Prosodic Bootstrapping of Clauses: Is it Language-Specific?

Introduction It has long been observed that prosody and syntax interact in our spoken language. For example, prosodic modulations such as pausing, final-syllable lengthening, and pitch resets occur reliably at clause and major phrase boundaries, particularly in infant-directed speech (Broen, 1972; Bernstein Ratner, 1986; Fernald et al., 1989; Soderstrom et al., 2008). Proponents of the prosodic bootstrapping hypothesis argue that infants may be able to take advantage of the correlations that exist between prosodic cues and syntactic boundaries to get a head-start in syntax acquisition (Gleitman & Wanner, 1982; Morgan, 1986). Since prosodic information is readily available in the acoustic signal and infants as young as 4.5 months have demonstrated sensitivity to correlations between prosodic information and syntactic boundaries (Jusczyk, 1989), prosody may facilitate the earliest stages of syntax acquisition. Recent work in this area has asked whether prosodic structure influences infants' memory for strings of words. Several studies (e.g., Mandel et al., 1994; Soderstrom et al., 2005) have found that infants more easily remember strings of words when they are initially presented with well-formed prosody; for example, as a sentence rather than in a list. These studies suggest that the prosodic grouping of words influences how we remember those words, but they leave open the question of whether prosodic grouping influences syntactic learning. In this paper, we explore two aspects of the prosodic bootstrapping hypothesis. First, we build on our previous work (Hawthorne & Gerken, 2011) and ask whether infants treat words within prosodic units as more cohesive and constituent-like than words that cross a prosodic boundary. More specifically, we ask whether infants recognize words with clause-like prosody (i.e., intonational phrase prosody) when they have moved to a new position in the sentence. One property of certain syntactic constituents in natural languages is that they can move, as seen in Examples (1) and (2) (brackets represent clausal constituents).