Prominence in Relative Clause Attachment: Evidence from Prosodic Priming

Abstract

This chapter presents two experiments utilizing prosodic adaptations of the structural priming paradigm. In each experiment, the goal was to explore the relation between the location of a prosodic boundary and the preferred parsing of a relative clause (RC) with ambiguous attachment to a preceding head noun. In Experiment 1, using read materials, ambiguous target sentences were preceded by prime sentences with RCs of different length: long, medium, and short. RC length was hypothesized to influence the location of an implicit prosodic boundary in the primes. However, no effect for this RC-length manipulation was found. In Experiment 2, the location of a boundary was manipulated in overt (spoken) prime sentences. For these auditorily-presented primes, the location of a prosodic boundary was found to influence attachment preference for targets. Interestingly, the effect was in the opposite direction as predicted: In the configuration NP1 NP2 RC, a boundary after NP2 resulted in more NP2 attachments. We propose that in the experimental materials, which contained equivalent accents on the two noun phrases (NPs), the boundary after NP2 leads to the accent on NP2 being interpreted as the nuclear pitch accent. Consequently, that accent was perceived as being more prominent than the accent on NP1, thus attracting RC attachment. The results suggest a close relationship between prosodic phrasing and prosodic prominence in English, and demonstrate a role for both in sentence processing.