Explaining the Final-over-final constraint: Formal and functional approaches
Michelle Sheehan
March 2012

Biberauer, Holmberg and Roberts (2007, 2008, 2010) posit the Final-over-Final Constraint, to capture a striking asymmetry in the attestation and grammaticality of disharmonic word orders. As an empirical generalization in its most abstract form, FOFC rules out the possibility of a head-final phrase immediately dominating a head-initial phrase. In this paper, two different explanations of this asymmetry are compared and contrasted: a ‘functional’ account derived from Hawkins’ (1994) Performance-Grammar Correspondence Hypothesis (PGCH) and a ‘formal’ account based on the syntax-PF interface, which attempts to derive FOFC from a version of the Linear Correspondence Axiom (LCA). The functional approach predicts FOFC to be a statistical effect, stemming from principles of efficient processing. While both accounts capture the basic asymmetry and (interestingly) overlap substantially in their predictions, they are by no means notational variants, and the focus of this paper is to test their empirical predictions. The first important difference concerns the status of apparent counterexamples to FOFC. It is proposed that while there is suggestive initial evidence that FOFC is merely a statistical effect, many of the apparent surface violations actually turn out to be illusory and the small number which remain cannot be considered a strong argument in favor of the PGCH. A second difference between the approaches concerns the correlation with harmony. It is proposed that instances where the relevant disharmonic gap does not correlate with a typological preference for ‘harmony’ create a serious problem for the PGCH-based analysis. The general conclusion is that harmonic tendencies and the FOFC gap cannot receive a unified explanation. This appears to disfavor an account of FOFC based on the PGCH, though it does not rule out a functional account per se. In fact, it is argued that the LCA might itself be partially ‘functional’ if the fact that it maps asymmetric c-command to precedence rather than subsequence is motivated by third-factor processing pressures. In these terms, FOFC is the arbitrary side effect of a partly functionally-determined linearization algorithm which uses narrow syntactic asymmetries to yield linearity.
Format: [ pdf ]
Reference: lingbuzz/001263
(please use that when you cite this article)
Published in: to appear in Theoretical approaches to disharmonic word orders
keywords: word order, lca, asymmetry, final-over-final constraint, syntax
previous versions: v1 [May 2011]
Downloaded:2352 times


[ edit this article | back to article list ]