Claudia Felser, Theodore Marinis & Harald Clahsen
University of Essex
The present study compares the way English-speaking children and adults process temporarily ambiguous sentences such as (1):
(1) The student photographed the fans of the actress who was/were looking happy.
One possibility of accounting for both cross-linguistic and within-language variation in the processing of relative clause attachment ambiguities is the suggestion that attachment preferences are determined by the relative strength of the two interacting locality principles of Late Closure/Recency, which favours attachment of the relative clause to the second NP, and Predicate Proximity, which favours NP1 attachment (Gibson et al., 1996). A number of non-structural factors such as lexical properties of the preposition joining the two potential host NPs are also known to affect relative clause attachment preferences in adults (Gilboy et al., 1995).
29 six to seven year-old monolingual English children and 37 adult native speakers of English participated both in an auditory questionnaire study and a self-paced listening experiment (Ferreira et al., 1996). The on-line experiment had a 2 x 2 design with the factors 'Preposition' (of vs. with) and 'Attachment' (NP1 vs. NP2). The experimental sentences were disambiguated by number marking on the auxiliary (was vs. were). Additionally, the children underwent (i) a grammaticality judgement test, to ensure that they had acquired the relevant grammatical knowledge, and (ii) a listening-span task (Gaulin & Campbell, 1994).
The results of the on-line task can be summarised as follows:
Adult speakers' attachment preferences were influenced by the type of preposition: They showed a preference for NP1 attachment for complex NPs joined by of, and a preference for NP2 attachment for NPs joined by with.
For the children, we found a significant interaction between listening span and attachment preference: While children with a relatively high listening span showed a preference for NP1 attachment irrespective of the type of preposition, the low-span children showed a general tendency towards NP2 disambiguation.
The results from the adults indicate that the type of linking preposition influences on-line processing in adult native speakers. The results from the children, on the other hand, suggest that children are guided primarily by structural information during processing, while disregarding lexical cues. Whereas the high-span children relied mainly on the Predicate Proximity principle, the low-span group preferentially applied the Recency strategy. We argue that the observed differences between children and adults, as well as those found between the two span groups, are due to working memory differences rather than reflecting qualitative differences in the parser.
References
Ferreira, F., J. Henderson, M. Anes, P. Weeks, & D. McFarlane (1996). Effects of lexical frequency and syntactic complexity in spoken language comprehension: Evidence from the auditory moving window technique. Journal of Experimental Psychology: Learning, Memory and Cognition 22, 324-335.
Gaulin, C. & T. Campbell (1994). Procedure for assessing verbal working memory in normal school-age children: Some preliminary data. Perceptual and Motor Skills 79, 55-64.
Gibson, E., N. Pearlmutter, E. Canseco-Gonzalez, & G. Hickok (1996). Cross-linguistic attachment preferences: Evidence from English and Spanish. Cognition 59, 23-59.
Gilboy, E., J. M. Sopena, C. Clifton, & L. Frazier (1995). Argument structure and association preferences in Spanish and English compound NPs. Cognition 54, 131-167.