Francey, Gillian Elaine and Cain, Kate (2019) The influence of the temporal characteristics of events on children's pronoun resolution. PhD thesis, Lancaster University.
2018FranceyPhD.pdf - Published Version
Available under License Creative Commons Attribution Non-commercial No Derivatives.
Download (2MB)
Abstract
The main goal of the thesis was to determine whether the temporal characteristics of events influence 7- to 11-year-olds’ pronoun resolution to identify if this might explain contradictory findings in the literature regarding the age at which children demonstrate a subject bias interpretation of a pronoun. A second goal was to determine whether children demonstrate the same influence of aspectual morphemes in narrative contexts as shown by adults in previous research. A third goal was to examine whether literature written for children might provide a potential causal explanation (environmental exposure) for the pattern of development observed in the experimental work. Chapter 1 reviews the literature that is considered relevant to the experimental work. The experiments in Chapter 2 examined pronoun resolution following SourceGoal transfer of possession events expressed with imperfective or perfective grammatical aspect. Experiment 1a (children) and Experiment 1b (adults) examined ambiguous pronoun resolution, Experiments 2a and 2b examined children’s pronoun resolution following transfer events with different gender protagonists. In all experiments an effect of grammatical aspect was observed with children and adults selecting the Source protagonist more frequently when the event was expressed with imperfective than perfective aspect. Despite demonstrating an influence of grammatical aspect, unlike adults, children did not demonstrate a subject bias interpretation of an ambiguous pronoun following the imperfective expression of a transfer event and children were more likely than adults to demonstrate a Goal interpretation of the pronoun following perfectively expressed transfer events. Transfer of possession events have endpoints (they are telic) and this pattern of results suggests that the presence of an endpoint in the stimulus sentences influenced children’s tendency to resolve a subsequent pronoun towards the previous subject protagonist. This possibility was examined in Chapter 3. The experiments in Chapter 3 examined the effect of grammatical aspect on adults’ (Experiment 3a) and children’s (Experiment 3b) resolution of ambiguous pronouns contained within events from two lexical aspect categories; those with inherent endpoints and those without. Adults and children were more likely to resolve the pronoun to the subject protagonist following events without endpoints than following events with endpoints. These findings support those reported in Chapter 2, showing that the inherent temporal characteristics of events influence adults’ and children’s pronoun resolution. Grammatical aspect had an influence on adults’ pronoun resolution only for events with endpoints. In contrast, grammatical aspect influenced children’s pronoun resolution for events from both lexical aspect categories. This was interpreted as a consequence of the smaller effect of grammatical aspect on children’s pronoun resolution which limited the opportunity to observe an interaction. The experiments in Chapter 4 examined the effect of grammatical aspect on children’s judgement of transfer events ongoingness (Experiment 4a) and on children’s pronoun resolution (Experiment 4b) when the events were embedded within narratives. In contrast to the previous experiments in the thesis where stimulus sentences were presented in isolation, the experiments in this chapter found an increase in children’s sensitivity to imperfective aspect with age. When asked immediately after the aspectual event, children’s judgement that imperfectively expressed events were ongoing increased with age as did their Source resolution of an ambiguous pronoun. When there was intervening text between the aspectual event and the question, children’s judgement that imperfectively expressed events were ongoing decreased with age, and similarly their Source resolution also decreased with age. Experiment 4c examined whether a short delay between the presentation of the aspectual sentence and the pronoun had the same effect as intervening text. This was confirmed: there was an age decrease in Source resolution for imperfectively expressed events but not for perfectively expressed events. These findings provide converging evidence that, like adults in previous research, older children take into account the temporal characteristics with which events are expressed and the inherent duration of events themselves in their construction of situation models during narrative comprehension. Chapter 5 examined the frequency with which verbs typically used to describe telic events (events with endpoints) and atelic events (events without endpoints) were expressed with imperfective and perfective aspect in a corpus of literature read by children within the age range 7- to 11-years. It also examined the coherence relations most likely to follow perfectively expressed transfer events and the grammatical subject of these coherence relations. Verbs were more often expressed with perfective than imperfective aspect. Verbs typically used to describe atelic events were expressed with imperfective aspect more frequently than verbs typically used to describe telic events. This pattern suggests that, within these two lexical aspect categories, a distributional bias in imperfective expression persists beyond young children’s language environment (Shirai & Andersen, 1995). Perfectively expressed transfer events were more often followed by Occasion than other coherence relations (Elaboration, Result, Parallel, Explanation, Other). There was no evidence of a Goal re-mention bias generally following transfer events, there was evidence of a Goal remention bias within Occasion relations. These findings support the proposal that the temporal characteristics of events are a stronger predictor of the type of coherence relation that will follow events than the re-mention of particular protagonists. They further suggest that re-mention biases are contingent on coherence relations. Taken together the findings demonstrate that children’s pronoun resolution is influenced by both the inherent temporal characteristics of events and the grammatical aspect with which events are expressed. How these findings explain previously reported contradictory findings in the literature is discussed. How the findings inform our understanding of the nature of the difficulties some children experience with pronoun resolution is also discussed.