Looking beyond information structure

Evidence from Magnitude Estimation Test Experiments for Weight Effects on Slovak Word Order

Abstract

The principle of end-weight (Quirk et al. 1985) posits that language users prefer to place short (“light”) before long (“heavy”) constituents because this is easier to cognitively process (Hawkins 1994). Weight effects on constituent order have been discussed for well over a century in languages like English and German (cf. Behaghel 1909, 1930). In Slavic languages, however, they have received little attention so far (cf. Kizach 2012: 251).

Rather, the focus has been on information structure (Short 2002: 494). However, Goldberg’s Tenet #5 predicts that general cognitive restraints such as weight effects apply across languages (Goldberg 2003: 219). This article presents the results of a pilot study that investigates a phenomenon known as Heavy NP Shift, in which the _VP[NP PP] pattern changes to _VP[PP NP] when the NP is heavier than the PP.

Employing the Magnitude Estimation Test method (Hoffmann 2013), grammaticality acceptability ratings from 39 L1 Slovak speakers were elicited. The results show that Slovak is susceptible to weight effects, such that placing short before long constituents is always preferred. Moreover, the results provide evidence for the existence of a _VP[V NP PP] pattern in Slovak that has been identified as “basic” for English (Hawkins 1994: 20). This supports Hawkins’s (2004) Performance-Grammar Correspondence Hypothesis, which posits that grammaticalized patterns in analytic languages are the result of performance preference and therefore preferred in synthetic languages.