A major purpose of this article is to examine first language (L1) reading theories from second language (L2) perspectives and, in so doing, to uncover significant research voids related to L2 problems. The article first considers the unique aspects of L2 reading in order to identify dimensions in which its theory must differ from accepted L1 constructs. It then discusses three fundamental distinctions that separate L1 and L2 reading: (a) the consequences of prior reading experience; (b) the effects of crosslinguistic processing; and (c) the compensatory devices stemming from the efforts of learners with limited linguistic knowledge to solve comprehension problems. Finally, several pedagogically oriented research themes which may be helpful in evaluating the instructional utility of current L2 reading theories are delineated.

This longitudinal study examines the acquisition of English articles by three 6-year-old, second language learning children whose native tongue is Chinese, a language without articles. Brown's coding scheme and an extended coding scheme were used in scoring the corpora of children's responses to a Syntax Elicitation Task. Results revealed that the Chinese children's acquisition of the definite article differed from- what had been previously found using Brown's coding scheme with English as first language learners and second language learning children of other native language origins. Chinese children's use of the definite article developed through an unmarked phase, a referential place-holding phase, a marked phase, and a referential substitution phase before the definite article was fully acquired. The acquisition of the indefinite article, on the other hand, was similar to the acquisition pattern already reported for children learning English as a first language or as a second language. It is suggested that referential place-holding, as well as referential substitution, might not be a Chinese-specific second language learning phenomenon; rather, they might be derived from a universal referential strategy for learning articles.

This article describes the development, use, and initial evaluation of a prototype computer system to enable nonspeaking persons with severe disabilities to engage in conversation on broad topics. The conversational aid produced (via a voice synthesizer) speech acts that were selected from a prestored menu, which was constructed by the user. Features of the system included facilities for switching the conversational perspective between the speaker and listener (i.e., “your experiences and views” vs. “my experiences and views”), providing a range of comments on what the other speaker had said, effecting repair when there was a conversational breakdown, and following predicted sequences of speech acts. The initial trials of the system produced dialogues that proceeded in a natural way and achieved encouraging conversational rates.

Two aphasics with a similar level of phonological production difficulty are compared to distinguish the properties of disruption to two stages in the phonological system for producing single words: activation of stored lexical-phonological representations versus construction of phonemic representations. A set of distinguishing behavioral features for breakdown at each stage is generated on the basis of a model of single word production. Important variables for analyzing output include: (a) the unit of phonological encoding (morpheme versus syllable), (b) the phonemic relationship between targets and responses, (c) the effects of target consonant-vowel (CV) structure, and (d) the level of pseudoword production. On a set of production tests, the expected behavioral pattern for impaired lexical-phonological activation was displayed by LW, while the expected behavioral pattern for impaired phonemic planning was displayed by CM.

Three studies were performed to explore the role of letter names in children's learning of correspondences between phonemes and graphemes. Some preschoolers and kindergartners were found to spell initial /w/ as y. They made this error because the name of y, /wai/, begins with /w/. Kindergartners in another study often said that y makes the sound /wdoubt/, even after they had been taught otherwise. Other results suggested that children benefit from letter-name knowledge to spell phonemes such as /b/, which occur at the beginning of a consonant-vowel letter name. Letter-name knowledge is less helpful for phonemes such as /1/, which occur at the end of a vowel-consonant letter name. Our findings suggest that children use their knowledge of letter names and their phonological segmentation skills when they learn the correspondences between phonemes and graphemes. They do not memorize these links in a rote, paired-associate manner.