The notion of deep structure has served to crystallize a number of recent controversies in syntactic theory. This paper examines several arguments which have appeared in the literature, chiefly in Chomsky's ‘Some empirical issues in the theory of transformational grammar’, in support of the notion. It concludes that these arguments are not compelling, and that the issue of the existence of deep structure is not an empirical one at present.

Data are presented to show that a grammar of English should generate NP's with internal structures like [NP [Det the][N planets][NP Mars and Venus]], as well as NP's with structures like [NP [NP the two moons of Mars][NP Phobos and Deimos]]. The transformation required by the analysis works in conjunction with Chomsky's Subjacency Condition to account for the semantic and syntactic properties of surface-structure forms.

Double nominative and discontinuous possessive structures in Chinese are investigated. It is demonstrated that base rules as postulated in current syntactic theories cannot adequately handle these cases. It is argued that predicates can consist of an entire sentence; i.e., VP can be written as S. With minor modifications, this proposed rule will also accommodate problematic ‘topic-comment’ sentences.

An Arabic theme–predicate structure may incorporate, as its predicate element, an embedded structure which itself consists of theme plus predicate. Authorities have not stated how far this process can be continued by further embeddings. Usage indicates, however, that a second embedding occurs only under certain limiting restrictions, and a third embedding not at all.

The basic formulation of a Telugu vowel syncope rule is discussed here, and certain peculiarities of its application to /CVC/ sequences containing a retroflex obstruent are pointed out. These peculiarities can be explained within the framework provided by the theory of natural assimilation in Schachter 1969. Situations arise where the syncope rule must choose the vowel to be deleted among several potentially deletable vowels. In certain cases of this sort, it is the output of a rule applying later than the syncope rule which provides a means of predicting which vowel is deleted. If we allow the syncope rule to look ahead in derivations at possible outputs, and impose on it a constraint derived partly from Schachter's assimilation theory, the phenomena of syncope can be explained.

Data from rural Southern white speech are compared with Vernacular Black English (VBE) in order to determine the relationship of these varieties. Copula absence and invariant be, the two features most often cited in arguments about black–white speech relations, are described from a viewpoint which admits structured variability as part of linguistic competence. Some lects of Southern white speech and VBE may use zero copula quite similarly, but there is considerable difference in the uses of invariant be; the ‘distributive’ function of be appears to be unique to VBE in this setting. There is evidence that certain aspects of copula absence in white Southern speech may have been taken from a decreolizing form of Black English.

This paper discusses the linguistic development of Genie, an adolescent girl who for most of her life underwent a degree of social isolation and experiential deprivation unparalleled in the reports of scientific investigation. This case touches on questions of profound interest to psychologists, philosophers, and linguists, including the relationship between cognition and language, the interdependence or autonomy of linguistic competence and performance, the mental abilities underlying language, proposed universal stages in language learning, the critical age for language acquisition, and the biological foundations of language.