Categories and relators: A new schema

Knowledge Organization

Volumn 21, Issue 4, 1994, Pages 189-198

  Perrcault, Jean ^a  

^a UNIVERSITY OF ALABAMA IN HUNTSVILLE   (United States)

Author keywords

[No Author keywords available]

Indexed keywords

EID: 21844499561     PISSN: 09437444     EISSN: None     Source Type: Journal    
DOI: None     Document Type: Article

References (27)

1. Taking this term in the broadest sense, to include all the structures comprehended in the various types 'analytico-syn-thetic', 'faceted', and 'free' -principally to avoid the strictures of J C Gardin's paper 'Free classifications and faceted classifications; their exploitation with computers' In: Classification Research: Proc. Int. Study Conf. on Classification Research, Elsinore, 1964, cd P Atherton, Copenhagen: Munksgaard 1965. 161-176.
2. Cf 'The need for a faceted classification as the basis of all methods of information retieval', reprinted in Proc. Int. Study Conf. on Classification for Information Retrieval (ISCCR), Dorking, 1957. London: ASLIB 1957. p. 137-147
3. Cf E de Grolier: A study of general categories applicable to classification and coding in documentation. Paris: UNESCO 1962.
4. E de Grolier: A study of general categories applicable to classification and coding in documentation. 1962. Ibidem, p.18-42 (Sect.11).
5. B C Vickery: Classification and indexing in science. London: Butterworth 1959. p. 186.
6. There is a fairly commonly held opinion that only in a truly enormous collection of documents does the need arise for relational terms (see for instance the comments by R A Fairthorne: Proc ISCCIR (cited in footnote 2) p. 107);
7. F W Lancaster: Some observations on the performance of EJC role indicators in a mechanised retrieval system. Spec. Libr. 1 (1964)No.10, p.696-701).
8. However, the Itek Laboratories' Summary of project activities (Program of Research on Information Searching Systems) (=IL-4000-17; NSF-C88), p.13, states that: 'Experiments were conducted where syntactic features of subject entries were ignored, and search was made only for co-occurrence of pertinent words within an entry. Results of searches made gave useful data. For example, (in one search) 60 percent of the responses were invalid, (in another) some 24 percent of the responses were invalid'. J.C Gardin states also In; SYNTOL.
9. New Brunswick, NJ: Rutgers University Graduate School Library Service 1965. p.54,
10. That: 'an earlier experiment showed that retrieval with unrelated descriptors in this same field leads to an appreciable percentage of false drops, ie, to a substantial fall in the relevance ratio'. He also cites R C Cros, J C Gardin, F Levy: Ľautomatisation des recherches documentaires. Paris: Gauthiers-Villars 1964. chapt. 5 and B, 3.1.
11. D Kervégant: Developpement de I 'analyse des relations dans la CDU. Quart. Bull. IAALD 3 (1958) p.l 11-116.
12. There is a good deal of similarity between this distinction and that of WCB Sayers between 'inner' and 'outer' 'forms' (see J Mills: A modern outline of library classification. London: Chapman & Hall 1960. p. 35.
13. See J M Perreaulťs essay A new devise for achieving hospitality in array. Amer. Doc. 16 (1965)No.3, p.245-246).
14. In J Mills: The Universal Decimal Classification. New Brunswick, NJ: Rutgers University Graduate School of Library Service, 1964), 61, an example is given of a four-element number representing 'Supersonic flow: Cones: Pressure gradient: Shear flow' -533.696.4 : 533.6.011.5 : 539.386 : 533.69.048.3 -comprising thirty numerical digits and twelve marks of punctuation. By use of the compounding apostrophe this could be reduced to twenty-three digits and ten marks -533.6'964.4'011.5'9.048.3 : 539.386. The fact that the order of the original elements (a:b:c:d) had been changed (to a:b:d:c) would make no difference in a mechanized search of a linear file, as long as the citation order was one determined by convenience alone and not by exigencies of meaning. (This device, of course, would be all the more likely to be suggested in the absence of a developed repertory of relators, since if several relators are appropriate to replace each of the colons in the original expression, it is wholly evident that the apostrophe could not replace them).
15. The two types are analogous to the two fashionable words 'roles' and 'links'. J C Gardin, commenting on the deficiencies arising in the use of simple roles, says (SYNTOL, p.27) that: 'A better answer is to do without roles altogether, and amplify links so that they convey the same information as roles and links taken together...' The cited passage came to my attention after the elaboration of the final form of the schema (figures 6-8), but it quite clearly expresses intentions identical to those that guided me. Footnotes 12-24 see under Fig.3
16. J C Gardin, SYNTOL. Cf ibidem, p. 135: 'Since each operator is in effect a category, each may express varieties of meaning.'
17. See Gardin in the work cited in footnotes 24 and 25, and his and R C Cros' Final report on a general system for the treatment of documentary data. Paris: Association Marc Bloch, 1963. p.1.
18. A more general statement of this tripartition might be affirmative, contrary, and contradictory; but affirmative is not actually a relation, but rather an attribute -and the same can be said of negative.
19. A convention must establish the position of such attributive usages with reference to the substantive code being modified; the examples given below, however, will refrain from such usage and hence from the need to establish such a convention.
20. As mentioned above, this schema was intended as the basis for a structural notation capable of forming complex classifications from a compound classificatory schedule, and the notating of it offered at least three choices: a: punctuation symbols, b: letters, c: numbers. The' first was attempted, but the results were so bizarre as to make optical scanning highly difficult. Letter-or number-combinations of the radix a-i or 1-9 are therefore recommended.
21. The UDC numbers used here are from the Trilingual Abridged Edition.
22. For the use of square brackets (as against the English usage), see K Fill: Einfühning in das Wesen der Dezimalklassifikation. 2nd ed. Berlin: Beuth 1960. p.20-21
23. A questionable point is whether it is necessary to specify classification and coding as areas within documentation; this inclusion is fairly nearly obvious.
24. See (as general background) J M Perreaulťs papers 'On bibliography and automation; or how to reinvent the catalog'. Libri 15 (1965)No.3, p.287-339 for a proposal oriented toward such a centralized activity.
25. The abstracts/codes used here to exemplify this problem were suggested by C David Batty.
26. Note that A-e-B may be taken to be properly permutable, since it means only that A and B are in some (indeterminate, in terms of content) relation, whereas A-ea-B is reversible into.B-ec-A, and A-eb-B indicates that/1 and B are interactive (ic, indeterminate in the orientational sense) relation (mutual therefore symmetrical).
27. This code should be added, as suggested by J C G Wesseling in On the Perreault schema, cited in the first footnote *, to mean 'passive presence'.