University of Virginia Library

III

How to reconcile the practices of descriptive bibliography and library cataloguing in the area where the two fields overlap is a problem which has exercised many people over the years. If the Anglo-American code reflects little awareness of the issue, there has nevertheless been a succession of librarians who have touched on the matter, generally in connection with the cataloguing of rare books. Whereas one tradition of bibliographical writing, developing through Pollard, Greg, McKerrow, and Bowers,[62] has taken up the problems of description for those concerned with the physical book (and the use of physical evidence in historical and literary studies), another tradition has consisted of librarians writing for other librarians about the physical details appropriate for inclusion in the catalogue entries for certain classes of material. The two traditions intertwine occasionally, and such men as Esdaile, Cowley, Dunkin, Alden, and Bennett, addressing themselves to library cataloguers, are fully aware of the other tradition of writing about the description of books. Their discussions are worth surveying, as significant attempts to bring the two approaches together, even though their attention is primarily directed toward rare books and special collections, not toward the larger problem of general cataloguing.

Arundell Esdaile, in A Student's Manual of Bibliography (1931), one of a series originally called "The Library Manuals," does attempt to encompass all kinds of cataloguing and begins his chapter on description with this statement: "Every catalogue-entry is a description of the book catalogued; but according to the purpose of the catalogue is the degree of elaboration of the description" (p. 248). What this pronouncement overlooks is the fact that a given purpose may be served by differing degrees of detail. By stating that "purpose" and "degree of elaboration" fluctuate together, it postulates a situation in which a shift in the aim of an entry entails a quantitative, rather than qualitative, shift in the annotation. Esdaile's four levels of description reflect the difficulties of this position. His "minimum entry" consists of nothing but title and author's name and is thus an

entry for a work, even though the work must necessarily be represented in the library being catalogued by a particular copy. The next higher level, called "short-entry" (p. 249), records place and date, but the intention is not so much to identify a book as to specify a text of the work. Esdaile's illustration includes the symbol "12°," though he admits that such notation, as well as a statement of pagination, "serves little purpose here"—except that pagination would "distinguish a pamphlet from a substantial work." And he adds that size notation "seems to be entirely useless." The concern of the entry, in other words, is still with a work; the increased detail does not stem from a different aim, though of course it allows the aim to be pursued with greater sophistication. The interest, clearly, is not in physical characteristics for their own sake. Esdaile's third level, the "short standard description," represents a "minimum standard" for all entries, since one "cannot be sure what book will become important, or what book will become rare" (p. 250). The entry now contains a quotation of the title with omissions noted and a record of format, signatures, pagination, and plates. Its emphasis is beginning to shift toward the physical and, one should observe, away from the individual copy, for the entry describes "firstly all copies, and secondly, the one copy" (p. 252). The highest level, the "full standard description," requires, among other things, quasi-facsimile transcription of the title page, a pagination statement showing which pages are numbered, information on type, and a detailed record of the contents (p. 253). It is designed to "anticipate as far as possible questions which may be asked about a book's physical and intellectual composition" (p. 250), yet the emphasis is definitely now on the side of the physical. Few people today would defend Esdaile's scheme,[63] largely because his inclusion of signature collation in a minimum entry is not a feasible requirement for routine library cataloguing. But a more basic weakness is theoretical: by shifting his emphasis as he moves to higher levels of detail, he is blurring the distinction between quantity of information (from simple enumeration to great detail) and orientation of approach (toward the work or the book). Nevertheless, he at least is confronting the problem of defining a framework which will embrace both a brief reference entry and an elaborate physical description.[64]

Eight years later, J. D. Cowley addressed another book, Bibliographical Description and Cataloguing, to library-school students, setting out specifically to help them "distinguish sufficiently between what is appropriate to cataloguing and what is suitable in a published bibliography" (p. v). In a carefully considered introductory chapter, he segregates, much more successfully than Esdaile, the aims, and resulting methods, of physical and subject bibliography. Recognizing that differing aims dictate the nature of the details to be included in an entry, not their quantity, he says, "Minute description of the physical form of the material is therefore out of place" in a subject listing; "Description of physical features need only be sufficient to secure identification of the work or the edition which is described" (pp. 6-7). That is, the number of physical details can be reduced not because a subject list is a less detailed descriptive bibliography but because it is not concerned with physical books, except as necessary to locate the physical embodiments of works; if annotation is to be provided, it should be of a different kind.[65] Having laid this admirable foundation, Cowley is disappointing in his ensuing recommendations, as they relate to the stated problem of "what is appropriate to cataloguing and what is suitable in a published bibliography." His real interest is in the latter, and virtually his whole work is devoted to detailed descriptive bibliography (it is the most thorough pre-Bowers exposition of the subject). As far as library cataloguing is concerned, he is content to accept uncritically the Anglo-American code then in effect. At the start, he says that for subject bibliography "a simple catalogue entry, constructed according to one of the recognized codes of rules, is the best form of description" (p. 7). And in his chapter on format and collation, after eighteen pages on a method of format designation and a formulary for the recording of pagination and signatures, he appends four lines: "In short entries or entries for subject bibliographies the

technical note should be reduced to the form adopted in the Anglo-American code, e.g. viii, 182, [10] p., illus., 10 plates, diagrs., 20 x 15 cm." (p. 106). It is clear that he equates library cataloguing in general with subject bibliography; but why the details for library cataloguing prescribed in the Anglo-American rules are appropriate for subject bibliography as defined in his opening chapter is never explored. Although his work makes a contribution to descriptive bibliography and offers a helpful theoretical distinction between that field and library cataloguing, he leaves the break between the two as ill-defined in practice as he found it.

In 1951 Paul S. Dunkin's booklet on How to Catalog a Rare Book was published by the American Library Association. Like Cowley, Dunkin does not attempt to correlate reference and physical bibliography. Routine cataloguing has to do with subject matter: "People want to find an ordinary book because they want to read it. Simplified cataloging serves well enough for such a book because it gives a call number and tells what the book is about" (p. 1). A rare book, on the other hand, is of interest as a physical object: "If people wanted only to read it, a microfilm or reprint would do. The fact that the rare book is valued as a material object must be the keynote of any useful approach to rare book cataloging." The booklet proceeds to offer an introduction to title-page transcription, determination of format, and the recording of signatures and pagination—the "cataloging problems peculiar to rare books" (p. 2)—without further considering what rationale underlies the segregating of certain books for this treatment.[66] Indeed, the cataloguer is not supposed to think about this question: "it is not," he says, "the cataloger's job to decide if a book is rare; that has been decided before the book reached his desk." But someone had to make a decision, for the dividing line between "rare books" and others is not self-evident, nor therefore is the division between books of interest for their content and books of interest as physical objects. Obviously any book can be of interest for either reason; but despite the title of his opening chapter, "Whys and Wherefores," Dunkin does not conceive of his task as involving any examination of such matters. Instead, he concentrates on particulars of form and is at pains to show that the recording of physical details need not

be complex, as in his recommendation for "collation by gatherings in simple language" (p. 82). He insists that what he is talking about is "simplified cataloging," not descriptive bibliography of the kind treated in Bowers's Principles.[67] "A cataloger," he adds, "tells only what a rare book looks like; a bibliographer tells how it came to look that way" (p. 1). In taking this position he is creating a false opposition, implying first that the difference between a catalogue and a bibliography lies in the amount of detail and then, rather confusingly, suggesting that the difference results from the presence or absence of analysis.[68] His later writings continue to make these points and show an increasing irritation with the practices of descriptive bibliographers. In the preface to the 1973 revised edition of this booklet he expresses "surprise" at the wide acceptance of Bowers's "highly complicated collation formula" and hopes that "catalogers will never . . . use the Bowers formula in their entries."[69] The new edition "tries to make the Bowers formula easy for catalogers to understand" only because they may have to consult printed bibliographies which employ it. Yet his comparison of Bowers's system and his own "simpler" one (in a section entitled "Collation: Cataloger and Bibliographer," pp. 94-97) succeeds only in demonstrating the superiority of Bowers's formula, not merely in conciseness but in clarity as well,[70] and in showing that the rumors of its difficulty had been greatly exaggerated.[71] More recently, in Bibliography: Tiger or Fat Cat? (1975), he has repeated his attacks on Bowers's system, concluding that "the space saved by shorthand notation is more apparent than real" (p. 28)—as if conciseness were the only reason for it. He also dismisses the descriptive bibliographer's definitions of "issue" and "state" as "jargon," though "harmless" (p. 18)—as if the lack of careful definitions would not lead to sloppy thinking. His resentment emerges in irony at times, as when he labels the cataloguer a "Country Cousin" and says, "In the hierarchy of Bibliography the cataloger stands lowest of the low" (p. 29). It is unfortunate that he chooses to pursue this unconstructive approach, which can do nothing to promote greater understanding and cooperation between cataloguers and bibliographers.

In contrast, John E. Alden, in his excellent essay on "Cataloging and Classification" for the Association of College and Research Libraries' Rare Book Collections (1965),[72] stresses the mutually fruitful relationship that can exist between the two groups. He calls rare-book cataloguing "bibliographical cataloging" (p. 68) and sees the cataloguer as a person with "a great opportunity to render a particular, not to say unique, service to the scholar—the opportunity to describe individual books analytically and to achieve significant patterns either by the correlation of these descriptions or, by means of classification, by the correlation of the books themselves" (p. 65).[73] But in order to play this creative role, Alden believes, the cataloguer must also learn from the scholar and keep abreast of developments in analytical bibliography and textual study. To say that the rare-book cataloguer can ignore those developments or that he can follow the standard cataloguing

codes is "a counsel of despair": "The more productive view is certainly that the 'new bibliography' is the province not only of the avowed bibliographer but also of the rare book cataloger, who in the course of his day-to-day activity has occasion to contribute to knowledge and to scholarship by his own discoveries or by making possible discoveries at the hands of others" (p. 67). Although Alden does make some specific recommendations, he places his emphasis "on ends rather than on means," feeling that cataloguers have given too little attention to "what purpose cataloging served"; formal matters will be handled imaginatively, in response to a given situation, by the cataloguer who is "a humanist before he is a technician" (p. 73). Alden does not therefore address himself to the evaluation of particular systems for recording details in physical and in reference bibliography; but his view that the card catalogue "is adaptable enough to serve the objectives of both rare book cataloging and general cataloging" (p. 68) presupposes a basic compatibility among entries prepared with different aims in mind.[74] The world of cataloguing which emerges from Alden's pages is far removed from the one Dunkin writes about; it is surely the more rewarding one to inhabit.

The approach which Alden describes in general terms is the one which underlies the detailed specifications set forth by Josiah Q. Bennett in his impressive booklet on The Cataloguing Requirements of the Book Division of a Rare Book Library (1969).[75] Bennett's thorough treatment of the form and nature of the elements required in an adequate catalogue entry for a "rare book" is concerned both with upholding scholarly standards and with recognizing practical realities. It is grounded in the belief, first, that library cataloguing of rare books—which necessarily emphasizes physical details—need provide only enough information for identification, not the greater quantity required for true description (e.g., pp. 8, 29); second, that careful initial cataloguing is more economical than the repeated investigation which would otherwise be necessary to answer inquiries or check

dealers' catalogues (p. 10); and, third, that standard Library of Congress entries can be converted, with a minimum of adjustment, to serve as informative basic entries for rare books, leaving plenty of space on the cards, in most cases, for the additional required data (e.g., pp. 41-42). His recommendations on particulars, therefore, naturally throw into relief the differences between ordinary LC entries and bibliographical cataloguing. For example, he underscores the absolute necessity of recording the characteristics of the actual copy at hand, not the ideal copy (pp. 19-20); he explains the usefulness of noting format, even for modern books; he rejects the measurement of the spine to the nearest centimeter in favor of measuring the sheets to the nearest millimeter (pp. 21-22); and he denounces the policy of confining "the page collation to the mere recording of the last page number found," which "has absolutely no place in rare book or special collections cataloguing" because it fails to take account of initial and terminal blanks or advertisements (pp. 18-19). Although he is making these criticisms of conventional library practice only in regard to rare-book cataloguing, his comments suggest the further question whether that practice is really appropriate to any cataloguing. A hint of this question comes to the surface when he speaks of "the decision to 'short catalogue,' not in the sense of eliminating detail unnecessary to the service of any particular type of library but in the sense of requiring inaccuracy for a presumed (one wonders if an actual) saving of time, as in 'last numbered page' collation and inaccurate measurement" (p. 45).[76]

It is beyond the scope of his essay to pursue the implications of this remark outside the rare-book field—that is, outside the area where interest in the physical book predominates. But his underlying concern with the split between the two ways of looking at books permeates his discussion, and the general problem is treated with understanding and insight in his opening section, on the "Rationale" of rare-book cataloguing, and in his closing section, on "The Indicated Symbiosis." The two groups that must learn to be symbiotic are of course those persons interested in information and text retrieval on the one hand and those interested in "bibliographical data retrieval," as Bennett puts it, on the other. The average library cataloguer holds a belief—

fostered by library schools[77] because of the "overwhelming preponderancy and influence of the informational library" (p. 46)—that "the basic purpose of cataloguing is simply the identification of a text in an informational series to the exclusion of all other factors" (p. 44). Bennett realizes that "serial identification" is also part of the task of the rare-book cataloguer, and thus the symbiosis between "serial" and "bibliographical" identification begins with the card entries prepared for the former purpose: This solution is not simply determined by the exigencies of economics; the basic card entry makes a positive contribution to the final expanded entry. The symbiosis, in practical terms, is feasible; what is more difficult to achieve is mutual understanding among persons with differing habits of thought. As Bennett points out, the rare-book cataloguer must have "a habit of mind flexible enough to approach each book brought before it as an entity in itself" (p. 42), not merely as a unit in a series. The development of this habit of mind is crucial to the symbiosis, and one of the merits of Bennett's treatment is that he recognizes questions of personality as well as of bibliographical theory.[77a]

It is natural that discussions of the relation between reference and physical bibliography have generally addressed themselves to problems of rare books and special collections. Cataloguing codes from Cutter to AACR have generally included a section on incunabula,[78] thus tacitly admitting to any user of the codes that certain situations demand fuller attention to physical details than the basic rules allow for. But since no clear dividing line separates "rare books" from other books, the thoughtful reader of these codes will be moved to reexamine those basic rules. Bennett has shown what can be done with a minimum of effort to convert a standard entry into a satisfactory entry for a "rare book"—a book, that is, the physical features of which are, or can be anticipated to be, of interest because it falls into a category of books frequently approached in that way. Any book, however, no matter how unlikely the choice, may be studied as a physical object, the product of a certain moment in printing and publishing history.

But if the conversion of a typical card entry into one appropriate for physical study entails the addition of several pieces of information, it is obviously unrealistic to ask that all books be catalogued as if they were "rare books." What can be done is to look at the matter from the other direction: instead of examining what needs to be added to a routine entry for the purposes of physical bibliography, one can ask to what extent the elements present in the routine entry are useful to the physical bibliographer. In other words, recognizing that for practical reasons the majority of books will be catalogued with a minimum of elaboration and that some essentially physical details will be included even in catalogue entries primarily intended to serve a reference function, one can ask whether those physical details are presented in a form which seems sensible—or at least is not misleading—to the physical bibliographer. To ask this is to ask whether it is not possible to refer to physical details unambiguously in all entries.

I should like to suggest a way of answering this question, taking the statement of pagination as my illustration—both because I concentrated on pagination in my comments on the Anglo-American rules and because the record of pagination is (among the elements included in an LC entry) of particular interest to physical bibliographers. Some of the problems involved in integrating the approaches of reference and physical bibliography toward pagination can be suggested by the practices of two bibliographers, one attempting to incorporate into a descriptive bibliography the pagination formula employed on LC cards and the other trying to devise a formula for brief catalogue entries that would have the precision expected by descriptive bibliographers. Donald Gallup, in his bibliographies of T. S. Eliot and Ezra Pound, gives pagination in a system adapted from the 1949 Library of Congress Rules. He says, "I have modified the system, which ignores blanks and leaves containing only advertisements, in order to account for all leaves, although blank pages are not mentioned."[79] When the preliminary pages add up to the total implied by the first numbered page, they are not specified; but when there would be a discrepancy in the numbering, the preliminaries are specified in leaves, with an indication of which are blank: "1 blank leaf, 3 leaves, 9-29 pp." At the end of a volume, a single blank verso is not mentioned ("29 pp." implies that page 30 is blank), but a single unnumbered verso with printed matter is indicated ("29, [1] pp."). When more leaves follow, the number is recorded, with those blank on both sides being labeled

"blank." Thus the notation "29 pp., 1 leaf" means, in Gallup's explanation, that "the text ends on page 29 and that additional material not a continuation of the text appears on a final unnumbered leaf, either recto or verso or both being printed." Because Gallup's system is intended for descriptive bibliography, it properly emphasizes physical details, recognizing that all leaves which form part of the printed gatherings must be recorded. But in two respects, influenced by the LC rules, it does not consistently stress physical form. First, the decision not to indicate which preliminary pages are unnumbered when their total matches the sequence established by the first numbered page means that the physical detail of printed page numbers is regarded as of secondary importance at the front of a book (whereas a notation like "29, [1] pp." suggests that it is primary at the end of a book, since no pagination is inferred there). Second, the distinction between "29 pp." and "29, [1] pp." reflects a concern with the content of a page (since the thirtieth page exists physically in either case), whereas for the later leaves or the preliminaries the only concern is with knowing the number of physical leaves, not which individual pages among those leaves are blank.[80] Gallup's experiment is interesting, but there remains in his formula a conflict between those details of physical and those of reference emphasis. The context is one of physical bibliography, and the intrusion of elements relating more to content than to form lessens the value of the formula as a physical record.

Rolf Du Rietz's experiment, in his Bibliotheca Polynesiana, illustrates a somewhat different problem. His goal is to work out a pagination formula for reference bibliography—not, like Gallup, for physical bibliography. Yet he believes, with Gallup, that all leaves must be accounted for; he believes, in other words, that a pagination statement for reference bibliography must be more than an approximate indication of the extent of a work—it must be an accurate representation of the physical book in terms of pagination. "One of the aims of a page formula," he says, "must always be that of making possible references to any pages in the unit described, and to meet this and the other purposes of a page formula, it is absolutely necessary to assume a purely analytical approach to the problem of collational formulas for the purposes and needs of reference bibliography" (p. xli). The formula he proposes (and uses in his catalogue) is made up of two elements:

a listing of the total number of pages in each sequence of pagination (inferring page numbers wherever possible), in the form "xii + 268 pp."; and parenthetical indications of which leaves are blank or devoted to advertisements, using a set of abbreviations, such as "BL" for "blank leaf (or leaves)," "FLA" for "first leaf (leaves) advertisements," "LLBA" for "last leaves blank (one or more) and advertisements (one or more)," and so on. A hypothetical collation which Du Rietz offers as an illustration takes this form: "[πIV] (BL) + xii (FLA) + 268 + 369-800 (3 LLB) + 232 (FLB, 2 LLA) + 398 (3 FLBA, 4 LLBA) + XVI + (8) pp." Of course, most actual books are not likely to be so complicated: entry 168, for instance, reads "xx + 344 (LLB) pp.," and entry 15 reads "280 (2 FLB, LLB) pp." The trouble with this system is that its complications arise from the inclusion of information not directly relevant to the reference function of the entries, yet not really sufficient for physical bibliography either. For reference purposes, the extent of the work is the central fact to be communicated by the pagination statement; therefore Du Rietz's inclusion in the record of all the leaves (except binder's leaves) in the physical book necessitates his insertion of explanatory parentheses to account for those leaves which are not part of the work.[81] But this elaboration of statement does not turn the record into one appropriate for physical bibliography, as Du Rietz recognizes: "a page-collation formula in reference bibliography is normally not concerned with the individual numbering of each page" (it records "sequences of pagination") and "should thus not be confused with pagination statements (or formulas) of the kind given in descriptive bibliography" (p. xliii). One of the principal expressions of his rationale illustrates the essential problem: As this passage reveals, his system falls between two stools: the extent or bulk of a work could be indicated much more simply, and in order to be a "real analytical formula" (p. xliv) still more information would have to be given.[82]

Du Rietz is right to observe that the pagination formulas in library catalogue entries frequently do not reflect any clear understanding of the purposes they are to serve.[83] But his own proposal seems to me, both on theoretical and on practical grounds, not to provide the answer. The theoretical problem results from his unwillingness to let the pagination formula in reference bibliography refer solely to the extent of the work. So long as one insists that the formula have a physical orientation, even in a reference context, and yet does not require it to conform to the practices of descriptive bibliography, the purposes of the formula are bound to be somewhat confused. Du Rietz asserts, "Since hardly any of the thousands of page collations for reference bibliographical purposes that are every week constructed all over the world are analytically conceived, it follows that they might as well have been left unwritten and that they mean a tremendous waste of time, labour and money" (p. xl). Then he adds, "I do not say that a page-collation formula in reference bibliography should conform to the standards required for the pagination formulas supplementing the analytical formulas in descriptive bibliography." Why not? Is it not also a waste of time to have people learn an intermediate system, when an unambiguous system already exists for recording pagination from a physical point of view? The theoretical consideration is thus linked to the practical one: any solution to the problem of appropriate pagination formulas must be based on a realistic assessment of the feasibility of its adoption. The farther it departs from standard practices, already widely understood, the less its chances of ready acceptance, unless what it proposes is so obviously necessary that no resistance

is conceivable. Du Rietz realizes that his system will require additional "training in analytical and descriptive bibliography" for cataloguers (p. xl), but he regards that fact as an advantage, since such training would help to promote rigorous standards in the field.[84] Of course, bibliographical training is an asset; but requiring an analytical approach to the pagination of every book and a special system (different from that in descriptive bibliography) for recording pagination is bound to work against the acceptability of the system. "I am aware," Du Rietz remarks, "that what has been said above may not be altogether agreeable to many librarians and reference bibliographers, but I assure them that there really is no choice" (p. xli). My belief is that a choice does exist, but Du Rietz's discussion is valuable in raising the issues and in pointing the way toward a more workable solution.[85]

Let us postulate a book which would have the following pagination formula in a descriptive bibliography (following Bowers's Principles): pp. [i-v] vi-viii [ix] x-xi [xii-xiii] xiv-xvii [xviii-xx], [1] 2-275 [276-277]

278-283 [284-288].[86] If one is interested in a physical record of pagination, it would be hard to improve on this approach. The great advantages of this formula are that it is readily understandable and that its construction does not require subjective or time-consuming decisions. Indeed, there are no special rules to be learned, except for an agreement never to link with a hyphen an inferred and a printed page number. When, in a reference bibliography, one wishes to pay particular attention to physical details, it seems pointless to construct a new system when an established one of such simplicity and clarity exists. This kind of formula may be somewhat longer for most books than the kind traditionally required for library catalogues, but whether its construction in most cases would take much (if any) longer is doubtful. And the reader who wishes to know only the extent of the work can tell from this formula just as readily as from "xvii, 283 p." what the approximate length is. Of course, the formula does not indicate the contents of any page, because in a descriptive bibliography a contents paragraph would follow; if this formula were used in a reference bibliography, the bibliographer might wish sometimes to specify separately the contents of certain pages, as "Checklist, pp. [277-284]; advertisements, pp. [285-286]; blank, pp. [i-ii], [287-288]."[87] Du Rietz's system, which would produce "xx (FLB) + 288 (2 LLBA) pp." is undeniably more concise, but it is far from self-explanatory and much less informative—for (like Gallup's system) it does not specify blank pages but only whole blank leaves,[88] and it does not attempt to show which pages are numbered. If the pagination record is going to be of use to physical bibliographers, the longer formula is so much more straightforward and precise as to outweigh whatever slight saving of time or space an abbreviated formula effects.[89] After all, the reason for complicating the pagination statement in the first place is to provide physical information, and it does not seem worthwhile to introduce complications without going far enough to achieve an unambiguous physical record.

I am not suggesting, however, that a full pagination formula of the kind employed in descriptive bibliography is appropriate for the majority of entries in a library catalogue or a checklist at the end of a book or any other bibliographical work emphasizing reference bibliography.[90] My point is that whenever a library cataloguer or reference bibliographer wishes (for whatever reason he finds persuasive) to include information about the physical details of pagination, he would be better advised to adopt the formula of descriptive bibliography than to settle for some intermediate scheme which only partially accomplishes his purpose. But the corollary is that in most instances, when he is concerned solely with reference bibliography, he need not complicate his statement at all with physical details. For the hypothetical book under discussion, why could he not simply say "xvii, 284 pp." (or, adding the two figures, "301 pp.")?[91] According to the Anglo-American rules he would have to say "xvii, 283 [1] p.," including the "[1]" only because the 284th page, containing part of a checklist, would be referred to in a note. Or if the printed matter in the preliminaries extended to page [xix], why should he not say "xix, 284 pp." (or "303 pp.")? If reference bibliography is a legitimate pursuit, as it plainly is, there is no reason to insist that entries devised for reference purposes should satisfy the demands of physical bibliography. Where library cataloguing rules for pagination have generally gone astray is in paying too much attention to the physical details of pagination when the interest of the entry is in the extent of the work. Why should a work be labeled as "283 p." merely because the last page number is "283," when the last page of printed matter is actually the 284th page? If the concern is with the extent of the work, that extent might as well be recorded as accurately as possible, especially since the presence or absence of a printed page number is a detail of

typographic design irrelevant in this context.[92] Or why should the work be listed as "283, [1] p.," giving the impression of a fidelity to physical detail which is not only beside the point but in fact not true? I do not see why a descriptive bibliographer should be bothered by an entry in a reference bibliography which simply records the total number of pages occupied by a work in a given embodiment. What is objectionable is a reference entry which appears to be concerned with the physical details of pagination but in fact is not fully committed to recording them and is therefore misleading and inaccurate. Either a pagination statement should aim solely at indicating the extent of a work, ignoring both the typographic form of the pagination and such nontextual features as blank pages, advertisements, and colophons (but not appendixes or indexes); or else it should offer a full and dependable accounting of all the pages in a book, showing exactly which ones are numbered and in what form. A simple convention would serve to distinguish the two: the former would follow the pattern "284 p." or "284 pp.," in which the abbreviation for "pages" follows a numerical total; the latter would take the form "pp. [1] 2-283 [284-288]," in which the abbreviation for "pages" precedes an indication of numerical sequence(s). Both systems are easy for the bibliographer to employ and easy for the user of a catalogue or bibliography to understand; each is appropriate in different situations and is recognizable for what it is. Neither reflects an indecisiveness as to its purpose: each is efficiently constructed in accordance with the aims of one branch of bibliography.

It is true, as Du Rietz says, that "One of the most urgent needs of reference bibliography is the working out of adequate rules for page collations" (p. xl). What I have tried to suggest here is that a sensible solution can only flow from a clear understanding of the aims of and relations between reference and physical bibliography. Pagination is not the only problem, and I hope that what I have said offers an approach, a way of thinking, which can be applied to other elements of an entry, such as the recording of dimensions or the quoting of

titles. The work of reference and physical bibliographers inevitably overlaps, and fruitful cooperation between them must rest on mutual respect and a recognition that each will need to draw on the expertise of the other. A physical bibliographer, in the aspects of his work which touch reference bibliography (such as establishing headings or classifications, indexing, citing particular libraries, and the like), should be glad to avail himself of the established practices which result from the accumulated experience of workers specializing in that field. Similarly, a reference bibliographer, in the parts of his work which impinge on physical bibliography (such as recording certain basically physical details like pagination), should turn for advice to those who specialize in studying the physical book and who have developed conventions for expressing their findings about physical form. In neither situation is it necessary to import a highly technical approach from the other field if it would be excessive in the context; but whatever is adopted should seem sensible to the specialists in the other field and be compatible with their approach. Library cataloguers, for instance, need not employ the full pagination formulas of descriptive bibliography; but their formulas should then be unambiguously focused on the content of the books, so that no one will mistake them for attempts to record physical facts. And when a need does arise for paying more attention to the physical book, they have the descriptive bibliographer's system to turn to.[93]

Much has been made of the difficulty of the formulas and terminology employed in descriptive bibliography and of the fact that a library cataloguer cannot speak to the user of the catalogue through a preface explaining his system.[94] But the standard formulas of descriptive

bibliography are not complex, except in the case of complicated books which would be likely to require more complex formulas under any system. An unfamiliar convention often appears to be more difficult than it really is, but the presence of a full descriptive pagination formula on a library card—without explanation—could not possibly prevent anyone unacquainted with the system from extracting the information he wished about the length of the work. To some extent a conflict is unavoidable between what A. Hugh Chaplin calls "tradition" and "principle."[95] The traditional practices of a field may restrict its receptivity to new approaches which are possibly more logical; yet tradition is important in maintaining stability and uniformity. Chaplin urges cataloguers—in their role as reference bibliographers—to be responsive to the requests for information from those who use their entries. He says that revisers of cataloguing codes "must adapt their rules to development and change in the user's tradition, the expectations of people using the catalogue, and where this would be hindered by their own habits, their own tradition, the cataloguer's tradition must be disregarded" (p. 11). But, as he recognizes, "The difficulty is that the expectations of the users of catalogues are inconsistent and variable, while the rules must conform to a system. The solution is to make general rules, which conform to generally prevalent expectations." When physical details are incorporated in a reference entry, the professional approach is to attempt to satisfy the prevalent expectations of physical bibliographers. And in the reverse situation, the physical bibliographer should of course meet the expectations of reference bibliographers. A century ago Charles A. Cutter concisely described the attitude which stands in the way of this kind of cooperation when he said, speaking of the reaction of catalogue-users to schemes of subject classification, "The reader at first glance is frightened by the appearance of a system to be learned and perversely regards it as a hinderance instead of an assistance."[96] All fields probably have some needless jargon; but for precise and efficient communication they must also have technical vocabularies and conventional forms. It is perhaps natural to be apprehensive about unfamiliar systems. But if we can learn to approach fields related to our own in an open-minded and positive spirit, assuming that they have someing to teach us which will be helpful in our own field, we will be well on the way toward achieving the advances which cooperation naturally produces.