foreign Lan- guages in the Soviet Union], Moscow, 1958 (The IVth Interna- tional Congress of Slavicists, Proceedings]. Rozents veyg, V. Yu. and Uman, L. M. , Interferentsiya i gramma- ticheskiye Kategorii [Interference and Grammatical Categor- ies], Collection entitled "Issledovaniya po strukturnoy tipologii" ["Investigations into structural Typology" J, Mos- cow, 1963. Rozents veyg, V. Yu. and Uman, L. M. , K probleme grammaticheskoy interferentsii [The Problem of Grammatical Interference], Collection entitled "Problemy strukturnoy lingvistiki" [Prob- lems in structural Linguistics"], Moscow, 1962. Rychkova, N., Lingvistika i matematika | Linguistics and Math- matics], "Nauka i zhizn." ["Science and Life" j, No. 9, 1961. Sobolev, L. N. , O mere tochnosti v perevode [The Measure of Accuracy in Translation], Collection entitled "Teoriya i metodika uchebhogo perevoda" ["The Theory and Method of In- structional Translation"], Moscow, 1950. Sobolev, L. N., Posobiye po perevodu s russkogo yazyka na frantsuzskiy [Textbook on Translation from Russian into French], Moscow, 1952. - Superanskaya, A. V., Protsess transkribirovaniya sobstvennykh imen i vozmozhnosti ego avtomatizatsii [The Process of Trans- - - 288 - 99. 100. 101. 102. lC3. 10||. 105. 106. 107. - ----- º-º- - º º: º - º - --- - - - ossibilities of Automating ---> --- -º-º-º: *** **** - -- this prºcedureſ, ºiletºn Sºedigeniya po probieńam mashin- nogo perevoda" ["Bulletin on Problems of Machine Transla– tion"], No. 6, Moscow, 1958. Suprun, A., Prepodavaniye russkogo yazyka i rodnoy yazyk uchashchikhsya [Instruction in Russian and the Native Lan- guage of the Student], "Russkiy yazyk v natsional noy shkole" ["The Russian Language in the National School"], No. 3, 1962. Tarskiy, A. , Vvedeniye v. logiku i metodologiyu deluktivnykh nauk [Introduction to Logic and the Methodology of Deductive Sciences], Translated from the English, Moscow, 1918. Tesniere L. O russko-frantsuzskom slovare L. V. Shcherby [The Russian-French Dictionary L. V. Shcherbal, "Woprosy yazykoznaniya" ["Problems in Linguistics" J, No. 6, 1958. Terpigorev, A. M. , Ob uporyadochenii tekhnicheskoy terminol- ogii [Ordering Technical Terminology], "Woprosy yazykoznan- iya" ["Problems in Linguistics"), I, 1953, pages 71–76. Tezisy konferentsii po mashinnomu perevodu [Theses from the Conference on Machine Translation] (15–21 May 1958), Moscow, 1958. - Toporov, V. N., 0 strukturnom izuchenii yazyka [The Structural Study of A Language], "Russkiy yazyk v natsional'noy shkole" ["The Russian Language International School"], No. 1, 1961. Toporov, V. N. , Vvedeniye i kommentarii k "Dkhammapade" [In- troduction and Commentary to the "Dhammapada" ), "Dkhammapada" ["Khammapada" j, Moscow, 1960. Trubetskoy, N. S. , Osnovy fonologii [Fundamentals of Phonol- ogy], Translated from the German, Moscow, 1961. Whorf. W. L. , Nauka i yazykoznaniye [Science and Linguistics), – 289 - - - - º ike" ["finguistic sentitled "Novoyev lingvist 108. Whorf B. L., Otnosheniya norm povedeniya i myshleniya k yāzyku [Relationships of Norms of Behavior and Thought to -- Language], Collection entitled "Novoye v. lingvistike" ["Lin- guistic News"), Issue l, Moscow, 1960. 109. Uspenskiy, W. A. , Itogi raboty sektsii algoritmov mashinnogo perevoda [Results from the Work done by the Section on Algor- ithms of Machine Translation], "Mashinnyy perevod i priklad— H naya lingvistika" ["Machine Translation and Applied Linguis- ties"], l (8), Moscow, 1959. - 110. Fedorov, A. V. , O zadachakh sopostavitel'nogo izucheniya --- yazykov [The Goals of Comparative Linguistic Studies j , "Nem– etsko-russkiye yazykovyye paralleli" ("German-Russian Lan- guage Parallels"], Compiled by A. V. Fedorov, N. N. Kuznetsova, Ye. N. Morozova, I. A. Tsyganova, Moscow, 1961. lll. Fedorov, A. W. , Vvedeniye V teoriyu perevoda [Introduction to the Theory of Translation, 2nd Edition, Moscow, 1958 (list Edition, Moscow, 1953). ll2. Firth D. R. , Tekhnika semantiki [The Technique of Semantics), Collection entitled "Novoye v. lingvistike" ["Linguistic News"], Issue 2, Moscow, 1962. 113. Shevoroshkin, V. V. , Drevnºe teksty i problema mashinnogo per- evoda [Ancient Texts and the Problem of Machine Translation], Tezisy konferentsii po MP [Theses from the Conference on Ma- chine Translation], Moscow, 1958. ll. Chomsky N. , Sintaksicheskiye struktury [Syntactic structures], | collection entitled "Novoye v. lingvistike" ["Linguistic News"), Issue 2, Moscow, 1962. - º - - 290 - - - ----------------- ºº: tº - 116. 117. 118. ll.9. 120. 121. --- - --- - * A. Yvedeniyev matematicheskuyu logº [Introduction ºf - - -------- *into Mathematical Logieſ, transiation from the English, Vol. I, Moscow, 1960. Chikobava, A. , Sopostavitel'noye izucheniye yazykov kak metod is sledovaniya i kak metod obucheniya [Comparative Study of Languages as a Method of Investigation and as a Method of Teaching], "Russkiy yazyk v natsional noy shkole" ["The Rus- sian Language in the National School"], No. 6, 1957. Shvanebakh, B. E. and Revzin, I. I. , Uchebnoye posobiye po teorii perevoda s nemetskogo yazyka na russkiy [A Textbook on the Theory of Translation from the German into the Russian], Moscow, 1960. Shcherba, L. V. , Opyt lingvisticheskogo tolkovaniya stikhot- voreniy [Experience in the Linguistic Interpretation of Poe- try], "Russkaya recht." ["Russian Speech" j, 1923. Shcherba, L. W. , Izbrannyye raboty po yazykoznaniyu i fonetike [Selected Works on Linguistics and Phonetics j, Vol. I, Lenin- grad, 1959. - Shcherba, L. V., Prepodavaniye inostrantsykh yazykov v Sred- ney Shkole [Teaching Foreign Languages in the Secondary School], "Obshchiye voprosy metodiki" ("General Problems of Method"], Izd-vo Akademi Pedagogicheskikh nauk RSFSR [Pub- lishing House of the Academy of Pedagogical Sciences of the RSFSR], Moscow, 1947. - Shcherba, L. V. , O troyakom aspekte yazykovykh yavleniy i ob eksperimente v yazykoznanii [The Triple Aspect of Lin- guistic Phenomena and an Experiment in Linguistics], Izves- tiya AN SSSR [Bulletin of the USSR Academy of Sciences], No. 1, 1931. - – 291 - 123. 12h. l25. 126. 127. 128. 129. 130. lºl. 132. l33. 1311. 135. - - - º º -- º - Tºbrannyye statº # seleášeňº º ------ º º * - --- - º --> -- - - - - --- - --- º icles i. - - - - Ashby U. Ross. , wedeniye V kibernetiku [Introduction to Cy- bernetics], Translation from the English, Moscow, 1959. Yaglom, I.M., Dobrushin, R. L. and Yaglom, A.M., Teoriya in- formatsii i lingvistika [The Theory of Information and Lin- guistics], "Woprosy yazykoznaniya" ["Problems in Linguistics" , No. 1, 1960. Bar-Hillel, Y., The Present state of Research on MT, "Ameri- can Documentation," Vol. 2, pages 229–237, 1951. Bergh, L. , Moyens d'exprimer en francais iſ idee de direction [Means of Expressing the Idea of Direction in French], Göte- borg, 1918. Booth, A., Brandwood, L. and Cleave, I., Mechanical Resolu- tion of Linguistic Problems, London, 1958. Brunot, F. , La pensee et la langue [Thought and Language], Paris, 1922. Buhler, K., Sprachtheorie [The Theory of Speech], Jena, 1931. Buyssens, E. , Les langages et le discours [Languages and Speech], Brussels, 1943. Carnap, R., Introduction to Semantics, New York, 1918. Cary, E., La traduction dans le monde moderne [Translation in the Modern World] , Geneva, 1956. --- Cary, E., Mecanismes et traduction [Machines and Translation], "Babel," Vol. II, No. 3, 1956b. Cauer, P., Die Kunst des jbersetzens [The Art of Translation], Berlin, 1911. Delavenay, E. , La Machine a traduire [The Translation Ma- -- chine], Paris, 1959. --- - - º 1110. 1111. 1/12. 1113. lili. 1115. 146. 1||7. --- - º Introduction to Massº º º-º-º-º: -- --- º - º Firth, J. R. , Linguistic Analysis and Translation, "For Roman Jakobson," The Haage, 1956. Frei, H., La grammaire des fautes [The Grammar of Mistakes], Paris – Geneva – Leipzig, 1929. Fries, Ch. , Preparation of Teaching Materials, Practical Grammars and Dictionaries, Especially for Foreign Languages. "Reports for the Eighth International Congress of Linguists," Vol. I, Oslo, 1957. ". Fries, Ch. , The Structure of English, New York, 1953. Fries, Ch. , Teaching and Learning English as a Foreign Lan- guage, Michigan, 1947. Golab, Zb. , Some Arumanian Macedonian Isogrammatism and the Social Background of Their Development, "Word," Vol. XV, 3, 1960. Gougenheim, G. , Systeme grammatical de la langue francaise [A system of Grammar for the French Language], Paris, 1939. Gougenheim, G., Structure Grammaticale et Traduction Automa- tº taramatical Structure and Automatic Translation], "Traduction automatique" ("Automatic Translation"], Paris, No. 2, 1960. Haugen, E. , The Analysis of Linguistic Borrowing, "Language," 26, 2, 1950. Haugen, E., Bilingualism in the Americas. A Bibliography and Research Guide. Published by the American Dialect Society, November, 1956. Hays, D. G., Linguistic Research at the RAND Corporation, "Proceedings of the National Symposium in Machine Translation," London, 1961. – 293 - - --------- --- --- l!!8. l!9. lBO. 15l. 152. lº, 3. lºl. 155. 156. lº,7. 158. lºj9. Humboldt, W. , Briersechsei. Brief an A.W. von Schlegel, vom 23. Juli 1796 [Correspondence. Letter to A. W. von Schlegel of 23 July 1796], Halle, 1908. Jakobson, R. , Beitrag zur allgemeinen Kasuslehre,"TCLP [Contributions to the General Teaching of Cases. "TCLP" ), No. 6, Prague, l936. Jakobson, R. , shirters, Verbal Categories and the Russian Verb. Harvard University, 1957 (Department of Slavic Lan- guages). Jakobson, R. and Halle, M., Fundamentals of Language. Gravenhague, 1956. Jakobson, R., On Linguistic Aspects of Translation, "On Translation," Cambridge, 1959. Jakobson, R., Poetics and Linguistics. "Style and Language." Sebeok, T. A. (ed.), New York, London, 1960. Kaplan, Abr. , An Experimental Study of Ambiguity and Context, "Mechanical Translation," Vol. 2, No. 2, November 1955. Koessler, M. and Derocquigny, J. , Les faux amis ou les trahisons du vocabulaire anglais [The Faux Amis or the Treachery of the English Vocabulary ), Conseils aux traducteurs [The Translators' Council], Paris, 1928. Lado, R. , Linguistics Across Cultures (Applied Linguistics for Language Teachers), Michigan, 1957. * Levi-Strauss, C. , Anthropologie Structurale [Structural An- thropology], Paris, 1958. Language in Culture. Conference on the Interrelations of Language and Other Aspects of Culture, Ed. Hoijer. Chicago, 195l. Linguistic Analysis and Programming for Mechanical Transla– – 294 – l6O. 16l. l62. l63. l6l. l65. l66. 167. l68. 169. tion (Mechanical Translation and Thought), Milan, 1960. Malºblanc, A. , Stylistique comparee du francais et de l' alle- mand [Comparative Stylistics for French and German], Paris, 1961. Malinowski, B. , A Scientific Theory of Culture and other Es- says, Chapel Hill, 1941. Masterman, M. , What is a Thesaurus?, Essays on and in MT by CLRU, London, 1959. Mauger, G. and Charon, J., Manuel de francais commercial a l'usage des etrangers [A Manual for Foreigners in the Usage of Commercial French], Paris, 1958. Meile, P. , Linguistique appliquee et traduction automatique [Applied Linguistics and Automatic Translation], "Les langues modernes" ["Modern Languages"), No. 2, 1961. Meillet, A. , Les interferences entre vocabulaires [Interfer- ence Between Vocabularies], "Linguistique historique et lin- guistique generale" ["Linguistic History and General Linguis- tics" ), Paris, 1926. Micklesen, I., Russian-English MT "American Contributions to IV International Congress of Slavists," s' Gravenhage, lgp8. Miller, G. A. , Human Memory and the Storage of Information "IRE Transactions on Information Theory," vol. IT-2, No. 3, September 1956. •. Morton, F. Rand. , The Teaching Machine and the Teaching of Languages. "PMLA," September 1960. Mounin, G. , A propos de "Language, Thought and Reality" de Benjamin Lee Whorf [With Reference to "Language, Thought and Reality" by Benjamin Lee Whorf), "Bulletin de la sociate de Linguistique (BSL) de Paris" ["Bulletin of the Linguistic – 295 — - 170. 172. 173. l'714. 175. 176. 177. , 178. l'79. 17l. l80. l8l. *Is society (BSL) of Paris" J, Vol. 56, 1961. Mounin, G., Les Belles Inriaeles [The Beautiful Infidels], Paris, 1955. Mounin, G. , Les Systemes de communication non-linguistique et leur place dans la vie due XXe Siecle, BSL de Paris [Nonlinguistic Communications Systems and Their Position in the Life of the XXth century, BSL of Paris), Vol. 51, Paris, 1959. Nägelsbach, K. F. von, Lateinische Stilistik für Deutsche [Latin stylistics for Germans], Nürnberg, 1888. Oettinger, A. G. , Automatic Language Translation, Cambridge, 1960. Oettinger, A. G. , Automatic (Transference, Translation, Remit- tance, Shunting), "On Translation," Cambridge, 1959. On Translation, Edited by Reuben A. , Brower, Cambridge, Mass. 1959. "O Sztuce tilumaczenia" ["The Art of Translation" ), Wroclaw, 1955. -- . Otto, Ernst. , Methodik und Didaktik des neus prachlichen Un- terrichts [Methodology and Didactics of Modern Language Teaching] , Bieiteia–Leipzig, 1925. Parker-Rhodes, A. F., Some Recent Work on Thesauric and Inter- - . à . . lingual Methods in Machine Translation, Cleveland, j959. Pimsleur, P., Semantic Frequency Counts, "Mechanical Trans- lation," Vol. 1, No. 1-2, November 1957. Pol, B. van der. , An Iterative Translation Text. "Information Theory-Third London Symposium," London, 1956. Proceedings of the National Symposium on Machine Translation, Los Angelos, 1961. *= – 296 - 182. . 183. l8l4. l85. l86. l87. 188. 189. 190. l91. 192. 193. 194. 195. **** ~. Quine, W.V., Methods of Logic, New York, 1950. Quine, W.V., From the Logical Point of view, New York, 1953. Reifler, E. , Mechanical Resolution of the Constituents of German Substantive Compounds, "Mechanical Translation," Vol. II, No. I, 1955. Rizescu, I., Contributii la studiul calcului lingvistic [Con- tributions to the Study of Linguistic Thought ] , Bucharest, 1958. - Sapir, E. , Conceptual Categories in Primitive Languages, "Science," Vol. 711, 1931. - Savory, Th., The Art of Translation, London, 1957. savory, Th., The Language Of Science, London, 1953. Spang-Hans Sen, H. , Recent Theories on the Nature of the Lan- guage Sign. "TCLC," Vol. IX, Copenhague, 1954. Straumann, H., Newspaper Headlines, London, l935. Symposium: Proceedings of Symposia in Applied Mathematics, Vol. XII, Structure of Language and Its Mathematical Aspects, Providence, l96l. - * : * Ramakrishna, B. S. and Subramanian, R., Relative Efficiency Of English and German Languages for Communication of Semantic content, "IRE. Transactions on Information Theory," Vol. 4, Nº. 3, New York, 1958. / Suszko, R. , Zarys elementarnej Skladni logicznej [Consolida- ted Elementary Logic Notation], Warsaw, 1957. - Tesniere, I., Elements de syntaxe structurale [Elements of structural Syntax), Paris, 1959. •. - . Thieme, K. , Die geshichtlichen Haupttypen des DolmetschenS [The Main Historical Types of Interpretation], "BABEL," Vol. I, No. 2. -- - - . . – 297 - FTD-TT–65–1597/1+2 196. , 198. 197. 199. 200. 2Ol. 202. tion : b) on machine translation: Vinay, J.P. and Barbelnet, J. , stylistique comparee du fran- cais et de l'anglais. Methode de traduction [Comparative Stylistics for French and English. A Method of Translation], Paris, 1958. Weber, H. , Das Tempus system des Deutschen und des Französis- Chen. tibersetzungs- und Strukturprobleme [The System of Tense in German and French. Translation and Structure Problems j , "Romanica Helvetica," Vol. 115, Bern, 1954. Yngve, W. H. , Sentence for Sentence Translation. "Mechanical Translation," Vol. 2, No. 2, November 1955. Yngve, W. H. , Gap Analysis and Syntax, IRE Transactions on Information Theory, Vol. JT-2, No. 3, 1956. Yngve, W. H. , The Outlook for Mechanical Translation, "BABEL," Vol. II, No. 3, October 1956(b). Yngve, W. H. , A Model and a Hypothesis for Language Structure. Research Laboratory of Electronics and Department of Modern Languages, Massachusetts Institute of Technology, Cambridge, Massachusetts, 1960. Weinreich, U. , Languages in Contact, New York, 1953. BIBLIOGRAPHY ON QUESTIONS OF TRANSLATION Bibliography on translation can be found in the following editions: a) on the general theory of translation and on literary transla– l. "Masterstvo perevoda" ["The Art of Translation" }, Collection of Articles, Moscow, pages 1187-508, 1959. 2. On Translation, Cambridge, pages 271-293, lººp9. 3. O Sztuce tilumaczenia [On the Art of Translation], Wroclaw, 1955. ... 4 - *S - l. Machine Translation l949–1960. Bibliographic Listing of the FTD-TT-65-1997/1+2 ITM and VT of the USSR Academy of Sciences, Moscow, 1962 (photoprint edition). 8 – 298 – 2. Ravich, R. D. : Bibliografiya zarubezhnykh rabot po mashinnomu. perevodu (1960–1961) [Bibliography of Foreign Research on Machine Translation (i960-19 l)], Edited by I. A. Mel' chuk and G. Ye. Tsveyg, Moscow, 1962. 3. Delavenay, E. and Delavenay, K. , Bibliography of Mechanical Translation, s' Gravenhage, 1960 (Janua Linguarum, No. ll). PERIODICALS DEALING WITH TRANSLATION Translators' Notes, Moscow, 1959 Machine Translation and Applied Linguistics, Moscow, l058 Index Translationum, Unesco, l–12, Paris. Babel, Bonn, 1955 Current Research and Development in Scientific Documentation, Washington, 1957 L'Interprete, Geneva, 1946 Journal des traducteurs, Montreal, 1955 Le Linguiste, Brussels, l955 MT: Mechanical Translation, Cambridge (USA), 1951, – 299 – FTD-TT–65–1597/1+2 _____. _ — — — —~~~~ ICHIGAN (O = O = _ == ;"öööö7 9839 [][] N0 I REMOWE []|R MUT||LATE GÅRD |- |-| |--|- |RECEIVED NOV/2 5 1066 HSRL ||}|\\ſ\\