While it was once uncontroversial to refer to a ‘Chomskyan revolution’ in linguistics, a number of commentators have recently questioned whether generativist theory truly represents a revolutionary departure from earlier approaches. This paper defends the view that a Chomskyan revolution has indeed occurred—a revolution which began with the publication of Syntactic structures (1957) and which has had profound effects, both intellectually for the study of language and sociologically for the field of linguistics. Paradoxically, however, the revolutionary success of generative grammarians has not resulted in their achieving ADMINISTRATIVE power in the field—a fact which is both documented and explained.

Despite recent claims to the contrary, noun incorporation has an undeniable syntactic reality in some languages. In Greenlandic Eskimo and Southern Tiwa, the incorporated nominal displays many syntactic, semantic, and discourse-functional features of independent nominals—features which much recent work would lead us to doubt could characterize proper subparts of words. Linguistic theory must therefore allow for some limited interpenetration of the two modules of syntax and morphology.

It has been suggested by Sadock 1986 that certain constructions in Greenlandic Eskimo and Southern Tiwa provide evidence against proposals of Mithun 1984 concerning the nature of noun incorporation. These objections are discussed, along with issues concerning the discourse salience, reference, and semantic interpretation of incorporated nouns.

The ‘stepping-stone’ model has been used to study genetic similarity of peoples as a function of the geographical distances separating them. A gene is represented by one of several alleles, much as a meaning is represented by one of many words. Based on this parallel, the stepping-stone model is here applied to a body of linguistic data from a chain of Micronesian islands. The logarithm of the lexical similarity, when plotted against geographical distance, shows a pronounced upper concavity not found in the genetic investigations. This deviation from linearity is largely caused by non-homogeneity of replacement rates in the words studied. Another contributing factor is that the effect of distance on lexical similarity is much greater for the eastern islands in the group.

Rates of replacement in space, computed here, show a significant positive correlation with those in time, computed for a related language or for different groups of languages.*

Morphological marking of grammatical relations may appear on either the head or the dependent member of the constituent (or on both, or on neither). Grammatical relations—and whole languages—may be classified according to their propensity for using one of these types of marking. Implicational relations among various marking patterns can be stated: languages display a tendency to use one type consistently throughout their grammar. The difference in patterns provides a typological metric and a functional explanation for certain word-order preferences. For historical linguistics, it provides a diagnostically conservative feature and a clue to genetic relatedness. Although the head-marked pattern is cross-linguistically favored, grammatical theory is strongly biased toward the dependent-marked patterns that happen to dominate in Indo-European.

This paper presents a theory of control (predication), in terms of thematic relations, which makes no use of the element PRO in the syntax. Important consequences of the theory are that the θ-Criterion must be relativized to particular local domains, and the Projection Principle cannot be maintained. A number of syntactic arguments against PRO are summarized, and the arguments of Koster & May 1982 in favor of PRO are addressed. It is concluded that, given a thematic relation-based account of predication, the Projection Principle in its current form is not a useful postulate in the theory of grammar.