The last entry in the list of books under review on page 85 was printed incorrectly. It should read as follows: Spiro, R.J., Bruce, B.C., & Brewer, W.F., (Eds.), Theoretical issues in reading comprehension. Hillsade, N.J.: Lawrence Erlbaum Associates, 1980

Groups of good and poor readers at second-grade level were tested for comprehension of adjectival constructions of the John is eager/easy to please types and of center-embedded relative clause constructions. The poor readers were inferior to good readers in understanding O-type adjectives (easy) but not S-type (eager). As well, they were poorer at comprehending embedded sentences, but only when the sentences described improbable events, ones which reversed the normal subject/object roles. When either noun could, on pragmatic grounds, assume either role, both groups fared equally well. The results are interpreted as casting doubt on recent assertions that deficient use of a phonetic memory code underlies the syntactic inferiority often seen in poor readers. A more pervasive linguistic immaturity is suggested as being involved.

Learning disabled children's pragmatic competence was examined on a task which allowed for the separate assessment of linguistic and social knowledge, that is, the ability to appropriately vary politeness and persuasive strategies as a function of listener status. Learning disabled and nondisabled children made requests of four imaginary listeners varying on the dimensions of intimacy and power. Learning disabled girls produced more polite requests than nondisabled girls to all listeners. Learning disabled boys produced a smaller variety of persuasive appeals and fewer appeals which represented sophisticated levels of listener perspective taking. Learning disabled boys' politeness and the elaboration of their persuasive appeals exhibited variations as a function of listener features, but attended to dimensions different from those reflected in nondisabled boys' requests. It appears that learning disabled boys produced less appropriate requesting strategies than nondisabled boys, even though their repertoire of linguistic forms for conveying politeness was not deficient. These results are discussed in terms of the interdependence of the social and linguistic knowledge underlying pragmatic development.

The Communicative Intention Inventory, an observational system for describing children's early gestural, vocal, and verbal communicative behaviors, is presented in this paper. The Inventory is comprised of eight intentional categories which are included because of their content and construct validity, their likelihood of occurrence in a clinical setting, and their high observerscorer reliability. The rank order of frequency of occurrence of the eight intentional categories is compared across 16 subjects resulting in a reliability coefficient of .66. The value of using the Inventory as a criterion-referenced measure of communicative intentions is explored.

Semantic factors affecting pronoun assignment were studied in a group of aphasie patients. Their ability to assign reference in sentences of the type John telephoned Bill because he needed some information was facilitated by the causal bias of the main verb known to influence the direction of pronoun assignment in normal speakers. The addition of gender constraints (e.g., John telephoned Sue because he needed some information) improved their performance as well. Explicit mention of the referent's name (e.g., John telephoned Bill because John needed some information) did not improve performance any more than did the addition of gender constraints. When these cues were eliminated so that pronoun assignment required a full syntactic analysis, their performance dropped to chance.

Memory for content, or gist recall, was studied in groups of normal and language-impaired children, matched on age and sex, using a story recall task. The children were pretested for their ability to comprehend individual sentences in the stories. Following a practice session, two test stories were read to each child and oral recall was requested immediately after each story. The content of the recall protocols was analyzed into propositions and scored for accuracy, organization, and temporal ordering. The results showed that the groups differed primarily in the amount of accurate recall, with the language-impaired group recalling considerably less than the normal group. It was concluded that language-impaired children do exhibit deficits in gist recall for material which is within their linguistic grasp.