A significant part of my Dewey life these past three years has been spent in Table 4 and the 400s. In that period of time, eight—count them, eight!—EPC exhibits (most developed with my colleague Juli Beall) have focused on aspects of Table 4 and the 400s. Work done in these two contexts (or maybe it’s just one context) has also spawned three blog entries during the same period. One, named Language, Language Everywhere, commented on the underlying importance of language to much of what we do. The other two solicited feedback on discussion papers in computational linguistics and language variations.
A number of changes in Table 4 and the 400s have recently been incorporated into WebDewey, some of which were announced in the September 2009 New and Changed Entries (in Word and PDF formats); some of which were announced in the December 2009 New and Changed Entries (in Word and PDF formats); and, as previously announced, some of which (changes with respect to language variations) will not be implemented until the publication of DDC 23. This entry will address changes in the September New and Changed Entries posting; a subsequent entry will address changes in the December New and Changed Entries posting.
The September 2009 New and Changed Entries posting included significant changes with respect to computational linguistics, corpus linguistics, and pragmatics. As indicated in the previous blog entry on computational linguistics, we have relocated computational linguistics from 410.285 Linguistics computer applications to 006.35 Natural language processing, where computational linguistics now appears in the class-here note. This (006.35) is now the number in which general works on computational linguistics should be classed. Examples of such works include Speech and language processing: an introduction to natural language processing, computational linguistics, and speech recognition and Survey of the state of the art in human language technology. According to the rule of application, works about the application of one subject (in this case, computational linguistics) to another subject, should be classed with the second subject. The relocation opens up the possibility of expressing computational linguistics applications through the addition of standard subdivision notation T1–0285635 Computational linguistic applications (built with T1—0285 Computer applications plus notation 635 from 006.35 Natural language processing, following instructions at T1—0285). Two such built numbers are given in the New and Changed Entries: 025.410285635 Automatic abstracting / automatic text summarization (built with 025.41 Abstracting plus notation 0285635 from Table 1, as analyzed above) and 401.430285635 Word sense disambiguation (built with 401.43 Semantics plus notation 0285635 from Table 1, as analyzed above).
We have expanded for corpus linguistics at 410.188 Corpus linguistics (under 410.18 Schools, theories, methodologies of linguistics, which itself is a previous expansion under 410.1 Philosophy and theory of linguistics). In the past, many works on corpus linguistics have been classed with works on computational linguistics, since much of the research in corpus linguistics is done automatically, using computational linguistics techniques. But this is not optimal, for multiple reasons. For example, corpus linguistics is a methodology (not a set of techniques), based on the premise that linguistic study should be carried out on language as actually used rather than on language as generated introspectively (“arm-chair linguistics”). Moreover, corpus linguistics includes aspects (e.g., selection of the language samples that will comprise the corpus) that may have nothing to do with computational linguistics. Comprehensive works on corpus linguistics (like Corpus linguistics and the web) should now be classed in 410.188. As a methodology, corpus linguistics is also subject to the rule of application. Hence a work on corpus linguistics applied to a specific topic in linguistics would be classed with the topic; for example, Discourse on the move: using corpus analysis to describe discourse structure should be classed in 401.41 Discourse analysis. And a work on corpus linguistics applied to a specific work or the works of a specific author should be classed with the work or author; for example, e-Lears: a Corpus Approach to Shakespeare and Tate should be classed in 822.33 William Shakespeare. (The Tate adaptation of King Lear would be classed in 822.4 Drama of Restoration period, 1625-1702; as the work treats Shakespeare and Tate equally, it is the first-of-two rule that tips the balance in favor of the Shakespeare number, which comes earlier in the schedules.)
We have also expanded at 401.45 for Pragmatics, with subdivisions under it for 401.452 Speech acts (“Class here illocutionary acts”), 401.454 Presupposition (“Class here implication, entailment”), and 401.456 Reference (“Class here anaphora, deixis”). In the past, the interdisciplinary number for pragmatics has been 306.44 Language (as an aspect of culture), because of the association between pragmatics and sociolinguistics. An analysis of interdisciplinary works on pragmatics revealed that the better home for such works is in linguistics. Therefore, 401.45 has now been designated as the appropriate number for interdisciplinary works on pragmatics. (In addition to the association of pragmatics with sociolinguistics, there is a strong association of some of the subtopics of pragmatics with philosophy. Class-elsewhere notes in the subdivisions of 401.45 remind the classifier to consider discipline first.) Examples of works that should class in 401.45 and its subdivisions are: Pragmatics and natural language understanding in 401.45 Pragmatics, Using language: The structures of speech acts in 401.452 Speech acts, Presupposition and implicature in compositional semantics in 401.454 Presupposition, and Studies in anaphora in 401.456 Reference.
Comparable changes, as appropriate, have been made in Table 4.