Analogy and English Adjective Comparison

Analogy and English Adjective Comparison,Dirk Elzinga

Analogy and English Adjective Comparison  
BibTex | RIS | RefWorks Download
There are two strategies for forming the comparative degree of adjectives in English: a synthetic strategy which suffixes -er to the adjective stem, and an analytical strategy which uses more in composition with the adjective. While there have been many descriptions of the choice between analytical and synthetic comparative con­ structions (e.g. Quirk, et. al. 1985, Biber, et. al. 1999, Huddleston and Pullum, 2002), English adjective comparison has defied an account of sufficient predictive power to cout as an explanation. Traditionally, such approaches have consisted of formulating rules or constraints which capture generalizations over a range of data. Such rule- based accounts can only approximate actual linguistic behavior; problems such as the variability seen in English adjective comparison are accounted for only with great difficulty in rule-based approaches, if at all. In this paper I propose that analogy is the mechanism which English speakers use to form the comparative degree of adjectives. I show that an explicit and rigorous for­ mulation of analogical principles, Analogical Modeling (AM; Skousen 1989, 1992, 1995; Skousen, et. al. 2002), can account for the distribution of analytical and syn­ thetic comparison as well as traditional rule-based approaches. AM is an instance- based approach to predicting linguistic behavior. Rather than extracting generaliza­ tions from data in the form of rules which are then applied to novel forms, an in­ stance-based approach uses a set of stored examples which serve as possible models to predict the shape or behavior of a novel form. Using a computational implementation of AM I show that it provides an excellent fit to usage information on English adjective comparison gleaned from the World Wide Web by correctly predicting the preferred comparative construction 92.6% of the time. This lends credence to the idea that analogy can be an explanatory principle for linguistic behavior beyond its usual rule in structuralist descriptions in merely ac­ counting for exceptions. I also report on the results of an adjective comparison sur­ vey which was conducted among English speakers. AM is shown to fit the range of responses very well, coinciding with the survey participants 87.9% of the time and making the same sorts of mistakes actually found in the survey results. This demon­ strates that the AM account is psychologically plausible as well as being descript­ ively adequate.
Cumulative Annual
View Publication
The following links allow you to view full publications. These links are maintained by other sources not affiliated with Microsoft Academic Search.