Why There Are No Clitics: An Alternative Perspective on Pronominal Allomorphy (review)

Abstract

Reviewed by: Why there are no clitics: An Alternative Perspective on Pronominal Allomorphy by Daniel L. Everett Joel A. Nevis Why there are no clitics: An alternative perspective on pronominal allomorphy. By Daniel L. Everett. (Summer Institute of Linguistics and The University of Texas at Arlington Publications in Linguistics 123.) Dallas, TX: Summer Institute of Linguistics & University of Texas at Arlington, 1996. Pp. xii, 188. $25.00. Everett offers a unified theory of pronouns, pronominal clitics, and agreement affixes in which there is a single semantic core with derivation of the syntax of each ‘allomorph’. E argues forcibly [End Page 162] that there is little need for a separate category for pronominal clitics and hints in his title that grammatical theory can do away with the category of clitic altogether. He argues that the lexicon contains no pronouns, clitics, or agreement affixes per se; instead, phi-features store the appropriate information. The differences among the three lie in the insertion of bundles of these phi-features into distinct syntactic positions, and syntax alone is responsible for their diverse morphosyntactic and morphological behaviors. E invokes the minimalist theory but utilizes terminology from principles and parameters for ease of exposition. He is explicit in his attempt to engage both theoreticians and fieldworkers, promising the latter a grasp of the issues and tools for analysis. In this theory-oriented approach, E’s monograph appears a noteworthy departure for its series. The book consists of nine chapters. After a short introduction (1–5), Ch. 2, ‘Storage, insertion and form of phi-features’, develops the basics of the model (7–43). Ch. 3, ‘Clitic doubling’, implements the model for detailed analyses of doubling in Spanish, Celtic, Pirahã, and Yagua (45–114). Ch. 4 is ‘Subject doubling in French and Northern Italian dialects’ (115–29). Chs. 5 and 6 explore two issues pertinent to Romance languages: ‘Romance se’ (131–54) and ‘Pronominal determiners in Romance’ (155–63). Ch. 7, entitled ‘Clitics and binding’ (165–71), addresses the left dislocation of clitics in Italian, Romanian wh-movement, and binding in Brazilian Portuguese. Ch. 8, ‘Acquisition of clitics’ (173–78), makes predictions about sequences in children’s acquisition of affixes and clitics. The book concludes with a succinct summary (179–80). There is a fairly comprehensive bibliography for Chomskyan works in the 1980s and up to 1992—but coverage is rather limited outside of this tradition. The book lacks an index. The second and third chapters comprise the greater part of E’s work. In Ch. 2 E develops his model, according to which clitic, pronoun, and agreement are not treated as grammatical primitives, lexical entries, or ontological categories. Instead, E invokes phi-features (i.e. inflectional categories such as case, person, number, and gender) to be the basic lexical entries inserted into D-structure. Phi-features differ from other lexical entries in that phi-features do not permit morphological subcategorization (23–25). Note that E employs the term ‘clitic’ as a convenient label throughout the book, even when he is arguing that the clitic does not exist as a separate category. This turns out to be rather deceptive to the reader and undermines a central point. Phi-features are inserted into syntax anywhere they are allowed, but there will be restrictions on the minimal number of phi-features for AGR (18). Cross-linguistic variation in the presence or absence of particular phi-features is reduced to differences in lexical entries, and since AGR and determiner are always filled by phi-features, other differences lie in whether phi-features are spelled out or not. Moreover, if a morphologically subcategorized AGR position is linked to a theta-role, a nominative pattern will be derived; if not, an ergative system will emerge. E predicts that a nominative-accusative language could have object clitics with subject agreement and an ergative-absolutive language could show subject clitics with object agreement, but the theory rules out the reverse. Allomorphy is significant here, defined by E as a relation of complementary distribution in configurational morphosyntax. That is, he, him, and third person singular -s are allomorphs under this scheme (though E notes that -s further offers tense, mood, and aspect information and constitutes a...