in classification appeared to wane. The printed book catalogue, ar-
ranged in classed order, was being replaced by the dictionary card
catalogue. Classification was retained for shelving books, but shelf
order did not require minute subdivisions. The day of a classification
developed for an individual library was passing; the age of conform-
ity was dawning.
1. Francis Bacon, The Advancement of Learning (London: Dent,
2. B.C. Vickery, Classification and Indexing in Science (London:
Butterworth, Scientific Publication, 1958), p. 119.
3. System of Conrad Gesner, 1548
arts and sciences
4. Jean Gamier, Systema Bibliothecae Collegii Parisiensis Socie-
tatis Jesu (Paris: 1678)
5. Ismael Bouillaud, Catalogus Bibliothecae Thuanae (Paris: 1679)
6. Gabriel Martin, Bibliotheca Bultelliana (Paris: 1711)
7. Guillaume de Bure, Catalogue des Livres de la Bibliotheque de
feu M. le Due de la Valliere (Paris: 1783)
8. Jacques -Charles Brunet, Manuel du Librairie et de I' Amateur de
Livres (Paris: Didot, 1810)
9. Leo R. La Montagne, "Historical Background of Classification,"
The Subject Analysis of Library Materials, ed. M.F. Tauber (New
York: Columbia University School of Library Science, 1953) p. 17
10. Catalog of the Library of Yale College in New Haven (New
London: Printed by T. Green, 1743) p. A2. Quoted in Joe Kraus.
"The Book Collections of American College Libraries," (unpublished
Ph.D. dissertation, Graduate School of Library Science, University
of Illinois, 1959.)
11. The classes as shown in the above source are:
Logic, Rhetoric, Oratory, Poetry
Ancient, Natural, and Moral Philosophy
Anatomy, Physicks, Chirugery
Pneumatology and Metaphysics
Geography, Voyages, and Travels
Lives of Famous Men
Plays and Books of Diversion
12. Catalogus Bibliothecae Loganiae: Being a Choice Collection of
Books as well in the Oriental, Greek and Latin as in English, Italian,
Spanish, French, and other Languages. Given by the late James Logan,
Esq., of Philadelphia, for the use of the Publick, Numbered as they
now stand in the Library Built by him in Sixth Street over against the
State-House Square (Philadelphia: Printed by Peter Miller & Comp.,
13. The classes as shown by headings were:
Divinity and Ecclesiastical History
History, Antiquities, Geography, Chronology, etc.
Voyages and Travels
Orators, Poets, Fables, Romances, etc.
Physick, Mathematicks, and Natural History
Arts, Liberal and Mechanical, Magick, etc.
Medicine, Surgery and Chymistry'
14. Elnathan Hammond (comp.), A Catalogue of Books Belonging to
the Company of the Redwood Library, in Newport, on Rhode-Island
(Newport: Printed by S. Hall, 1764.)
15. The classes used for octavos were:
Divinity and Morality
Natural History, Mathematics, etc.
Arts, Liberal and Mechanic
Miscellanies, Politics, etc.
16. James Ranz. (Ph.D. dissertation in progress. Graduate School
of Library Science, University of Illinois)
17. Letter from Thomas Jefferson, Charlottesville to John Page,
Charlottesville, Feb. 21, 1770, The Writings of Thomas Jefferson
(Washington: Issued under the auspices of the Jefferson Memorial
Association, 1903) v. 5, p. 18.
18. Dumas Malone, Jefferson, the Virginian. (Boston: Little, Brown
& Co., 1948) v. 1, p. 401.
19. Letter from Thomas Jefferson, to James Ogilvie, Jan. 1806.
Quoted in La Montagne, op. cit., p. 19.
20. Ibid. "1. Antient history. 2. Modern. 3. Physics. 4. Nat. Hist,
proper. 5. Technical arts. 6. Ethics. 7. Jurisprudence. 8. Mathe-
matics. 9. Gardening, architecture, sculpture, music, poetry.
10. Oratory 11. Criticism. 12. Polygraphical."
21. J.A. Service, Thomas Jefferson and His Bibliographic Classifi-
cation (Chicago: University of Chicago, 1950) Microfilm, p. 70.
22. Catalogue of the Library of the United States (Washington:
Printed by Jonathan Elliot, 1815)
23. Benjamin Pierce (comp.), A Catalogue of Harvard University in
Cambridge, Massachusetts. (Cambridge: E.W. Metcalf and Co.,
24. The main classes were:
I Class I Theology
Class II Jurisprudence, Government, and Politics
| Class III Sciences and Arts
Class IV Belles Lettres
Class V History
Class VI Works Relating to America
25. Catalogue of Books Belonging to the Saint-Louis Mercantile Li-
brary Association (St. Louis: The Association, January 1950) pp. 226-
26. Ibid. p. v.
: . 27
27. Edward Wm. Johnson (comp.), Catalogue, Systematic and Analy-
tical of the Books of the Saint Louis Mercantile Association, Prepared
for the Board of Directors (St. Louis: Printed for the Association,
1858) p. viii.
28. Classified Catalogue of the Saint Louis Mercantile Library, with
index of authors (St. Louis: The Association, 1874)
29. A Classified Catalogue of the Mercantile Library of San Francis-
co: with an Index of Authors and Subjects; Consisting of About Fourteen
Thousand Volumes. Made by the Librarian. January, 1861 (San Fran-
cisco: Published by the Association, 1861)
30. The main classes were:
Class I Novels
Class II Religion
Class III Jurisprudence & Government
Class IV Philosophy, Science & the Arts
Class V Voyages, Travels, and Personal Adventures
Class VI History
Class VII Miscellaneous
Class VIE Polygraphy
31. N.B. Shurtleff, A Decimal System for the Arrangement and Ad-
ministration of Libraries (Boston: Privately printed, 1856)
32. Jacob Schwartz (comp.), Catalogue of the Apprentices' Library
Established and Supported by the General Society of Mechanics and
Tradesmen of the City of New York (New York: Chatterton and
Parker, Printers, 1874)
33. William Torrey Harris, "Book Classification." Journal of
Speculative Philosophy, IV (April, 1870), pp. 114-129.
34. William Torrey Harris, "Essay on the System of Classification."
Catalogue, Classified, and Alphabetical, of the Books of the St. Louis
Public Library (St. Louis: Missouri Democrat Book and Job Printing
House, 1870) p. xi.
35. Ibid. p. xiii
36. Catalogue of the Peoria Mercantile Library, Alphabetical and
Classified (Peoria: Published by the Library Association, 1872) p. xi.
37. Peoria Public Library Classified Catalogue (Peoria: Edward
Hine & Co., Printers, 1899.)
38. Ibid. p. i.
39. A Classification and Subject Index for Cataloging and Arranging
the Books and Pamphlets of a Library (Amherst, Mass.: Printed by
Carr, Lockwood & Brainard Co., 1870) p. 10.
40. Letter to Katharine Sharp from C.A. Cutter. Stamped date of
receipt, Nov. 14, 1898. (Original letter uncatalogued in University of
41. Charles Ammi Cutter, "The Expansive Classification", Tran-
sactions and Proceedings of the Second International Library Con-
gress, held in London, July 13-16, 1897 (London: 1898) p. 87.
42. Horace Kephart, "Classification," Papers Prepared for the
World's Library Congress, held at the Columbian Exposition, ed.
Melvil Dewey (Washington: Government Printing Office, 1896)
43. American Library Association, A Survey of Libraries in the
United States (Chicago: A.L.A., 1924-27) Vol. IV, p. 7.
44. William Parker Cutter, Charles Ammi Cutter (Chicago: Amer-
ican Library Association, 1931) p. 44.
45. Thelma Eaton, "Classification in College and University Li-
braries," College and Research Libraries, XVI (April, 1955), pp. 168-
46. Thelma Eaton, "The Classification of Books in Public Libraries,"
Classification in Theory and Practice (Champaign, 111.: Illini Union
47. Charles Ammi Cutter, Rules for a Printed Dictionary Catalogue
(Washington: Government Printing Office, 1876). Today's students
use the following edition of Cutter's Rules: Rules for a Dictionary
Catalog (4th ed., rewritten; Washington: Government Printing Office,
1904; Republished by The Library Association, London, 1948)
48. Charles Ammi Cutter, Two-figure Author Table (Boston: Li-
brary Bureau, 1906)
Charles Ammi Cutter, Three-figure Alfabetic Order Table
(Boston: Library Bureau, 1902.
49. Frederick B. Perkins and George Rankin (comps.), Catalogue
of the Public Library of the City of Fall River (Fall River: Press of
Fiske & Munroe, 1882)
50. Frederick B. Perkins, A Rational Classification of Literature
for Shelving and Cataloguing Books in a Library. With alphabetical
index. Revised ed. (San Francisco: Francis Valentine & Co., Printers,
51. Ibid. p. iii.
52. Grant 1850 Perkins 1882
I Mental and Moral Science Philosophy
Political Science Society
[ History and Geography History
\ Mathematics Biography
Natural Sciences Science
Medical Science Literature
53. Perkins, A Rational Classification .... p. iii.
54. Ibid. p. iv .
55. A Catalogue of the Books Belonging to the Library Company of
Philadelphia (Philadelphia: Printed by Zacharia Paulson, jr., 1789).
56. Lloyd P. Smith. A Plan for the Classification of Books as Well
in a Printed Catalogue as on the Shelves of a Library (Boston: Li-
brary Bureau, 1882).
57. Verner W. Clapp. "A.L.A. Member Number 13: A First Glance
at John Edmands." Library Quarterly, XXVI (January, 1956), p. 1-22.
58. John Edmands. Explanation of the New System of Classification,
Devised for the Mercantile Library of Philadelphia. (Philadelphia:
Mercantile Library of Philadelphia, 1882.)
59. Elizabeth V. Benyon. Class K. Law (Washington: Library of
60. System of Classification Index and Scheme for Numbering Books
(Minneapolis: Public Library, 1889).
61. Lloyd P. Smith. "Classification of Books," Library Journal,
VH (May, 1882), p. 173.
62. Clapp, op. cit., pp. 4-5.
63. J.N. Lamed. "A Nomenclature of Classification," Library
Journal, IX (April, 1884), pp. 62-69.
64. Joseph C. Rowell. Classification of Books in the Library (Berk-
eley, Calif.: University of California, 1894).
65. Letter from J.C. Rowell, Library of the University of California,
Berkeley, to Katharine L. Sharp, Director, Department of Library
Science, Armour Institute, Chicago. Nov. 16, 1896. (Original letter
bound in University of Illinois copy number 1, of title listed in note 64).
66. W.I. Fletcher, Library Classification (Boston: Roberts, 1894) p. 3
67. W.I. Fletcher, Public Libraries in America (London: Low, 1894).
68. Fletcher, Library Classification.
69. Letter, W.I. Fletcher, Amherst College, Massachusetts to
Katharine L. Sharp, Director, Dept. of Library Science, Armour In-
stitute, Chicago, Nov. 14, 1896. (Original letter bound in University
of Illinois copy of Fletcher, Library Classification],
70. Kephart,op. cit., p. 862.
71. Only three libraries reported the use of Poole's scheme, but
Indianapolis, which reported its scheme as movable, broad, later
referred to it as Poole's scheme. In reporting to Kephart, Omaha said
of its scheme "same as Indianapolis." This accounts for the number
of five Poole schemes given in the text.
Shadow or Substance
President, Documentation, Inc., Washington
The topic which has been assigned to me, "Classification Today-
Shadow or Substance," might more appropriately have come at the end
of the Institute, rather than the beginning. If I could convince you that
our pursuit of valid classifications was the pursuit of a shadow, there
would be no reason to listen to the papers on the remaining part of the
program. We could all pack up and go home. Hence, I must conclude
that when those who planned this Institute gave me this topic, they as-
sumed that regardless of what I might say about classification, I would
certainly be unable to demonstrate its ephemeral or shadowy nature
and that I would conclude that classification had substantial value for
librarianship and related information activities.
Confronted with this dilemma, it occurred to me that the way out
for an erstwhile student of logic like myself might be found in the first
instance not in examining the nature of shadows nor the nature of sub-
stances, but in examining the meaning of the connective between them,
namely, the logical operator "or." Most of us, when we think of the
word "or," think of it in the exclusive sense as meaning "either or,"
that is, the word used in this title, "Shadow or Substance," would or-
dinarily be interpreted to mean that if classification were substantial
it could not be shadowy, or if it were shadowy, it could not be substan-
tial. There is, however, another meaning of "or" which is the usual
meaning attributed to it in works of logic, where the "or" is taken as
meaning logical disjunction with reference to propositions and logical
sum with reference to classes. In this sense "or" means "and/or"
rather than "either or." Thus if I say "It will rain tomorrow or I will
stay home," both statements could be true; that is, it might rain tomor-
row and I could still stay home. Similarly, if I say of an item that it is
a member of the class A or B, it could be a member of A, a member
of B, or a member of AB, and the general proposition "X is a member
of A or B" is true in all three cases. This general proposition is only
false when the item is a member of neither A nor B. This logical re-
lation can be illustrated by the truth table for disjunction at the top of
the following page.
Now then, if we assume that the "or" in the title is the "or" of
logical disjunction, then it is possible for me to take the line that clas-
sification in some sense is substantial, in some sense is shadowy, and
A B A B
T T T
F T T
F F F
in some sense is both. My text, then, becomes one of indicating the
sense in which it is substantial and in warning against the sense in
which it is shadowy, so that you will be able to judge subsequent papers
in this Institute in this context.
We must admit in the beginning that the concern of librarianship
with problems of classification represents one of the oldest and strong-
est links of librarianship with basic intellectual and theoretical ques-
tions. As a first year student in library school many years ago, John
Lund and I found that questions of classification constituted an intellec-
tual oasis in a barren waste of learning how many spaces should go
between the author and title in descriptive cataloguing, or how one col-
lates a book when the publisher has gotten mixed up in his numbering
procedure. Hence, the earliest contribution I attempted to make to the
art or science of librarianship was a paper on classification. Some of
you may have read it. It was called "A Non- Expansive Classification
System" and it appeared in the Library Quarterly 1 over twenty years
ago. In this paper we took the line that a classification system cover-
ing all knowledge for all time was certainly chimerical or, as the title
assigned to me has it, "shadowy." Hence, we felt that in order to save
classification as an intellectual activity for librarianship, it would be
necessary to set up our major classes in terms of time divisions; that
is to say, the major classes we recommended, instead of being such
things as science, literature, art, etc., were historical epochs. We
thought it possible that within these hjstnrira.1 epochs one might con-
struct adequate classifications; and by this we meant not classifica-
tions of knowledge, but .c lass if i cations of library material itself. One
of the major considerations which led us to this conclusion was some-
thing we were taught very early in library school. We were taught that
the Dewey system was a theoretical system which attempted to legis-
late for books and that its pigeonholes were created independently of
a concern with the content of the pigeonholes. We were taught at the
same time that the great advantage of the Library of Congress classi-
fication system was that it was made from the books themselves and
based upon an empirical study of the material at the Library of Con-
gress. Hence one could say that the pigeonholes or classes in the Li-
brary of Congress classification system were actually designed to con-
tain the material in the Library and therefore one could predict an ex-
cellent fit. From such an argument, however, it is a simple matter to
draw the conclusion that the Library of Congress classification might
have fitted the contents of the Library of Congress at the time it was
made, but that for future materials to be received by the Library of
Congress, the classification system took on the same theoretical and
predictive character as the Dewey system; that is to say, once the pi-
geonholes were set up, all new material would have to be fitted into
the pre-existing pigeonholes. Hence, with reference to new material,
the Library of Congress system differs from the Dewey system only
in having different pigeonholes. It was this predictive character that
a classification system based on temporal epochs was designed to
avoid. We felt that the great virtue of the Library of Congress system,
namely, its development from an actual examination of the material to \
be classified, could always be retained if new classifications were set / \
up as required by the changing pattern of literature.
In the twenty years that have elapsed since this paper, I have seen
no reason to weaken its conclusions but I am now convinced that Dr.
Lund and I did not go far enough. At that time we did recognize a
changing pattern of literature. What we overlooked were the different
interests which might exist in the same historical epoch. Now we
would say that not only is it necessary to make classifications for dif-
ferent periods of time but that it is necessary to make classifications
for different special purposes.
It is not my intention to give you a biographical sketch at this time
based upon the various papers that I have written about classification,
but Anatole France once described literary criticism as "the adven-
tures of a soul among masterpieces." He meant by this expression
that the important thing about literary criticism was not the book crit-
icized, but the nature of the critic. Hence, I feel that I can best carry
out my assignment at this Institute by telling you of the various consid-
erations and the steps along the way which have led to my present con-
clusions about classification in librarianship.
In 1950 I was privileged to give an opening paper at a similar Insti-
tute, although the title of the Institute was different. I refer to the In-
stitute on Bibliographical Organization held at the University of
Chicago at the Fifteenth Annual Conference of the Graduate Library
School. At that Conference I was assigned the topic "The Functional
Approach to Bibliographical Organization." 2 In preparing that paper
I felt that my first obligation was to define the concept of function, and
following suggestions from the biological sciences, I concluded that
a function could only be defined in terms of purposes. Hence, a func-
tional approach to bibliographical organization could only mean an
analysis of bibliographical organization in terms of its purposes. I
concluded, then, that there were no universal purposes and hence that
there could not be any universal bibliographical organizations. This \
indicated that the Universal Decimal Classification is certainly "shad-
owy" since it assumes that the scientific and intellectual enterprises
of all men everywhere could be subsumed under a universal purpose.
Certainly we can say that an increase of knowledge is the universal
purpose of all scientific and intellectual endeavor, but what is requirec
here is not such a general and vague universal purpose but a universal
purpose in terms of which we can design and construct an elaborate
system of major classes, sub- classes, sub-sub-classes, etc., into |
which we can organize the products of all intellectual endeavor. I did
not believe then and I do not believe now that this is even a remotely
feasible enterprise. Hence our conclusion at that time was that differ-
ent individuals or different groups should determine the specific bibli-
ographical organization necessary to their own purposes. I suggested
that these special purposes might be related by having each special
group make its selection of major class, sub- class, sub- sub- class,
etc., from a common vocabulary. Let me say at this time, parentheti-
cally, that at the present time I despair of even such a universal appa-
ratus as a common vocabulary for all sciences.
Aside from my theoretical interest in the problem of classification,
I had learned a good deal about its nature and utility from working in
libraries, and one of the things I did learn from working in both large
and small libraries was that for most such institutions, classification
has become a method of shelving books and has ceased to be, if it ever
was, a way of organizing the information in such books. This was
brought home to me most clearly in my years at the Library of Con-
gress. The Library of Congress, as you know, has closed stacks. The
approach to the content of these stacks is through the standard type of
dictionary catalogue. Beyond the dictionary catalogue, those who con-
sult its collections use the standard type of printed bibliographies,
e.g., Chemical Abstracts, Physical Abstracts, the publications of H. W.
Wilson and Company, Public Affairs Information Service, etc. It there-
fore seemed that the effort expended in setting up and maintaining an
elaborate system of close classification is wasteful, since it has no
real impact on the users of the Library. Of course, I knew that there
remained within the system of American libraries a number of institu-
tions, such as the Crerar Library and the Engineering Society's Li-
brary in New York, which utilize classed catalogues, but it still re-
mains true that in general, classification is not a major tool for the
use of contents of libraries. Certainly I was also aware of the very
great value of classification as a method of arranging books in open
shelf libraries, mainly public libraries or small academic libraries;
but I felt in this instance that these classifications had a special pur-
pose, namely, making available to the general reader a rough break-
down of books which reflected a similar rough breakdown of the inter-
ests of the general reader, i.e., fiction, travel, science, religion, etc.
I should like at this time to refer to just one other previous paper
which we did on this subject. In 1953 we prepared a report for the
Office of Naval Research on "Machines and Classification in the Or-
ganization of Information." This report was published in Volume II,
Studies in Coordinate Indexing. In this report I raised the following
question: Why, in the face of a general decline in interest of problems
of classification in regular library organizations, was there such a re-
newed interest in the problem among documentation people and people
who were concerned with machine searching of information?
How then do we account for the renewed interest in classification as
a method of information control? Within the last few years, we have
witnessed the birth (and in some cases, the rapid death) of dozens of
new classification systems, among which we can name, The Story of
Classification for the Army Technical Reference Service; the Office of
Naval Research Project Status Classification; the Research and Devel-
opment Board Classification of Research Projects; the American Soci-
ety for Metals- Special Libraries Association Metallurgical Literature
Classification, and the Standard Aeronautical Indexing System. There
has been a revival of interest in the Universal Decimal Classification,
in the Patent Office Classification, and in Ranganathan's Colon Classi-
fication. Western Reserve University has labored for several years
and is still laboring on the development of "abstraction ladders" and
This renewed search for the solution to an unsolvable problem re-
sults from a paradox, namely, the promise of machine organization