In their abstract, Branigan & Pickering (B&P) assert that the advent of syntactic priming should “end the current reliance on acceptability judgments.” We argue that this claim is too strong; although syntactic priming is useful for some research questions, acceptability judgments are still needed for other important questions.
B&P anticipate one of our two major objections: that studying syntactic priming in production requires an alternative syntactic structure that expresses a semantically equivalent idea. Setting aside reasonable objections from construction grammarians, who may question whether different structures could ever be semantic equivalents (e.g. Goldberg Reference Goldberg1995), such a restriction dramatically narrows the scope of syntactic inquiry, both within and across languages. B&P's proposed solution is to measure syntactic priming in comprehension, where an alternative structure is not necessary. They also acknowledge, however, that the evidence for syntactic priming in comprehension has been much more scarce and fragile; this appears to contrast with their belief that syntactic priming research has reached maturation. Even if there were robust and reliable methods of measuring syntactic priming in comprehension, that would not reach the goal of achieving a unified theory of syntactic representation, in the same way that current research on language production mechanisms has not been particularly informative about theories of language comprehension.
Our second major objection is that syntactic priming studies require (1) that groups of participants be linguistically homogenous and (2) that their hypothesized grammar already be known to the researcher. These requirements would eliminate two critical areas of inquiry: (1) the study of the mental grammars of deaf individuals who were not exposed to language until after early childhood, and (2) traditional linguistic fieldwork.
Deaf individuals with delayed access to language cannot be assumed to share the internal grammar of native users or even of other late learners. Indeed, the nature of their mental grammar is often the very question under investigation. Furthermore, it is uncommon for these individuals to be available to participate in the numbers that are typically necessary to detect priming effects. What these questions require, then, is not a method that averages across many participants, but one that allows for the description of syntactic performance within and between individuals. Acceptability judgments allow this; syntactic priming does not.
Previous work, much of it accomplished using acceptability judgments, has shown that deaf individuals who are first exposed to an accessible language after early childhood exhibit variation in both phonology and morphosyntax compared with individuals immersed in language from birth, particularly with respect to more complex syntactic structures (Boudreault & Mayberry Reference Boudreault and Mayberry2006; Mayberry et al. Reference Mayberry, Lock and Kazmi2002). Low syntactic performance, revealed by acceptability judgments, led us to ask whether Deaf non-native signers of American Sign Language (i.e., those first exposed to ASL after early childhood) would exhibit patterns of syntactic priming that varied from those of Deaf native ASL signers, or of hearing individuals who learned ASL as an L2 in adulthood (Hall et al. Reference Hall, Ferreira and Mayberry2015). Contrary to our expectation, Deaf adults who had learned ASL after early childhood showed patterns of syntactic priming that were identical to those of Deaf native signers and hearing L2 signers.
If syntactic priming were our only tool, we might conclude that late exposure to a first language does not perturb syntactic representation. We view such a claim as premature, given substantial evidence that late learners struggle with syntax (especially with complex syntax) if they were not exposed to any linguistic system in childhood. For example, an fMRI study revealed substantial differences in the neural organization of ASL sentence processing in such individuals using acceptability judgments (Mayberry et al. Reference Mayberry, Chen, Witcher and Klein2011). Although we did not administer an acceptability judgment task to the participants in our ASL syntactic priming study, the Deaf late learners did show unusual patterns in phonological similarity judgments (Hall et al. Reference Hall, Ferreira and Mayberry2012), suggesting that the results of the priming study may have failed to detect differences in syntactic processing.
One possible explanation relates to our first objection above: the type of structures that we can study with a priming paradigm is limited. ASL grammar is only partially described and syntactic processing differences among Deaf, atypical, learners are most pronounced on more complex structures, such as object relative clauses (Boudreault & Mayberry Reference Boudreault and Mayberry2006; Mayberry & Lock Reference Mayberry and Lock2003). However, ASL does not have an alternative structure for expressing object relative clauses; therefore, we were restricted to studying a simpler structure (prenominal vs. postnominal adjectives; cf., Cleland & Pickering Reference Cleland and Pickering2003), where prior research has not detected differences between early and later learners (Boudreault & Mayberry Reference Boudreault and Mayberry2006; Mayberry & Lock Reference Mayberry and Lock2003).
One way to investigate syntactic acquisition in this population is the case study approach using well-established methods to study language acquisition longitudinally in a few individuals (Ferjan Ramirez et al. Reference Ferjan Ramirez, Lieberman and Mayberry2013). Syntactic priming is simply not a viable methodology for single-subject or small-group designs. This limitation applies to any case in which there is concern that an individual's grammar deviates from that of the broader linguistic community.
The second research area that would be jeopardized by exclusive reliance on syntactic priming is field linguistics. In many situations of language documentation, it is common for very few informants to be available (let alone cooperative). Gathering acceptability judgments from a small number of speakers enables documentation of the range of structures used by the world's languages, which, in turn, is essential to our understanding of linguistic representation. Syntactic priming is an insufficiently powerful tool to investigate these theoretically important situations.
Acceptability judgments can reveal what an individual implicitly knows about syntax in a way that syntactic priming cannot. In our view, both methods have their place, and more progress will be made by judiciously deploying each tool in its proper context.
You have
Access
In their abstract, Branigan & Pickering (B&P) assert that the advent of syntactic priming should “end the current reliance on acceptability judgments.” We argue that this claim is too strong; although syntactic priming is useful for some research questions, acceptability judgments are still needed for other important questions.
B&P anticipate one of our two major objections: that studying syntactic priming in production requires an alternative syntactic structure that expresses a semantically equivalent idea. Setting aside reasonable objections from construction grammarians, who may question whether different structures could ever be semantic equivalents (e.g. Goldberg Reference Goldberg1995), such a restriction dramatically narrows the scope of syntactic inquiry, both within and across languages. B&P's proposed solution is to measure syntactic priming in comprehension, where an alternative structure is not necessary. They also acknowledge, however, that the evidence for syntactic priming in comprehension has been much more scarce and fragile; this appears to contrast with their belief that syntactic priming research has reached maturation. Even if there were robust and reliable methods of measuring syntactic priming in comprehension, that would not reach the goal of achieving a unified theory of syntactic representation, in the same way that current research on language production mechanisms has not been particularly informative about theories of language comprehension.
Our second major objection is that syntactic priming studies require (1) that groups of participants be linguistically homogenous and (2) that their hypothesized grammar already be known to the researcher. These requirements would eliminate two critical areas of inquiry: (1) the study of the mental grammars of deaf individuals who were not exposed to language until after early childhood, and (2) traditional linguistic fieldwork.
Deaf individuals with delayed access to language cannot be assumed to share the internal grammar of native users or even of other late learners. Indeed, the nature of their mental grammar is often the very question under investigation. Furthermore, it is uncommon for these individuals to be available to participate in the numbers that are typically necessary to detect priming effects. What these questions require, then, is not a method that averages across many participants, but one that allows for the description of syntactic performance within and between individuals. Acceptability judgments allow this; syntactic priming does not.
Previous work, much of it accomplished using acceptability judgments, has shown that deaf individuals who are first exposed to an accessible language after early childhood exhibit variation in both phonology and morphosyntax compared with individuals immersed in language from birth, particularly with respect to more complex syntactic structures (Boudreault & Mayberry Reference Boudreault and Mayberry2006; Mayberry et al. Reference Mayberry, Lock and Kazmi2002). Low syntactic performance, revealed by acceptability judgments, led us to ask whether Deaf non-native signers of American Sign Language (i.e., those first exposed to ASL after early childhood) would exhibit patterns of syntactic priming that varied from those of Deaf native ASL signers, or of hearing individuals who learned ASL as an L2 in adulthood (Hall et al. Reference Hall, Ferreira and Mayberry2015). Contrary to our expectation, Deaf adults who had learned ASL after early childhood showed patterns of syntactic priming that were identical to those of Deaf native signers and hearing L2 signers.
If syntactic priming were our only tool, we might conclude that late exposure to a first language does not perturb syntactic representation. We view such a claim as premature, given substantial evidence that late learners struggle with syntax (especially with complex syntax) if they were not exposed to any linguistic system in childhood. For example, an fMRI study revealed substantial differences in the neural organization of ASL sentence processing in such individuals using acceptability judgments (Mayberry et al. Reference Mayberry, Chen, Witcher and Klein2011). Although we did not administer an acceptability judgment task to the participants in our ASL syntactic priming study, the Deaf late learners did show unusual patterns in phonological similarity judgments (Hall et al. Reference Hall, Ferreira and Mayberry2012), suggesting that the results of the priming study may have failed to detect differences in syntactic processing.
One possible explanation relates to our first objection above: the type of structures that we can study with a priming paradigm is limited. ASL grammar is only partially described and syntactic processing differences among Deaf, atypical, learners are most pronounced on more complex structures, such as object relative clauses (Boudreault & Mayberry Reference Boudreault and Mayberry2006; Mayberry & Lock Reference Mayberry and Lock2003). However, ASL does not have an alternative structure for expressing object relative clauses; therefore, we were restricted to studying a simpler structure (prenominal vs. postnominal adjectives; cf., Cleland & Pickering Reference Cleland and Pickering2003), where prior research has not detected differences between early and later learners (Boudreault & Mayberry Reference Boudreault and Mayberry2006; Mayberry & Lock Reference Mayberry and Lock2003).
One way to investigate syntactic acquisition in this population is the case study approach using well-established methods to study language acquisition longitudinally in a few individuals (Ferjan Ramirez et al. Reference Ferjan Ramirez, Lieberman and Mayberry2013). Syntactic priming is simply not a viable methodology for single-subject or small-group designs. This limitation applies to any case in which there is concern that an individual's grammar deviates from that of the broader linguistic community.
The second research area that would be jeopardized by exclusive reliance on syntactic priming is field linguistics. In many situations of language documentation, it is common for very few informants to be available (let alone cooperative). Gathering acceptability judgments from a small number of speakers enables documentation of the range of structures used by the world's languages, which, in turn, is essential to our understanding of linguistic representation. Syntactic priming is an insufficiently powerful tool to investigate these theoretically important situations.
Acceptability judgments can reveal what an individual implicitly knows about syntax in a way that syntactic priming cannot. In our view, both methods have their place, and more progress will be made by judiciously deploying each tool in its proper context.