Computational Lexicology is the use of computers in the study of the lexicon. It has been more narrowly described by Amsler [1980] as the use of computers in the study of machine-readable dictionaries. It is distinguished from Computational Lexicography, which more properly would be the use of computers in the construction of dictionaries, though some researchers have used Computational Lexicography as synonymous with Computational Lexicology.

In any case, it was the appearance of machine-readable dictionaries (MRDs) that gave Computational Lexicology its start as a separate discipline within Computational Linguistics. The first widely distributed MRDs were the Merriam-Webster Seventh Collegiate (W7) and the Merriam-Webster New Pocket Dictionary (MPD). Both were produced by a government-funded project at Systems Development Corporation under the direction of John Olney. They were manually keyboarded as no typesetting tapes of either book were available. Originally each was distributed on multiple reels of magnetic tape as card images with each separate word of each definition on a separate punch card with numerous special codes indicating the details of its usage in the printed dictionary. Olney outlined a grand plan for the analysis of the definitions in the dictionary, but his project expired before the anslysis could be carried out. Robert Amsler at the University of Texas at Austin resumed the analysis and completed a taxonomic description of the Pocket Dictionary under NSF funding, however his project expired before the taxonomic data could be distributed. Roy Byrd et al. at IBM Yorktown Heights resumed analysis of the Webster's Seventh Collegiate following Amsler's work. Finally, in the 1980s at Bellcore, George Miller and Christiene Fellbaum completed the creation and wide distribution of a dictionary's in the WordNet project, which today stands as among the most widely distributed computational lexicology resource.

Computational lexicology has contributed to our understanding of the content and limitations of print dictionaries for computational purposes. Basically, almost every portion of a print dictionary entry has been studied ranging from what constitutes a headword, what variants and inflections it forms, how it is delimited into syllables, how it is pronunced, the parts of speech it takes on, any special subject or usage codes assigned to the headword, the headword's definitions and their syntax, the etymology and its use to characterize vocabulary by languages of origin, the example sentences, the run-ons (additional words and multi-word expressions) that are formed from the headword, and related words such as synonyms and antonyms.

Speech researchers looked at the use of the pronunciations in machine-readable dictionaries for a source of spoken language. Following the work on English language machine-readable dictionaries, researchers looked at bilingual dictionaries and pairing of multiple dictionaries to assist in machine translation.

Many computational linguists were disenchanted with print dictionaries as a resource for computational linguists because they lack sufficient syntactic and semantic information for computer programs. Speech generation systems did make use of pronunciations from machine-readable dictionaries and text content anslysis systems were built that used the subject codes of the Longman Dictionary of Contemporary English (LDOCE) to analysis document subject content. [There were also pre-computational lexicology projects such as the General Inquirer and others that performed content analysis of texts using hand-crafted subject tags associated with words in text, but as these didn't derive from machine-readable copies of general print dictionaries, they are not 'computational lexicology' as defined here]. The computational linguistic community has undertaken to create its own dictionary resources through projects such as the SEMCOR dictionary and the work of Fillmore.

2006-11-07T14:54:46Z

RAM:

'''Word Sense Disambiguation'''

Word sense disambiguation (WSD) is the ability of software to distinguish what sense of a word is being used in a textual context. In modern WSD systems, the senses are typically taken from some specified dictionary. In earlier systems the senses were more typically generic senses selected by the originators of the system. These days WordNet is the usual dictionary in question. WSD has been investigated in computational linguistics as a specific task for well over 40 years, though the acronym is newer. The SENSEVAL conferences have attempted to put Word Sense Disambiguation on an empirically measurable basis by hosting evaluations in which a given corpus of tagged word senses are created using WordNet's senses and participants attempt to recognize those senses after tuning their systems with a corpus of training data.

Word Sense Disambiguation has several debates within the field as to whether the senses offered in existing dictionaries are adequate to distinguish the subtle meanings used in text contexts and how to evaluate the overall performance of a WSD system. For example, does it make sense to describe an overall percentage accuracy to a WSD system or does evaluation require specific comparison of system performance on a word by word basis.

Among the earliest efforts at word sense disambiguation was the work of Kelly & Stone (Ref needed) who published a book explicitly listing their rules for disambiguation of word senses.