College and Research Libraries Of the Arrangement of Books IN THE YEAR 465 A.D. or thereabouts Sidonius Apollinaris described in a let- ter the private library in the country house of his friend Tonantius Ferreolus: ". . . books in abundance ready to your hand; you might have imagined you~self among the shelves of some grammanan, or the tiers of the Athenaeum, or a book- seller's towering cases."1 But pagan writ- ers were shelved in one place, Christian in another. "You had to consult them on different sides of the room," Sidonius complained. "The arrangement had this defect, that it separated certain books by certain authors in manner as near to each other as in matter they are far apart. Thus Augustine writes like Varro, and Horace like Prudentius." The chance which preserved his letter thus made the observant Sidonius per- haps the first recorded critic of the clas- sification and arrangement of the books in a library. But had they been preserved, I am sure that complaints were made about the catalog engraved upqn the walls of the library at Edfu in Egypt, 2 and that the Babylonian Amid-anu, who lived some seventeen hundred years be- fore Christ and is perhaps the first li- brarian whose name is recorded, had to give ear from time to time to unhappy scholars who wanted to rearrange the clay tablets in their own-conflicting- systems.3 For it seems to be a library axiom that no arrangement of books can please all of the people all of the time. And this fact, although it is a nuisance, has func- 1 0. M. Dalton (trans.), Th e Letters of Sidon ius (Oxford, 1915), I, 50. 2 James Westfall Thompson , Ancient Libraries (Berkeley, 1940), p.3. 3 Ibid., p.12. MARCH 1964 BY WILLIAMS. DIX Dr. Di:x is Librarian of Princeton Uni- versity. This paper was read to the Univer- sity of Washington Library Convocation on October 18, 1963. tional roots. A library is a collection of books arranged for use. The trouble is that different people use books in dif- ferent ways for different purposes. One of the categories posted in stone at Edfu was "The Book of how to repulse the: crocodile," an admirable classification, I am sure, for the priest whose specialty was repulsing crocodiles; but I suspect that his colleague in the Department of Protecting against Serpents grumbled when a tricky case of serpent-protecting. forced him to look up a few esoteric: facts which happened to be in one of the crocodile books. But even though all arrangements rep- resent compromise, some are better than others. When one stops to think about it, there is a surprising variety of ways in which books can be arranged, and it is even more surprising that we apply por- tions of many of these systems in our li- braries today. My predecessor at Prince- ton, Ernest Cushing Richardson, when_ he published his little book on the sub- ject in 1901, listed nineteen basic kin~s of arrangements which he had seen 1n use. 4 One could add others. Aside from those systems designed to bring together books which are like each other in sub- ject, which we are inclined to take for granted as the way of arranging books, think for a moment about some of the- other ways in which books could be ar- ranged. 4 Ernest Cushing Richardson, Classification-The- oretical and Practical (New York, 1901). 85 1. By color. A useful mnemonic de- vice which we all use unconsciously, but of more importance to the interior deco- rator than the scholar. (But I did have once from a part of the campus which shall be nameless a request for ten feet of green books from our duplicates!) 2. By height. We all necessarily make some divisions by height for the sake of economy, and height is a basic principle of. compact storage, which many of us are forced to adopt in some degree. 3. By alphabet. An arrangement ac- cording to the surname of the author has obvious merits as a self-finding device (as well as demerits) and is an element of most systems. 4. By date. To be able to see together in one place all of the books published in England, say, in 1688 would have ob- vious attraction for the historian of ideas; one of our economists makes periodic journeys to New Haven, where there is a special collection arranged in just this fashion. But this is obviously not the best arrangement for all users. 5. By thickness. We commonly treat broadsides in some special way. 6. By fragility. Papyri obviously can- not be shelved indiscriminately. 7. By weight. A bit esoteric, but we do have about ten tons of marble or plaster casts with carved inscriptions for epigraphic study which cause us endless shelving problems. A consideration new to me is sug- gested by an etiquette pamphlet of 1863, quoted in the May I June issue of that lively little sheet distributed by the Co- lumbia University Press, The Pleasures of Publishing: "The perfect hostess will see to it that the works of male and female authors be properly segregated on her book shelves. Their proximity, unless they happen to be married, should not be tolerated." There is no point in continuing this list. Most of us decided long ago that the best compromise is some system of ar- rangement based on subject, which at- tempts to bring together those books which somehow resemble each other in content.· To this end formal systems of subject classification have been devised which attempt to provide a logical loca- tion for each book which may be added to the collection and usually a symbol by which it may be located. Mr. Richardson lists 170 of these systems which had been described in some detail by 190 1, begin- ning with the classification of the Alex- andrian library as expounded by Calli- machus. The devising of the perfect uni- versal classification can become a fas- cinating intellectual exercise, and new proposals have continued to appear dur- ing the past fifty years, perhaps most notably in India under the stimulus of Ranganathan. These exercises, intriguing as they are, may have more to do with metaphysics than with practical librarian- ship, and they need not concern us here. Indeed I venture to be so heretical as to suggest that any reasonably logical and complete classification, applied consist- ently, will work, admitting of course the economic advantages of cataloging and classification done centrally for many li- braries using the same system. I except of course the personalized system of a tall and angular colleague, who hated to stoop and who insisted upon reshelving a book, once he had found it, at eye level, where he-but only he--could always find it again. Thus I am not talking today about schemes of formal classification but about what we do with oooks once we have classified them and given them call num- bers. There are, of course, great libraries in which the books are not classified, ex- cept by size, in which each book is sim- ply placed on the shelf following the one which arrived just before it, to be fol- lowed by the one which arrives after it, regardless of subject. This is a marvel- lously efficient way to pack books, with nearly every cubic inch used and no need 86 COLLEGE AND RESEARCH LIBRARIES to allow any open space except at the end of the file. Perhaps the only more effi- cient ways are to take millions of tiny pictures of books, then throw the books away, or to tear the contents of books apart into bits of information suitable for a computer to digest and regurgitate upon . demand. All of these methods of storing books are in some ways admirable, but they all have serious defects. The computer as yet is not really very useful as a storage bin except for facts, and who wants a library composed of nothing but World Almanam? Very often both the student and the scholar want not facts but prose or poetry, those admirable arrangements of words in sequence, which can record and transmit love or hate, fear or forti- tude. Microreproductions are fine, too, for some things, and no doubt we shall be using more of them, for their develop- ment is obviously far from complete. But they too present problems; they are un- comfortable to handle physically, it still seems to be cheaper to store a book than take pictures of it, and sometimes when I want a book, my need cannot be satis- fied by a picture of a book but only by the book itself, complete with all that the paper, the ink, the binding have to say to me. Compact storage, with no con- cessions to subject, author, date of pub- lication, or any consideration except size, does save space, and space costs money. But it seems to me folly for the univer- sity library to arrange a very large pro- portion of the collection in this way and unwise to arrange any books in this way unless there are very pressing space re- strictions indeed. This is why. The basic test of the quality of any university library is its ability to get into the hands of the reader the book he wants when he wants it. I suggest that this is the first principle, and that from it stem nearly all of the things we do or ought to do in libraries. (I say "nearly all" because there are other things. For MARCH 1964 example, we do certain things, and ought to do more, to lead people to want more books, such as display them attractively or provide comfortable chairs in which to read them. But nearly all library ac- tivities lead toward this one end: to get into the hands of the reader the book he wants when he wants it.) Like all simple generalizations this statement turns out to be more complex than it seems at first. It hardly has the significance of E = mc2 , but in its own sphere it has a certain relevance to a great deal. No library in the world can pass the test perfectly, supplying all po- tential wants of all potential readers. But we should do what we can to come as near the mark as possible. To this end we go through a complex series of rituals to improve our percent- age of hits. Since the odds of success are obviously better if we have more books; we raise all the money we can to acquire and house all the books we can. We at- tempt, by one sort of divination or an- other, to improve the odds by selecting the books believed, by someone, to be required most often. As we get these books we catalog and classify them some- how, for they must be found to meet our test. We bind and package them in one fashion or another to preserve them phys- ically. We develop elaborate systems of records and of notices and fines to snatch them from the hands of one reader so that they may be ready for the next one. (Our objective is only to get the book into the reader's hands, not to leave it there!) Ironically, the things we do to achieve our goal often seem the very things which deflect us; the book which is wanted always seems to be in the proc- ess of cataloging or at the bindery. Consider for a moment the implica- tions of this basic test of library quality, this First Law of Bibliodynamics, for the arrangement of books. Assuming the existence of a reasonably comprehensive collection, how should these books be 87 arranged to get into the reader's hand the book he wants when he wants it? The relevant phrase here is, I think, "the book he wants." Look at several situations. If he knows he wants a particular book in a particular edition and knows the author and title, it probably does not make too much difference where the book is shelved, provided there is some sort . of author catalog giving an index number or symbol to that particular vol- ume by which it can be located and placed in his hands. Even if the book is stored electronically in a computer or on microfilm or in a random location in compact storage the particular book can be found, although in a large, closed- stack collection the reader may grow im- patient of the wait while it is being brought, even by a page on roller skates. Even if he wants a score or more of books which he can identify precisely in the catalog, say to check a series of ref- erences-a not uncommon situation- any of these systems will work. But it does begin to seem that an inordinate amount of manpower, albeit low-level manpower, is being expended to retrieve these books if they are scattered at ran- dom through a million-volume compact storage or fixed-location library, espe- cially when each of them is wanted for perhaps a minute only. As the number of books increases, the advantage of some system of classifica- tion by subject becomes apparent, for the odds are that m(:lny of these books will be of the same type and will thus be found in the same part of the build- ing, if not immediately adjacent to each other. If they are shelved in some kind of subject classification, this type of use situation suggests further that it might be much simpler and more economical to let the reader go to the shelves himself and consult the books at some convenient nearby table, thus avoiding a great deal of filling out of call slips, waiting, and running about. I suspect however that this situation in which the reader kriows precisely the book he wants, whether one or twenty, is by no means the most common one in the university library. Rather, is not the reader more likely to know only in gen- eral what he wants and to be fully satis- fied only after examining briefly the books themselves and then happily taking the right one away to read? It may be a good text of Hamlet, or a good history of the Crusades, or a good study of business re- cessions, or the Oedipus complex, or aerodynamic turbulence, or the flora of the Sierra Nevada. It seems to me that a quick look at the subject headings in the card catalog, followed by a quick exam- ination of the shelves of a well classified collection to which the catalog has sent him is more likely to place in his hands the book he wants than any other system. No catalog card, however complete, no electronic console for scanning a biblio- graphic store, can quite do the whole job. I recognize that this may seem an anti- intellectual ap.proach, that we librarians are constantly telling students to learn to use the card catalog and the standard bibliographic tools to identify the books they want. Perhaps one should make a list of possibly useful books, then read the reviews of each, then weigh the various merits and demerits, then finally send for the one best book. But few of us can work that way. One should of course use all the tools available, but should he not also cultivate by practice the marvellous flair of the true bookman and scholar for skimming quickly through a series of volumes and then almost by instinct find- ing the one which fits exactly his needs of the moment? The library that facilitates this practice is the open-stack, classified collection. Only in this sort of collection can one get at one other type of book he wants, that which he did not know he wanted until he found it. It can happen to a fresh- man who, hunting for a novel by Stewart 88 COLLEGE AND RESEARCH LIBRARIES ·• Edward White which he had read in high school, comes for the first time upon the delights of E. B. White, or T. H. White, or even the subtler ·delights of old Gilbert White of Selburne. It can happen to the mature scholar who, working on the theory of taxation, stumbles upon an ob- scure seventeenth century sermon which opens up for him a whole new line of in- quiry. This kind of browsing is a by-no- means-unimportant by-product of the kind of arrangement of books about which I speak. One of the things which worries me most about the random com- pact storage arrangement which lack of space is forcing many of us to adopt for parts of our collections is that the seldom- used book which is the obvious first can- didate for relegation to a compact storage collection is precisely the book which may never be found and used except by dis- covery on the open shelves of a classified collection. It may appear in no bibliog- raphy, its author may be unknown, its true importance may not be brought out fully by the subject headings assigned to it by the cataloger; yet its discovery by the right scholar may bring to light a point of view or a trend or a literary style that deserves attention. It is the possi- bility of discovering such a book which makes research in a great library more exciting than work in a collection con- taining only the standard works. Assuming that the local situation, al- ways the determining factor, permits us to establish a system of subject classifica- tion as the basic organizing principle of the library and that we are so fortunate as to be able to permit our books to be placed on open shelves for anyone to con- sult, another problem arises. Do we put all of the books on the campus in one building, or do we lift our great chunks of books which cohere by subject and disperse these chunks around the cam- pus? After considerable reflection on the subject and a fair amount of abrasion, I for one have concluded that there is no MARCH 1964 one right system, no sacrosanct ideology applicable to all situations. So long as the collection is thought of as a single university library, existing for the greatest good of the greatest number of readers and administered with as much tolerant concern for the special interests of the individual as circumstances permit, a con- siderable variety of local geography can be tolerated by the academic community. I must confess to an increasing aversion to the phrase "departmental library," hav- ing seen otherwise respectable collections lose a great deal of their utility by deploy- ment along administrative lines rather than on the basis of the intellectual con- tent of the books. I must confess also to feeling a certain rightness in the concept of a great li- brary pulled together as a unit, with every book standing in its proper place accord- ing to some clearly understood coherent system, all open to every user, with a per- haps austere but comfortable and well- lighted chair within twenty-five feet of it. No book is missing because it is on re- serve or in some special reading room; no book is c_harged out to one of my col- leagues who wants to read at home; and especially no book which I happen to want is a mile away across the campus in the Institute of Numismatics library. There is no great library which satisfies all of these requirements. But I recognize that some concessions must be made. It is hardly practical to make the 862 freshmen in Professor Jones' History 100 course who must read chapters 5 to 10 of Gibbon by next Thursday fight over that one copy in its, to me, proper place on the shelves. We do have a responsibility to take fairly good care of the only copy of Fanny Hill, Lon- don, 17 4 7, in wrappers, between the Huntington and the Houghton. And, above all, my colleagues in the Institute of Numismatics on the North Campus across the river do have a particularly high incidence of varicose veins which 89 makes it difficult for them to walk, and they do advance persuasive reasons why there are pedagogical and scholarly bene- fits in having the books they need daily located somewhere nearby, pointing out the irrefutable fact that it is exactly the same distance from the Institute of Nu- mismatics to the main library as it is the other way. A good case can be made on the basis of local geography and other obvious considerations for some deployment of a large collection in various campus loca- tions, and little harm is probably done if the basic concepts of the single university library with central cataloging is pre- served and if the natural lines of cleavage are followed in separating the various parts. What seems to me fatal is the series of separately-budgeted, separately-con- trolled faculty libraries common in Con- tinental universities, which rob the uni- versity library of any meaning or plan. Equally bad is any arrangement along some arbitrary administrative pattern which separates the collections in history from those in the classics or some other abomination of the same sort. If the collections must be split, I sug- gest that lines of cleavage do exist be- tween some of the highly applied subjects and their relatively "pure" manifestations and that with comparatively little duplica- tion one can make the applied practition- ers happy without destroying the basic unity of the library. Similarly, it may not please Sir Charles Snow, but one can split the natural sciences from the social sciences and the humanities without causing too much pain, except to the historians of science, who are a hardy race anyway, accustomed to inconveni- ences. Of course, if mathematics goes with the sciences, the mathematical logi- cians in the philosophy department will cry out in anguish; if it stays with the humanities, the scientists will scream. The exotic languages of the Near and Far East can be safely segregated from the re- mainder of the collection, but English language books on those areas will then cause a problem. In the long run, is it not better for everyone, increasing his chances of find- ing the book he wants when he wants it, to keep as many books as possible to- gether? If it is necessary to serve some subjects by branch libraries in remote areas, the wise librarian will search dili- gently for the most appropriate lines of cleavage and will resort to somewhat more duplication of titles than his natural frugality quite endorses in order to pre- serve as much as possible the basic coher- ence of the library. You have a very wise librarian indeed, and I am not surprised that my reflec- tions upon the more satisfactory arrange- ments of books seem to have ended up by describing a system substantially like the one I have seen here. My views seem to be echoed by the most important ele- ment of the community, the customer. The Washington Daily editorial this morning says, "It really is easy to find a book." What finer verdict could you want? Library buildings are obviously less important than what they contain, but it is equally obvious that they are of vital importance when they permit and encourage an optimum arrangement of books, one which comes closest to getting into the hands of the reader the book he wants when he wants it. I con- gratulate you on the progress you have made toward this goal. • • 90 COLLEGE AND RESEARCH LIBRARIES