Clitic doubling and M-chains in Pirahã

From the introduction: "Recent studies of syntax have shown clitics to be a rich source of insights into a variety of principles governing the well-formedness of sentences (cf., inter alia, Kayne (1975); Steele, et al. (1981); Jaeggli (1982); Borer (1981); Kaisse (1982)). One particularly interesting focus of discussion continues to be the relationship between pronominal clitics and coreferential NPs in the same clause, so called clitic doubling constructions. [...]

"At least two points of consensus have been reached by studies on clitic doubling in the Government and Binding Theory (henceforth GBT) of Chomsky (1981; 1982). These are : (i) the basic properties of this phenomenon are derivable from the subtheories of Case, government, thematic θ roles, and binding and, closely related to this, (ii) clitic doubling is possible only when a special Case assigner appears to attribute Case to the doubled NP, since clitics "absorb" Case.

"The present paper may be seen as a contribution to this general discussion to the degree that we are able to establish that while the basic properties of clitic doubling in Pirahã are indeed derivable from subtheories of GBT, additional principles are required to explain why languages such as Pirahã can allow clitic doubling without a special Case assigner. [...]

"We will suggest that the analysis of clitic configurations in Pirahã offers important new insights into parameters governing clitic configurations in Universal Grammar (UG).

"The discussion is organized as follows. First, a basic sketch of Pirahã surface syntax is provided, focusing especially on intrasentential reference configurations such as those exemplified by [data] above. This is followed by a brief introduction to the relevant principles of GBT. Next, we propose an analysis of clitic configurations in Pirahã based on the notion of M-chains, by which, it is argued, clitics in languages like Pirahã transmit Case, θ roles, and other features to their doubled NPs. In the last section the predictions of this idea are tested with regard to Pirahã and shown to be superior to those of other recent analyses of clitics. Finally, an appendix is added in which we speculate on the implications of this analysis for the understanding of how clitics relate to the development of verbal affixes."