- Contiguity theory by Norvin Richards
This book can only be described as a tour de force. In Contiguity theory, Norvin Richards juggles a mind-boggling amount of data from a diverse variety of languages and develops a complex theoretical framework with far-reaching implications for the theory of grammar. The result is by no means an easy read: the theory is elaborate and unconventional, and fully appreciating the dense discussion requires proficiency in syntactic and phonological theory.
R’s point of departure is the observation that current syntactic theory ‘offers no real answer to the question of why … movements are distributed as they are among languages’ (1). He rejects the common practice of invoking formal features as ‘triggers’ of movement, which ‘have no detectable properties other than their ability to trigger overt movement’ (1) and hence offer little in the way of explanation. Whatever one’s evaluation of the theory that unfolds over the following 300+ pages, R is to be applauded for addressing head-on this glaring gap in current theorizing, while much work in syntax—ironically self-identifying as ‘minimalist’ in many cases—continues to content itself with restating generalizations about surface word order in terms of ad hoc features and phrase-structural templates, with little concern for the adverse implications of this rank growth of stipulations for the theory of universal grammar (Chomsky et al. 2017).
R’s book is an attempt to overcome this unsatisfying state of affairs. The central idea is that ‘apparent syntactic differences between languages are always the consequence of more fundamental [End Page 720] phonological and morphological parameters’ (2), so that ‘a complete description of a language is also a complete description of its syntax’ (344). The theory developed in this book, dubbed Contiguity theory after one of its central notions, is based on the assumption that prosodic structure and syntactic structure are built cyclically and in parallel; certain kinds of phonological information are visible to the narrow syntax and can effect displacement and other operations where required to derive a well-formed prosodic representation.
Chs. 2 and 3 are the central pillars of the monograph. Ch. 2 attempts to derive classical ‘EPP’ (extended projection principle) effects from principles of phonology that require affixes to be prosodically supported: the requirement affix support demands the presence of a metrical boundary in the direction in which an affix attaches. In languages such as Italian and Spanish, verbal stress reliably provides such a boundary to the immediate left of the suffixal tense morpheme. This is not the case in languages like English and French, R argues, which only compute the metrical structure of the verb once it is inflected; consequently, affix support must be satisfied by merging some XP to the edge of T. R shows how affix support allows for a reduction of EPP effects to independent parameters: head-directionality, the morphological status of T, and the presence or absence of metrical boundaries between T and the verb. For instance, in a head-initial language with prefixal T (e.g. Greek), no EPP effects obtain, since nothing is required to precede T. By contrast, in a head-initial language with suffixal T (e.g. French), EPP effects will obtain unless verbs provide the relevant metrical support (e.g. Spanish). A head-final suffixal T, like final heads in R’s model generally, requires ‘untethering’ (delinearization) of T and vP, which abrogates the metrical support for T otherwise provided by vP. This, R argues, motivates EPP effects in Japanese.
Ch. 3 turns to wh-movement and develops a theory of its crosslinguistic distribution, modifying earlier proposals in Richards 2010. The core idea is that wh-phrases must be prosodically contiguous to the clause-typing C, such that the wh-phrase and its associated C-head must be contained within a single phonological phrase, within which the wh-phrase is placed adjacent to a prosodically active edge. Contiguity can be established in different ways, depending on the directionality of active edges in a language and the placement of C relative to the wh-phrase. Languages...