Orthographic Regularities in English Words

Abstract

The feature that most clearly distinguishes present day investigation of visual word recognition from studies of the same phenomenon done at the turn of the century is the current emphasis on within-word familiarity. Beginning with the sequential probability notions formalized by Claude Shannon in the late 1940s, a variety of different and often contradictory descriptions of within-word familiarity have been utilized in psychological studies. All such descriptions can be divided into two major classes: a statistical redundancy class which consists of quantitative approximations based on random samples of text words (word tokens), and a rule-governed regularity class consisting of rulebased approximations generalized from dictionary lists (word types). In the redundancy class are sequential probabilities, single-letter positional frequencies, and bigram and trigram counts. In the structure class are letter-sound rules and orthographic structure rules. Analysis of the differences within and across descriptions, especially as they relate to the generation of pseudowords, yields testable hypotheses for building information processing models of visual word recognition. Included in the purview of the paper will be those descriptions of English regularity which are based on the identity, as opposed to the graphic shape, of letters or letter sequences. This excludes, for example, descriptions based on word shape or on the graphic forms of particular letters. What remains will be labeled orthographic regularity, a phrase which is etymologically suited to the subject, but nevertheless ambiguous in its present day applications. Central to all descriptions of orthographic regularity is the assumption that through repeated exposure to printed words, readers acquire expectations for letters or letter sequences which normally occur at different word positions. Descriptions differ in the manner in which these expectations are defined, and often, by implication, the role they play in word recognition.