University Studies. VOL. i. JULY, 1888. A, ^ No - 1 I. — On the Transparency of the Ether . By DeWITT B. BRACE. Whether light coming from the remotest members of the visible universe has not been enfeebled to a greater extent than the variation of distance would require, is still an open question. If there be absorption at all, it must be exceed- ingly small through spaces comparable with the dimensions of the solar system, in order that the light of these distant bodies may be perceived. It is proposed in the present paper to investigate the phe- nomena which would occur if the energy were absorbed by the ether itself through frictional forces or imperfect elastic- ity. If absorption does take place, there must be a differen- tial effect for varying wave-lengths, if the ether satisfies the equations of motion of elastic bodies. Several arguments have been advanced as proving that such an absorption takes place, of which those of Cheseaux, Olbers, and Struve are the most celebrated. Considerations on other grounds would seem to suggest such a conclusion. Cheseaux and Olbers, arguing from insufficient data as to stellar distri- bution, have shown that if the number of stars is infinite and distributed with anything like uniformity in space there must be absorption of light, as otherwise the sky would appear all over of a brightness approaching that of the sun, University Studies, Vol. I., No. 1, July, 1888. I 2 / / De Witt B. Brace , since tlie brightness at any point depends on the depth of the luminous layer and the solid &ngle which it subtends at that point. I The researches of both Herschel and Struve prove a non® uniformity of distribution in all directions, but a concentration^^ of stars toward the medial plane of the Galaxy, more marked the smaller the magnitude of the star. Struve, from conclu- f sions based on the supposition of an average uniformity of I stellar distribution in layers parallel to this central plane, and on the assumption that the brightness is a measure of the relative distance, attempted to prove that absorption must I take place. In fact, for a uniform distribution the number of ( calculated stars of different magnitude should vary inversely as their brightness. Now the number of calculated stars of ! any magnitude exceeds slightly the number observed, this excess being greater with the diminution of the magnitude. Hence it is concluded that absorption must take place to explain this increasing discrepancy, and that there must be a limit to the space-penetrating power of a telescope much lower than the enfeeblement of light with the distance would require. Later investigations regarding the constitution of the visible universe show that Struve’s assumptions were false, and that no law of uniformity in distribution or in intrinsic brightness can be accepted. While it is at present impossible to ascertain with much accuracy the real magni- tudes of the stars, there are sufficient data to show that botl their volume and their intrinsic brightness, per unit surface vary between wide limits. The annual parallax of severa stars, as measured by different observers, gives approximate!] consistent results. A comparison of the brightness and dis tances of these stars with the intrinsic brightness and dis tance of the sun, as made by Zollner, shows that the sun’: volume is but a small fraction of that of these stars, suppos ing equal intrinsic brightness per -unit of surface. Thi: method of comparison, when applied to Sirius, gives a mucl greater volume than other methods would warrant, thougl far exceeding that of the sun in any case. It must henc< 2 3 O On the Transparency of the Ether. be concluded that the intrinsic brightness of a unit of its sur- face is much greater than that of the sun. The results of spectrum analysis point to a wide variation in the age and temperature of different stars, some being very much farther advanced in the process of cooling than others. It seems certain from these considerations that not only the absolute size but the intrinsic brilliancy vary within very wide limits, some stars emitting several thousand times as much light as others. The observations on stellar distribution indicate a much more complicated law than the earlier observers supposed. The more or less marked crowding together of stars in cer- tain .regions, with the existing intermediate voids, and the only partial resolvability of these aggregations, show a ten- dency to some system of clustering in which the various orders of magnitudes are actually intermingled. In certain regions the more minute stars are much more sparsely scat- tered than in others, while the distribution should approach more marked uniformity with diminishing magnitude. The absence of vast numbers of stars, with excessive crowding of the smallest magnitudes which such a distribution would require, shows that the telescope can penetrate to the bounds of the system in these regions. The observations on the immense extent of the orbits of certain binary stars furnish evidence that there exists a connection between certain stars which have not heretofore been suspected as being members of the same system. The fact that the stars are gathered together in clusters princi- pally in or near the Galactic zone indicates that they must form a part of the Galaxy, since there is no reason why, if they were outside our stellar system, they should not be more uniformly distributed toward the poles of this zone. These evidences of the complexity of the laws of distribution in magnitude and distance furnish strong proof that the present stellar system is finite, and that it does not appear so from the ultimate absorption of the light of the remoter members of an infinite system. Nothing but a cosmical veil of vary- 3 4 De Witt B. Brace , ing tenuity existing in interstellar spaces and closing our view more or less effectually from the infinite expanses beyond, could possibly explain these appearances, — a supposition which is wholly unallowable. If there be absorption in space, it must be determined by other methods than the one by which Struve attempted to prove it. If the law of the Dissipation of Energy is absolutely uni- versal, then it must be allowed that no distortion of the ether can take place without a certain loss of energy however small, so that the luminiferous vibrations would be gradually frittered down, and after an almost infinite number of such distortions be dissipated away so as to escape perception. On this hypothesis, from analogy with all known phenomena connected with ponderable bodies under similar conditions, a differential effect should be produced for different periods of vibrations, which would give a perceptible coloration in dis- tant stars. If an excessively diffused material substance be supposed scattered through space in a gaseous state, such a body could only absorb selectively through its atoms, its molecules being too widely scattered to allow of any transformation of energy into molecular friction. Hence the only loss, other than by selective absorption, would be in the ether itself. The ab- sorption would then take place according to the same laws which determine it when such a substance is not present. In a medium in which there were dissipative forces propor- tional to the rate of distortion, there would be a relative change in the velocity of propagation of transverse vibrations of different periods which, for sufficiently great distances, might be detected in the coloration produced by any sudden outburst or extinction of starlight. If the absorption were small, such a difference in the velocities of different rays would be exceedingly small, even for distances comparable with the greatest dimensions of the stellar system, so that the coloration could only last for a very short time. The luminiferous medium bears a close analogy to the ponderable substances of nature in respect to its rigidity for 4 On the Transparency of the Ether. 5 high rates of distortion and its apparently perfect fluidity for motions of distortion of low rates. JTe existence of such apparently incompatible qualities does not seem so difficult to understand, when a material substance subjected to rates of distortion of far less range than the extreme limits at which these two qualities are observed to exist in the case of the ether appears in the one case like a rigid solid, and in the other like a very mobile fluid. Maxwell found that, by rotating a cylinder rapidly in a liquid and passing a ray of polarized light close to its surface, the plane of polarization was altered, proving clearly a state of strain for ordinary liquids when the rate of distortion is sufficently high. Sir William Thomson has also shown how wax or pitch may, in the one 'case, vibrate like ordinary solids, and, in the other, allow bodies to pass very slowly through them without appreciable resistance. The luminif- erous medium presents similar phenomena. For periods of vibration comparable with those of light, it acts like a very elastic solid. For low rates of distortion like those which the motions of the planets and comets as -well as those which the molecules of a gas produce, there is no sensible resistance, and the medium seems to act like a perfect fluid. That this resistance is exceedingly small is shown by the fact that the comets, which are in general of extreme tenuity, give no definite indications of a resisting medium in space. While the properties of the ethereal medium manifestly transcend those of ordinary matter, yet it seems to fulfil, in the qualities of elasticity and fluidity, the conditions o& natural bodies. Very strong analogy to the ether is furnished by viscous substances, and these substances always dissipate more or less rapidly any vibrations to which they are sub- jected, proportionally to the rate of distortion, — at least for small rates. If the ether has a corpuscular structure, — and it is difficult to conceive of absorption otherwise, — and the analogy in respect to viscosity is extended to it, as well as the analogy in respect to its elasticity and fluidity, there should be a loss or transformation of radiant energy. 5 6 De Witt B. Brace , Loss of energy may also take place in other ways depend- ing on imperfect elasticity alone, or the loss may arise both from viscous forces and from imperfect elasticity. In the one case we have the stress varying with the rate of distor- tion, and in the other, with the duration and magnitude of the strain. The existence of either will give a differential effect for the absorption of different rays. Suppose that absorption does take place, the amplitude of a periodic motion would be some function of the distance, wave-length or period, and of the viscosity and imperfect elasticity. Let it be required to find the form of the function for parallel rays of light, propagated in the direction of the j/-axis and with a displacement f parallel to the 4r-axis. Let i = Ae-^F(y,\, (i) where Ae~&* represents a periodic motion at the origin, of amplitude A, and p the coefficient of viscosity. When dissi- pative forces proportional to the relative velocities are present, the form of the function F is readily obtained. If the ether is perfectly elastic, the equation of motion for parallel rays is 2 | 3 2 £ » — - = n — dt 2 dy~ O) where p is the density and n the rigidity of the ether. Now Stokes has shown in his celebrated paper “On the Friction of Fluids in Motion” 1 that the expressions for the stresses *in an isotropic solid may be obtained directly from those found for the case of a viscous fluid in motion by merely substituting the displacements f, rj, f for the velocities u, v, w, and the rigidity n for the coefficient of viscosity p. In the case under consideration, we have not only the rigidity n , but a viscous coefficient p, each of which produces a shearing stress independently, so that the resulting stress will be the sum of the two, and the equation of motion becomes 1 Collected Papers, Vol. I. 6 On the Transparency of the Ether. 7 or, since dfp df n (- ix ^,2 1 ^ d 2 u t df’ (3) c^_ d 2 f d 3 j ^ dt 2 dy 2 ^ dtdy 2 The left-hand side represents the force of acceleration per unit of volume ; the first term of the right-hand side expresses the force arising from the distortion of the surrounding ether, and the second term the dissipative force arising from the rate of distortion. A particular solution of this equation is £ = ( 4 ) Substituting in equation ( 3 ), we have — pp 2 = n/3 2 — ipp/3 2 . (5) Since we are dealing with simple periodic motions, p must be real, and /3 must therefore be complex. Let (3 = — K + iy, ( 6 ) p where y = -• Substituting in ( 5 ), and separating the real and the imaginary terms, we have whence 0 = pp 2 + M< 2 — ny 2 — 2 ppuK*b'+ are approximately indepen- dent of the amplitude, they are interchangeable as regards sequence in absorption, and we may suppose the atmospheric absorption to have taken place first.. Hence in every case, we can leave out of consideration this effect and simply apply I., II., and III. to spectra as they are seen, to determine the relative coloration produced by absorption in space alone. We have now to apply I., II., and III. to a normal spectrum to determine the amount of energy absorbed when coloration is perceptible. In plate I., the curve A 2 represents approxi- mately the distribution of energy in the visible portion of the normal solar spectrum for different wave-lengths at high sun, according to Langley. 1 The effect of space-absorption on the solar spectrum would be inappreciable. Let now such a spectrum be carried to a very great distance ; suppose the rays parallel, and absorption present. The loss of energy can be represented graphically by plotting curves with values obtained from I., II., and III. The intensity is proportional to the square of the amplitude or in I. and II. to 1 Researches on Solar Heat, Plate I. Prof, papers of U. S. S. S., No. XV. 13 14 De Witt B. Brace , I a. £1 == A 2 e 11 a. Curve I 1 represents the distribution of energy according to I a. after the amplitude of a wave corresponding to .80 in the diagram has been diminished .01 of its original value. Curve I 2 and II 2 represent this distribution according to I a. and II a. respectively after a diminution in amplitude of .16 of its original value for the same wave. The law representing absorption according to III. would be a curve between these two. The curve A ? represents what the normal distribution would be if no energy had been absorbed and the amplitude had been uniformly diminished by .10. From these curves we are able to determine the proportion of the rays lacking in the different parts of the spectrum which would when added give the original spectrum. Thus from 1 1 of the red rays about .006 are lacking ; of the orange and yellow, about .020 ; of the green, nearly .030 ; and of the violet, about .050. In the same way for curve I 2 , about .06 of the red rays would be lacking, .15 of the yellow, .20 of the green, and about .50 of the violet. If the law of absorption is according to curve II 2 , about .03 of the red would have been absorbed to .06 of the yellow, nearly .10 of the green, and nearly .40 of the extreme violet. It is thus evident that the greater the absorption, the redder the spectrum will appear. Aubert has shown that less than one per cent of red mixed with white is perceptible. For sufficient intensity, curve I 1 would be within this limit, so that a hue near the orange-red would be perceptible. Either I 2 or 1 1 2 would evidently give a very perceptible reddish tinge. Thus from the hue it is possible, for a given intensity, to determine the total loss in intensity and the diminution in amplitude. In the curve Ij the amplitude of the yellow rays has been diminished from two to three per cent. In curves I 2 and II 2 the diminution has been about twenty per cent and fifteen per cent respectively. 14 On the Transparency of the Ether. 15 In the case of the heavenly bodies there should then be a coloration, becoming more marked with the distance, this col- oration also depending in part on the intensity. No regular gradation in hue is perceptible, and hence it may be concluded that the loss of energy is small, if any. From what is known of the spectra of incandescent bodies, the effect of increase of temperature is to displace slightly the position of maximum energy up the spectrum. Any irregular distribution of stars as regards temperature would not cause the average light of a certain number in one part of the heavens to differ materi- ally from that of another number taken anywhere else in the heavens. To carry the test for absorption to the utmost limit possible, we have only to consider those milky patches of light visible to the eye in the Galaxy ; or, better, those star clusters which are barely resolvable with the best telescopes ; or, going still further, to consider those nebulae whose spectra resemble the stellar spectra, and which consequently are probably resolvable into stars. From the vast number of stars which must constitute such a stellar mass, it may be concluded that if there were no absorption, the light with which such a mass would shine would be white. In travers- ing such vast distances, the absorption must be infinitesimal, not to produce a perceptible coloration. The general absence of gradation in color, even in the remotest visible bodies, shows that but a small per cent of their light can have been lost in space. This shows that e~ 2Ky cannot differ from unity by more than a small quantity. Hence ny will in general be less than unity, and k will not be greater than -, which for the ' y distances we have been considering, is excessively small. Referring to equation (10), we see that w is nearly unity, and hence the difference in time of propagation is a very small quantity, even for the remotest visible bodies. Taking now our complete equation as it would be for plane polarized light propagated in spherical waves, we have for the intensity at any point e = A 2 er 2K y ■ 09) 5 6 DeWitt B. Brace. Thus the variation in intensity with distance becomes e -2 *y / approximately, (20) when y is taken as the dimension of the visible universe. In order that the effect of absorption might equal that due to the variation in distance, we should have to take a distance ny, such that or (21) if the diminution in amplitude from absorption were ten per cent for a distance y. Thus n would have to be very great, and the system would be of dimensions n times as great as those of our own stellar system. To a close approximation, the system should have the same appearance whether absorp- tion were present or not. The apparent finiteness of the stellar universe cannot thus be due to absorption, as Struve supposed, his assumption of uniform distribution requiring a loss of as much as one-third the light of stars of the ninth magnitude. Either, then, the universe must be finite, or, if infinite in extent, the average density of distribution of self-luminous bodies outside our own system must be exceedingly small, as otherwise the sky would appear of a uniform brightness, approximating that of the sun. 16 700 - II. — On the Propriety of Retaining the Eighth Verb-Class in Sanskrit. By A. H. EDGREN. Doubts concerning the propriety of . retaining the Hindu classification of the so-called tan-verbs in a special conjuga- tion date as far back as Bopp. The first, however, to devote to the subject a careful investigation was Brugman in his article Die achte conjugations-classe des altindischen und Hire entsprechung im griechischen (Kuhn’s Zeitschr. XXIV.), where he tried to prove — except for kar — the identity of the tan (VIII.) and sn- (V.) classes, on principles of which I shall speak later. He was followed by a Belgian savant, Professor Van den Gheyn, who also tried to establish the same identity, but on principles wholly different in nature from those adopted by Brugman. Having myself, for the preparation of my brief Sanskrit grammar (Triibner, 1884), made an independent in- vestigation of this subject of the tan-verbs, and for the first time in any similar work classified them with the su-ve rbs as forming with these one class with the present-sign -no, I put my notes together in a brief paper (‘ On the verbs of the so-called tan- class in Sanskrit ’), which I presented at the meeting of the American Oriental Society in May, 1885, and which was subsequently reported in the proceedings of that society. My short remarks, in which I took issue espe- cially with Van den Gheyn, called forth from him, as a reply, a special paper in Bulletins de V Acade'mie royale de Belgique (XL, 1886), in which he tried to refute, with all fairness and courtesy, the arguments adduced by myself, in so far as they differed from his own, and to reaffirm the position he had already taken. 17 2 A. H. Edgren , However improbable I judged Van den Gheyn’s mode of explanation, I was satisfied to dismiss the question for the time being. Byt as no agreement is yet reached with regard to the treatment of the tan-ve rbs, — some, as especially Sanskrit grammarians, adhering to the Hindu classification, while others disagree concerning the principle on which the Eighth class should be given up, — it has seemed that an amplified review of my own arguments may not be useless as a contribution towards a definite settlement of the disputed question. The root-verbs of the Sanskrit language are by Hindu grammarians, as is well known, classified, according to the various forms of their present-stems, into ten groups or con- jugational classes, which are designated by the root heading each group in the native lists. The present-stems of the su and the tan - class ( i.e . of classes Fifth and Eighth) are said to be formed respectively by adding to the root the suffixes -nu ( su-nu -) and -u (< tan-u -), which are gunated in strong forms. The Hindu system of classification, and along with it the distinction made between the ^-class and the tan- class, was naturally enough adopted in the earlier grammars published in Europe. Bopp, however, who did not fail to notice that all the roots of the tan - class — kar alone excepted — termi- nate in -77, doubted the propriety of separating in principle the to-class from the su- class, and suggested that the tense- sign for both the classes was originally -777/, and that this sign in the ^77-class then lost its initial nasal after the nasal of the root. Yet his doubts did not lead him to deviate in his own grammar from the Hindu classification. Benfey, like- wise, was inclined to combine the two classes into one, and suggested, though, like Bopp, without special investigation of the subject, or practical application of the principle, that the nasal of the tan-verbs may in fact have been artificially trans- ferred to them from the suffix. Such suggestions, however, as were made by Bopp and Benfey were generally left unheeded by later grammarians as 18 3 On the Eighth Verb-Class in Sanskrit. resting on no proofs ; and the Hindu system was reproduced by them, either absolutely (as by Muller, Williams, Wester- gaard, Kielhorn, etc.), or modified in such a way that the tan- class, always with the suffix -u, was arranged as a sub-class under the .sw-class (by Whitney and Harlez). It is really outside of the pale of the Sanskrit grammarians that some attempts have been made to pro$e by the facts of the language and of comparative philology the identity, in the main at least, of the two classes in question. Brugman in his above-named article, Die achte conjugations-classe des altindischen etc., assumes without farther argument, and as a generally admitted fact, that the tan-ve rbs terminating in - n have received that nasal by artificial transfer from the suffix, and turns the force of his argument on the verbs in -an, which he considers as forming their present system by the suffix -no {nil), before which an is weakened to a (through n), according to his well-known theory of a nasal vowel. As for kar, he refers it, as irregular, to the second class (its root-formation being yet discernible in kur-vas , kur-mds, while karo-mi, etc., are formed after the analogy of kuru-tha{s ), whose second u, however, is a mere phonetic addition, not a suffix). Quite a different theory was put forth by Van den Gheyn in an article on the verbs of the Eighth class, published in Bulletins de V Acade'mie royale de Belgique (L., 1880), and further supported by two new articles in the same publication (VII., 1884, and XI., 1886). Van den Gheyn endeavors to show in these articles by the facts of Sanskrit, and of cog- nate languages as well, that the final nasal of the tan-verbs is not original, but a later accretion, a transfer from the present- sign to the root, and that these verbs properly belong to the Sixth or ^-class. As for kar, he adopts the hypothesis of Harlez, who considers kur to be the ‘ thbme principal ’ of the verb (cf. kurmi, epic, kurvas, kurmas), and attributes its later changes to analogy. As already noticed, Benfey had before suggested the theory of ta instead of taproots, and Gustav Meyer had likewise pointed out the analogy of forms like ta-td : ra-To?, re-ra-ra, etc., in support of his hypothesis that 19 4 A. H. Edgren , the original root of these forms was ta , not tan. But Van den Gheyn was the first to present a detailed argument in favor of this view, and may be considered as its chief advocate. The following considerations based, in the main, on an examination of the inflectional and derivative forms of the tan-ve rbs in Sanskrit, but also on the evidence of cognate forms in other languages, would seem to confirm, in the main, the position taken by Brugman, although its cor- rectness is not made dependent on the theory of a nasal vowel. The Hindus classified with the tan- class ten verbs, viz., three in -rn : arn , gharn , tarn ; one in -in : ksin ; five in -an : ksan (ksan), tan , man, van, san ; and one in -ar : kar , - — the only one that does not terminate in a nasal. To these have been added, on more or less convincing evidence from the Sanskrit literature, in, han, and tar. As regards those three roots that terminate in -rn, the fol- lowing considerations are to be noticed. The mere fact that the root arn is said to be inflected exactly like the well-authenticated root ar (r), ■ — both form- ing the strong and weak stems rno and rnu , — and that the meaning assigned to each is the same, suggests that the for- mer is nothing but an artificial extension of the latter, having no independent existence whatever, — a suggestion which is corroborated by the facts in the case. In the first place, no verb-form whatever outside of the present system has been made from a radical ant. Further, among all the derivatives (not less than 35) that must be referred to either of the two roots in question, only three, arnas, drna, rnd, contain a nasal. But that nasal may here be explained as belonging to the suffix. The primitive suffix -nas, though rare, occurs beyond question in some words (cf. dp-nas ‘possession,’ -bhar-nas ‘ offering,’ etc.) formed precisely like ar-nas . A primitive suffix -na is, indeed, hardly met with, except in participles. But drna ‘agitated,’ as a noun (m.) ‘flood,’ if not a participle, is too evidently of the same stock with arnas ‘ flood ’ to be 20 5 On the Eighth Verb-Class in Sanskrit. separated from it (cf. apna- : dpnas ‘ possession/ dravina : dravinas ‘ chattel ’). As for rna ‘ debt,’ its derivation is quite uncertain, but if we must resort to ar or arn, nothing better can be suggested tharf the participial form r-nd ‘ hurt, bur- dened, in-debted, (n.) debt.’ Finally, not one of the kindred words in the sister-tongues — and the number of such words is quite considerable — shows any trace of a root-nasal, unless, indeed, as suggested by some, it be opvvpu, whose v, however, is much more likely to belong to the suffix -vv (cf. cop-op-ov, etc.). As regards gharn , which is said to mean ‘shine,’ no verb- form that could be referred to such a root has been met with in the extant literature. It might then be left out of consid- eration here, were it not that a couple of nominal forms, ghrnd ‘heat’ and ghrni ‘heat,’ seem referable to the root gharn. As the native root-lists give also the root ghar (ghr) ‘ shine,’ as belonging to the su- class, it is evident, however, that this root is • the only acceptable form, and that gharn sustains to ghar precisely the same relation as arn to ar (ghr-nd, ghr-ni being quite regular formations). This sup- position is decidedly favored by kindred words in related tongues (cf. Zend gar-ema , Gr. Oep-opicu, Oep-pur), 6ep-o$, etc. ; Lat . for-mus, etc. .; Goth, var-in-jam ; Slav, gr-e-ti, etc.). With reference to tarn , said to mean ‘graze,’ it is, like gharn , entirely unauthenticated ; and as there seems to be no support for the acceptation of any similar root in any other Indo-Germanic tongue, it may well be considered as wholly fictitious, and invented to furnish an etymology for the noun trna (trna) ‘grass.’ This noun may possibly be a participial form (parallel with trna ) of tar in the sense of ‘broken through,’ — viz. the soil (cf. tirna-padi name of a plant). We come next to the root ksin ‘destroy,’ which has no more right to appear in the root-list than arn , above. Even here we find a shorter and well-authenticated root ksi ‘de- stroy ’ inflected according to the j&-class ( ksi-no-mi , etc.). No certain or authenticated example of a nasal is found either 21 6 A. H. Edgren , outside of the present-system 1 or in derivative words. To be sure, we have beside the participial form ksita also ksind ‘destroyed,’ but the latter is usually understood to be one of the common participles in -na. The root of the correspond- ing Greek verb c^Olvco ‘perish’ is which occurs every- where outside the present-system. Also ktlvvv/jll ‘kill’ has been suggested as a parallel form, and Curtius supposes its root to be rcrtv, a weakened form of ktclv = Skr. ksan. If the words are connected, the double nasal of ktlvvviu may, how- ever, be explained as owing to a phonetic doubling between two vowels, — a process that is. not uncommon, — or else, as Curtius suggests, to a special weakening in Greek. And even if ksin, in spite of strong evidence to the contrary, be considered as a genuine root, its conjugation in analogy with ksan, of kindred form and meaning, would not be any more anomalous than the inflection of nominal zzz-stems in analogy with an-stems ; and the present formation of ksan , as will be shown, is not ksan-o-mi, but ksa-no-mi. It is evident from these facts that arn, gharn, and ksin are, in all probability, mere figments or pseudo-roots, which the Hindu grammarians have foisted into their root-lists with- out any good reason, and which consequently should be cancelled altogether, — as, indeed, they are, very rationally, in Whitney’s grammar, in the enumeration of the tan-ve rbs (§ 713, and Root-supplement). We come next to the five roots in -an: ksan ‘destroy,’ tan ‘stretch,’ man ‘think,’ van ‘win, like,’ san ‘reach.’ As their nasal occurs not only in the present-system, but also, with few exceptions, outside of that system and in derivatives, it has been considered both by native and western grammarians as pertaining to the radical forms underlying the whole con- jugation-system of each of the verbs. Serious objections having been made, however, to this view by a few scholars, it will be necessary, before explaining the formation of the present stems of above verbs, to try to determine the nature of their nasal. 1 Brugman quotes ksenisydti , for which I can find no authority. 22 On the Eighth Verb-Class in Sanskrit. 7 * It is of some importance here to distinguish between that stage of the Indo-European language when the radicals which are now deduced by comparative analysis as those underlying its formal and inflectional development were already evolved, and a yet earlier stage when these radicals had not assumed the form they then had. There is, indeed, every reason to believe that such an evolution of roots from earlier germs, or perhaps, in some cases, from earlier polysyllabic and com- pounded entymons, took - place long before the language passed into its inflectional stage, and in some instances we may even yet discover the probable or evident traces of such a development. Thus, it seems probable enough that coup- lets or groups of roots like i : in ‘go,’ ci : cit ‘observe,’ mar ‘ grind ’ : warn ‘ crush ’ : mai'd ‘ grind ’ : march ‘ hurt,’ etc., are cognates of the same origin. If their original germ is actually represented by any one of the forms preserved to us, or if it is entirely lost, cannot be decided. There are some faint indications, indeed, that the shortest form may, ordinarily, be the most original, but they are after all uncertain . 1 Now it so happens that for every one of the roots in -an enumerated above has been suggested also, on more or less convincing evidence, a co-ordinate radical lacking the nasal and terminating in -a or, usually, in -a. Thus, cf. ksan ‘hurt’ : *ksa (in ksapay- ‘destroy,’ tuvi-ksa ‘much destroy- ing,’ etc.), and, perhaps, ksa ‘burn’; — - tan ‘stretch’ : ta (in the pass, tayate, and in tayate ‘ stretches ’) ; — man ‘ think ’ : (?) ma ‘ measure ’ ; — van ‘ win, like ’ : va * desire ’ (in the parti- ciple vatd and the desid. vivasati), va (only in vasimahi) ; — san ‘ procure ’ : sa ‘ procure ’ (in sata , sis-asati, etc., and in agva-sa ‘ horse-acquiring,’ etc.), sa (only in sa-sa-vahs). It is evident that the assumption of <£-roots finds a very meagre and doubt- ful support in these comparisons (about which later). But a host of forms, participles or derivatives, with a radical in -a, x Cf H. Edgren, “On the Verbal Roots of the Sanskrit Language and the San- skrit Grammarians,” Journal of American Oriental Society , XI., p. 5, etc. 23 A. H. Edgren , instead of -an ( ma-ta , ma-ti , ta-td , ta-tva , etc.), which are by all grammarians and lexicographers referred directly to the roots in -zzzz, have, farther, been explained by a few investiga- tors as made from zz-roots instead. This apparent variation between -zzzz, -a, and -zz-roots has been made use of to prove that the present tanomi , etc., was made from an zz-root by adding the suffix -no ( ta-no-mi ), not, as assumed by the grammarians, from tan by the suffix 6 (tan-6-mi), nor, as advocated by Brugman, from tan by the suffix -no, through * tn-no-mi, and that the suffixal n has pene- trated outside the present system, or else that the so-called ‘ general tenses ’ are formed from a root tan , existing at the side of ta . To this view some grave objections may certainly be made. Thus : a. The root-forms required by the advocates of formations like ta-no-mi , etc., are not ta , etc., but ta , etc. There may, indeed, seem to be some plausible reasons for accepting the existence of the former (as do Delbriick and Brugman), in so far as their occurrence in various forms is not yet satisfac- torily explained on the basis of a phonetical change of tan , etc., and as the language after all has quite a number of zz-roots. But there is no very plausible, and yet less con- vincing, reason for assuming any roots in -zz, like ta , etc. The quoted verb-forms ksapayati, vaslmahi , sasavans are nowise convincing. Causatives with p are not satisfactorily explained, and may come from roots in -p ; vaslmahi , occur- ring only once (R. V. XI. 72), is probably for vahslmahi ; and sa-sa-vahs, which occurs a few times in different cases in the Rig and the Atharva Veda, though hard to explain satisfac- torily (Grassmann and Delbriick suggest that it stands for sasanvdhs , and Saussure for sasavans ), cannot very plausibly be derived from sa , since no perfect-form, and in fact no tense-form whatever, of such a root is found in the language. Outside of the above forms, where analogy would require the nasal, if they are to be derived from roots in -n, the -a radical hardly occurs except where analogy would require a weaken- ed On the Eighth Verb-Class in Sanskrit. 9 ing of the root, especially on account of the displacement of the accent, as in verbal nouns in -td -tva, -tyia {ta-ta, ta-tva , ta-tya , etc.), and in nouns in -ti (formed in analogy with participles in -td). In all other cases, and they are very numerous, we find the full -^.-radical ( tdn-a , tan-as , tan-u , tdn-tn , tan-tra , etc.). It seems evident that we are here dealing with principles per- fectly parallel with those that have produced such root- variations as in han-ti : ha-tha ; hdn-am : ha-bhis ; rajan-am : raja-bhis ; ds-ti : s-tha ; pitar-am : pitr-su : pi-tr-n, pitr-a , and so on ; and it would be late in the day now to try to substitute for the principle of accentual influence as causing these vari- ations, the confusing principle of a variety of independent roots and stems. b. Since there is no evidence that the so-called ‘ general tenses * are of a more recent formation than those of the present-system, there is also no good reason to assume that the -tf-roots have formed first the latter, and then with an un- paralleled accretion, the others. We should expect at any rate to find in the general tenses some trace of an earlier -a-root, but there is absolutely none. As for the assumption that the whole conjugation-system of the various -an-v erbs is consist- ently made up of two independent roots, which, if existing at all, must have been distinguished by some shade of meaning, peculiar to each, it would be hard to find anything parallel. To be sure, sporadic cases of root-mixture in the make-up of verb-systems occur in all languages, but they are usually of a different kind, consisting in the often traceable supplanting of an earlier form by one of a different verb (as in French of Ire by etais ), and, in all events, they never take place so systematically in a whole class of verbs. c. If the facts of the Sanskrit language itself suggest clearly enough that the nasal of the ta/z-verbs is genuine, the evidence of cognate tongues also favors this view. To be sure, we find even in them radical elements corresponding to these roots with and without the nasal. Thus, cf. Skr. ■y/ksan: Gr. rcT€iv-co, ktov-os ; e-icra-Tco, 6 -ktcl-v, KTci-fievo^ \ 25 IO A. H. Edgren , Goth, ska-tkaQ) ; — Skr. i /tan.: Gr. relv-w, tl-tcllv-w, rev-cov ; ra-ros, Te-ra-fca, e-rdOrjv ; Lat. ten-do , ten-eo , ten-ax ; Goth. than-jan ; — Skr. -[/ man : Gr. /lev-co, ixe-fiov-a, /lev-os ; fxav-ia\ /jbe-fjLa-fjLev, iie-fjid-TCD ) Lat . mon-eo ; Goth, mun-um, ga-mun-an ; Skr. -y/van: d-fd-co, d-aa-ros ] Lat. ven-ia , ven-us, ven-ustus ; Goth, vnn-an , vinn-an ; — Skr. j/ san : Gr. ev-w (Fick, I. 226), avvw, avco ; e-ro? ; Lat. sinit, sen-ex ; Goth, sin-ista , sin-teino ; and so on. This variation, however, nowise favors the theory of independent -an and -^-roots. The nasal is rarely lacking except in Greek, and its absence there is frequently explain- able on precisely the same grounds as in Sanskrit, viz., as owing to an original weakening of the root : cf. tata : ; tati : rao-Ls ; mata : -/jlcltos ; satd : eVo?, etc. The very fact that cognate forms in other languages usually retain the nasal where it is lost in Sanskrit or Greek, tends to show that it is original, having disappeared in Sanskrit and Greek under certain circumstances which in the cognate tongues have produced different results : cf. Skr. mata or mati: Gr. yu-aro? : Lat. menti-y Goth, ga-mundi- ; Skr. menima (for *me-m n-i-ma : n preserved by the following i ) : Gr. fie/ia/iev : Goth, munum , and so on. Analogy has in all languages, as especially in Greek, wrought many changes, and must account for some of the seemingly irregular non-nasal radicals. Concerning the Greek forms, compare especially Brugman’s article already referred to above. If, then, contrary to the opinion of Van den Gheyn and others, it seems incontestable that the nasal of the roots ksan , tan , many van , and san is genuine and original, at least with reference to the time when the inflectional system of the Indo-European language was developed, does it necessarily follow that in a stem like tano -, the suffix must be -0 and not -no? Not at all. The stem would have precisely the same appearance were we to suppose either, with Bopp, that the n of the original -^ * tan-tva , etc. Then, such a theory completely ignores any influence on the root on account of the accentual shift in tanomi , though such an influence is directly required not only in analogy with words of the kind quoted above, but also by the analogy of other su-verbs (cf. star : str- no-ti ; kar : kr-no-ti). And it is precisely in consideration of this required weak- ening of the root that we are forced to explain the formation of tanomi as arising from a weakening of the unaccented tan to ta before the accented suffix -no. Whether we are to con- sider this weakening as consisting simply in the direct loss of the final nasal of the root or in its vocalization after the loss of the preceding #-vowel, is immaterial to the argument, and need not here be V^cussed. It may be said, en passant , however, that the objections made by Van den Gheyn in his third paper against Brugman’s ^-theory are nowise convinc- ing. It is true that no written language has left a trace of an zz-vowel ; but written languages never perfectly represent all the sounds of the spoken. No modern language has any sign for either a liquid or a nasal vowel, and yet such vowels are often met with (cf. English sabre , sable , fatten y button , etc., pronounced, like sabr, sabl, fattn , buttii). They may have existed jus as well, though imperfectly represented, in older dialects, the skilful Hindu phoneticians being the only ones to recognize in writing any of them (r and l). Van den Gheyn also objects that while the r-vowel always leaves a 27 12 A. H. Edgren , trace of its r y the ^-vowel has entirely sacrificed its nasal in Sanskrit and Greek. Even here English has a lesson to teach. Dialectically the r-element of such words as father , mother , arbor , etc., is often omitted in both America and England (cf. Whitney, Orient, and Ling. Stud., II., p. 236). But that means virtually that those words often, instead of being pronounced regularly as fathr, mothr , arbr y are pro- nounced as fatho , motho , arbo ( like kurmi , i.e. as borrowed from the root-class. Owing, no doubt, to the frequency of its use, kar shows a great mixture of forms (cf. auxiliaries, etc., in other languages). No other verb has a greater variety of stem-forms : kar- {kr-), kara -, krno- ( krnu -), karo -, kuru -, kur- ; and hence there is certainly no necessity to expect its conjugation in the classical period of the language to be of a homogeneous nature. Having thus disposed of the verbs assigned to the Eighth class by the Hindu grammarians, little need be said of those arranged along with them by later discoveries, viz. : han , in, and tar. As for han , if the form hanomi , occurring only once, is correct, it is subject to precisely the same treatment as tart, and must be removed with it to the Fifth class. In retains its nasal even outside of the present-system. If this nasal is to be considered as genuine, the conjugation of in in analogy with roots in -an would not be any more anomalous than the inflection of nominal zVz-stems in analogy with an- stems. As for tar, we have only the doubtful and anomalous form tarute , occurring once in the Rig -Veda. It is too prob- lematic to offer any grounds for a serious argument. The result of the preceding investigation would then be that of the thirteen roots which have been referred to the 29 A. H. Edgren. H Eighth class, arn, gharn , tarn , and ksin are to be struck out as fictitious, and tar as of wholly problematic relation to the questionable form tarute , while the remaining roots must be referred to the F ifth class, — fcsan, tan , man , van , san, han , as formed perfectly regularly with the suffix but kar and m as conjugated in analogy with verbs of that class. 30 C 3 III. — On the Auxiliary Verbs in the Romance Languages. By JOSEPH A. FONTAINE. Diez in his well-known “ Grammatik der Romanischen Sprachen ” has treated the question of the use of the auxiliary verbs in the Romance languages only in a general way, and much must be added, especially as concerns their history. Mr. Gessner in the Jahrbuch fur romanische und englische Sprache und Literatur : neue Folge , III. Band , 2. Heft , has made a very interesting but somewhat complicated study of esse, considered as an auxiliary verb, and has, moreover, no- ticed important facts that had escaped the attention of Diez. M. Camille Chabanneau in his “ Histoire et Theorie de la con- jugaison franqaise ” has devoted to the auxiliaries a few pages, containing valuable suggestions as to their use ; but as a general remark it may be said that he has treated this ques- tion too briefly. It may be simply because a full treatment was not directly included in the plan of his work. It may be said also that, in a certain way, M. Chabanneau has explained the use of the auxiliaries according to tendencies prevailing in modern French, and has fallen into the error, common to most grammarians, of trying to explain the inconsistent use of the auxiliaries, especially with the so-called neuter verbs, in accordance with modern usage. So far as I am aware, no one has yet tried to explain the difficult problem by a thorough comparison of modern usage with that of Old French. The main idea of M. Chabanneau is that the auxiliary is nothing but the inflectional part of the main verb. Granting this to be true, such a suggestion is not an historical explana- tion, and does not account fully for the various and often in- consistent uses of auxiliaries. Let it be admitted that ai and suis are merely inflectional University Studies, Vol. I., No. I., July, 1888. 3 I 2 J. A. Fontaine , parts of verbs. In “je suis venu ,” suis is considered as the inflectional part indicating the person, the number, and the tense of venir. But why do we use suis rather than ai, which is another inflectional part exercising the same function as suis in some other Romance languages ? (Cf. Sp. He venido as equivalent to Fr. Je sins venui) This is what needs explanation. Whether it is possible to solve this problem I shall not presume to say, but shall offer on this subject the result of my own researches. My subject naturally falls under three heads, viz. : — Chapter I. — Auxiliaries used with Transitive Verbs ; Chapter II. — Auxiliaries used with Intransitive Verbs ; Chapter III. — Auxiliaries used with Reflexive Verbs. Chapter I. — Auxiliaries used with Transitive Verbs. In comparing the conjugational system of the Romance languages, it becomes evident that in the active voice the use of auxiliaries is very nearly identical. Fr. J'ai chante. Pg. Hei cantado (tenho cantado). It. Ho cantato. Pr. Ai cantat. < Sp. He cantado {tengo cantado ) Wall. Am cantat. All the Romance languages use the auxiliary habere ; two of them may exchange habere for tenere. This interchange of auxiliary verbs, unknown to the other Romance languages, does not take place under the same circumstances in Spanish and Portuguese. The Spanish uses tener with transitive verbs when a certain stress is laid on the action expressed by' the verb, whereas in Portuguese ter has almost completely superseded haver. This discrepancy must not be overlooked, since it shows how independently languages develop. How is this production of a new auxiliary in the western group of Romance languages to be explained ? I do not believe it safe to think with Diez that the Spanish and Portuguese have introduced their new auxiliaries from a 32 Use of Auxiliary Verbs in Romance Languages . 3 desire to avoid the repetition of habere. If such had been the cause in Portuguese and Spanish of that innovation, probably the same cause would have exerted its influence over other Romance languages and produced in them parallel changes. A strong objection to the view of Diez is found in the very history of habere in Spanish and* Portuguese. Why is it that, during several centuries, the Spanish did not feel the neces- sity of using tener for haber , when the latter was used as a principal verb, and of reducing it to a mere inflectional part of other verbs, or to a mere auxiliary, as subsequently hap- pened ? If the use of haber from the first l^nown period of the Spanish language be exhibited, it will be seen that this verb was treated as a principal, or true verb, side by side with tener , and that the instances of its occurrence as such by far outnumbered those of tener ; that its use went on decreasing steadily, but slowly, till it disappeared (except when used im- personally, or with the force of debed) to play the part of a mere verbal inflectional ending. Statistics here will not be out of place, and they will show better than any general statement the fate of haber. Haber. Tener. El libro de los Reyes d’ Orient. E hovieron gozo por mira. Buena casa e fuerte tenemos. Grant ira avia. Yo tengo tan manya cuita. E non ayamos de ellos duelos. Que nunqua mas fin non habra. Vida de Santa Maria Egipciaca. Si ayades de Dios pardon. E los que de Dios non an cura. Nos ende avremos grant lacerio. Que debes haber honor. Doce anyos hovo de edat. Sol que aya algo quel dar. Elios avien grant sabor. Ella avie cinquanta vivos (amigos). Redondas avie las orejas. La faz tenie colorada. Si de Dios ayas amor. Non ho talente d’aqui estar. 33 4 J. A. Fontaine , Haber. Tener. Yo dieze, he buen corpo. Que non he oro ni argento. Non he conmigo mas que un dinero. Bien se que habre pardon. Un nombre avemos yo e ti. Ave mercet de mi. Non es home que paraiso hoviesse. Dos panes e medio ha en todo su poder. Tres panes hovo non grandes mucho. Grandes avian las coronas. Non avian cura. Querie haber proprietat. Malas intenciones havien. Grant pavor havran. No he vestidura ninguna. He yo gran repintencia. Si la gracia non he. En Dios he mi creyenza. Ayas tu duelo de mi. Piedat de mi cuerpo non avre. Que mucho mester lo avemos. Che non tengo mas d’un dobro. Mas non tenemos amas una via. Tu tienes un tal tesoro. El Poema del Cid. Avie grandes cuydados . . . . 6 Tanto avien el dolor . . . . . 18 Si oviesse buen senor . . . . 20 Del RRey non avie gragia . . . 50 A las fijas que ha . . . . . 384 Tiene dos areas lenas .... II 3 Avien los de ganancia . . . . 465 Dozientos marcos que tenie el rey Myedo yva aviendo . . . . . 1079 Alfonso 3246 Valencia que avemos por heredad, 1401 Cinco escuderos tiene don Martino, , 187 Todo el bien que yo he . E de mi abra perdon . . . . 1899 Todo lo tengo delant .... 1634 Non auredes my amor . . . . 2029 Tantos avemos de averes . . . 2529 Tienen buenos cavallos .... 602- Romancero del Cid. Hayais la muerte que el hubo 1 . 99 Que tenia mil amigos .... 5 Que magiier que haya razon . . IOI> Si tengo razon 6 non .... !5 Caballeros Castellanos. Que como otro bien non tengo . 39 Mudafar consigo habia . . . . 105 Y cinco hijos que tenie .... 7 i Tiene la culpa e no el dueno . . 85 El rostro tiene turbado .... 87 Que no tienen piedad .... 97 34 Use of Auxiliary Verbs in Romance Languages. 5 Tener. Con quien amistad tenia. Los de a pie no tienen cabo . • 143 Y tienan gran presuncion . . . 1 68 Testigos tiengo presentes . . . 169 El Libre d' Appolonio. Avian ventos derechos. Buena fija avemos. Aviemos tal senyor. Un vestido he solo. Non avie el poder de veyer. Avia grant repintencia. Avia placer. 1 Poema de Alexandro Magno. Tenyen viento bueno. Tenemos un buen home. Senyor destas companyas. Que no tiene vestido. El mar que mengua tuvo leyaltad. Avra de mi solas 3 Non tenie todas oras encobadas Avian gran alegria 13 las manos. Avie grant corazon 14 Tienes gran mejoria .... 47 Non he cura 38 Mas yo en mi non tengo el cor Avie grandes virtudes .... 83 que vos tenedes 2123 Non avria pavor 92 Si lo avia el brazo 96 A tales a los pelos cuemo faz un leon 138 Avie grant bontat. Avian buenos agueros .... 274 Tornada non avremos .... 847 t Los nuestros pensamientos non han stabilidat 940 Ca non avria tal vulto . . . ombre Tenie cara alegre. nacido 1104 Avran de vos venganza .... H 55 Yo por seso lo avria 880 No lo tengo por seso .... 879 Vida de Santo Domingo de Silos. Avie cuerpo fermoso 128 Non avemos dinero nin oro . . 364 Un caballo tenemos en casa Non avras nul trabaio .... 663 Avien los companneros grand rancura 293 1 I quote from “ The Book of Appolonius ” these few instances only, having placed opposite each other instances in which haber and tener are used with equivalent meaning, and where the one could have been used for the other. This will illustrate my position more forcibly. In “ The Book of Appolonius ” tener is already used at least as often as haber. 35 6 J. A. Fontaine , Haber. Tener. Martyrio de San Lorenzo. Ovo grant alegria 18 Avie en la cabeza enfermedat 35 La Estoria de Senor Sant Millan. Avie una azemila 27 El non tenie que dar li . . . . 239 Signos del Juicio. Si io gran set avia 33 Averan fambre e frio .... 33 Cronica del Rey don Alfonso decimo. Non abrian tiendas ningunas . . 9 Porque tenia derecho e tenia ver- E si poder avian para facer dad 69 emienda 26 Teniendo gran sentimiento del mal, 69 Donde avran ellos caballos e Todo el pan que tenian .... 72 donde los avremos nos . . . 5 2 El pesar que tenia por el fijo . . 77 Avia grande amistad con don Lope . 65 La gran guerra que ovo con el rey Aben 75 Hurtado de Mendoza. (La Vida del Lazarillo de Formes.) El tenia una area vieja .... I Sin duda debia tener spiritu de profecia 27 Tenia cargo de proveer. Otra cosa no tenia. Que casi tiene forma de loro. Tenia otras mil formas. El gran miedo que tenia ... 25 Tenia poco caridad. Mejor vida tienes que el Papa. Yo no tengo dinerosf .... 38 Mala medra teneif. Harto miedo la tengo. 49 36 Fingindo haber frio. Yo hube miedo que con aquellas diligencias J i Use of Auxiliary Verbs in Romance Languages. / Haber. Tener. Antonio de Solis. (. Historia de la Conquista de Mexico .) Porque tenia rara viveza de spfritu, 2 1 Tenia otras . . . de su proprio natural 9 El poder que tenia el Cardinal . 3 El re don Fernando solo tenia este titulo 3 Tenia en ella tres 6 quatro hijos . 16 Donde tenia el vidrio tanta esti- macion 5 De que tenemos algunos exemplos, 4 Don Quixote de la Mancha. Tenie en su casa una ama ... I. Tenie el sobre nombre de quijada, I. En que no tengo aqui dineros . IV. La hermosura que tengo . . XIV. Tengo riquezas proprias . . XIV. No tengas pena, amigo . . . XVII. El miedo que tienes. Yo no tengo la culpa .... XX. Tengo aficion, tengo misericordia, etc., etc. Las Mocedades del Cid : . En mi tendra . . . un fiel vasallo, 120 Tiene la cura a su cargo . . . 157 Tiene razon 272 Honra tiene para todos . . . . 321 Pulso tengo todavia. No tengo mas virtud 2254 Tiene prudencia y valor. Tiene la leche en los labios . . 785 We see that in Don Quixote and Las Mocedades del Cid haber is no longer used to indicate possession. It is used only as an impersonal verb. Cf. Don Quixote : Habiendo infinitos anos de lo uno a lo otro Asegurando le que no hdbria cosa que mas justo le diese que saberlo . 37 XLVIII. XLVII.j 8 J. A. Fontaine , or, in the sense of deber , cf. Las Moceclades del Cid : Que he de poner, 710. On glancing through the foregoing references it becomes evident that haber has ceased to be an independent verb in Don Quixote and Las Mocedaaes del Cid, and is used as such only in a few instances in Hurtado de Mendoza, so that the sixteenth century may be considered as the time of the dis- appearance of haber as a principal verb. In the above quota- tions I have placed side by side similar sentences, in some of which Jiaber is used, and in others tener , thus showing that haber and tener could be used indiscriminately, the one for the other, and that no essential difference existed then between those two verbs. Compare: — Haber. Tener. Avien ventos derechos. Tenyen viento bueno. Buena fija avemos. Tenemos un buen home. Un vestido he solo. Que no tiene vestido, etc. But at the same time it must be noticed that whenever an abstract idea is to be expressed, as for instance in duelo , dolor , pardon , placer , miedo , etc., the preference is given, in most cases, to haber. In fact, I have only noted seventeen instances in which haber was used to express a concrete idea, and these mostly in the earliest documents. Tener was the verb to express concrete conceptions and material possession, though it was also frequently used to express abstract ideas ; so that, in fact, its use was more extensive and varied than that of its rival haber. This being the case, tener grew in power and in favor, while the reverse was true of haber. Little by little haber was deprived even of the power of expressing an abstract conception, and left to play the part of a pure inflectional end- ing. Even here it is rivalled by tener ; for whenever a par- ticular stress is laid on the verbal notion, tener takes the place of haber. Thus, instead of attributing the introduction of new auxili- ary verbs in Spanish and Portuguese to a dislike of the repetition of the same verb, it seems more natural to attri- 38 Use of Auxiliary Verbs in Romance Languages. 9 bute it, first, to the close relatio?i of meaning existing between tenere and habere , which accounts for the indiscriminate use of the one for the other in the beginning ; secondly, to the fact that haber , by its very meaning, was adapted to the ex- pression of abstract ideas ; because, taking into consideration the great importance of that development of new auxiliaries in Spanish and Portuguese, it cannot be admitted that a law of mere euphony, as Diez suggests, has occasioned such a development. New auxiliaries are not produced to avoid too frequent a repetition of one already existing, but really because new ones; more expressive and more convenient to convey the thoughts, offer themselves. In tracing the history of haber and tener in Spanish, the period of the language at which the two verbs had nearly the same meaning and were used almost indifferently in the same sentences has been shown ; as also the period at which they became distinct in use, one of them even losing its independ- ent existence. The same phenomenon may be observed in Portuguese. A few instances will suffice. Haber. Tener. II canzoniere Portoghese ( MonacVs edition). Poys mal ne bem de vos no ey . 18 Tenheu por gram maravilha . . no Quanto ben auya perdi .... 33 Que nu ca teneste por ben . . . 128 E po mhavedes g m des amor 100 Poys no auedes merce de mi . 108 Q no ey en my forza ni poder 130 Cura no auedes 366 Des q mespertey ouiu gra peser. Si vos pecado avedes .... Per sonho mu g m vergonca ave- 470 Mais tenho q xha errou 0 iograr. des 982 In the Canzoniere, as we see, haver is by far the most in use ; ter has rather the meaning of ‘ esteeming, thinking,’ than * possessing.’ Hardunf s Roma?iceiro Porttcguez. E tres irmaos que havia . . . 159 Uma so filha que tendes .... 16 Chorava e razao havia . . . . 161 Uma irma que eu tinha .... 31 Ja se foram as galleras. As tres azenhas que tenho. Que Dom Duardos havia . . . 161 Tendes los olhos bonitos .• . . 97 39 IO J. A. Fontaine , Haber. Tener. Antologia Portuguesa. Ey gran cuidado . . . 7 Quern frores d’amor tern . . . • H Que gram saber eu havya . * 9 Se mais tevesse mais daria . • 33 Avemos majores coidados . . . 20 Tendes pae e tendes mae . . . 65 Nom ha torto .... • • • 33 Avian dinheiros . • • • 37 Averia gran prazer migo ... 48 Senhoras nam ajaes medo . . . 156 Has medo que morrerei? Hei . . . 164 Up to this period haver is used more frequently than ter , as was the case in Spanish. Both bore about the same meaning, and could be used the one for the other in similar sentences. In Sa de Miranda and in Camoens, the reverse is true. Ter prevails, and haver becomes more nearly obsolete, and is reduced to play the part corresponding to that of the French verb avoir in “ il y a ,” or to be a substitute for convir or the French falloir: Hei de par tir (il faut que je parte). Sa de Miranda ( Os Fstrangeiros ') . Deixay me passar que nao Ey contigo nada 102 Tu tern cuydado de meu . . . 93 Nao ajaes vos medo 1 16 E tern razao 94 Lucrecia avia a minha filha nome 142 E essas nao tern spirito ”5 Bons pes tenho 124 Pouca confian?a tens en Lucrecia 129 Os Vilhalpandos . Mas ey miedo que nos fuya 0 tempo E averemos todos conselho • 273 Nao temos tempos . . . , . 273 Cuidado avia en casa . 258 Tenho inimigos . * . 257 Quando aviamos mister mil olhos 226 Nao tive mais paciencia . . . 240 Tenho grande necessidade de ti . 192 Tinha algun sentimiento de homen 196 Aiuda tu tens boas pernas . . . 220 Os Lusiadas. Que tambem della hao. medo , . II.47 Victorias que tiveram .... 1-3 Portugal houve em sorte . III. 25 Tenham inveja ...... 1-3 Este porhaver fama sempiterna IV. 60 40 Use of Auxiliary Verbs in Romance Languages. ir Haber. Tener. Quando juntas com subita alegria Poder nao teve a morte . . . I. 14 Houveram vista da ilha namo- Por armas tern adargas . . . . I. 47 rada IX. 51 Piloto aqui tereis. Tempo concertado eventos tinha I. 95 Nao teve resistencia . ... II. 69 O tu que so tiveste piedade . II. 105 . O tu que tens humano o gesto III. 127 As I have not had the opportunity to gather instances from a later period of the Portuguese language, I cannot give the exact date of the complete disappearance of haver as used to indicate possession. The sixteenth century and the beginning of the seventeeth may be assigned both for Spanish and Portuguese as the declining period of haber , though the change of this verb appears to have been slower in Portu- guese than in Spanish. Such has been the history of haber and haver , tener and ter considered by themselves. It will now be shown that, con- sidered as mere auxiliaries, they have passed through nearly the same process of development ; that is to say, tener is used to form the compound tenses in Spanish alongside with haber ; and ter supersedes almost entirely haver in Portuguese. Vida de Santa Maria Egipciaca . Tanto la avia el diablo comprisa. Dexare aquesta vida. Que mucha la he mantenida. Quando hovo fecho su jornada. El Libre d'Appolonio. Por el amor que yo tengo estab- lecido contigo. Des aqui adelante lograr quiero lo que tengo ganado .... 649 Poema de Alexandro Magno. El infante quando ovo su cosa Pora en Pentapolim lo tengo es- acabada 147 leido. Quando todas tierras ovo en paz Que esso que tu dices tenia yo es- tornadas 223 mado 1840 41 2 J. A. Fontaine , Haber. Et yo si non oviesse abiertos los caminos 236 Tovieron que havian fecho bona conquista 1698 La Estoria de Senor Sant Millan . Secund esta noticia que avemos contado 364 Que todos estos signos que vos visto avedes 403 El Poema del Cid. Bien los ovo bastidos .... 68 Quando tal batalla avemos aran- cado 793 Desta batalla que avemos aran- cada 814 Tener. Ya tenie aguisado naves e mari- neros. Antonio de Solis. Per que habia muerto su compa- nero XIV. Las lineas que tenia tiradas . III. Cuyo imperio tenia el cielo des- tinado par engrandecer . . V. Ya tenia comprados algunos ba- xeles VIII. La confederacion que tenian hecha XVII. Don Quixote . Tambien como otro que haya go- bernado insulas X. Has tu visto mas valeroso Cabal- lero que yo X. Ya no he leido ninguna historia X. De cinco que habia dejado en ella XXIII. Pero ya tenia abierto uno el barbero VI. Lo que de la insula me tiene promedido VII. No se paguen donde tengo dicho X. Como tenie tambien conecido el humor XLV. Thus we see that in Spanish tener is used in compound tenses wherever there is a stress laid upon the verbal idea, or when the action expressed by the verb is considered as not belonging merely to the past, but as continuing in the pres- ent. In a word, the tenses compounded with tener denote a 42 Use of Auxiliary Verbs in Romance Languages. 13 more lasting and more emphatic conception. This is, of course, the natural result of the stronger and more concrete meaning of tener , as compared with the weaker and more abstract one of haber. A greater predominance has been accorded to the new auxiliary in Portuguese than in Spanish ; for while in the latter tener allows haber to play a very important part in the conjugation of verbs, or replaces it only in special cases, ter in Portuguese has become the true auxiliary, and haver is very seldom met. The following quotations will show ter superseding the auxiliary haver , as ter principal verb had superseded haver principal verb. Haber. Tener. II canzoniere Se qr do q lhy auia emp stado. Nostro senhor se averez guydado Gram pecado avedes de mi coy- 130 tado 131 P r qua te mal ey levado . . . 145 Que muyto mel avia jurado *45 Mays p r q mha mentido . . . 250 Q lheu avya mandado .... 413 Muytos anos avemos passados 455 Pois me avedes preguntado . . Ca hu iudeu avedes enganado. 903 Portoghese. Ora senhor tenho muyt agisado . 220 Ora tenho guysado de marchar . 952 Hurdling' s Romanceiro Portuguez. 0 falcao perdido havia .... 47 Tenho feito juramento . . . . 2 5 Que nunca me has mentido 106 Aonde nunca tinha ido. Ou grande traicao ha sido . . . no Acorda ja bella infanta. Triste sommo tens dormido . . in Pois tudo tendes ouvido . . . 11 3 Antologia Portuguesa. Um cantar d’amig ha feyto . . 10 D’aver coita muita ten’eu guisado . 46 Mays pois m’o houveran dito . . 10 Tinha 0 cavallo sellado .... 63 Ca muitos annos avemos passados 20 A morte tenho passada .... 69 Preguntal-o-ey porque me ha des- pagado 2 5 43 J. A. Fontaine , H Haber. Tener. Ca langa ha torta d’un ramo . • 32 "E tantos padecimientos tenho pas- Que 0 aormir ja 0 ey perdido . 44 sados Eo ouve bem servido . . . . 84 Senhor que nos ha livrado . . . 109 Haver is more frequently used than ter as an auxiliary in the Romanceiro, the Antologia, and especially in the Canzo- niere, but the balance is soon changed. Sa de Miranda (Os Vilhalpandos). Temos gastado muito do tempo . 78 Quantos exercitos tenho eu so por num desbaratados 161 Milagros que ja tinhao feitos . . 166 Nao sabes tu que tens mudado o costume 216 Escudos que tinha recebidos . . 249 Quantas vezes tenho dito. Seu conselho que eu avido tenho 124 O rosto do pay a que tens errado. Cuida que me tern alugado . . 144 Os Lusiadas. Que havendo tanto ja que as partes vindo I. 27 Do licor que Lyeo prantado havia. Enchem vasos de vidro . . I. 49 Hajam os Portuguezes alcan- ?ado 1-74 Nao nego que ha comtudo . VIII. 42 Nas aguas tern passado o duro inverno I. 28 E tendo guarnecida a lassa frota I. 29 Ouvido tinha aos Fados . . I. 31 Da determinagao que tens to- mada I. 40 Diversos ceos e terras temos visto I. §1 Como entendido tenho ... 1-79 Tem determinado de vir por agua I. 80 Nao menos tem mostrado es- forgo e manha VII. 71 Que Amor por grao mercS lhe tera dado V. 46 Ter in the Lusiadas is used in eighty-eight per cent of the instances of compound tenses. 44 Use of Auxiliary Verbs in Romance Languages. i 5 Haber. Tener. Colleccao de Tratados ( sixteenth century) por Biker. Algunos portugueses a que vos alteza tem dado credito ... 7 Tinha ganado Urmuz 25 Enquiry coes que sobre a dita morte tenho tiradas 52 Os portos que tynheis ganhado aos Mouros 25 A ilha de Maquiem que tinhao to- rnado • . 52 I have found no other instances of haver in the first part of the Tratados. Chapter II. — Auxiliaries used with Intransitive Verbs. The question concerning the use of habere and tenere as independent verbs or as auxiliaries to transitive verbs, is easy to study and to solve. In Italian, French, and Provencal, habere is now and has always been the only auxiliary. In the preceding rapid sketch, it has been shown that the use of two auxiliaries in Spanish and Portuguese is due to the antagonism and particular development of two words having originally a similar meaning. But the question of the use of auxiliaries with intransitive verbs is of a quite different kind, and fraught with many difficulties. An historical treatment of that question in French and Italian needs be quite exten- sive. The task is easier in Spanish and Portuguese, since these languages, yielding more to the power of analogy in their conjugational system than either French or Italian, have employed a single auxiliary for their intransitive conju- gation, — at least, during the last three or four centuries. Let us examine the question in each of the Romance languages. I. Auxiliaries used with Intransitive Verbs in French. Two auxiliaries are used in compound tenses of the French intransitive verbs : avoir and etre. The derivation of avoir 45 1 6 J. A. Fontaine , from Latin habere is perfectly clear and well understood. The different tenses of the French verb etre have been taken from those of the corresponding Latin verb, esse, sum,fui, etc., save the imperfect e'tais from stabam, the present participle etant from stantem, and the past participle e'te from statum. The Old French possessed two imperfects, ere and estoie. The first was used more frequently in early French docu- ments, and very likely continued to be employed during the next two centuries with decided preference, and we find it still in the prose of Villehardouin and Joinville; but at the end of the thirteenth century estoie had become the more important, and in the next hundred years rose to be a rival of ere, and even began to usurp its place. Now why did the French reject eram for the sake of staba7n ? The reason is similar to that given in explanation of the preference given by the Spanish to tenere as contrasted with habere. Stare bore about the same relation to esse as tenere to habere. Stare in Latin could be used and was used to indicate exist- ence/being thus synonymous with esse. Esse has an abso- lutely abstract meaning, and expresses a permanent and inti- mate state or condition of existence of the subject, while stare implies a more descriptive, more external, and more transient one. It is especially to this last meaning of stare that the Spanish and Portuguese have given a particular development. In French, stare has entirely supplanted esse in its imperfect use, the imperfect being the descriptive tense par excellence ; the keeping of etant and ete is due rather to a lack of corresponding forms in the Latin verb esse than to anything else. The future ser-ai is generally taken from essere habeo. But essere habeo ought to have given us estr-ai. This form is actually found in Saint Leger, tos consilier ia non esti-ai (16), and in the Alexis, Chambre, dist ele ia mais vfestras parede (29). Could not, then, serai be derived from other sources ? We shall see when we come to the Spanish. 46 Use of Auxiliary Verbs in Romance Languages. ij After these general remarks on esse in French, let us con- sider the use of the auxiliaries in the intransitive verbs in French. Grammarians have laid down the following rule concerning the use of these auxiliaries, viz. : With intransitive or neuter verbs the auxiliary avoir is used when the verb is considered as expressing an action, and the auxiliary etre is used when the verb is considered as expressing more particularly a state or condition. M. Chabanneau adds : “ II n’y a pas a cet egard de regie absolue, non plus que d’usage uniforme dans les langues romanes.” Does this principle hold good ? When I say, Je sins venu vous voir ; Quand je suis entre le theatre com- menqait . . . / Les allies sont entres dans la ville , vers les sept heures , I use the auxiliary etre , although the verbs venir , entrer., do not denote state or condition, but express clearly the action of coming and entering. Such a law as that stated above is not satisfactory. M. Chabanneau tries to illustrate the rule by the two following examples : Cette femme a accouche ce matin ; Cette femme est accouchee heureusement. In the first example, says our author, we have in mind the action rather than the result of accoucher ; but in the second, the result rather than the action. The dictionaries of the French Academy and of Littre say the same thing. It would seem bold to oppose two such authorities. Their statement may be true as regards present usage, but it is not always true in regard to Old French. Before showing what was the case with the Old French, I shall remark that M. Chaban- neau, in the two instances above given, limits the first verb by an adverbial modifier, 11 ce matinf which time limitation obliges us to think rather of the action than of the condition, and modifies the second verb by an adverb of manner, “ heur- eusement f pointing out a condition rather than an action. The example given is evidently chosen to fit the rule. But take the same example, omitting the modifiers referred to above: Cette femme a accouche d'une file ; Cette femme est accouchee d'une file. Both of these expressions are, gram- 47 i8 J. A. Fontaine , matically speaking, correct, and both express an action, show- ing that there ought to be no difference between the two auxiliaries. Littre in his dictionary says that “ Elle a accouche heu- reusement,” “ Elle est accouchee depuis un mois,” are faulty locutions. Why ? Because the Academy says so, and be- cause M. Littre asserts that a accouche expresses the action, and est accouchee the condition. Compare parallel expres- sions in Old French : — Ci-dessous gist estendue et couchee Une qu’amour si bien vaincue avoit Que plusieurs fois elle en fust accouchee. — St. Gelai, 197. According to M. Littre, such a locution also is faulty, because plusieurs fois elle en fut accouchee expresses an action, as well as elle a accouche depuis un mois. Yet we have a more strik- ing instance of what M. Littre could have called a faulty locu- tion in Joinville, Histoire de St. Louis , p. 218: “La Royne acoucha d’un fil qui ot non Jehan. Le jour meism que elle fu acouchief etc. Is not the verb accoucher used here in the same way as in Littre’s “vicieuse locution,” quoted above, and in Chabanneau’s “elle a accouche ce matin ” t Hence it is apparent that the rule requiring etre to denote condition and avoir action is based on modern usage, and, I dare say, has nothing to do with the historical development of the auxiliary usage itself. It would take too long to quote in detail what other gram- marians have said about the use of auxiliaries in the intransi- tive verbs in French. Suffice it to say that they have divided the neuter verbs of the French conjugation, numbering about six hundred, in the following three classes, viz. : first class, comprising verbs conjugated regularly with avoir (about five hundred and forty) ; second class, comprising verbs conju- gated with etre exclusively, viz. : — aller avenir deceder disconvenir entrer partir rancir tomber arriver avorter dechoir echoir mourir provenir redevenir retomber atterrir choir devenir emaner naitre rabougrir ressortir venir 48 Use of Auxiliary Verbs in Romance Languages. 19 and third class, comprising verbs conjugated sometimes with etre and sometimes with avoir , such as f aborder apparaitre contrevenir demeurer expirer passer resoudre accoucher baisser croitre descendre grandir perir resulter accourir cesser degenerer disparaltre monter rester vieillir and some others. As to the verbs conjugated with etre alone, grammarians maintain that they cannot take the auxiliary avoir , because they express nothing but a state or condition. As to the verbs conjugated with avoir or etre , grammarians have used all their critical acumen to explain the use of the one or the other. They have even censured illustrious'writers for not adhering to the rules formulated by them. For in- stance, Racine said in Berenice, A. II. Sc. L, II en etait sorti lorsque fy suis couru. If we listen to grammarians, Racine was wrong, and ought to have said, fy ai couru. We shall see later whether Racine was wrong or not in saying j'y suis couru. Compare courir with its compound accourir. Courir must take the auxiliary avoir , and accourir may take either avoir or etre. Why ? M. Girault Durivier, in his “ Grammaire des Grammairesf p. 472, says : “La raison pour laquelle courir prend toujours. l’auxiliaire avoir, et accourir tantot l’auxiliaire avoir et tantot l’auxiliaire etre est que courir n’exprime qu’un mouvement,. qu’une action, au lieu que dans accourir qui signifie se mettre en mouvement pour arriver promptement a son but on dis- tingue deux choses, l’action de se mettre en mouvement pour courir vers un but, et l’etat qui resulte de cette action faite. Des que je V ai entendu se p lain dr e, j' ai account a son secours. Je suis accouru a son secoitrs (would mean, says the author) j' etais dans V e tat qui resulte de V action d' accourir au secours de quelqu un." The correctness of the author is not to be doubted, but the explanation is questionable. M. Girault Durivier is right, according to the modern tendency which dictates the use of avoir or etre in cases similar to those quoted. But I hardly think that the use of the one or the other auxiliary essentially modifies the verbal meaning. Other Romance languages have, in their early period, made 49 20 J. A. Fontaine , use of two auxiliaries also ; but later on they were confined to one, without losing the power of expressing with one auxiliary shades of meaning that seem to require in French the discriminating use of two auxiliaries. Voltaire said in Orphelin de la Chine, II. 3 : — Ou serais-je, Grand Dieu ! si ma credulite Etit tombe dans le piege k mes pas presente ! Here Voltaire seems to be wrong, because he makes use of the auxiliary avoir with tomber , and tomber cannot take such an auxiliary. So J. J. Rousseau, according to the same gram- marian, was wrong when he said C' est ainsi que la modestie du sexe est disparue peu a pen ; because here disparaitre expresses an action, and not a state or condition. The same remarks have been made about the use of auxiliaries with other intransitive verbs, , such as perir , e'chouer , accoucher, cesser , demeurer , apparaitre , croitre , partir, rester, etc. It would take too long to consider these verbs, one by one, and to discuss their special meaning according to the auxiliary with which they are conjugated. I shall only remark that the rules given by the grammarians are not observed in popular usage, where II a descendu l’escalier en courant or II est descendu l’escalier en courant are used indiscriminately, and convey the same meaning. The inconsistency of grammarians and the unsatisfactory explanation they have given concerning auxiliaries may serve as an excuse for my trying another explanation, based entirely on the comparison of the modern with the Old French, and for presenting a few considerations on, first, different kinds of verbs ; secondly, the origin of the so-called neuter or intransitive verbs ; and thirdly, the different kinds of neuter verbs. First : Different Kinds of Verbs. The classification here made of verbs is based on the fol- lowing principle : Every verb is active in its original and 50 Use of Auxiliaiy Verbs in Romance Languages. 21 primary meaning, and expresses an action. Every verb has the inherent power of being used as transitive or intransitive; consequently verbs intransitive in modern French may have been formerly, and can be used in the future as, transitive ; verbs intransitive in some of the Romance languages have been, are, or may be used as transitive in some other Romance languages. Compare jouir, prosperer, courir, obeir, dormir, renoncer ( intransitive dn French) , and godere, prosperare, correre, ubbidfre, domire, rinunciare ( transitive in Italian ). Ex. — Ed allora gode la fortiria (Bocc.). Ex. — Mai hai i tuoi maestri ubbidfti, etc. A verb is not necessarily transitive or intransitive ; it is made the one or the other according to the development of its own meaning, and according to the peculiar genius, stage, and tendency of the language in which it is used. Hence all verbs in their nature are active verbs, and express an action ; in their use they are divided into two classes, viz. : transi- tives, affecting an external object; and intransitives, or semi- transitives, affecting the subject. For brevity’s sake, the verbs of the first class may be called objective-transitives ; and those of the second, subjective-transitives. The verbs of the first class, expressing an activity directed towards an external object, are conjugated with avoir ; the verbs of the second class, expressing an action affecting the subject itself, partake of the nature of passive verbs, and thus take etre. But later on the second class of verbs was developed into two classes : subjunctive transitives fully expressed : Je me repens; and subjunctive transitives elliptically expressed : Je meurs (for Je me meurs). Compare Spanish, Italian, Portu- guese. Secondly : Neuter Verbs. Their Origin. To the second of the above classes belong the so-called neuter verbs ; they are nothing but elliptical reflexive verbs, or subjunctive transitives elliptically expressed. As a conse- quence, they form in several cases their compound tenses with etre , according to the principle described above. 5i 22 J. A. Fontaine , Thirdly : Different Classes of Neuter Verbs. With reference to conjugation, neuter verbs must be divided into three classes : — First, neuter verbs, the radical significance of which being still felt, are conjugated with etre ; not at all because they express a state or condition rather than an action, but be- cause they are elliptical reflexive verbs. Secondly, neuter verbs still under the influence of their origin, but yielding in a great measure to the effect of what I should call the Romance tendency. Under that tendency Romance languages have been substituting, in a less or greater degree, active and reflexive for passive expressions, and active verbs for passive verbs ; hence active auxiliary for passive auxiliary. As a consequence, the verbs under the influence of these two principles will have a wavering use of the auxiliaries etre and avoir , with a slight preference for avoir. Thirdly, neuter verbs subject to the influence of the Ro- mance active tendency, and taking the auxiliary avoir exclu- sively. This third class of neuter verbs is the only class to be found at present in some of the Romance languages. I will now take a certain number of verbs belonging to the first class, and prove that they have been used in the Old French as true reflexive verbs. For instance, with alter, Je sins alle is equivalent to Je me suis alle. The study of the verb aller in Old French literature is very interesting. The frequency of its use may be said not to be exceeded by that of any other verb. When the different ways in which aller is conjugated are observed, the student is at no loss to explain its auxiliary, and can no longer say with grammarians that aller takes the auxiliary etre because it expresses a state or condition ; as if any verb could express more clearly an action than the verb aller. This verb was used, first, in the reflexive form with en : s' en aller ; secondly, in the reflexive form without en : s' alter ; thirdly, in the so-called neuter form with en: en aller ; fourthly, in the so-called neuter form without en : aller. 52 Use of Auxiliary Verbs in Romance Languages. 23 Examples of the Firs'!. Petrus dalo fors s’en aled Pass. Chr. . . . 197 Tot s’en vait declinant Alexis .... 9 Puis s’en alat en Alsis la cite Alexis .... 86 Com s’en alat et com il s’en revint Alexis .... 285 Franc s’en iront Roland ... 50 Li empereire s’envait desuz un pin Roland . . . 168 Que en France m’en alge Roland . . . 187 S’en vait a son ostel Roland . . . 342 S’en est alez li reis Roland . . . 501 Et lors li consaus s’en rala parler an Soudanc .... Joinville . . . 186 Et vous commandons que vous ralez vers notre Seignour . Joinville . . . 248 Signor, je m’en vois outre mer Joinville ... 64 Examples of the Second. Que a pou se ala que il ne nous afondrerent en l’yaue . Joinville . . . 168 Examples of the Third. Et en alat en Alsis la citet Alexis .... 382 Alez en est en un vergier Roland ... 1 1 Desuz un pin en est li roi alez Roland . . . 165 Seignurs vous en irez Roland . . . 360 Li Rois d’Ermine en ala au roy des Tartarins .... Joinville ... 78 Tu en iras a ton roy . ‘ . Joinville . . . 264 Il en ala grant pas par mi son vessel Joinville ... 88 Et bien toute la voie que li connestable et je en alames amont Joinville . . . 136 Si nous en irons tuit en paradis Joinville . . . 174 Qui en fut alez apres les Sarrazins Joinville , . . 108 Examples of the Fourth. 1 st de la nef e vait edrant a Rome Alexis . . . . 21 1 Est alez conquerant Roland . . . 553 Si l’est alez ferir Tant que il veist que sa chevalerie feroit, qui aloit a terre . Joinville ... 6 Mais je dont se je aloie vers vous Joinville . . . 216 Car nous sommes alei contre le commandement Mahomet Joinville . . . 202 Car vous estes alei la sus sans mon commandement . . Joinville . . . 316 Qui avoit appelei contre li et estoit alez a Rome . . . Joinville . . . 370 Cf. Italian andarsene , andarne , andarse , andare. Examples of the First. Tu te ne andasti e si rimase seco Petrar. S. . . . 204 Secando se ne va 1 ’ antica prora Dante Inf. VIII. 29 Ora sen va per uno stretto calle Dante Inf. X. 1 53 24 J. A. Fontaine , Vattene omai Dante Purg. XIX. 139 L’ una gente s’ en va l’altra s’ en viene Dante Purg. XXVI. 46 Di Tebaida andati se n’ erano Bocc. Dec. . . 3. 9 Examples of the Second. E in su una sua nave . . . n’ ando in Cipri Bocc. Dec. . . 3. 7 II peregrino tantosto n’ ando a quattro fratelli .... Bocc. Dec. . . 3. 7 Sappiendo verso che parte n’ era la fregata andata . . Bocc. Dec. . . 5. 6 Examples of the Third. E s’ andarono tutti alio cortiglio Ma vassi alia via sua . . . Examples of the Fourth. Da Foggia andao le Re Matteo Spinelli 1093 Remark. Alter was sometimes also conjugated with the. auxiliary avoir , just as formerly in Italian, and now in Span- ish. Compare Quant j’oi un poi avant ale Roman de la Rose 5 (Ital.) Egli a andato a lui (common speech). With arriver , je suis arrive is equivalent to je me suis arrive. (Arriver used as active verb.} Cil a sa nef apareilee; entrent dedenz II les mena tant qu’al rocher les arriva Greg, le Gr. . . 104 (Arriver used as reflexive verb.} S’en alia outremer . . . et s’ariva a Acre Chron. d’outr. Et ayant dit cela s’arriva contre le corps en la fosse . . Matt, de Coucy . 738 Compare Italian arrivarsi. With entrer, je suis entre is equivalent to je me suis entre. Si s’ en intrat en un moustier St. Leger ... 66 As porz d’Espeigne s’en est entre Rolant Ronceval ... 14 Compare Italian : — Non potea riveder d’ ond’ io m’ entrassi Purgatorio XXVIII. 24 Da lui ne dall’ amor che in lor s’ entrea Parad. XIII. . . 57 Ignudi amenduni se n’ entrarono nel bagno Bocc. Dec. . . 8. 10 54 Matteo Spinelli 1093 Purg. XXV. . . 5 Use of Auxiliary Verbs in Romance Languages. 25 Compare Provencal : — Ab aquestas paraulas lo rey s’en es intratz. With mourir , je suis mort is equivalent to je me suis mort. Por 0 s furet morte a grant honestet Eulalie. 11 se fut morz, dam i fud granz 51 Et disoit que li ennemis si soutilz que quand les gens se meurent, il se travaille tant comme il puet que il les puisse faire mourir Joinville . . . 24 E alors elle se pasma, et cuida l’on qu’elle fut morte et li Roys qui cuida qu’elle se mourust retourna . . Joinville . . . 332 Qu’ele li dist qu’il li donroit S’amor, ou ele se morroit Roman de la Rose 48 Les uns mouroient sans parler les autres se mourroient en parlant Rabelais Here it may be remarked that the Old French uses se mourir in Eulalie and St. Leger to express the very act of dying, whereas in modern French se mourir means 'to be agonizing ’ ( Madame se meurt , Madame est morte. — Bossuet). Both meanings are found in Joinville (see above). A very important fact to notice is that mourir could be used not only as a reflexive verb, but also as an active verb ; and that not only in the French, but in all the Romance languages. Fittre, in his long article on mourir> says nothing about the active meaning of this verb. Or veez vous bien que je vous eusse bien mort, si je vousisse Toute voiz ce ne leur eust riens valu que li Turc ne les eussent touz mors ou champ Mes ce m’a mort que poi me dure Or se ce non vous m’avez morte Compare Italian : — Ne necessity conviene que la gentilissima Beatrice alcuna volte si muoi Per una donna que s’ era morta Per paura morte s’ erano E dopo alcun di arrabiato si mori Ettore avendo morti grandissima quantita di Greci . . I quali furono morti et cacciati Muorsi si subito nelle sue braccia Joinville . . . 348 Joinville . . . 152 Roman de la Rose 81 Roman de la Rose 1 88 Vita Nuova XXIII. Vita Nuova XXXIII. Bocc. Dec. . . 2. 7 Dino Compagni . 506 Ricardo Mai. . 885 Ricardo Mai. . 996 Bocc. Dec. . . 3. 6 55 26 J. A. Fontaine , Nella sua vista, e cotal si moria Purg. XVII. Sanar le piaghe c’ hanno Italia morta Purg. VII. Per lo giusto disdegno che v’ ha morti Parad. XVI. Compare Provencal : — Quar non puesc vezer qui t’a mort. 27 95 [ 37 With naitrc, je sins ne is equivalent to je me snis n£. Et s’est nee et estraite de si bonne lignee Bert. . . . 7 2 Compare the Wallachian, where naitre is only reflexive. Eu m’am nascut la anul i860. With partir, je suis parti is equivalent to je me suis parti. Et quant li frere s’en furent parti Que oncques ne s’osa partir tant que il fut accordet au conte de la marche Et lors je me parti de Joinville sanz rentrer ou chastel jusques a ma revenue Mai apertement se partirent li Turc de Diamiete . . Mort le trebuche, l’ame s’en est partie La parole est finee et li conseil se part Se partit du diet lieu Se furent de Israel partiz Dont il s’estoient folement parti Et se partent des cors les ames Compare Italian : — Poiche la gentilissima donna fu partita . . . Messer Francesco de Loffredo partio de Tarento Alcuni si sono partiti del suo proprio parlare . Che riguardasse se partito si fosse Rimanete con Dio che io mi parto Quando tre ombre insieme si partiro .... With tomber , je suis tombe is equivalent to je me suis tombe. I was unable to find a reflexive example of tomber. Du Cange says: “ tumber vero active sumitur pro dejicere, vulgo faire tomber.” Hence, if tomber was used formerly as a transitive verb, there is no reason to suppose that it could not have been used as a reflexive verb. 56 Joinville . . . 14 Joinville . . . 56 Joinville . . . 68 Joinville . . . 90 Rone 58 Saxe XXIX. Pantagruel II. . 5 Livre des R. . . 48 Guillaume de Tyr. 103 Roman de la Rose 246 Vita Nuova XXX. Matteo Spinelli . 1605 Della Volg. Eloq. 73 1 Bocc. Dec. . . i- 7 Bocc. Dec. . . 4.. 1 Infer. XVI. . . 4 Use of Auxiliary Verbs in Romance Languages . 27 Instances of tomber used as a transitive verb : — ■ Mes la contraire et la perverse Quant de leur grant estat les verse Et les tombe autor de sa roe Du Cange. Icellui Giraut donna au dit Manson un si grand coup sur l’espaule qu’il le tomba par trois fois en la charriere Du Cange. Puis le tombent en un fosse Roman de la Rose 52 Now it is clear that Voltaire, in saying, “Eut tombe dans le piege,” was not wrong, since the verb tomber is conjugated with the auxiliary avoir in Old French, and bears an active meaning. The sentence of Voltaire, though obsolete in the time he wrote, had nothing wrong in itself. With venir , je suis venu is equivalent to je me suis venu . The reflexive form of this verb is generally found in con- nection with en. But se venir and s’ en venir must be re- garded, just as s en alter and s alter were regarded ; that is to say, en does not change the reflexive nature of venir and aller , but when that en is found in a sentence in connection with venir and alter, it controls and necessitates in most cases the use of the reflexive form, which, however, is per- fectly independent of it. It is only necessary to compare the instances cited above, of s' aller, and the Italian venirse, or venirne, and venirsene. Ma viensi per veder le vostre pene Inf. XII. ... 21 Ma vienne omai che gia tiene ’1 confine Inf. XX. . . . 124 Venir sen deve giu tra’ miei meschini Inf. XXVII. . . 115 L’ un poco sovra noi a star si venne Purg. VIII. . . 31 Sen venne suso ed io per le su’ orme Purg. IX. . . 60 Io mi vengo a star un po teco Si vinse il peggiore Ricardo Mai. . 990 Ove in Leone ad incontrar si venne Orlando .... Orl. Furi. . . 46. 21 Let us now examine a few of those verbs that take either etre or avoir, according to their peculiar meaning, as gramma- rians say. I hold, on the contrary, that when those neuter verbs take etre, they are purely elliptical reflexive verbs, just as those of the above list ; and that when they take avoir, they follow the general tendency of development that pushed the Romance verbal-system towards an active auxiliary. The Old French used such forms as etre apparu and etre peri very 57 2 8 J. A. Fontaine , frequently, whereas the modern French avoids them and generally prefers the active auxiliary form. With apparaitre , je suis apparu is equivalent to je me suis apparu. Si s’aparust et sor mon chief me mist sa main .... Rom. delaR. 10,347 Tout fust il ainsi que nus ne se fust aparus contre eux . Beau 54 La deesse Vesta s’apparust a lui Amyot. Rom. . 4 Jesus Christ glorieux s’apparut Flechier. L’ange du seigneur s’apparut a lui Volt. Phil. V. .110 De rechef s’apparut Dieu en Silo Livre de Rois I. Saint Andrius apareuz a lui Guill. de Tyr. . 208 With cesser , la pluie est cesse'e is equivalent to la pluie s' est cessee. Quand ce cri fut repandu parmi l’ost tous se cesserent . Froissard II. . . 215 Je me veux cesser de parler de faitz d’Angleterre . . . Commine . . . 3. 7 With croitre , la riviere est erne is equivalent to la riviere s est erne. The only example found of se croitre is in Lacurne’s Dic- tionary. Se croistre, he says, is used for s' accroitre in Perard’s Histoire de Bourgogne. Very likely the reflexive use of croistre has been transmitted to the ’compound s' accroitre. Compare the active use of crescere in Italian. Madama voi dalla poverth di mio padre . . . come figliuola cresciuta m’ avete Bocc. Dec. . . 2. 8 E che piii volte v’ ha cresciuta doglia Inf. IX. ... 96 With demeurer , je suis demeure is equivalent to je me suis d erne ure. The first meaning of demeurer was to ‘stop at,’ to ‘delay.’ (See instances of se demeurer in Lacurne’s Dictionary.) E si li estoient chil doy roy si prochain que a envis s’en mesloit et a envis s’en demoroit . . . «, . . Froissard II. . . 481 Compare Italian dimorarsi : — Mi sono dimorato in Parigi sei anni. 58 Use of Auxiliary Verbs in Romance Languages. 29 t With descendre , je suis descendu is equivalent to je me suis descendu. Si se descendirent et se firent leurs logis sur ces beaux pres sur la riviere de Dordogne Froissard II. . . 3 Et loerent au roy que il se descendit de la nef Ik ou il estoit Joinville . . . 342 II n’y avoit en la mer ilecques pres ancun port la oii il se peut descendre Joinville ... 28 With courir, je suis couru is equivalent to je me suis count. We have seen that Racine’s expression,/ 'y suis couru , has been condemned by grammarians. But M. Littre, in his Dictionary, has justified the great poet in the following man- ner : “ Les Grammairiens condamnent cet emploi et disent que courir exprimant une action ne peut recevoir l’auxiliaire etre. Mais venir exprime aussi une action et ne s’en con- jugue pas moins avec l’auxiliaire etre. Ici encore l’usage est pour l’auxiliaire avoir. L’auxiliaire etre est tres peu usite, mais il est egalement correct, dans l’ancienne langue il etait de plein usage.” Here M. Littre recognizes the fact that the auxiliary etre is as justifiable as the auxiliary avoir in that neuter verb, and that everything depends on usage, — rather, I should say, on the development of the language, as it is very clearly seen in the case of venir and coitrir. Conse- quently M. Littre ought not to defend, as he does everywhere in his Dictionary, the idea of grammarians who say that neuter verbs are conjugated with etre when they express a state or condition, and with avoir when they express an action. The rules of grammarians are rules of assertion and not of investigation, made for the present stage of the lan- guage without any reference to the past. Hence such rules are not always observed by great writers, nor by the people either. Racine was right in that particular instance, and, according to our own theory, j'y suis count stands for je my suis couru. Compare the Old French : — Roland regarde puis (se) lui est couru Roland . . . 153 Chascuns y est couru la merveille esgarder Berthe III. Ils se coururent sus l’espee au poing Montaigne I. . . 256 II s’en court en disant : A Diep me recommande . . . Regnier St. XI. Le pauvre homme s’en courut La Fontaine. 59 30 J. A. Fontaine , 4 With disparaitre , je snis disparu is equivalent to je me suis disparu. J. J. Rousseau’s La modestie du sexe est disparue pen a peu is equivalent to s est disparue peu d peu. Compare also Calvin : — Jesus Christ ne s’est point fait invisible mais seulement s’est disparu. Hence se disparaitre was used even in the Middle French, and the grammarians have condemned Rousseau wrongly, as they ought to have seen in his supposed mistake a remnant of the old style. With monter , je suis monte is equivalent to je me s?iis monte. Si vint tout ik pied Messire Herve jusqu’a Abbeville, la se monterent Froissard II. . . 214 With passer , je suis passe' is equivalent to je me suis passe. Outre s’en passe . . Rone 65 Si tost comme il le sot il 1 ’alla querre, il s’en passa sans amende Beau. XX. . . 57 With perir, je suis peri is equivalent to je me suis peri. Perir was used in the Old French in the active voice, as well as mourir. Compare : — Pour Dieu ne perissons mie la grant honeur que notre sire nous a faite Villehard. LXXXIX. Tellement qu’elles perissent tout ce qui se trouve en ce destroit Du Bellay IX. . 296 Sire, ces seigneurs qui ci sont arcevesques, evesques m’ont dit que je vous deisse que la crestiente se perit . Joinville . . . 200 Thus have I tried in the preceding pages to find instances in which verbs now used as neuter were formerly reflexive and therefore conjugated with the auxiliary etre. Such in- stances certainly suggest a more satisfactory solution of the question of the use of auxiliaries in neuter verbs than the one proposed by grammarians and M. Littre, especially by the latter when he says of perir , “avec l’auxiliaire Hre, perir exprime plus particulierement l’etat, mais cette nuance n’est 60 Use of Auxiliary Verbs in Romance Languages. 31 pas toujours observee.” It appears, on the contrary, that with the auxiliary etre, perir , as well as all the other verbs examined above, is of reflexive origin, and that with avoir they constitute a development towards active forms. All shades' of meaning attributed to neuter verbs in modern French, according as they are conjugated with avoir or with etre , may be assigned not only to a refinement of thought and of style, but also to the subtle distinctions of grammarians. In the above considerations I have taken account only of such neuter verbs as either retain etre exclusively, or take both etre and avoir , making no mention whatever of those conjugated with avoir alone. A couple of instances will show that those also were used reflexively. Dormir : — Charles se dort qu’il ne s’esveillet mie Rol 724 Pleurer : — Amarement mult se ploret Pass, du Christ . 198 II. A uxiliaries used with Intransitive Verbs in Italian. Since the Italian system of conjugating the neuter verbs presents about the same character as that of the French, I have thought proper in the preceding to bring such parallel instances of Italian as I might find directly under those of the French. This was done to illustrate the one language by the other, and to bring stronger proofs to my assertion that neuter verbs conjugated with esse in either of these lan- guages are merely remnants of reflexive verbs elliptically ex- pressed. I may add here that in Italian authors like Matteo Spinelli, Ricordamo Malespini, Dino Compagni, Dante, Boc- caccio, and Ariosto, this alternating use of elliptical reflexives with full reflexive verbs is more frequent than in any French author I have read. The Italian still uses the auxiliary essere with a few neuter verbs, where, the French now uses the auxiliary avoir. I have found two striking instances of such - difference in fuggire and vivere. 6 1 32 J. A. Fontaine , Sono fnggito is equivalent to mi sono faggito. Compare : — S’ erano fuggiti Hist. Fior. Bk. IV. 40 La o altrove si fosse fuggito Bocc. Dec. . . 2. 2 Ridendo gli contarono perche s’ eran fuggiti .... Bocc. Dec. . . 2. 5 Compare : — S’en estait fulz Guill. de Tyr. . 468 Lesquels estoient fuis Rab. Garg. . 41.276 Sono vissuto is equivalent to mi sono vissuto. Compare : — Numitore si vivette nel suo campo Ricord. Mai. . . 886 Gli nomini si vivono quietamente Machiavelli del Pr. III. La donna onestamente con lui si visse Bocc. Dec. . *3-9 Gia mi vivea di mia sorte felice Orlando F. . 13. 5 Non so, risposi lui, quant’ io mi viva Purg. XXIV. . 76 Compare Old French : — Par les pasture de quoi ils se vivent Guill. de Tyr. . 77 S’il n’a dont il se puisse vivre Roman de la Rose 19 Et se font povre et si se vivent Roman dela Rose 5 III. A uxiliaries used with Intransitive Verbs in Spanish . Just as we saw the Spanish using two verbs to indicate possession, so we shall find that it makes use of two different verbs to indicate existence : esse and stare. As to the rela- tion of esse to stare , I shall refer to page 16, where the rela- tion between the F rench imperfects ere and estoie was spoken of. The Latin stare furnished all the forms of the Spanish estar, but esse did not furnish those of ser in the same meas- ure. The infinite esse and the subjunctive sim have been lost in Spanish, and the corresponding forms, sedere and sedeam , have been taken from a verb whose transferred meaning denotes existence, just as stare does. In the oldest documents of the Spanish no traces of esse and sim can be found, and if those forms were used at all, it must have been at a period antedating those documents. 62 Use of Auxiliary Verbs in Romance Languages. 33 Compare : — Bien seya castigado Vida de Santa M. Eygp. Seya dixeron todos puesto e otorgado Devie seyer en vida tal ome adorado Libre d’App 91 Devemos seyer todos firmes en la sua tenencia . Libre d’App 93 It is, in fact, a question whether, in Old Spanish, the whole verb sedere Was not used in all its tenses to denote existence, for besides the infinitive sedere and the present subjunctive sedeam and also the gerund sedendo , I have found instances of the indicative present sedeo , and especially of the imperfect sedebam. De san Millan criado en la su merced seo .... Vida de St. Dom. . . 757 Sennores e amigos quantos aqui seemos Est. de St. Millan . 317 Mientre que esta duenna en tal caita sedie .... Vide d. S. D. d. S. . 562 La casa de Onorio que sedie escarnida Est. de S. Mill. . . 198 Como era mal apriso sedie fuert embargado . . . Est. de S. Mill. . . 333 Sedie una eglesia non mueho aredrada Milag. de N. S. . . 415 Sus gentes mui devotas sedien en oracion .... Milag. de N. S. . . 415 Quando esta palabra udieron los trufanos Que sedien mas rabioscs que carniceros canes . . Duelo de la Virg. . 39 Seyendo aun el rey en la batalla Chro. del Rey Alp. X. 52 Seven do estas companias llegadas Chro. del Rey Alp. X. 27 Hy sedie una mesa de cobre bien labrada .... Sacrif. de la Misa . 9- Now, if there can be no doubt that the Spanish ser is a contracted form of sedere , and that the future sere is for sedere habeo , could not the French serai be derived from the same source ; and ought not estrai to be considered as the only form coming from essere habeo f Compare page 1 6. After these few remarks about ser and estar , it will be shown how they were used in the oldest period of the Spanish language. It is recognized that in Spanish ser expresses what is essen- tial and permanent, whereas estar expresses an accidental and transitory state. Indeed, this distinction may be observed to have existed from the recorded beginning of the language. There may be some instances in which the difference is not observed ; but, I dare say, they are very few. 63 34 J« A. Fontaine , Ser. Estar. El Libro de los Reyes d'' Orient. Tu que major e mejor eres. Dios es sin dubdanza. Vida de Santa Maria Egipciaca. Fija tu eres de gran natura. Porque estas en mala natura. Di me donde eres. O como estas. El Libre d’ Appolonio . Tu eres la raiz, tu fija el cimal. Antiocho estando en tamanya error. Rev yo fui esse e fuy verdadero. Estaba en tal guisa. Mientre ellos estaban en esta encencia. Dixo el marinero que en somo estaba. Las ondas mas pagadas estar no podien. Desuyo le sangre que estaba enagada. No sabie do estaba. Poema d' Alexandro Magno. Tornal como se fusse su mortal Que escusa non ayas porque estas enemiga 2193 desarmado 123 Quando estas irado as fiera cata- dura 212 Que estaban lidiando a una gran pressura 500 El Poema del Cid. Yo so Ruiz Dias 721 Estando en la cruz 351 Tres reyes veo de moros derredor do mi estar. Cronica del Rey Alfonso X. Veyendo como era de tan poco e a los que estaban y con el . . X. poder II. a Toledo do estaba el Rey Alfonso. Entonces era vivo el rey don Jaimes VIII. Que era tuerto de un ojo . . . XIII. La Vida del Lazarillo de Tormes. Quien ere su padre. para los que estaban de parto. Era'huerfano. pensando que yo estaba enten- Ser la misma avaricia, diendo 14 64 Use of Auxiliary Verbs in Romance Languages. 35 Ser. Don Quixote. Estar. Que era hombre docto .... I. Aquel idolo de Mahoma que era todo de oro I. Hombre que por ser muy gordo era muy pacifico II. Estaba confuso mirando lo . . III. El estaba alii pronto paro obede- cerle. cuando estaba muy cansado . . V. Now let us consider ser and habere intransitive auxiliaries. The Spanish conjugated its neuter verbs or its elliptical re- flexives with ser and haber , mostly with ser , till the beginning of the thirteenth century, when haber began, little by little, to supersede ser , and at last, about the middle of the sixteenth century, completely displaced it. It was about the same period of time that haber ceased to be considered as a prin- cipal verb ; so that we may consider the beginning of the sixteenth century as the period when the tendencies of the Spanish language to use haber became exclusive. Ser. Haber. El Libro de los Reyes d' Orient. a Jesu Christo que era nado. Quando Erodes sopo. El angel fue a el venido. Que por hi non le han venido. Vida de Santa Maria Egipciaca. En Alexandria es venida. En tal hora hi fue entrada. Quando fue passada. A 1 monasterio son tornados. El Libre Ya es del siglo passado. porque era hi venido. Si entonces fuesse mortuo. n Appolonio. Porque por muchas tierras no avia an- dado. Avian de la marina gran partida an- dada. Quando toda la hove la ribera andada. Poema de Alexandro Magno. El infante fue venido .... 107 Avie tan fiera lluvia ante noche passada Alegre fue el rey quando fue ar- ribado 272 65 1883 3 6 J. A. Fontaine , Ser. Haber. El Poem a del Cid. Car por il agua a passado . . . 150 Cronica del Rey don Alfonso X. E soy aqui venida a pedirle ayuda XVII. Era venido a recebir .... XVIII. Despues que Don Nuno fue Dos caballeros hermanos que partido ....... 24 havian passado 63 Seyendo estas companias lle- Aquel camino que aveva ido antes 77 gadas 27 Supo de commo Aben Yuzaf era passado 62 Romancero del Cid. A tiempo eres venido. Cuatro veces he venido .... 16 Amigos salidos somos . . . . 119 Un romero habia llegado. Por su tiempo es pasado . . • 93 Y lides do habeis entrado . . . 46 Nuevas al Cid son venidas . . . 187 Venido han en perdicion . . . 46 Until now, neuter verbs conjugated with haber and ser have been found, but hereafter the auxiliary haber is univer- sally met in compound tenses of neuter verbs. Hurtado de Mendoza . El qual habia muerto en la batalla. 10 Mucha gente que le habia ido a socorrer 30 los dias que no habia muerto. de mi amo habia ido fuera del lugar 37 porque no podria menos de haber caido 47 Etc. Antonio de Solis. Creyendo que ... la voz que habia corrido III. Cuyo suceso habia llegado ya a stf VI. la brujula y carta que habian decaido XIV. 66 Use of Auxiliary Verbs in Romance Languages. 37 Ser. Haber. Don Quixote . No habia pasado de hidalgo . V. Esta noche me ha sucedido una de las mas XVII. El miedo que habia entrado en su corazon. he ya que hemos caido en sos- pecho XXIII. al cual ya habia venido . . XXIV. Las Mocedades del Cid. He venido. Ha venido? He venido. Ya he caido en tu pesar . . . 1574 To come back to my theory and prove that the above neuter verbs are merely elliptical reflexives, I shall give a few instances where the same verbs are used as full reflec- tives. For example: — Estos reyes son se tornados Quando desto te avras partido No pudo estar que non se iria El obispo don iheronimo se entrava Entraron se en la cibdad Y de secreto se ha ido Encendieron su fuego que se les era muerto De mi mal se parte Ya se iban las naves Si de nos non te partes avras mal ventura Que se iba para Burgos Commo do Fernando se venia Los moros que se iban con el Reyes d’Orient. Santa M. Eg. Santa M. Eg. iglesia Que de mi y de ellos se habio ido Para el gafe se venia. En terra se descendia . Mas en castillo se entraba A vos me vengo Se habia venido a recoger aquel su castello . Donde vieron que se habia entrado .... . Poema del Cid . . 1579 . Poema del Cid . . 2896 . Romancero del Cid . VII. . Libre d’Ap. . . . K/\ 00 . Alexandro .... 118 . Alexandro. . Alexandro .... 120 . Alf. X 8 . Alf. X 66 . Alf. X 74 la . Vida del Lazar. . . 55 La vida del Lazar. . 85 . Rom. del Cid . . . XXII. . Rom. del Cid . . . 3 i . Rom. del Cid. . Rom. del Cid. . Rom. del Cid . . . 15 67 38 J. A. Fontaine , IV. Auxiliaries used with Intransitive Verbs in Portuguese. It has been shown above that in Spanish ser expresses what is essential and permanent, whereas estar expresses an accidental and transitory state. The same thing is true of the corresponding auxiliaries in Portuguese, and a rapid glance at the earliest and most important written documents of the Portuguese language will be proof sufficient that such is the case. Ser. Estar. Canzoniere Portoghese. Si foss eu rey 37 Poys ante vos estou aqui . . . 141 mays de tato seede sabedor . . 1220 Har dung's Romanceiro Portuguez. Era esse dom Beltrao 6 como estas bem assentada. Pensando que era verdade . . .31 Chegando aonde elle estava . . 10 tu es uma mana minha .... 64 Zamora estava cercada .... 17 ja esta vain en mar largo. Antologia Portuguesa. E se era vos c’ant’o prazo saido . 15 Yo me estaba em Coimbra . . . 76 Melhor e de seer traedor ... 34 Despois de estar ja vestido . . 258 E em que sempre cuidando seyo . 43 Sd de Miranda. E este teu amigo he tao meren- Estas tao demudado 192 corio vil 144 Estava como fora de mim . . . 201 Os Lusiadas. O mouro astuto esta confuso . I. 62 O capitao que a tudo estava attento I. 98 Now ser and ter will be considered as intransitive auxil- iaries. The Portuguese auxiliaries follow the same development and undergo the same changes as their Spanish equivalent, and, as far as I can make out, at about the same period. 68 Use of Auxiliary Verbs in Romance Languages. 39 Ser. Canzoniere Portoghese. Por em mais en q mal dia fui nado 17 Por semp sodes de mi partido . 834 E se no fosse antexpo nado . . 1013 Hardung's Romanceiro . Chegadas sao as gall eras . . . 13 Saiban quantos sao nascidos . *5 Que meu pae que era morto . . 100 Mas anno e dia e passado . . . 113 Agora a saber son vindo ... 7 2 Sd de Miranda (Os Estrangeiros). A tanto sao chegados .... 78 Pois tu es vindo a salumento . . 118 ja tudo esso he passado a Por- tugal 147 Os Lusiadas. Sao chegadas . I. 78 meio caminho a noite tinha an dado II. 60 Sendo 0 capitao chegado . . I. 104 ja tinha vindo Henrique da conquista III. 27 A Melinde foi chegado . . . . . antes que chegado . II. 57 Chegado tinha 0 prazo promet- tido Ou que partes do mar corrido HI. 37 Seya este capitao . . . . I. 76 tinhan Do mar temos corrido e navi- gado toda a parte .... I. 50 I. 51 The Lusiads represent the period when ser and ter could be used equally well with neuter verbs. After that time ter must have gotten the upper hand very rapidly. I shall say nothing of the Provencal, since it agrees, so far as I know, with the French. Chapter III. — Auxiliaries used with Reflexive Verbs. The question concerning the use of auxiliaries in the com- pound tenses of the reflexive verbs has been extensively treated, but whether the best explanation has yet been found 69 40 J. A. Fontaine , is very doubtful. It would be too long a task to review fully, one after the other, the authors that have written on that subject. The question has been fully examined by Gessner ( Jahrbuck fur romanische und englische Sprache und Litera- tur , XV, p. 20 1 ) and by A. Mercier (. De V Histoire des parti- cipes fran^ais). Gessner is the one author who seems to have treated the debated question more fully than any other. The principal thing to be considered in this author’s work, and by far the most important (Mr. Gessner himself calling it the “Kernpunkt” of the whole question), is the assertion that the reflexive pronoun^ accompanying the verb is in the accu- sative, and yet not the direct object of the verb. Mr. Gess- ner tries to illustrate his theory for the German language by selecting as an instance two expressions, one reflexive and the other active, and showing that the former is more vivid than the latter, describing more intensively the actual situation or feeling of the subject. Who will deny that the reflexive pronoun adds a mild, poetical meaning to the verbal expres- sion ? But at the same time, Mr. Gessner tries to prove that logically the reflexive pronoun, though in the accusative, is not the direct object of the verb. Let us take two French sentences directly representing those given by Gessner in German : — II craint le danger ; II s' effr aye du danger. In both cases, he says, danger is the direct object of the verb ; accordingly the reflexive pronoun cannot be the direct object (“ dieser Accusativ kein Object ist ”). As far as the meaning is concerned, nothing is changed in adopting that view, and Mr. Gessner is right ; but if we take into consideration the syntactical connection, he is wrong. If I say simply, Cet homme s effr aye facilement , undoubtedly we have to consider the reflective se as being the direct object of eff'/ayer , just as the first personal pronoun me would be in Cet homme m ejfraye. Now if to our first sentence cet homme s' ejfraye, I add dti danger , will these few words change the relation of the reflex- 70 Use of Auxiliary Verbs in Romance Languages. 41 ive pronoun to its verb so completely as to prevent it from being any longer the direct object of the verb ? I do not think so. The reflexive pronoun will remain in the accusa- tive case and be the direct object of the verb. The words du danger merely explain why the subject is under the control of fear, and could be replaced by en face du danger or a cause du danger. This way of looking at the logical construction of words seems to me the natural one. Mr. Gessner is too obscure when he pretends that in expressions like the above the re- flexive pronoun is in the accusative case and yet not the object of the verb, and that this accusative adds to the reflex- ive idea an interior intenseness, and has a tendency to de- prive the verbal notion of all exterior activity, reducing it to an especially interior expression. M. Littre, I think, was on the way towards a more plausible explanation of the difficulty, when he said in his Histoire de la langue franqaise, II, 317, “ Se erant convers, of the Frag- ment de Valenciennes, presupposes the low Latin sc erant conversi." The explanation he gives of the connection be- tween convertor passive and convertor reflexive is good, but he fails to see that the pronouns me , te , se, etc., had been added in the low Latin to the perfect of convertor through an ana- lytic tendency, and in analogy with convertor reduced to me converto in Romance, and he concludes in saying that se is not an accusative case, but that it represents all the cases with the exception of the nominative, being “a regime inde- termine sans cas determine and that as such “Sea pu se joindre a des verbes neiUres, tels que s' en aller , s' enfuir, se taire , s' eerier." M. Chabanneau, in his Histoire de la conjugaison franqaise, considers the auxiliaries as mere inflectional endings of the verbs in their compound tenses. Whether that inflectional ending be the auxiliary etre or avoir, it does not change the nature of the relation of the verb to its object. So M. Chabanneau recognizes in the reflexive pronoun not only an accusative case, but also a direct object of the verb ; for he 7 l 42 J. A. Fontaine , says, “ Dans je me suis frappe, par exemple, me est le com- plement de suis frappe, comme il le serait de ai fj'appe dans la phrase supposee plus correcte je m’ai frappe.” We agree thoroughly with M. Chabanneau, and indeed a rigorous pars- ing of the sentence cannot be made to yield to the reflexive pronoun me or se any other office (compare the “direct object ” of the Latin deponent verb). It seems to me that M. Chabanneau is very near solving the question when he says (p. 5 of the work quoted above) : “A l’epoque du haut moyen age tous les verbes deponents (du latin) suivaient dans la langue parlee la conjugaison active, du moins quant a leurs temps simples, car leurs temps composes etaient trop d’accord avec les tendances des langues nouvelles qui se for- maient pour ne pas etre maintenus, et nous les retrouvons parfaitement conserves dans la conjugaison de nos verbes reflechis et d’un grand nombre de nos verbes intransitifs.” Although I prefer the theory of M. Chabanneau to that of Mr. Gessner, as being more complete and giving a safer clue to the question, I by no means intend to depreciate that of Mr. Gessner. The poetical meaning, the emphatic expres- siveness added to the verb by the use of the reflexive pronoun se cannot be doubted, and this is felt by every one who is well acquainted with the French language ; for it is very easy to notice that between two expressions, the one active and the other reflexive, the French generally chooses the latter, be- cause there is in it something agreeing better with the genius of the language. But all this is far from explaining to us the difficulty presented by the reflexive verb system of the Romance languages, especially when we take into account the different tendency of some of these languages to use -the auxiliary esse rather than the auxiliary habere , and that with either one of these auxiliaries they could convey their reflex- ive meaning, sometimes making use of the poetical reflexive pronoun and sometimes leaving it out, without causing the verb to undergo any change in its meaning, probably also without its losing any of its poetical coloring. This is espe- cially striking in the first centuries of the development of the 72 Use of Auxiliary Verbs in Romance Languages. 43 Romance languages, where very often one is at a loss to find any difference in the meaning of a verb when accompanied by a reflexive pronoun and when without it. Wishing now, after these observations, to discover why the reflexive pro- noun is sometimes omitted and sometimes expressed, and the reasons why some languages have adopted the auxiliary esse and some others the auxiliary habere , I must present, to begin with, a few considerations on the fat$ of the Latin verb itself. On considering what has become of the deponent and neuter verbs of the Latin in coming down to the Romance languages, and how they have been used, especially in the old stage of these languages as purely transitive verbs (compare se mourir , moui'ir quelqu un, imiter quelqu! un, etudier quelque chose, perir, se perir , perir quelque chose , se naitre , se venir ), it may be assumed that all the Latin verbs that have survived in the Romance languages have been handed down in an active form, having an active meaning and capable of expressing a transitive action. Among these verbs there was a certain number that contained in themselves a reflexive idea, that is, that the subject was doing the action for its own seif. Such, for instance, are *morio , ambulo , venio , vado , vivo , descendo , ascendo , etc. The idea expressed by these verbs might be rendered by morio me, ambulo me, venio me, vado me, vivo me, etc. We know that a strong analytical tendency pre- sided over the formative period of the verbal system in the Romance languages, and it was, no doubt, this tendency that caused them to give to Latih verbs that were intransitive in their synthetical state, complements, just as other comple- ments were given to other verbs more intransitive in their outward appearance. But since the verbs of the first cate- gory, on account of the very essence of their internal mean- ing (which meaning was arrived at by the disintegrating- or analytic genius of the new languages) were left, by general consensus undoubtedly, to express an action especially for che benefit of the subject itself, without going outside of it ; so that no other complement could be given them but a reflexive complement expressed by the pronouns me, te, se. 73 44 J. A. Fontaine , Hence we have the name of reflexive verbs given to me morio , me vivo , me vado, me descendo, me venio, etc., which verbs have been used in a reflexive form in all the Romance languages, and are the legitimate growth and legitimate rep- resentatives according to the analytical tendency of the Latin verbs vivo , venio , morior , etc. Grammarians say that j' ai ve'cu, je suis mort,je suis allege suis venn , je suis descendu , etc., are neuter verbs. Appar- ently they are ; but before these forms arose we had je me suis vecu,je me suis mort, je me suis alle, je me suis venu , je me suis descendu :, etc. ; and these verbs are nothing but re- flexive verbs. When the question of the conjugation of such verbs as me vivo , me morio , me venio arose, or, to speak more plainly, when these verbs came to be used in all their tenses, it was very easy to conjugate them in their simple tenses. But in compound tenses (and here let us remember what M. Chabanneau said in his Flistoire , p. 5) the difficulty was greatly increased. There must have been, at the time of the formative period, two tendencies working in the Romance lan- guages : one, the analytical, resolving the compound tenses into periphrastical by combination of the auxiliary habere and of the past participle of the conjugated verb. The indicative present being je me vais, je me descends , je me pars , the com- pound tenses of the preterite regularly became je m ai alljje mai descendu , je m ai parti. This is the most natural ex- planation, and this accounts for the well-known fact that a great number of French dialects still use the auxiliary avoir in combination with compound tenses of reflexive verbs, and that children and uneducated people do the very same thing. It would be of great importance to know the relative use of esse and habere ' in the different Romance dialects. Very likely in all of them instances of the use of both auxiliaries are to be found, with this difference, that they are more or less abundant, according to each dialect. M. Chabanneau has mentioned the fact that several dialects of France make use of avoir a,s well as etre. I may say,, that in the Parler Sancerrois , and in the Berry generally, this use is very com- 74 Use of Auxiliary Verbs in Roma?ice Languages. 45 mon. For instance, it would not be rare to hear some one say, Je m avais trompe, il s avait sauve \ je m avais plaint , il sa bien donne de la peine , etc. The use of avoir is extended also to neuter verbs, like il a tombe \ etc. Compare also the Franco -Venetian in Romania, XIV., 177. Avec lui m’avero corucer . . . . Que a lui plu s’avoit aprosmer . . Ni an Milon no se soit consoler Dapois que de Fran?a m’avi sevrer Mai vero l’ore q’i s’en aura sevrer . Mere fait il porqe vos ert envier Berta e Milone • 251 Berta e Milone • 274 Berta e Milone Orlandino . . • 399 Orlandino . . • 234 Orlandino . . . 232 But besides that tendency of analysis on the part of modern languages, there must have been another, not less powerful, which may be termed the Latinistic, by which is meant the tendency to conjugate deponent or reflexive verbs in accord- ance with the system prevailing in the Latin language. Par- tior had given in the preterite under the influence of the ana- lytical tendency, je m ai parti ; the same verb will give under the Latin influence, je suis parti ( partitzts sum). The table on pp. 46 and 47 will illustrate the effect of the two tendencies. This table shows us the analytical or active tendency of the modern languages, which tendency, had it not been checked by another, would have caused that all the com- pound tenses of Romance reflexive and neuter verbs should be conjugated with habere , just as they are at present in Spanish and Portuguese. These two languages have carried out thoroughly the analytical tendency, though not without at first yielding to the syntactical influence of the Latin verb- system. Thus we see that the analytical and conservative tendencies have been working side by side, the one overrul- ing the other in different languages ; and as a result of that struggle we have a double system of conjugating reflexive verbs in the Romance languages, just as the struggle between the strong Latin conjugation and the weak Romance conju- gation resulted in the development of two classes of verbs. Which one of these tendencies was the stronger at the begin- ning is difficult to say ; but I am inclined to think that the 75 46 J. A. Fontaine , Latin tendency was the stronger, at least in French and Italian. In the Fragment de Valenciennes we find etre alone Latin. Romance Analysis. Latin. .tnorior morio me mortuus sum French Italian Spanish je me meurs (or je meurs) mi muoio (or muoio) me muero (or muero) Portuguese . me morro (or morro) parti or partio me pa rtitus sum French . Italian . Spanish . je me pars (or je pars) mi parto (or parto) me parto (or parto) Portuguese . me parto (or parto) venio , *venior Cf. venitur venio me *venitus sum or *venutus sum French . •Italian . Spanish . je me viens (or je viens) mi vengo (or vengo) me vengo (or vengo) Portuguese . me venho (or venho) intro, *intror intro me *intratus sum French . Italian . . Spanish . . Portuguese . je m’ entre (or j’entre) mi entro (or entro) me entro (or entro) me entro (or entro) used. In the Alexis and Chanson de Roland we find avoir only three times, but etre very often. It is very difficult to say how long avoir continued to be used in the conjugation 76 Use of Auxiliary Verbs in Romance Languages. 4 7 of reflexive verbs in French proper. The latest instances I have found occur in the Roman de la Rose and Villehardouin ; Romance Analogy. Full Reflexive Verbs. Latin Influence. Elliptical Reflex. V’bs. Romance Creation, or Active Tendency. Reduced Expression. me morluus sum morluus sum me mortuum habeo mortuum habeo je me suis mort je suis mort mi sono morto sono morto * me soy muerto soy muerto me he muerto he muejrto me sou morto sou morto me tenho morto tenho morto me partitus sum partitus sum me partitum habeo partitum habeo je me suis parti je suis parti mi sono partito sono partito me soy partido soy partido me he partido he partido me sou partido sou partido me tenho partido tenho partido me venitus sum or me venutus sum venitus sum or venutus sum me venitum habeo venitum habeo je me suis venu je suis venu mi sono venuto sono venuto me soy venido soy venido me he venido he venido me sou vindo sou vindo me tenho vindo tenho vindo me intratus sum intratus sum me intratum habeo intratum habeo je me suis entre je suis entre mi sono entrato sono entrato me soy entrado soy entrado me he entrado he entrado me sou entrado sou entrado me tenho entrado tenho entrado but I dare say that avoir was used with reflexive verbs in French about as long as ser was used in Spanish with the same reflexive verbs. We shall see further on when the 77 48 J. A. Fontaine , latest traces of the Spanish reflexive verb conjugated with ser are to be found. We see, then, that as long as the French and Spanish languages can be compared, the opposite pro- cess of development in this particular took place. Thus, while the Latin or conservative influence was predominating in the French, the modern tendency was predominating in Spanish. The explanation of this is plain enough. We know that the southwest languages of the Romance family (the Spanish and Portuguese) have a development totally inde- pendent of that of the northeast group (the French and Ital- ian). One of these characteristics is found in the way these different languages have treated their verbal system. Where- as, on the one hand, the French, and yet more the Italian, have striven against the influence of analogy to keep alive the strong Latin conjugation, whether by retaining original Latin strong verbs, or by making weak Latin verbs strong, the Spanish and Portuguese have, on the other hand, transferred, we may say, the whole of the Latin strong conjugation to the weak conjugation, thus yielding to the unifying power of anal- ogy. Thus one may see how independently each language or each group of languages develops. And we must not wonder that the Spanish and Portuguese use haber and ter with their reflexive verbs, and the French and Italian esse, Even in these two languages the development of the use of esse was not totally accordant. From the very beginning the French made no difference whether the reflexive pronoun that accom- panied the verb was in the accusative case or dative ; every- where etre was made use of. The Italian, on the contrary, used at first avere whenever the reflexive pronoun was in a dative case, and essere when it was in the accusative case ; but later on avere yielded to essere in the dative case also. I have already mentioned above the great freedom of the Romance languages in older times to express or to leave out the reflexive pronoun without altering the meaning of the verb. How shall we account for that peculiarity, which is common to all the Romance languages ? Here, again, I attribute it to two tendencies : the conservative, which was 78 Use of Auxiliary Verbs in Romance Languages. 49 to represent the compound tenses with sum (as in Latin) and the past participle, without adding to it any reflexive pro- noun ; and the analytical tendency, which was to apply to compound tenses the analysis made of simple tenses. The former tendency had given : — morior ( morio me), mortuus sum, partior (partio me), partitus sum; the latter tendency gave : — morior, morio me, (me) mortuus sum, partior, partio me, (me) partitus sum, thus carrying the use of reflexive pronouns from simple tenses to compound tenses. Hence there was a conflict between these two tendencies, the one leaving out the ana- logical reflexive pronoun of the compound tenses, while the other had introduced it. But, of course, the meaning was perfectly preserved, and remained the same in both cases. In this way could be explained the apparent inconsistencies of the reflexive Romance conjugation. The Latin tendency, which was weaker than the analytical, has gained the upper hand in one class of verbs, called above “elliptical reflexive verbs,” and consequently the pronoun has been omitted : Je me suis parti , je me suis alle, je me suis venu , have settled definitely into je suis parti , je suis alle , je suis venu. But the tendency that had pushed the Spanish language towards adopting the auxiliary haber for the conjugation of all the neuter verbs was also working in French, and our first class of verbs having adopted, through Latin influence, etre, the second class of neuter verbs adopted the auxiliary avoir. In yet a third class of verbs neither tendency prevailed, and to this very day they make use of the two auxiliaries. Compare : — Je suis monte 01* fai monte ; Je suis descendu or fai descendu; Je suis reste or fai reste ; Je suis passe or fai passe, etc. Grammarians have decided that etre must be employed when the verb marks state or condition, and avoir when it 79 ' 50 J. A. Fontaine , indicates an action, and generally our feeling is influenced by this distinction ; but I have tried to show that such was not the case in the Old French period. It has already been assumed, in the first part of this trea- tise, that every verb is active, and consequently all the verbs that have come down to the Romance languages from the Latin mother tongue must have been verbs belonging to the active voice. To explain this phenomenon we must go back to the early period of the Latin, just as to explain the Italian forms struggere and traggo we have to go back to the period when- the classical forms struere and traho were represented by the more primitive forms strugere , trag(Ji)o. The g of these latter forms is still preserved in the sigma perfects, traxi, struxi standing for trag-si, striLg-si. By going back to the formative period of the Latin, we shall see that the active voice was the only voice this language then possessed. It has been said (cf. Die Verbal Flexion der lateinischen Sprache of Westphal and others) that the Greek constructed its middle and passive voices (the aorist and future excepted) with the same inflexional endings, jjlcu, crcu, tcu, and these endings were originally, as the comparison with other lan- guages proves, endings of the middle voice, which afterwards, through a transfer of meaning, were used as passive endings. The Latin verb must have once possessed endings similar to the fiat, (tcu, tcu of the Greek, and with a parallel meaning ; but they were afterwards lost, a periphrastic expression being introduced in its stead. That periphrastic expression was made up of an. active form, eg. amo, and the reflexive se. But the Latin did not keep this new formation in its primi- tive state, and a fusion of the periphrastic expression into a single word took place, and amo se became amor{e), amas-se , amar(i)s, etc. Just as the endings fiat, acu, tcu had served as inflexional endings to the middle and passive voices, so the parallel Latin endings formed its passive and deponent forms. Thus four periods might be distinguished in the verbal growth of the Latin, viz. : The first period, when the Latin had forms similar to that of the Greek in /jlcu, ecu, Tat ; the second 80 Use of Auxiliary Verbs in Romance Languages. 51 period , when the active voice became middle by the independ- ent use of se (eg. amo-se) ; the third period, when the verb and se combined into a single word (aino + se > amo'd), and the reflexive meaning was changed to a passive. Such trans- fer of a reflexive to a passive meaning was the regular devel - opment of human thought, which with a reflexive idea necessa- rily and logically associates a passive idea ; so that in the classical period of the Latin language the same endings had to express a reflexive and a passive verbal notion. The new period having been created by a tendency towards a passive notion or idea, it is natural that the passive verbs should have been more developed in the classical Latin than the reflexive verbs ; but.^ at the same time the tendency was not so strong as to push all the more ancient reflexive verbs into passives. Hence we have a certain number of them reserved by classi- cal authors to a reflexive use exclusively ; but even these con- tained in themselves the power of being used passively as well as the others, and that power had been given them by the general tendency that had pushed Latin verbs from reflex- ives to passives. This seems the most plausible way to explain how deponent verbs like adhortari , admirari , consolari, dilargiri , meditari , partiri , sortiri , reserved by classical writers to deponent use, are found in the passive voice, especially in the past participle (compare Livy : — Partitis divenditisque reliquiis XXI. 21 Ex malignitate praedae partitae V. 20 Compare also Draeger, Historiche Syntax, I., p. 156) ; and how passive verbs retained the deponent meaning and were used as deponent verbs (compare Livy : — Sed ruinae' maxi^ne modo jumenta cum oneribus devolvebantur . . . XXI. 33 Ut idem in singulos annos orbis volveretur III. 10 Priusquam hostes moverentur XXXVII. 18). The fourth period, when the passive meaning was given up through popular influence, and the original reflexive mean- ing was restored by means of decomposition resulting in the independent expression of the verb-form and the reflexive pronoun. This last change, beginning at a time when the 81 52 J. A. Fontaine , Latin language had no longer power to restrain the inde- pendent growth of the popular idioms, was carried out by the Romance languages. That decomposition was so thorough that abstract notions and inanimate things that can only express a passive idea, were clad with active forms in such a way that they seem to be the agents of the verbal notion. Compare the French expressions : Ce livre se trouve sur la table ; ce bois se fend difficilement ; cette maison se bdtit lente- ment ; cette terre se desseche ; cette expression s emploie ; cette chose se dit ; ce pays se mine , etc., etc. Why the Latin abandoned its middle endings would be difficult to say, since we have lost nearly all traces of the languages that surrounded the Latin in its formative period, but that transformation is no more surprising than a great number of other linguistic phenomena. Nor is the use of se for all the persons of passive verbs in Latin inexplainable. The idea contained in the word se (one’s self) is of an inde- finite character, and hence se can be connected with any person in any number. This same phenomenon is to be found in Scandinavian dialects, as well as among Romance- speaking people of a certain portion of Switzerland, the Rhato-romonsch. Compare Grammatica elementara dil Inn- gatg Rhato-romonsch , scritta da J. A. Btihler, p. 64 : “ II pronom 1 se ’ ei en tuttas formas, persunas, modas tuts temps ligiaus vid il verb . . . ridicul ei de voler declinar quei pronom ‘ se ’ a la moda italiana e franzosa, sco p. t.j'eu mi fidel , tu te fidas , la flexiun de quei pronom ei en el lungatg romonsch buca veguida cultivada e nos ureglia romonscha sa buca vertir pei quei Italianismus.” Thus we see that in this particular language the reflexive verb takes the reflexive pronoun se in all the persons, in both numbers. The “ se” is now yielding to me, te, se, and that under the influence of the Italian and of the French. Consequently the reflexive verbs of the Romance languages are nothing but the legitimate representatives of the first Latin type amo-se, with the difference that they have under- gone from the very beginning the change which the Ro- 82 Use of Auxiliary Verbs in Romance Languages. 53 monsch is undergoing now. Their analytic tendency caused them to substitute for se alone the different pronouns me , te, se, nos , vos. After these considerations let us review the history of the reflexive conjugation in the different Romance languages. This history has been extensively treated by Gessner and others in the works quoted above, for French especially ; and as far as the last language is concerned, we have but' very little to add. I. Auxiliaries with Reflexive Verbs in French , without Direct Object. It has been already stated in the foregoing part of this article, that French, making no difference whether the pro- noun was in the accusative or in the dative case, used the auxiliary etre in conjugating its reflexive verbs. As an excep- tion to the above rule is found a certain number of compound tenses of reflexive verbs conjugated with the auxiliary avoir. A list of s'uch instances is found in Chabanneau’s and Gess- ner’s works, and in some others. Two instances nowhere found quoted are given here, since it is desirable that such a list should be as complete as possible : S’ai moi dedens Varchiere mis (Roman de la Rose, 22. 616); E quand chascun s'ot a sa terre assene (Villeh. Conq. de C. 100). It is very probable that avoir has not been Used in compound) tenses of French reflexive verbs since the beginning of the fourteenth century. II. Auxiliaries with Reflexive Verbs in French , with Direct Object. M. Littre, in his Histoire de la langue fran^aise , p. 321, says that “je me suis coupe le doigt” ought to be considered as a solecism, and that the correct expression would be u je-mai coupe le doigt But since we have considered the reflexive pronoun in such cases as je me suis coupe as being the direct object of the verb, we shall easily understand that here; the 83 54 J. A. Fontaine , personal pronoun me , being accompanied by a more explica- tive and descriptive object upon which finally the action expressed by the verb falls, yields its place to it and assumes a dative case. When we say je me suis coupe , me receives entirely the action expressed by the verb ; but when we specify more by adding le doigt , the pronoun receives the action only indirectly, and consequently assumes the place of ‘an indirect object. It is, as we see, a mere change of case, which does not entirely break the relation of the reflexive pronoun with the verb, but still leaves £he latter under the indirect influence of the former ; and hence there is not a sufficient ground to allow the verb to change its auxiliary. Indeed, in such cases the Italian, as a rule, used to change its auxiliary and take avere ; but this fact only proves that the tendency towards a single type of conjugation, when a reflex- ive pronoun accompanied the verb, had a more powerful sway in French than in Italian. But if M. Littre’s assertions that the French had sacrificed the rule of the grammar for the sake of euphony, and had been imposed upon with -a solecism, were correct, the same thing could be said of the Italian of to- day, since the latter follows the same rules as the French in the reflexive conjugation. This will be clearly shown later on. Speaking of the supposed anomalous expression ,“je me suis coupe le doigt” M. Littre says : “ Je ne sais si elle est ancienne, je suis porte a croire que non, mais je n’ai la-dessus aucun renseignement.” I have tried to collect a certain amount of material to prove that it was ancient, and as a result find plenty of reflexive verbs with direct object in Livre de Job, Sermons St. Bernard, Livre des Rois, Guillaume de Tyr, Join- ville, Villehardouin, but none in compound tenses. But there is no reason to suppose that the Old French, taking a direct object in the single tenses of reflexive verbs, should have avoided it in compound tenses. However, we find in Rabe- lais and Montaigne so many instances of reflexive verbs accompanied by a direct object in compound tenses, that it is impossible not to think that the same thing had been done a long time before them. 84 Use of Auxiliary 'Verbs in Romance Languages. 55 Compare : — II se frottait ordinairement le ventre d'un panier (Rab. 123); Elle s'etait fait aucun mal (Rab. 396); S’est rompu le coul (Rab. 74 2 ); S’etre avec une plume tire un ciron de la main (Rab. 117 2 ); S’etant un peu frotte le front et secoue les oreilles (Rab. 294 s ) ; Nous estant been a point saboures I’estomac, etc., etc. III. Auxiliaries with Reflexive Verbs in Italian. In the most ancient writers of the Italian language, Matteo Spinelli, Ricordamo Malespini, Dino Compagni, and Dante, I do not find a single exception to the rules of the accusative and dative cases. That is to say, reflexive verbs accompanied by a personal pronoun in the dative case take the auxiliary avere , and those accompanied by the personal pronoun in the accusative case take the auxiliary essere. Matteo Spinelli. Dative Case: — Ma che imprestassero alio re chilli denari che se avevano portati per le spese (1101). — Che s’ avesse fatta la tassa delle spese a se et a soi famigli (1069). — E cha isso se 1 ’ havia recuperata per viva forza da maiio di due Papi (1087). — Per che isso se 1 ’ haveria tenuto come a figlio (1087). — Lo re disse cha non volea fare perdere la ventura a quella zitella que per la belleza si se P havia procacciata (1095). Accusative Case: — ; Che se ne erano fuggiti in Schiavonia (1079). — Che Napoli si era arrenduta (1085). — Li frati della zitella se ne sariano contentati (1093). — Dove s’ era ritirato lo comte di Tricario (1107). — Recuperando quelle terre che s’erano ribellate (1107). Ricordamo Malespini. Accusative Case : — II centurione si era dilungato (893). — La quale cittk non s’ era rifatta (907). — I quali s’ erano recati in contado de Firenze (908). — E di poco s’ erano levati d’ uno poggio (931). — Allora s’ era retta la citta sotto signoria de’ consoli (942). — S’ erano posti ad assedio al castello di Capraja (970). — II suo fratello bastardo Manfredi se n’ era fatto vicario (976). — Manfredi s’ era coronato re di Ci cilia (978). — S’ erano messi ad assalire tutta 1’ oste dei Fiorentini (989). — Molta buona gente del regno di Sicilia s’ erano partiti (1024% — II re Carlo no s’ era voluto imparentare con lei (1024). 85 56 J. A. Fontaine , Cronaca di Dino Compagni. Accusative Case: Cavalieri novelli s’ erano fatti (473); Niuno se sarebbe campato (503); Ma poi che i Bianchi si furono partiti (517); Reggio e Modena s’ erano rubellate (520). Dante. Dative Case : — Quando s’ebbe scoperta la gran bocca Inf. XII. 79 Dali’ altra gia m’ avea lasciata Setta . . XXVI. 1 1 1 Tre Frison s’ averian dato mal vanto XXXI. 64 Gualandi con Sismondi e con Lanfranchi, S’ avea messi dinanzi dalla fronte XXXIII. 33 Rivolsersi alia luce che promessa Tanto s’ avea Parad. VIII. 44 Dell’ anime, che Dio s’ ha fatte amiche XXV. 90 Dell’ eterno Valor poscia que tanti Speculi fatti s’ ha XXIX. 144 Accusative Case : — Considereremo come gli uomini molto onorati si siano da esse loro proprie partiti (Della Vol. El. 750). — Alcuni si sono partiti dal suo proprio materno parlare (752). Ed io, che del color mi fui accorto Inf. IV. 15 E poi ch’ alia man destra si fu volto IX. 132 Vedi \h. Farinata che s’ e dritto X. 32 Credo que s’ era inginocchion levata X. 54 Restato m’ era X. 74 E quel medesmo, che si fue accorto XIV. 49 Gittato mi sarei tra lor disotto XVI 47 Ma perch’ io mi sarei bruciato e cotto XVI. 49 Che questa, per la quale io mi son mosso Purg. I. 63 Ma s’ io fossi fuggito ni v£r la Mira V. 79 L’ ombra, che s’ era al giudice raccolta VIII. 109 Con Beatrice m’ era suso in cielo Cotanto gloriosamente accolto Parad. XI. 13 E ch’ io non m’ era li rivolto a quelli XIV. 135 Boccaccio. Dative Case : — Li quali come vestiti s’ ebbe a suo dosso fatti parevano (Decam. 2. 2). — Par- lando s’ arebbe vitupero recato (3. 2). — Hommi posto in cuore di fargliele alcuna (3. 3). — S’ avea posto in cuore di non lasciarla mai (3. 7). — In tanto che pa- rente ne amico lasciato s’ avea che (4. 10). — Avendo si prima tirato il cappuccio (6. 10) — Per venire a costui che non pensa cui egli s’ ha menata a casa (7. 2). — Avendosel tirato un poco innanzi (7. 5). — Egli s’ avesse molto messo il cap 86 Use of Auxiliary Verbs in Romance Languages. 57 pucio innanzi (7. 5). — Messer lo geloso s’ avea messe alcune petruzze in bocca (7.5). — Ma tuttavia che egli s’ abbia di me detto (7. 8). — Avendo se adunque questa promession fatta (7. 10). — Per non poterti vedere t’ avresti cavati gli occhi (8. 6). — Avendo si P* anello di lei messo in bocca (10.9). — I panni che spogliati s’ avea (10. 10). — Que gli che la mi diede se 1 ’ ha ritolta (La Fiam- metta, V.). Accusative Case : — D’ ogni cosa opportuna a dovere . . . fornito s’ era (Decam. 3. 5). — E poi che egli in diverse maniere si fu molto ingegnato (3. 7). — S’ era per paura gittato nel canale (4. 2). — Fuggita si sarebbe del padre e venuto se ne a Gerbino (4. 4). — Alcuna posta vicina al cuore gli s’ era rotta (4.6). — Della sua novella s’ era deliberata (6. 9). — Tu ti sei ingannato di dimostrarmi (1. 2). — Che riguardasse si partito si fosse (1. 7). — Altrui s’ e di beffare ingegnato. La altrove si fosse fuggito (2. 2). — Si io mi fossi di cio accorto (2. 3).-r-Che maggiore non si saria potuta portare (2. 8). — Io non avrei mai creduto che . . . ti fossi guardato (2. 8). — E quando ella si sarebbe voluta dormire (3. 4). Two exceptions to the rule of the accusative case are found in Boccaccio: — Che alia gelosia tu t’ hai lasciato accecare (7. 5) ; Poiche la donna s’ ebbe as- sai fatta pregare. * Machiavelli. Dative Case : — Trovando si ingannati . . . di quel futuro bene que si avevano presupposto (II Principe, III.). — I gentiluomini Romani si aveva guadagnato (VII. 37). — E tanto erano validi i fondamenti che in si poco tempo si aveva fatti (VII.). — Li gradi della milizia quali ... s’ aveva guadagnato (VIII.). — Onde avendo si creato odio (XIX.). — Tutta la gloria che si avevano' nel principio acquistata (Isto- rie F. II. 281). — Le quale per la morte di Messer Niccolo si avevano acquistate (II.). — Parendo gli aversi tirato addosso troppo importante nimico (VI. 229). — Cosimo avendo si alia sua potenza la publica et la privata via aperta (VII. 239). — Quella citta che non s’ avevano saputa conservare (VII. 276). — Pensando di godersi . . . quello stato che s’ avevano stabilito (VII. 293). Accusative Case : — E quelli che se li erano gittati in grembo (II Principe, III. 17). — La virtu dell’ animo loro si saria spenta (IV. 28). — Si non si fusse lasciato ingannare da Cesare Borgia (VIII.). — Si era del regno di Napoli insignorito (Istorie F. II. 281). — Mold ghibellini che si erano con loro accostati (II.). — E se da Francesco s’ era avuto poco (VII. 263). Just as were found in Boccaccio exceptions to the rule given about the use of the auxiliary when the pronoun was in the accusative case, so in Machiavelli was found one ex- 87 58 J. A. Fontaine , ception to the rule given when the pronoun was in the dative. Here, for the first time, I met with esse in the com- pound tense of a reflexive verb preceded by a pronoun in the dative case : — Per non gli potere satisfare in quel modo que si erano presupposto (II Principe, II.). Dative Case : — • Ariosto. Che piede o braccio s’ abbia rotto e smosso 1-59 E piu volte s’ averan rotta la fronte 2. 66 Polinesso che gih s’ avea proposto Di far Ginevra al suo amator nemica 5-36 E composto fra te t’ hai queste cose 5-39 Che per dolor s’ avea dato la morte 5. 61 Che s’ avea per non esser conosciuto Cambiati i panni e nascose le chiome 5. 65 Che a difender Ginevra s’ avea tolto 5-77 Dove li dui guerrier dato e risposto molto s’ aveano 5. 80 (Compare <_n 00 O 7-6 4; c )• 73 ; 10. 6 ; IO. 99-; 12. 61 ; 15 - 30; i 5 - 29 ; 16. 60 ; 16. 79 ; 17. no; 18. II 7; 20. 10; 2 J. 2 ; 23- IO; 23 - 53 ; 24. 21 ; 25- 51 ; 25 - 85; 26. 91 ; 27. 71 ; 27. 132 ; 28. 36 ; 28. 65 to vp. 00 29. 37 ; 31. 41 ; 3 i* 79; 3 i- 88 ; 32 . 30; 32 . 40; 32. 64; 32. 59; 32. 72 ; 32 . 79 ; 33 - 43 ; 33 - 71 ; 34 - 41 ; 37 - 35 ; 37 - 40; 38. 78; 39 - 57 ; 40. 21 ; 40. 56 ; 40. 00 41. 41 ; 43 - 85; 44. 74; 45 - 69 ; 41. 49 -) Accusative Case : — Piu volte s’ eran gih non pur veduti . Che s’ era in mar sommerso Ariodante . Ed armato con lui s’ era condutto . D’ all’ onde Idaspe udita si saria . . E perche molto dilungata s’ era . . Che s’ erano serbati in quegli affanni Ma Farrau che sin qui mai non s’ era Col re Marsilio suo troppo disgiunto Che ’1 pagan s’ era tratto in quella parte A caminar se gli era messo a lato Malgrado di Christian rimesso s’ era. I. 16 5-57 • 5 - 9 i 7 - 36 8. 32 14. 101 , 16. 71 , 16. 89 18. 30 (Compare 18. 76 ; 18. 91 ; 18. 103 ; 21. 64 ; 23. 17 ; 23. 50 ; 26. 108; 27. 56; 27. 75; 27. 1 12; 27. 1 15; 27. 137; 29. 58 ; 30. 80 ; 33. 65 ; 34. 41 ; 36. 54 ; 37. 4 7 \ 38. 72 ; 40 - 62 ; 88 Use of Auxiliary Verbs in Romance Languages. 59 40. 71; 48. 8; 42. 39; 43. 125; 43. 187; 44. 89; 45. 8; 46. 26 ; 46. 50 ; 46. 63 ; 46. 77 ; 46. 108 ; 46. 120.) I found also in Ariosto two exceptions to the rule for the accusative, the first of which is undoubtedly due to the rhyme : — Non cosi strettamente edera preme Pianta ove intorno abbarbicata s’ abbia ; Come si stringon li due amanti insieme; Cogliendo dello spirto in sulle labbia ; Suave fior 7. 29 E poi che ’1 Sol s’ ebbe nel mar rinchiuso 34. 68 Francesco Guicciardini. During the time of Guicciardini, that is to say in the latter part of the fifteenth century and the beginning of the six- teenth, several important rules must have become general in Italian, at least in prose-writers ; for one does not find in Guicciardini forms in ade , ate , for a, and conditionals in aria like avria for avrebbe. And it must be also about that period that the auxiliary essere was used to conjugate reflexive verbs with a pronoun in the dative case, just as it was used to con- jugate reflexive verbs with a pronoun in the accusative case. Of course avere was also used, and plenty of instances are found in Guicciardini himself ; but we may say that from that time on avere was decreasing all the time in its use as auxiliary of reflective verbs with dative case pronouns. From the eighteenth century on, I do not find traces of it in the authors examined. Istoria d' Italia. Dative Case : — Pero poi che lungamente si ebbe rivolto per 1 ’ animo lo stato delle cose (I. 25). — Cominciassero cosi presto a non corrispondere a quel che di lui s’ aveva pro- messo (I. 69). — Che una famiglia sola s’ avesse arrogata la potesta (I. 136). — I quali Alessandro con doni s’ aveva fatti benevoli (I. 166). — Mold che s’ avevano proposta maggior larghezza (II. 200). — S’ avevano astutamente insino allora lasciata libera la facoltk di fare il contrario (II. 224). Accusative Case: — Si la morte non si fosse interposta a’ consigli suoi (I. 18). — Non si sarebbe per avventura la pace d’ Italia perturbata (I. 26). — Desideroso di ricorreggere 89 6o J. A. Fontaine , quel che ... s’ era fatto (I. 29). — Nel qual modo si erano altre volte abboc- cati insieme (II. 333). — II quale quella mattina s’ era unito con Ercole, fu morto (III. 94). — E se ne s^rebbero fuggiti molti piu (IV. 270)'. Dative Case with Essere : — Quando ben si facessero poi* effetti molto maggiori di quegli che gli uomini prima si erano promessi (III. 68). — Si erano promessi molto prima la vittoria degl’ inimici (IV. 226). Thus we see that in the first four books of the Istoria cT Italia we find two instances that make an exception to the dative case rule. Torquato Tasso. Dative Case : — Ne trattane colei ch’ alia partita Scelta s’ avea compagna Gerusalemma, lib. VI. 9 Le belle arme si cinge e soppravvesta Nuova ed strania di color s’ ha presa XVIII. 1 1 E insanguinati 1’ aquila gli artigli E ’1 rostro s’ abbia XX. 113 Accusative Case : — Che s’ e d’ Egitto il Re gia posto in via Gerusalemma, lib. 1. 67 Gia 1’ aura messaggiera erasi desta III. Recato s’ era in atto di battaglia Gia la guerriera III. 26 S’ erano all’ alte mura avvicinati III. 33 Del suo avaro pensier non m’ era avvisto V. 48 S’ era del lor partir Goffredo accorto V. 85 Mentre con tal valor s’ erano strette L’ audaci schiere XI. 41 A suoi liberator s’ era condutto XX. 6 ( Aminta .) Accusative Case : — Poi che s’ e posto in uso il grano e 1’ uva Atto I. Sc. I. Ahi ! che s’ e certo ucciso III. I, Che s’ era tutto abbandonato IV. II. io che m’ era nascoso , III. I. Accusative Case with Avere : — Per uccider se stesso e s’ avra ucciso IV. I. This is the last exception I found in Italian to the rule given for the accusative case. And in this particular in- 90 Use of Auxiliary Verbs in Romance Languages. 61 stance the auxiliary avere must have been used to suit the metre. From the time of Beccaria on, the use of avere as auxiliary of reflexive verbs must have been mostly confined to popular usage, and avoided by writers as being inelegant ; for not a single instance of it is to be found in Beccaria, Goldoni, Silvio Pelico, Ugo Foscolo, Manzoni, Cantu, etc. Beccaria. Dative Case : — Si e riserbato a se solo il diritto di essere legislatore e giudice nel medesimo tempore (Dei delitti e delle pene). Accusative Case: — Per sostenere questa vana metafora molte vittime si sono sacrificate (VIII.). — Alcune societa si sieno astenute dal dare la morte (XV.). — Gli uomini si siano voluti assoggettare ai minori (XIX.). Ugo Foscolo. Dative Case : — Mi sono assai volte dimenticato il mio Linneo sopri i sedili del giardino (Jacopo Ortis, 36). — Una lunga treccia di capelli che Teresa, alcuni giorni prima delle sue nozze s’ era tagliati senza che . . . (no). — S’ era piantato un pugnale sotto la mammella sinistra ( 1 1 7) . — Ma se 1 ’ era cavato dalla ferita, e gli era caduto a terra ( 1 1 7) . — Poi non potendo ne volendo ritrarla egli solo delle rapine e degli incendj, si contento si farsi consigliero de’ capi che la moltitudine s’ era eletti, e s’ armo con essi (Commentario Pol. III. XV.). Manzoni. In the remaining quotations I shall only mention reflexive verbs accompanied with a pronoun in the dative case, and show that they take the auxiliary essere as well as when ac- companied with a pronoun in the accusative : — Dative Case : — Cosi dicendo, s’ era levata la chiave di tasca e andava ad aprire (i Promessi Sposi, II. 22). — Ah! gli disse poi vi siete perd fatto tagliare il ciuffo (III. 31). — Come un materialone dopo essersi cacciata in bocca stoppa (III. 33). — Le donne nella sua assenza dopo essersi tristamente levate il vestito delle feste e messo quello del giorno di lavoro (III. 34). — Non lasciava mai sfuggire un’ occasione d’ esercitarne due altre che s’ era imposti da se (IV. 49). — Ma non c’ 9 1 62 J. A. Fontaine , era quell’ allegria che la vista del desinare suol pur dare a chi se i’ e meritato con la fatica (VI. 70). — Cosi impegnandosi a ogni delitto che gli venisse comandato, colui si era assicurata 1 ’ impunita del prime (VII. 81). — Dopo aver sofferto ed essersi morse le labbra un pezzo (X. 136). — Un piccol sentiero indicava che altri passeggieri s’ eran. fatta una strada ne’ campi (XI. 48). — Quando Renzo si fu levato il farsetto (XV. 290). — Facendo tutta via litigar le dita co’ bottom de’ panni che non s’ era ancor potuto levare (XV. 190). — Si trovava ancora indosso quegli stessi vestiti que s’ era messi per andare a nozze in quattro salti (XVII. 215). — E proprio del vestro paese quello che se 1 ’ e battuta per non essere impiccato (XVIII. 229).. — S’ e fatto scrupulo di darle una briga di piu (XVIII. 236). — ma appena partito costui sentendo scemare quella fermezza che s’ era comandata per promettere (XX. 253). — L’ innominato penso subito a rispondere a questa che s’ era fatta lui stesso (XXI. 269). — Dopo essersi cacciate le mani ne’ capelli (XXIV. 306). — E di paragonarlo con 1 ’ idea che da longo tempo s’ eran fatta del perso- naggio (XXIV. 309). — Si mille volte se n’ eran fatti beffe, non era gia (XXIV. 314). — L’ avrebbe preteso e se ne sarebbe fatto render conto (XXV. 317). — Della fllosofia naturale s’ era fatto piu un passa tempo che . . . (XXVII. 348). — Questo aveva sempre continuato a far cio che ... s’ era proposto (XXIX. 376). — Nessuno scrittore d’ epoca posteriore s’ e proposto d’ esaminare (XXXI. 389.). — Andasser facendo di quegli atti che s’ erano figurati che dovessero fare gli untari (XXII. 413). — S’ eran promesse di non uscir dal lazzaretto (XXXVI. 470). Cesare Cantu. Ora mangiavano di quel che s’ erano preparato (Margherita Pusterla, I. 6). — Sotto i piedi dei destrieri s’ era per alcuni minuti vista la morte ad un pelo (II. 22). — Margherita erasi recato in mano un libriccino (III. 38). — Franciscolo . . . si era assunta la esibita ambasceria a Mastino (IV. 58).' — non mostrava d’ averli in quel conto ch’ e’ s’ erano ripromesso (V. 80). — Onde erasi formato un modo proprio di vederle (V. 83). — E s’ era fatto premura di recarli quella sera a Mar- gherita (V. 82). — II cui padre lavorando s’ era acquistato pel paese un triste nome (VIII. 144). — i soldati eransi tolta in mezzo la Margherita (IX. 160). — Ma s’ erano lasciato fuggire Franciscolo (XIII. 221). — Per riparare la quale erasi cavato la giubba (XII. 223). — Quivi entrando Luchino, sebbene gia si fosse messa in- torno al cuore la calcolata freddezza (XX. 365). - — E due gran mustacchi che s’ era acconci (conclusione). Nuova Antologia ( Volume 48). Arrigo il Savio. E quando mi ha scritto che aveva bisogno di me, si figuri, mi sono augurato u-n bel paio d’ ali (478). — Guasti pur troppo la bella immagine che io m’ era for- mata dell’ amor tuo (494). — Il conte Guido non si era proposto di andare (Vol. 49. 98). — Te lo diro un’ altra volta quando mi saro formato una vera certezza intorno a certe cose (233). 92 Use of Auxiliary Verbs in Romance Languages. 63 Vita e avventure di Riccardo Joanna { Volume 52). Riccardo si era annodato dietro la nuca, con molta disinvoltura il tovagliolo bianco (690). Dalla culla alia tomba {Volume 54). Gli uomini s’ eran presi un pezzo di crescia sotto al braccio (460). — S’ eran messo 1’ abito piu bello benche non dovesse accompagnare il battesinio (467). Montegu {Volume 46). Egli in vece si era dato ogni premura per soddisfare a tutte le richieste di danaro che gli venivan fatte (89). — Il rincrescimento di dover rinunciare a un tratto alia fama di giovinetto elegante e alia moda ch’ egli s’ era conquistata (276). — Chi avrebbe avuto fiducia in un uomo che s’ era mangiato il suo (280). — E che per conto mio me ne sarei lavate le mani (460). — Piena di freddo, sebbene si fosse tirate le coltri fin quasi sopra il capo (463). — Anche quella volta s’ era preso i guanti e il cappello (469). — Il tenente Aschieri per aver detto cio, s’ era presa una sciaftolata (479). — Bianca si era creata un mondo a parte (651). — Parendo le oramai di esserselo guadagnato il suo Leonardo (Vol. 47, 296). — La tutta Milano si era dato convegno alia corte d’ assise (87). Scuola nor male feminile ( Volume 49). L’ alunna sedette soddisfatta perche almeno il suo zero se 1 ’ era quadagnato. IV. Auxiliaries with Reflexive Verbs in Spanish. In the earliest documents of the Spanish language reflexive verbs are found conjugated with the auxiliaries ser and haber. El Libro de los Reyes d’ Orient. Ser . . Estos reyes complieron sus mandados E son se tornados 46 Vida de Santa Maria Egipciaca. Haber . Quando desto te avras partido Nos te daremos buen marido. Romancero del Cid. Ser . . E de el me soy escapado . Haber . Y de secreto se ha ido . . 66 7 93 64 J. A. Fontaine , Libre d' Appolonio. Ser . . Ancoraron las naves in ribera del puerto Encendieron su fuego que se les era muerto 00 w-> Ser . . Los que solia tener por amigos leyales Torftados se le son enemigos mortales • • 59 Poema de Alexandro Magno. Ser . . Sennor dixo 1 griego tengo me por tu pagado Quando vassallo tuyo me soe tornado ro 00 Ser . . Cuemo sabial falso que se fues arrancado . • 152 Ser . . Que se se fusse ende estovies bien recaudada Era se ya tornado La Estoria de Senor Sant Millan. Ser . . Que se era probado por sancto muy complido . . 3 22 Milagros de nuestra Senor a. * Ser . . Mucho mas li valiera si se fuesse quedado • • 73 i Cronica del Rey don Alfonso. In this cronica reflexive verbs are conjugated with haber , except in the two following cases : — Ser . . E que dejase los arrayaces porque el oviese dellos emienda E cobrase la tierra con que se le eran alzados . . 16 Ser . . E el era se ido dfinde a le facer guerra de los castillos . . . . 76 Hurtado de Mendoza. All the reflexive verbs take the auxiliary haber in this author. Haber . Que de mi y de ellos se habia ido 85 Don Quixote. I have found only one instance of a reflexive verb conju- gated with ser. Compare : — Ya se es ido el caballero (XXI. 1 1 7) . La demas gente de casa toda se habia ido a comer. 94 Use of Auxiliary Verbs in Romance Languages. 65 Don Quixote is the last work in which any trace of a reflexive verb conjugated with the auxiliary ser is to be found. V. Auxiliaries with Reflexive Verbs in Portuguese. In leading such old Portuguese texts as were obtainable, I found only doubtful cases of reflexive verbs having ser in their compound tenses. Haver is the auxiliary met with in Hardung’s Romanceiro and other Portuguese texts. Cf. A el-rei se ha queixado (Romanceiro, 123). VI. Auxiliaries with Reflexive Verbs in Provencal. The verbal system of the Provencal follows very closely that of the French. In Le Moine de Montaudon, Pierre Vidal, Bertrand de Born, Roman de Flamenca, I found the auxiliary esser always used to conjugate the compound tenses of reflexive verbs. Compare Le Moine de Montaudon. Qu’ieu m’en sui tant defendutz e loignatz. For de mon cor que s’es en vos mudatz. Que de drut s’es tornatz maritz. Pierre Vidal. De chantar m’era laissatz. Bertrand de Born. Quand lo reis Richartz s’en fou passatz outra mar. Dizon que trop me sui cochatz. Roman de Flamenca. Et al rei si son presentat E quan si fou agenolhatz. Roman de Jaufre. El reis es se meravillatz. Cant la vi pueis es se seinatz. E Jaufres s’es apareillatz. S’en es vengutz a Melian. 95 66 J. A. Fontaine. Gerard de Rossil/on. S’en es issiht lo como de grand iror. Chronique des Albigeois . Car lo due de Bergonha s’en es ladoncs crozat. Que el cap des castel se son tuit amagatz. E s’es vengutz a Roma. Lo reis P. d’Arago felos s’en es tornatz. Roman de Fierabras. Devas Contastinoble s’es lo rey regardatz. Ab aquestas paraulas lo rey s’en es intratz. Mireio. En-lio jamai s’es plus fa veire I. 42 S’ei roustido lou comi e la cara au souleu II. 58 . . . Un trau S’ero fa de l’emboucaduro II. 64 The lack both of proper texts and time makes it necessary to end here the investigation of the use of the Auxiliary Verbs in the Romance languages. Should this article be found of any use or interest to the student of Romance com- parative grammar, it is the intention of the writer to complete it by further examination in the Wallachian, Catalan, Rhato- romonsch, and other minor Romance dialects. 96