University of Virginia Library

Search this document 


expand section 
expand section 
expand section 
expand section 
expand section 
expand section 
expand section 
expand section 
expand section 
expand section 
expand section 
expand section 
expand section 
expand section 
expand section 
expand section 
expand section 
expand section 
expand section 

expand section 



And continued at present by Oxford University Press. (Some of the volumes published by Hart-Davis are being reprinted—often with addenda—by St. Paul's Bibliographies.)


In reviewing the revised edition (1971) of the Fifoot bibliography, B. C. Bloomfield remarks, "It is doubtful whether there ever was a Soho formula, but there was, and is, a general style"; he then specifies some features of it that bear reconsidering (Library, 5th ser., 28 [1973], 76-77).


This point of view is fundamental to Fredson Bowers's Principles of Bibliographical Description (1949); see esp. pp. 3-34, 355-370 passim. I have made a few further comments on this matter, as in the opening pages of "The Descriptive Bibliography of American Authors," Studies in Bibliography, 21 (1968), 1-24, in "The Descriptive Bibliography of Eighteenth-Century Books," in Eighteenth-Century English Books Considered by Librarians and Booksellers, Bibliographers and Collectors (1976), pp. 22-33, and in "The State of Bibliography Today," PBSA, 73 (1979), 289-304 (esp. p. 300).


J. D. Cowley was approaching the same point (in his chapter on "Arrangement and Headings" in Bibliographical Description and Cataloguing, 1939) when he said, "A mere collection of notes thrown together in any convenient order is not a bibliography, though it may be a sale catalogue or a finding list. Bibliography is not content with the assembling of descriptions, but endeavours to relate one with another. . . . In other words it must tell a story" (p. 179).


One cannot even say that it should be presented in chronological order, for effective works of history and biography have followed other plans.


I trust it is clear that I am talking about the selection and arrangement of material, not about the specific form for recording a given detail, once chosen. There is good reason, for instance, to follow in general the standard formulary for signature collation, as codified by Bowers in the Principles, rather than to invent a new system of one's own. Procedures for recording particular details are tools of writing; I am speaking here of the arrangement of blocks of material that have been recorded using these tools. On the choice of details to record and the degree of accuracy to be employed in recording them, see my "Tolerances in Bibliographical Description," Library, 5th ser., 23 (1968), 1-12.


A. W. Pollard, writing in 1909 on "The Arrangement of Bibliographies" (Library, 2nd ser., 10, 168-187), stated this point forcefully: "To make a good bibliography of any subject (even if the goodness be confined to the arrangement, without critical notes) postulates a very intimate knowledge of that subject in all its parts" (p. 168). Later in the same passage he emphasized the point: "the best bibliographies will always be made . . . by the people who are steeped in a subject and the literature of it" (p. 169). Although Pollard's essay is primarily concerned with subject checklists, it includes several statements—like these—equally applicable to descriptive bibliographies. The most thoughtful successor to Pollard as a commentator on enumerative lists is D. W. Krummel; see his chapter on "Organization" in his forthcoming book on bibliographic lists.


Bowers, in the Principles, takes the same view: "No rigid rules for the arrangement of material in printed bibliographies can be laid down," he says. "Each author will offer distinct problems to be solved on the basis of the material itself" (p. 382). He also points out that "One of the most difficult problems facing the writer of a bibliography is the decision about the precise arrangement he will employ" (p. 383). Donald Gallup similarly says, on the "question of arrangement" (which "should be as simple and logical as possible"), "Every bibliographer will of course run into problems that are to some degree unique with the author he has chosen" (On Contemporary Bibliography with Particular Reference to Ezra Pound [1970], p. 26).


My comments below refer specifically to author bibliographies; but many of them apply also to other descriptive bibliographies, such as those recording the output of printing and publishing firms.


How a division into sections is determined—or whether there are to be sections at all—is taken up below.


This annalistic approach is favored by Rolf E. Du Rietz in "Thoughts on Author Bibliography," Text, 1 (1974), 203-216 (see p. 209). It was also preferred by Cowley (see note 4 above), p. 183.


Pollard in 1909 (see note 7 above) expressed his strong disapproval of this "hybrid" approach (p. 178). But half a century later W. W. Greg was able to say, in the fourth volume of his A Bibliography of the English Printed Drama to the Restoration (1959), a work in which plays are entered chronologically, "It was of course axiomatic in such a work as this that all editions of each piece should be grouped together"; "the procedure finally adopted," he added, "was to list all plays individually in the order of printing of their first extant editions" (p. xxiii).


See Bowers, Principles, pp. 37-123, 371-426; and G. T. Tanselle, "The Bibliographical Concepts of Issue and State," PBSA, 69 (1975), 17-66.


One attempt at such revision is James B. Meriwether and Joseph Katz, "A Redefinition of 'Issue,'" Proof, 2 (1972), 61-70. My criticism of their effort appears in the article cited in note 13 above.


This phrase is meant to include such pieces as the products of private presses; the point is not whether they were placed on public sale but that they may be considered "published" when they left their printers' hands.


It is a publishing decision that causes, for example, 2000 copies of an edition to be printed in two printings of 1000 copies each rather than in a single press run.


Presumably only the main entry would receive a number, however, and thus the scheme of reference numbering does represent one approach at the expense of the other; but the basic purpose of reference numbering is to point to the physical location of material, not to indicate its classification. (See part III below.)


Some classification schemes—such as those often used in libraries—have a single unambiguous arrangement as their purpose; but such an ordering can be achieved only by making some arbitrary decisions about the location to be assigned to certain items. Bibliographical classification, on the other hand, is not primarily intended to suggest a physical arrangement (though any specific classification plan to some extent does so) but rather to furnish a theoretical structure for thinking about certain kinds of historical relationships.


That subedition does not specify the set of plates is suggested by Bowers: "When the original plates of a book, or extra sets of plates, are sold or leased by the publisher to another, and a new impression is struck off under this second publisher's imprint, we are forced to consider that a sub-edition to the parent has been constituted" (p. 389). From the point of view of publishing history, the publisher's imprint may be the dominant fact; but whether the impression with the second publisher's imprint was printed from the original, or a second, set of plates is of significance in printing history and textual study.


"The Bibliographical Concept of Plating," SB, 36 (1983), 252-266. West employed this approach in his bibliography of William Styron (1977). In both the bibliography and the article he credits Joseph Katz with the idea, and in the article he further acknowledges David Farmer. William W. Kelly, in his 1964 bibliography of Ellen Glasgow, had previously grouped impressions into "plate sets" (see pp. 14-17, 36-38, 79-82).


With the prefix "sub-", suggesting subordination, it is hard to see how subedition could be taken to imply "that type has been reset," simply because "edition" involves resetting. Rather, a term subordinate to edition but yet containing the word "edition," and thus encompassing the possibility of separate impressions, seems ideal to stand for discrete groups of impressions within an edition.


He does, however, mention the possibility that another level, "shootings," should also be taken into account, since more than one set of offset plates can be made from the same photographic shooting (p. 266).


If a bibliographer, he continues, "suspects that replating has occurred but cannot prove it, he should use the four basic terms and record his suspicions in a note."


Subedition cannot, however, be limited to such readily recognizable differences. If, for instance, a firm distributes a whole impression, or a series of impressions, of an edition originally published by a different firm and if (as might happen in book-club distribution) it does not alter the imprint or call attention to itself through a series title, the result is nevertheless a subedition.


That the bibliographer is concerned with published objects is further reflected in the fact that chronological order (when it is adopted) usually means the order of publication, not of composition.


Cf. G. T. Tanselle, "The Description of Non-Letterpress Material in Books," SB, 35 (1982), 1-42.


It seems, from entry A3.I.b.1 in West's Styron bibliography, that the name of the Book Find Club does not appear in copies of this impression—which is therefore an example of a subedition not immediately recognizable as such (see note 24 above).


A further occasion for mixing publishers together would be the situation in which more than one publisher used a single set of plates, a situation that does not arise in this Styron example. If it did, the name of the publisher would have to be associated with the impression, not with the plating as in the outline above (which follows West.)


Traditionally the term subedition is not applied to the impressions of the originating publisher, from which all the other "subsidiary" groups of impressions must ultimately derive. Logically, however, when an edition (as the totality of all the impressions from a single act of typographic composition) is divided into subeditions, that first division, like any other group of impressions linked by publishing auspices, would have to be considered a subedition as well.


Obviously this problem reaches significant proportions only when two or more countries with well-developed publishing industries use the same language. It is, therefore, of great importance in the bibliographical and textual study of works written in English.


States need not be brought into the discussion, for they are variations in parts of books and do not refer to books as wholes. Any states within an issue or an impression must naturally be taken up within the entries for those issues or impressions; a state cannot be given an entry of its own, for it does not refer to a published entity. See the definition of state in PBSA, 69 (1975), 65-66, and my comments on the relation of state to ideal copy in "The Concept of Ideal Copy," SB, 33 (1980), 18-53 (esp. pp. 29-30).


This information would presumably be retrievable in any case from the index—where one would normally expect to turn for similar references to particular printers. The necessity of a thorough index—containing all personal and firm names as well as the titles of all works—goes without saying.


I have made further comments along the same lines in the fourth section of "The Bibliographical Concepts of Issue and State" (see note 13 above), pp. 56-65, which includes discussion of the relation of this approach to that of Bowers in the Principles.


Arranging whole subeditions by country under each edition would serve little purpose, unless there were no transatlantic issues of any of the impressions and no other editions of the same work. One could similarly argue that ordering whole editions of a given work by country of origin would be of limited utility if any of those editions included impressions published in a different country.


After reference numbers are assigned to entries, such cross references can be made more explicit by adding the relevant entry numbers.


That is, any separately published work or group of works.


There are exceptions, of course, as in William W. Kelly's bibliography of Ellen Glasgow (1964) and Edwin T. Bowden's bibliography of James Thurber (1968), in which the record of periodical contributions precedes that of contributions to books.


Roy Stokes, in his second revision (1967) of Arundell Esdaile's A Student's Manual of Bibliography (1931)—which from the beginning has had a chapter on "The Arrangement of Bibliographies"—cites and accepts unquestioningly the Soho division of material into categories: he asserts that "the physical nature of the material dictates certain aspects of the arrangement" (p. 285) and claims that the difference in treatment reflects "the varying importance of different kinds of material" (p. 284). Cf. his chapter on arrangement in The Function of Bibliography (1969), esp. p. 126.


On periodicals and anthologies, see G. T. Tanselle, "Non-Firsts," in Collectible Books: Some New Paths, ed. Jean Peters (1979), pp. 1-31 (esp. pp. 20-26).


In contrast, Greg, in his pre-Restoration bibliography (see note 12 above), enters the appearances of plays in collections in the same chronological sequence with plays printed separately ("since the object was to collect the whole of the textual evidence concerning each play").


There are other possibilities, of course. One could, for instance, extend the chronological principle further and deal with each new typesetting under the year in which it appeared, thus splitting up the treatment of different editions of the same piece. (On the annalistic approach, see part I above.)


Sometimes it is helpful to provide in a separate section a record of the collected editions of an author's works, so long as individual volume-length works are also treated at the appropriate places in the main section. See, for example, the handling of this matter in Joel Myerson's bibliography of Emerson (1982).


That is, in a volume other than one composed entirely of the author's own collected pieces—which would obviously be described in the section recording the author's books.


The textual importance of periodical appearances—in particular, the various appearances of syndicated newspaper pieces (and the different editions of the newspapers in which they appeared)—is made clear in Fredson Bowers's "Four Faces of Bibliography," Papers of the Bibliographical Society of Canada, 10 (1971), 33-45 (esp. pp. 34-38).


This point is also made (in a somewhat different way) by James L. W. West III in "'Section B' and the Bibliographer," forthcoming in Analytical & Enumerative Bibliography.


In their bibliography of Hart Crane (1972), Joseph Schwartz and Robert C. Schweik divide "Works Not Published Separately" into "Poetry," "Prose," and "Letters," with periodicals as well as books included under each heading. (The order of the entries in these sections is alphabetical by the title of the work, and the entries include reference to the later printings of each work.)


"Degressive Bibliography," Transactions of the Bibliographical Society, 9 (1906-9), 53-65. The way the idea is now applied, however, is often quite different from what Madan had in mind.


See Fredson Bowers's thoughtful treatment in "Bibliography Revisited," Library, 5th ser., 24 (1969), 89-128 (reprinted in his Essays in Bibliography, Text, and Editing [1975], pp. 151-195). (His discussion includes comments on a series of letters on "The Degressive Principle" in the Times Literary Supplement on 4 and 11 August and 1 and 22 September 1966 [pp. 716, 732, 781, 884].)


These questions have occasionally—but not often enough—been asked before. R. J. Roberts makes a forceful statement on this subject in his review of George Lilley's bibliography (1974) of J. M. Murry (Book Collector, 25 [1976], 117-118): "One of the logical flaws of bibliographical presentation in the 'Soho' tradition was that the full range of descriptive techniques was applied to anything that appeared separately in its own covers—an obvious concession to bibliophilic interests—while something which appeared in a periodical, though of equal importance textually . . . only received check-list treatment."


Such bibliographies, with their organizing emphasis on publications rather than authors, would be analogous to bibliographies of the output of printing and publishing firms. (Whereas descriptions in such bibliographies would naturally include full listings of contents, the extended physical description of periodicals and authologies in author bibliographies could not be expected in many cases to include detailed records of contents.)


Bowers effectively expresses this position in the Principles: "From the point of view of a descriptive bibliography which should be concerned with books as material objects as well as with the literary, textual, and publishing history of an author, the tenth impression may be as important as the first; and the details of its physical appearance, contents, and the number of copies printed are just as interesting. . . . Subsequent editions are all a part of the life history of an author and his books" (pp. 379-380). Cf. G. T. Tanselle, "Non-Firsts" (note 39 above) and "Bibliographers and the Library," Library Trends, 25 (1976-77), 745-762 (esp. pp. 757-759).


This system still occurs, as in the recent (1982) Soho Bibliography of Jane Austen by David Gilson.


He also indicated states with subscript figures (e.g., "A8.1.b2"). Since state refers to a part of a book and since copies of an impression or issue may contain different mixtures of states of individual sheets, one may question whether states should be the subject of separate entries and hence whether entry numbers for them are required. (See note 31 above.) Of course, whether states can be the subject of separate entries is a different question from whether it may be convenient to have a shorthand means for referring to copies containing particular states.


"A System of Bibliographical Reference Numbering," PBSA, 69 (1975), 67-74.


This practice introduces some uncertainty into the reference, since the item with which it is simultaneous is not specified; there is also a potential ambiguity when more than one set of simultaneous editions or issues exist.


Which he says that Joseph Katz created and allowed him to use first (p. xxxii).


As Patricia C. Willis and Clive E. Driver point out in more detail in their criticism of this aspect of Abbott's proposal; see their "Bibliographical Numbering and Marianne Moore," PBSA, 70 (1976), 261-263. (They do not, however, question the idea that a system of reference numbering "should accurately represent the history of a text.") Actually the nature of the reference numbering is irrelevant to Abbott's suggestion, since any kind of numbers could be joined with plus and equals signs to signify the presence or absence of revision in the items referred to.


In Computers and Early Books (1974), pp. 95-99, and revised in Fingerprint Newsletter, No. 1 (1981), where a fingerprint is defined as "four groups of two pairs of symbols taken from the last and penultimate lines of four specified pages, one pair of symbols per line, taken from each bibliographical unit" (p. 9).


A distinction already blurred enough in library cataloguing practice. See my comments in "Descriptive Bibliography and Library Cataloguing," SB, 30 (1977), 1-56 (reprinted in Selected Studies in Bibliography [1979], pp. 37-92), and in "The Concept of Ideal Copy" (note 31 above), pp. 47-53.


One cannot exaggerate the importance of this section of a bibliographical description. A record of copies examined, with defects noted, makes clear the specific body of evidence on which the bibliographical account is based.


In Joel Myerson's bibliography of Emerson (1982), section A occupies 537 out of 802 pages.


Indeed, most of the proposed schemes for bibliographical reference numbering are based on this fact, even if they favor a consistent function for a prefixed letter, since they posit more elaborate numbers for sections of a bibliography involving fuller descriptions.


A statement in words (often including still more information, such as the name of the publisher of a subedition) would of course also appear in the heading of each entry.


If entry numbers were assigned to issues, it would be hard to avoid adding lower-case letters to the impression numbers; but the whole designation would then be cumbersome enough to raise doubts about its value. Sometimes one might feel that entry numbers did not need to be given to issues: under the entry for an impression, there could simply be subheadings in words for issues (and the index would then not distinguish issues, or would include issue indicators in words).


Or, if the number of editions is too large, using numbers (following periods) for that function also. (There is no reason why the use of prefixed letters could not be limited to the later sections.) An example of the use of letters for a category that includes more than 26 members occurs in Donald Gallup's revised bibliography of Ezra Pound (1983), where references "E2za" through "E2zi" follow "E2z"; one could argue that references on the order of "E2.27" would be neater and more convenient (or, since section E has only seven subdivisions, perhaps "Eb27"; or yet again, simply "F27").