BEHAVIORAL AND BRAIN SCIENCES (2017), Page 1 of 61 doi:10.1017/S0140525X16002028, e282 An experimental approach to linguistic representation Holly P. Branigan Department of Psychology, University of Edinburgh, Edinburgh, EH8 9JZ, United Kingdom [email protected] http://www.ed.ac.uk/profile/holly-branigan Martin J. Pickering Department of Psychology, University of Edinburgh, Edinburgh, EH8 9JZ, United Kingdom [email protected] http://www.ed.ac.uk/profile/martin-pickering Abstract: Within the cognitive sciences, most researchers assume that it is the job of linguists to investigate how language is represented, and that they do so largely by building theories based on explicit judgments about patterns of acceptability – whereas it is the task of psychologists to determine how language is processed, and that in doing so, they do not typically question the linguists’ representational assumptions. We challenge this division of labor by arguing that structural priming provides an implicit method of investigating linguistic representations that should end the current reliance on acceptability judgments. Moreover, structural priming has now reached sufficient methodological maturity to provide substantial evidence about such representations. We argue that evidence from speakers’ tendency to repeat their own and others’ structural choices supports a linguistic architecture involving a single shallow level of syntax connected to a semantic level containing information about quantification, thematic relations, and information structure, as well as to a phonological level. Many of the linguistic distinctions often used to support complex (or multilevel) syntactic structure are instead captured by semantics; however, the syntactic level includes some specification of “missing” elements that are not realized at the phonological level. We also show that structural priming provides evidence about the consistency of representations across languages and about language development. In sum, we propose that structural priming provides a new basis for understanding the nature of language. Keywords: language production; linguistics; mental representation; psycholinguistics; semantics; structural priming; syntax The cognitive science of language is concerned with both about linguistic structure relating to syntax and semantics, linguistic representations and how those representations so that it can be used to develop linguistic theory and dis- are used in processing. All researchers, whether nominally criminate among competing accounts. Thus, the domi- psychologists or linguists, should seek to address both ques- nance of acceptability judgments can be ended, and the tions. In practice, however, linguists have focused largely understanding of linguistic representation can develop to on representation and used a single method (acceptability a greater extent than before. judgments) to investigate it, whereas psychologists have This paper describes our theoretical claims, linguistic not investigated representation but instead imported lin- account, and applications. In section 1, we motivate the guistic theories into their accounts. In this paper, we use of structural priming to investigate mental representa- argue instead that researchers need not, and should not, tion and present the advantages of structural priming over be restricted to acceptability judgments when investigating acceptability judgments. In section 2, we consider what the linguistic representation. extensive evidence using this method tells us about linguis- This proposal is not new but was previously just a theo- tic representation. Section 3 discusses the implications of retical possibility. It now appears, however, that structural our account. priming – the tendency to repeat linguistic structure across utterances – allows researchers to investigate linguis- tic representations in a way that has many advantages over 1. Why a psychological account of linguistic acceptability judgments. Most important, it has now structure is now possible reached maturity, in that hundreds of studies have used this method; and many of them are informative, not A complete theory of human language requires characteri- merely about language processing, but also about linguistic zation of both people’s linguistic representations and the representations themselves. In fact, we argue that evidence processes that operate over those representations. There- from structural priming supports quite specific proposals fore, issues of representation and processing are in © Cambridge University Press 2017 0140-525X/17 Downloaded from https://www.cambridge.org/core. Max-Planck-Institut fuer Psycholinguistik, on 16 Nov 2017 at 07:29:33, subject to the Cambridge Core terms of use, available at 1 https://www.cambridge.org/core/terms. https://doi.org/10.1017/S0140525X17000528 Branigan and Pickering: An experimental approach to linguistic representation principle of interest to both linguists and psychologists, (One of the fans punched the referee). As subsequent albeit from different perspectives. In practice, however, studies have demonstrated, these effects appear to arise linguistic representation (in particular, with respect to syn- from repetition of aspects of abstract linguistic structure tactic and semantic structure) has for the last four decades and occur largely outside of awareness (Pickering & Ferre- been largely the domain of linguists and has been studied ira 2008). primarily using a single approach in which linguists or They cannot be explained in terms of repetition of partic- their informants make explicit metalinguistic judgments ular words. Bock (1989) showed that participants tended to about the grammatical (or semantic) acceptability of indi- use prepositional object (PO) dative sentences (The girl is vidual sentences – henceforth, acceptability judgments. handing a paintbrush to the man) rather than double Such judgments constitute the dataset upon which theories object (DO) sentences (The girl is handing the man a paint- of linguistic representation are based. brush) after a dative sentence that did not include to (The In this paper, we propose that the representations under- secretary baked the cake for her boss). Therefore, lying language use need not and, in fact, should not be priming could not be due to word repetition, because the investigated only via such judgments. Rather, we suggest PO and DO target sentences share all words except to. that they can be examined directly through a behavioral We also can rule out explanations couched entirely in measure that has been used widely in psychological terms of meaning. First, the alternative responses (e.g., research to investigate the representation of a range of PO and DO, or active and passive) denote the same information types. This method is priming: If processing events, in that they both can be used to describe the one stimulus affects the subsequent processing of another same picture. Second, Messenger et al. (2012b) found stimulus, then these stimuli share some aspect of their rep- priming between sentences describing different event resentation. Hence, structural priming effects, where pro- types (e.g., Experiencer-Theme: The king is being cessing one utterance affects the processing of another ignored by the bear, and Agent-Patient: The doctor gets utterance that shares an aspect of linguistic structure but licked by the cow). Additionally, Hartsuiker and Westen- is otherwise unrelated, provide evidence about linguistic berg (2000) found that Dutch participants repeated the representation. In the classic demonstration, Bock (1986) order of auxiliary and main verb (was geblokkeerd [“was had participants repeat active or passive sentences and blocked”] vs. geblokkeerd was [“blocked was”]), even then describe pictures depicting transitive events. She though they do not differ in meaning. Moreover, the found that they were more likely to use a passive target sen- effects cannot be explained by repetition of metrical struc- tence (e.g., The church is being struck by lightning) after ture, because The girl is handing a paintbrush to the man repeating a passive prime sentence (The referee was was not primed by Susan brought a book to study, punched by one of the fans) than after repeating an active though it was primed by the metrically equivalent Susan brought a book to Stella (Bock & Loebell 1990). Overall, these results are consistent with priming of representations that are specified for syntactic information but not seman- tic, lexical, or phonological information. This conclusion is supported by studies showing priming of many other syn- HOLLY BRANIGAN is a Professor of Psychology of tactic constructions, such as noun-phrase structure Language and Cognition at the University of Edin- (Cleland & Pickering 2003) and verb-particle placement burgh. She has published more than 65 journal articles (Konopka & Bock 2009). and in numerous other publications on topics relating to language processing. Her research investigates the Priming, however, is also informative about other aspects of linguistic structure, including many components of cognitive processes and representations that underlie fi language production and interactive language use, semantics including thematic roles, quanti cation, and with a particular focus on syntactic structure, and information structure. It occurs in diverse languages (e.g., aspects of the relationship between psycholinguistics English, Mandarin, Basque) and between languages, and and linguistics.