Structural priming in the grammatical network: A study of English argument structure constructions
Item statusRestricted Access
Embargo end date14/06/2023
Recent cognitive-linguistic approaches view grammar as a mental network of stored knowledge. The present study investigates to what extent psycholinguistic evidence from structural priming can inform one of the crucial linking mechanisms in this network: similarity relations between formally and/or functionally related constructions. Structural priming occurs when processing a multi-word unit affects speakers’ subsequent processing of the same or a similar construction. Focusing on a range of English argument structure constructions, it is shown that priming can help address theoretical claims about the structure of the grammatical network, and that the network-based view in turn motivates reinterpretations of previous priming results as well as novel extensions of the paradigm. Chapter 1 introduces the topic and outlines the goals of the investigation. It is proposed that structural priming can provide a rich source of evidence for cognitive-linguistic models of the grammatical network, but that the ways in which the two research areas can inform each other have not yet been sufficiently explored. Chapter 2 outlines the central role of similarity relations in cognitive-linguistic network models, combining elements from Construction Grammar, Cognitive Grammar, Word Grammar and other related frameworks. Different types of similarity relations between argument structure constructions are illustrated with detailed linguistic analyses, including functionally similar (alternating) constructions (e.g., the double-object and the to-dative construction) and formally similar constructions (e.g., the resultative and the depictive construction). It is shown that previous theoretical accounts raise questions about the nature and degree of these similarities, which can be addressed with the help of the priming evidence provided in the later chapters. Moreover, the discussion provides an overview of corpus-based and experimental methods that can be used to investigate constructional similarities. Chapter 3 describes how structural priming can be used to study similarity relations in the grammatical network via their implicit effects on speakers’ processing. Evidence is provided that priming is sensitive to both formal and functional similarities between constructions. Based on an overview of production and comprehension priming methods, it is argued that different methods are needed to study different types of similarity relations. In addition, the discussion outlines three relevant distinctions for the interpretation of structural priming effects: the difference between within-construction and cross-constructional priming; between symmetric and asymmetric priming; and between facilitatory and inhibitory priming. Chapter 4 reinterprets previous structural priming findings about alternating constructions in the context of a cognitive-linguistic network model of grammar. First, based on within-construction priming effects between instances of the same construction, it is suggested that speakers’ grammatical networks comprise both abstract argument structure constructions and verb-specific representations. Second, it is shown that the evidence for cross-constructional priming between members of the same alternation (e.g., the double-object and the to-dative construction) is weaker than previously assumed, casting doubt on the central role that alternations play in some grammatical theories. Third, varying degrees of cross-constructional priming are observed between members of different alternations (e.g., the dative and the benefactive constructions) and related to previous theoretical claims about their respective similarities. Chapter 5 reports novel comprehension experiments that extend structural priming to previously understudied non-alternating argument structure constructions. In two experiments, a variant of self-paced reading known as the ‘maze task’ is used to investigate priming between the resultative and the object-oriented depictive construction. The results indicate that depictives facilitate both depictive targets (within-construction priming) and resultative targets (cross-constructional priming), but that resultatives do not give rise to priming. Frequency-related and other possible explanations for the asymmetric effects are provided. The discussion then addresses another recent experiment that tests priming between the caused-motion and the resultative construction. The observed symmetric priming effects do not support previous claims about an asymmetric metaphorical relation between the constructions. Moreover, an explanation is offered for the occurrence of inhibitory priming in the experiment, which has seldom been observed in structural priming. Finally, Chapter 6 summarises the key findings of the study and outlines three implications that merit further investigation. These concern the source ambiguity of priming effects, the use of priming for investigating other types of network relations, and a novel view of ‘constructionhood’ as a gradient rather than a categorical property. Together, the discussions and remaining questions illustrate that cognitive-linguistic theory and structural priming research complement each other in their efforts to uncover the structure of the grammatical network.