Browsing by Author "Yan, Hao"

Now showing 1 - 2 of 2
  • Thumbnail Image
    Item
    Canvas Survey and User Testing, Summer 2015
    (2015) Kolah, Debra; Yan, Hao
  • Thumbnail Image
    Item
    Sentence processing in aphasic speakers with short-term memory deficits: Interactions between structural and lexical processing
    (2017-08-10) Yan, Hao; Martin, Randi
    The current research investigates interactions between lexical and structural processing in the construction and interpretation of transitive and dative sentence structures, and relates speakers’ choice of structure in sentence production to STM processes and recent/long-term language experience conditioned on lexical information. Study 1 investigated whether STM is related to the increased tendency for speakers to reuse a structure when the verb is also repeated. Speakers show syntactic priming – that is, a tendency to repeat syntactic constructions they have recently comprehended or produced – and this tendency is even stronger when adjacent utterances share the same main verb, termed the lexical boost. Some have suggested that abstract syntactic priming (i.e., with no lexical overlap) derives from implicit learning, whereas the lexical boost derives from explicit short-term memory (STM) for the prime (e.g., Chang, Dell, & Bock, 2006). To address this issue, Experiment 1.1 assessed twelve aphasic patients with varying degrees of STM deficits and eleven age-matched healthy control speakers in a syntactic priming experiment using a picture description paradigm. Experiment 1.2 assessed eight patients and ten control speakers in another syntactic priming task using written word arrays for target trials. Despite patients’ difficulty in maintaining phonological, semantic, and structural information, as evidenced by various STM and sentence repetition measures, most of them showed lexical boost effects comparable to those of healthy speakers. Moreover, the size of the lexical boost was unrelated to the degree of STM deficit, suggesting that the lexical boost does not rely on explicit memory. Alternative explanations for the differing patterns for syntactic priming with and without lexical overlap are discussed. Compared to young speakers, many patients and some older control speakers had very limited production of double object (DO) dative sentences in the picture description task in Experiment 1.1 and predominantly produced prepositional dative (PD) sentences. Study 2 included four experiments to explore factors behind the choice of structure in dative sentence production. There are various possible accounts of speakers’ preference for the PD structure over the DO structure. DO could be syntactically more complex (Chomsky, 1975; Beck & Johnson, 2004), or more difficult to parse without explicit markers of grammatical/thematic roles of post-verbal nouns (Stowell, 1981). The word order in DO could be different from a default order of conceptual planning, which plans the verb and the theme together, before a recipient (O’Grady & Lee, 2005). Accessibility of nouns after the verb may also affect word order and the choice of structure in dative sentence production (Bresnan, Cueni, Nikitina, & Baayen, 2007). In accordance with these hypotheses, the factors investigated included syntactic complexity (DO vs. PD), function word processing (functions of prepositions in dative sentences), semantic factors (ease of semantic integration between the verb and the theme), processing factors (accessibility of the recipient), as well as the relation between these factors and STM capacities. In a sentence repetition task, Study 2.1a found that DO sentences were more difficult to repeat than PD sentences for patients, and this structural effect was related to percent DO production in Experiment 1.1. Pronoun recipients facilitated repetition of DO sentences, and easy semantic integration between the verb and the theme improved repetition of PD sentences. A grammaticality judgment task in Study 2.1b found some evidence for a weak pronoun recipient effect, but not a semantic integration effect. Only a small structural effect was found, which was not related to percent DO production in Experiment 1.1, suggesting that the limited DO production was not caused by any potential weakness in the syntactic knowledge of the DO structure. Study 2.1c found limited evidence for a structural effect in a sentence anomaly judgment task, and some evidence that prepositions in PD sentences and provide-with sentences serve as explicit markers of grammatical functions to facilitate parsing of post-verbal objects. Finally, Study 2.2 found that both patient and control speakers were more likely to choose the DO structure in picture description when using pronoun recipients or more accessible full NP recipients, reflecting a tendency to produce the more accessible word earlier in a sentence. This accessibility effect and the pronoun recipient in Experiment 2.1a and 2.2 were consistent with findings in corpus studies that a DO structure with a pronoun recipient is a highly frequent dative construction (Bresnan, Cueni, Nikitina, & Baayen, 2007), suggesting an important role of language experience in sentence processing. Importantly, STM capacities were related to percent DO production in Experiments 1.1, 1.2, and 2.2, and the structural effect in dative sentence repetition in Experiment 2.1a, indicating that the processing difficulty with DO sentences is related to STM costs. Taken together, these findings in Study 1 and 2 support a multi-factorial, usage-based account of sentence processing (Gahl & Menn, 2016; Menn & Bastiaanse, 2016), which is achieved via the interactions between structural processing and lexical processing, with the potential support of STM resources for some aspects of these processes.