How Adults and Children Produce Relative Clauses when Animacy and Thematic Role Expectations Clash

• In Event: Integrating Morphosyntactic and Discourse Cues to Process Sentences: Insights from Turkish, Cantonese, and English

Abstract

Subject relative clauses (RCs) have been reported in many studies on various languages to be easier to process than object RCs. Adults process subject RCs with greater efficiency and less difficulty; children often acquired them earlier, comprehended more accurately and produced more easily than object RCs. At the same time, there is another set of studies in the literature demonstrating that object RCs are not necessarily inherently difficult, the difficulty induced from processing is often due to a mismatch of expected features to the RC construction. If an object RC is formulated with the expected features, its processing disadvantage can be disappeared or even reversed. For example, Mak et al. (2002, 2006) found that with the right animacy feature(s) for the arguments involved in a transitive event – animate for agent, inanimate for patient - that match to our thematic expectations, object RCs will become comparable in terms of processing ease as subject RCs. Similar observations have been found with children in Kidd et al. (2007).
This study examined the production of RCs in Cantonese children and adults. We investigated the strateg(ies) adopted by adults and children in their RC construction when the thematic configuration runs counter to their expectation. An elicited production paradigm using pictures of a transitive event of two entities was adopted (see Figure 1). Participants were asked to describe which entity was highlighted in the picture. By manipulating the animacy features of the entities, all four animacy configurations were tested: animate agent-animate patient (AA), animate agent-inanimate patient (AI), inanimate agent-animate patient (IA), and inanimate agent-inanimate patient (II).
The types of construction children and adults used in their RC responses for different animacy configuration are given in Table 1. Adults had presented a very strong preference for the animate entity to be in the subject position, hence the passivized subject RCs in the DO condition for AA and AI configuration. They also seemed to have a preference for the entity in discourse focus (i.e. the relativized NP) to be the subject in a sentence, as evident by the significant amount of passivized subject RCs in the DO condition for II configuration. When the animacy configuration matched well with expectations, adults were happy to produce object RCs, as seen in the AI configuration.
Children showed similar traits of preferences as adults, but instead of using passivized subject RCs, they used alternative structures such as internally-headed RCs (a presumably flat structure using the full transitive clause in the function of RC, with the relative marker appended to the end of the clause) and subject RCs with a patient subject (the patient entity in the event has become the subject of the clause, giving a middle-voice like structure). This paper will discuss the motivations underlying these alternative constructions in children’s production.

Author

• Elaine Lau
  Presenting Author