Off-line resolution of adjunct ambiguity in a structural priming paradigm: Implications for the production-comprehension parallel hypothesis

Syntactic or structural priming is the facilitation in processing a linguistic structure due to prior processing of the same structure. The first experimental study on the effect (Bock, 1986) was inspired by naturalistic evidence that language users tend to choose the same syntactic structure repeatedly across successive utterances (Levelt and Kelter, 1982). This tendency has been also confirmed by quantitative corpus analyses (Gries, 2005). While the effect of structural priming has been studied mainly in language production, recently researchers’ interest has shifted to its study in comprehension. Most research on structural priming in comprehension has investigated the effect of reading a temporarily ambiguous sentence (prime) on the reading of a following, semantically unrelated sentence (target) containing the same kind of structural ambiguity. Experiments involving a self-paced reading task and eye-tracking techniques have shown that processing of the target sentence is facilitated, relative to a baseline control condition. However, until recently (Traxler, 2008), this type of priming has been found only with the lexical repetition of the main verb between the prime and target sentences. The present study aims at showing the effect of structural priming on the comprehension of the relative clause when it can be attached either high (HA) or low (LA) in Italian. Target sentences contain a complex noun phrase and a fully ambiguous relative clause (e.g., Il sarto della sposa che ha cambiato indirizzo starnutisce “The taylor of the bride who changed address sneezes”), which can have either a HA (sarto) or a LA (sposa) interpretation. Prime sentences have the same structure, but the attachment site for the relative clause is unambiguously determined by the presence of a third person clitic pronoun marked for gender (e.g., L’ammiratore della diva che gli ha mandato un bacio sospira, (LA) vs. L’ammiratore della diva che le ha mandato un bacio sospira, (HA)). In addition, in this study, no open-class lexical item was repeated between the prime and target sentences. The following two hypotheses were tested: - if structural priming in comprehension is independent of lexical repetition, thus confirming Traxler’ s (2008) results; - if there is an interaction between the effects of the eventual preferred interpretation of the relative clause (HA in Italian according to De Vincenzi and Job, 1993, 1995) and structural priming. The study consisted in two phases. In the first, a pre-test was carried out with 40 participants evaluating the 24 target sentences for comprehensibility and 21 participants answering a comprehension question in order to assess the eventual preference direction (HA vs. LA). In the second phase 32 participants were presented with the 24 experimental prime-target pairs of sentences each on a different sheet of paper. They had to reply to a question on the target sentence which showed their disambiguation choice (HA vs. LA). The data thus obtained are under analysis.