The Independence of Letter, Word, and Meaning Identification in Reading

Abstract

ed from context. Feature analytic model of meaning identification We suggest the term meaning identification to embrace both comprehension and the utilization of semantic redundancy in sequences of words because the process by which such information is extracted from text by the skilled reader may be regarded as essentially no different from the process by which he identifies letters or words. Featural information from the visual array is analyzed and allocated to reduce uncertainty in: a] the domain of letter categories, for letter identification b] the domain of word categories, for word identification or c] the domain of cognitive structure, for meaning identification. Obviously it cannot be asserted that the domain of meaning to which visual information is applied in meaning identification is a domain of categories in the same way that sets of letters or of words constitute domains. Rather, comprehension may be conceptualized as either a i)ermanent or temporary reorganization of cognitive structures or the rearrangement of feature specifications on semantic feature lists. It is also not easy to identify the mechanisms by which meaning identification reduces the set of word categories (or reduces the size and increases the number of criterial sets for a number of categories in the word domain) to produce the lowered information threshold for words in sentences. However, reluctance to speculate 3. We are aware that some of the sequential constraint in a sequence of words may often be syntactic as well as semantic, but the syntactic component is subordinate in normal language (Treisman, 1965) as well as difficult to disentangle. It is convenient to subsume syntactic and semantic information in the terms meaning and compre-

Cite this work

No public copy is stored for this record, so only the metadata is cited.

Stored only when both papers are in the corpus — 34 citations exist upstream.

Papers that cite the same references — the relational fallback while embeddings are still backfilling.