A Study On Cross-domain Structural Priming

Cognition is often held to consist of a set of distinct modules,each specialized to a specific domain and processing mode.For example,the manner that people process language is intuitively different from when they do arithmetic or music.However,in the context of structure building resources,a range of studies has demonstrated a potential link between linguistic cognition and other cognitive domains.Although several studies explored the shared structuring system across different cognitive domains,relevant issues remain in deliberation and are far from being resolved.For instance,most previous related studies have focused on synthetic languages,such as English and German,more cross-linguistic evidence is required to confirm the universality of this domain-general structuring system.Thus,this motivates us to explore whether there are shared structural representations and/or operations for processing between language and other non-linguistic cognitive domains with a series of cross-domain structural priming experiments in this dissertation.First,we adopted two offline tasks to investigate the cross-domain structural priming from arithmetic to language on Chinese adults.Experiment 1a was conducted through a sentence completion task,where arithmetic expressions were adopted as primes,and Chinese structure NP₁+YOU+NP₂+HEN+AP was selected as targets.Results showed that there were cross-domain structural priming effects from arithmetic expressions to this specific Chinese structure in production.Experiment 1b utilized a sentence judgement task that applied similar primes and targets.Results showed that there were also cross-domain structural priming effects in comprehension.Therefore,we argued that similar to the effects observed in English,cross-domain structural priming effects from arithmetic expressions to language could also be detected in Chinese and observed in both comprehension and production modalities.These findings confirmed the existence of shared structure building resources between mathematics and language.Additionally,cross-domain structural priming effects in English were observed by using relative clause attachments as targets.In this case,attachment priming effects were located within the domain of the complex noun phrase.Findings in Chinese structures suggested that such effects were not constrained by the syntactic domain of a complex noun phrase and could be achieved across different syntactic domains.Then,to investigate the developmental trajectory of the shared structuring system and the priming effects in children,we conducted another set of two offline experiments.Experiment 2a used a sentence completion task similar to Experiment 1a,and Experiment 2b used a sentence completion task similar to Experiment 1b.Results demonstrated that such cross-domain structural priming effects were also observable in children.More importantly,these children seemed to be more susceptible to these priming effects than adults.Our suggestions highlighted its attribution to the different computational process between children and adults,such as the longer calculation time,the stringent calculation strategies,and the fewer computational resources available in the calculation.Since the cross-domain structural priming effects from arithmetic expressions to languages could be steadily observed in both production and comprehension,we investigated the mechanisms of such effects.More specifically,we conducted Experiment 3a and 3b to verify the incremental-procedural account and the representational account through the online eye-tracking technique.The fixation sequences during arithmetic calculation reflected the hierarchy-based processing patterns of arithmetic expressions,providing evidence to refute assumptions of the incremental-procedural account.Alternatively,combined with differentiation between priming from left-and right-branching arithmetic expressions,we argued that the representational account might be more plausible in explaining cross-domain priming effects from arithmetic expressions to linguistic structures.There was a probability that the representations of the structural configurations of arithmetic expressions,which specified structural information of attachment height and linear order of composition,primed the structural representations of target sentences.Furthermore,eye-movement data during Chinese sentence comprehension in the forms of NP₁+YOU+NP₂+HEN+AP and NP₁+Kangzhe/Wangzhe+ADJ+NP₂+Verbal idiom indicated that priming effects from arithmetic primes could not only influence the way participants interpreted linguistic targets,but also facilitate processing times of targets when parsing of targets were structurally consistent with preceding arithmetic primes.Such facilitation effects were more frequently observed in late measures,including total reading time and fixation count,indicating that the processing of arithmetic primes might have a late effect on the subsequent processing of linguistic targets.Then,to investigate whether structural information could be represented at a rather abstract level for L2 learners as that for L1 speakers,we conducted Experiment 4a and 4b to explore cross-domain structural priming from arithmetic expressions to English sentences containing ambiguous relative attachments and English sentences in the form of NP₁+VP+NP₂+PP+Post-PP in Chinese-English learners.Priming effects were found on both structures,not only in their interpretation choices of target sentences but also in processing times.These outcomes suggested that L2 learners could also represent the structural knowledge of their second language at a relatively abstract level.The occurrence of structural priming from arithmetic expressions to PP attachments further suggested that the cross-syntactic-domain priming effects,which were observed in Chinese structure NP₁+YOU+NP₂+HEN+AP,could also be found in English.Similar to effects observed in L1,the facilitation effects in L2 were also more frequently observed in later processes involved in the comprehension of targets.Cross-domain structural priming effects found in both L1 and L2 comprehension further lend support for the hierarchical structural tree representation account,which ascribes the priming effects to the isomorphic abstract hierarchical configurations of arithmetic primes and linguistic targets,which specify structural information like attachment height and linear order of composition.Next,we conducted Experiment 5a and 5b to investigate whether the cross-domain structural priming effects could be extended to other cognitive domains such as visuo-spatial hierarchies.Specifically,we constructed high-attachment,low-attachment,and baseline visuo-spatial hierarchies by modifying the position and presentation sequence of element integration.Priming effects were found from visuo-spatial hierarchies to both L1 and L2 comprehension,but the magnitude of priming from these hierarchies was weaker than arithmetic expressions.We reasoned that such difference could have resulted from the fact that the visuo-spatial hierarchies adopted in the present study were organized in simple grouping patterns rather than by well-defined rules.And their structural configurations might not be as stringently represented as arithmetic structures.Moreover,to explain the domain-generality of structural priming from the perspective of structural representation,we established the temporal hierarchical representation account based on the hierarchical structural tree representation account.In this case,in addition to structural information of attachment height and linear order of composition,we argued for the requirement of imposing a well-defined sequential order on elements of a specific structure.Finally,this dissertation culminates with its implications for psycholinguistic theories and methodologies and limitations and suggestions for future research.