The reign of typicality in semantic memory

Abstract

This paper begins with a brief description of a theoretical framework for semantic memory, in which processing is inherently sensitive to the varying typicality of its representations. The approach is then elaborated with particular regard to evidence from semantic dementia, a disorder resulting in relatively selective deterioration of conceptual knowledge, in which cognitive performance reveals ubiquitous effects of typicality. This applies to frankly semantic tasks (like object naming), where typicality can be gauged by the extent to which an object or concept is characterized by shared features in its category. It also applies in tasks apparently requiring only access to a 'surface' representation (such as lexical decision) or translation from one surface representation to another (like reading words aloud), where typicality is defined in terms of the structure of the surface domain(s). The effects of surface-domain typicality also appear early in the time course of word and object processing by normal participants, as revealed in event-related potential studies. These results suggest that perceptual and conceptual processing form an interactive continuum rather than distinct stages, and that typicality effects reign throughout this continuum.

Figure 1

Results demonstrating how SD patients' success in two verbal tasks, one productive and one receptive, varies as a function of word frequency, stimulus typicality and patient severity. For both tasks, severity refers to the fact that the 14 patients have been divided into two equal subgroups on the basis of scores on a word+picture comprehension test. For the task of reading aloud (on the left), all stimuli were words, and the condition labels refer to a combination of the words’ frequency (HiF, high frequency; LoF, low frequency) and their spelling–sound typicality (Typ means that the word's pronunciation is typical for its spelling and Atyp that the pronunciation is atypical for or unpredictable from the word's spelling). For the task of lexical decision (on the right), each trial consisted of a word+non-word pair. HiF and LoF in the condition labels refer to the frequency of the real word. The remaining part of each label refers to the relative orthographic typicality of the word/non-word pairs in that condition: W>NW means that the word had more typical orthographic structure than the non-word, W

Figure 2

SD patients' performance in three productive language tasks: (a) reading words aloud; (b) generating the past tense of verbs from their stem (present-tense) forms; (c) spelling words to dictation. For each of the four conditions formed by crossing word frequency with domain typicality, the figure shows, across the group of 14 patients, the proportions of responses that were correct, LARC errors or some other type of error. LARC stands for legitimate alternative rendering of components, and indicates that, although the patient's response to the stimulus is incorrect, it would be correct if the stimulus were a similar one from the same domain. Such errors mainly occur to atypical items, where the response represents a more typical rendering of the stimulus (e.g. reading the word hood as if it rhymed with ‘food’); but LARC errors can and occasionally do occur to typical items (e.g. reading the word food to rhyme with ‘hood’).

Figure 3

Drawings by an SD patient (DS) of (a) the stimulus picture of a rhinoceros: (b) the drawing is DS's copy of the picture with it present and (c) the drawing is his copy of the picture approximately 10 s after it had been removed from view.