Review of Towards a Biolinguistic Understanding of Grammar
This volume containing 14 papers attempts to develop a theory of interface of the language system with systems external to the language faculty, particularly semantics and phonetic implementation. It attempts to do so primarily within the theoretical orientation of the Minimalist Program. The papers are organized into five sections: (1) three essays on the syntax-semantics interface; (2) three essays on the role of syntactic features in derivations and interfaces; (3) two papers on phonology and the phonology-syntax interface; (4) two papers on language development and acquisition; and (5) four papers on experimental studies. The book is not about biological foundations of language, but rather aims to formulate Minimalist accounts of interface with non-linguistic faculties that might inform future attempts to integrate linguistic theory with other fields.
Howard Lasnick’s paper, ‘Single cycle syntax and a constraint on quantifier lowering,’ deals with the problem of lowered quantifiers, e.g., “Nobody is certain to pass the test” is not necessarily equivalent to “It is certain that nobody will pass the test.” The lowered readings do not work with negatives and other sentences, but rather with indefinites, e.g., “Some politician is likely to address John’s constituency.” Lasnick argues that single cycle syntax prevents quantifier lowering rather than allowing it; quantifier scope is determined cyclically, and lowered readings of indefinites come from scope assignment without quantifier raising.
Tim Hunter’s article, ‘A constraint on remnant movement,’ proposes a “Just Outside Constraint” (JOC) in island effects to rule out extraction from adjuncts and moved constituents. For example, “Who did you send a big heavy a picture of [t] to London?” does not allow for extraction from heavy NPs, and hence, “*Who did you sent [t] to London a big heavy picture of [t]?” is ruled out. Hunter also discusses the JOC with regard to fronting in German and word order in SVO, SOV and VSO languages.
Paul M. Pietroski’s article on ‘Language and conceptual reanalysis’ argues that natural languages juxtapose words into phrases to introduce semantically monadic concepts, e.g., the phrase “chase cows” is regarded as an instruction to form a single concept of “chase + cows-PATIENT” rather than just a conjunct of “cow” and “chase.” Pietroski develops arguments for this from Fregean logical and Chomskyan syntax. Thus, natural languages essentially use phrasing and lexicalization to freely convey distinctive concepts, rather than for communicating abstract and conventionalized phrases.
The next section on features and interfaces begins with Daniela Isac’s work on ‘Decomposing force,’ which deals with pragmatic force types of sentences, such as declarative, interrogative, imperative, or exclamatory. She argues that force is not encoded as a simple, primitive syntactic feature, but rather derives from other components such as modality. She argues that imperative force derives from modality and the second person, as seen in the behavior of true imperative forms and surrogate imperatives, i.e., subjunctives and infinitives used as imperatives in Romance languages, Greek, German, Slavic languages, and others. This is compared with the root meanings of modal verbs (e.g. “You must go now”), whose interpretation is in reference to the sentence subject, and non-root meanings (e.g. “He must be happy”), which involve reference to the speaker’s mental state. In this sense, imperatives pattern with non-root modals and true and surrogate imperatives derive their force differently from different syntactic Merge operations.
The next paper, ‘Function without content,’ by Christina Christodoulou and Martina Wiltschko, proposes a unified analysis of the Greek subjunctive marker ‘na,’ which is associated with many different contexts, such as realis and counterfactual conditionals, future oriented clauses, requests, orders, wishes, and others. This is done by having ‘na’ spell out INFL (Inflection) before being associated with any content, and thus, the functions of functional categories can be independent of their contents and associated features.
Atushi Fujimori’s essay on ‘The association of sound with meaning’ argues for a correspondence of the non-low vowels /e/ and /u/ with telicity in Japanese verbs, and /i/ and /o/ with atelicity. This is shown with native Japanese verbs, by carefully controlling for context, and also by means of an experiment with Japanese subjects and their intuitions about nonce words containing these vowels. The experiment shows binary features for lexical aspect in Japanese, and the author proposes instances of such alternations in other languages.
The next section on phonology and syntax begins with Charles Reiss’ paper, ‘Towards a bottom up approach to phonological typology,’ which argues for underspecification in phonology, in accord with Chomsky’s recent, simpler bottom-up approach to Universal Grammar (UG). These trends represent an attempt to simplify linguistic theory so that a more powerful theory with underspecification requires less detail that is genetically predetermined, in order to more readily integrate linguistic theory into a biolinguistic program. For example, a simpler UG would align with so-called evo-devo biology, or the recent attempts to better integrate developmental biology and genetics into an evolutionary framework. In this vein, Reiss shows mathematically that a simpler, underspecified system leads to greater possible combinations of features and phenomena, i.e., less is more. For example, starting with four binary features (and leaving the remaining features unspecified) leads to seven septillion possible inventories. However, this does not impose an undue computational demand on language learners, since they make parsing decisions as they encounter each segment with a modest number of associated features.
Bridget Samuel’s paper, ‘The emergence of phonological forms,’ argues that phonology, in a sense, preceded syntax or general language abilities in human evolution in that phonological production would have been first developed and used for other functions before being paired with syntax in language development. Samuels draws upon animal cognition and sign language for evidence that phonological signals predate external language development. This includes studies on the development of perceptual categories and category perception in young children and animals, as well as phonological category perception in humans (e.g. voice onset time categories). According to the author, the evolution of the vocal tract and larger sensorimotor areas in the human brain were developments that were subsequently exploited by sign and gesture systems. Contrary to what some innatist and generative linguists have assumed, linguistic features may have arisen from phonetic cues or abilities rather than vice versa. That is, these phonetic functions came first and were co-opted by the language faculty.
The section on language development begins with Calixto Aguero-Bautista’s paper on ‘Non-native acquisition and language design,’ which addresses the age-old controversy regarding whether final attainment of a second language (L2) is constrained by UG, which rejects the critical period hypothesis (CPH). For example, past studies have shown that learners can successfully learn an L2 with the Obligatory Pronoun Constraint (OPC), even though it is lacking in their L1, which some have used as evidence against the CPH. The author argues that the OPC arises as a computational and interface effect of the weak crossover effect (WCO), and is thus not part of the UG, but rather independent of the L2 principles to be learned.
The next paper, ‘Interface ingredients of dialect design,’ by Kleanthes K. Grohmann and Evelina Leivada, deals with acquisition of dialects. This study deals with young learners of Cypriot Greek as their native language who later encounter standard Greek in formal schooling, where they encounter issues because the two forms differ in direct object clitic placement. The authors reported their previous studies with Greek speaking children in Cyprus using picture elicitation tasks. Children’s productions of clitics in these studies showed variation in clitic choice, a progression based on grade levels, code-mixing patterns in clitic placement in the prestige dialect (standard Greek), and variation in their clitic use based on test version (i.e. standard or Cypriot Greek). They propose a socio-syntax of development hypothesis (SSDH) to describe this sociolinguistic variation.
The next section on experimental studies begins with ‘What sign languages show,’ by Evie Malaia and Ronnie B. Wilbur. They review studies on whether motion differences in signs for telic and atelic verbs in systems like American Sign Language are phonological features, as these involve differences in velocity and acceleration of hand gestures. They also look at neuroimaging studies showing that such signs are processed in a manner neurologically similar to phonological processing. They conclude that signers process such signs as abstract phonological features, similar to prosodic lengthening, such that motion changes in the signs can signal telicity. This points to grammaticalization of distinctions in physical motion in language.
‘Indeterminacy and coercion effects,’ by Roberto G. de Almeida and Levi Riven, examines the comprehension of indeterminate but grammatical sentences like “The man began the book,” in terms of whether the complement is necessarily construed as an event (coercion), and how much pragmatics contributes to the interpretation. Past psycholinguistic processing studies found that indeterminate sentences are processed differently from more determinate ones (e.g. “The man read a book”) in that the former takes longer to process. Related structures involving structural or semantic gaps include indeterminate middle voice verbs and compounds with potentially ambiguous aspectual structures (e.g. “He is a book-reader on Saturdays”). The authors summarize their previous fMRI study, showing that neural activation patterns in processing indeterminates are similar to and in between the patterns for determinate sentences and pragmatically infelicitous sentences. Comprehension of indeterminates patterns neurologically with pragmatic processes, but also involves brain regions involved in decision-making. Linguistic cues and literal readings are used, and these in turn invoke extralinguistic processes for inferencing.
‘Computation with doubling constituents,’ by Sandiway Fong and Jason Ginsburg, describes a computational implementation of classical Binding Theory in a Minimalist framework for pronoun-antecedent relations. They propose a model for coreference that they claim is simpler, more efficient, and able to account for more data than previous models. Simple, multi-clausal, possessive, and picture determiner phrase (DP) sentences are discussed.
In ‘Concealed reference-set computation,’ by Thomas Graf, reference constraints -- well-formedness constraints on transderivational comparisons -- are appealed to in order to disconfirm a strong hypothesis that syntax is uniquely determined by interface constraints. Some reference constraints can be used in syntax without being too computationally costly, and syntax can avoid obeying some interface requirements without necessarily violating them.
The general theme of many of the articles in this volume is interface, and as such, this book will be of interest to those interested in the interface of linguistic components, particularly within a Minimalist framework. The volume overall deals fairly consistently with this general theme. This book may not be of interest to those interested in evolutionary psychology, evolutionary biology, neurolinguistics, or experimental psycholinguistics, since it is more theoretically oriented and is specifically from a Minimalist framework.
Although grammaticalization is not specifically addressed in this volume, some articles will be of interest to those working in this field, and there exists good fodder here for future research. For example, the Fujimori article on telicity in Japanese, the Pietroski article on lexical concepts, the Isac article on imperative force, the Grohmann and Leivada article on Greek clitics, the Christodoulou and Wiltschko article on the Greek subjunctive, and the Malaia and Wilbur article on sign language phonology all have implications for grammaticalization, lexicalization, and related processes in language development.
However, in terms of a theory of interface, a noticeable omission is the failure to consider fairly well developed models of interface in previous literature, most notably, Jackendoff’s ideas of language architecture and interface (Jackendoff 1997, 2002). The Jackendoff model describes interface components in the architecture and makes use of a parallel processing architecture, which is consistent with what is known in neuroscience about brain functioning and processing, as are some connectionist models (e.g. Elman et al. 1996). For interface theory, more processing studies would also be in order. One of the few processing studies here, the Almeida and Riven article on semantic indeterminacy, interestingly shows the pragmatics-semantics interface as seen in effects on processing times. This fits with a number of other psycholinguistic studies showing processing times (1) for semantically anomalous sentences, and (2) when so-called bridging inferences must be made in discourse interpretation (e.g. Kintsch 1993).
Several computational experiments in this volume aim to show the computational feasibility of particular analyses. One caution to bear in mind is that demonstrating one form of computational feasibility shows that a particular model is computationally possible. It does not necessarily mean that that is how the human mind or language faculty does things; the mental architecture might take a different route than what a mathematical model assumes. This potential pitfall can be clearly seen, for example, in some earlier amodular connectionist experiments, that is, in connectionist models that assume no modularity within the language faculty and that the language faculty is distributed and unidimensional (as in Elman et al. 1996). Such experiments lead to successful machine learning or implementation of small bits of syntax or phonology, but without empirical validation or analysis of what happens inside the network, the results are not definitive, as they may not reflect what the brain actually does. As with the computational studies in this volume, empirical validation would be needed with controlled experiments of language production and processing in human subjects.
A few articles in this volume attempt to address language evolution, particularly the Samuels article (‘The emergence of phonological forms’), which posits that vocal or phonetic capabilities evolved first, which were later co-opted by the grammatical system for linguistic expression. It seems plausible that phonology and syntax co-evolved in tandem, or that vocalization developed first and was co-opted by syntax. Or it is plausible that vocalization arose first, but the process of being co-opted by the language faculty caused vocalization to co-evolve into a formal phonological component in the language architecture. Language evolution research is difficult, since little hard evidence is available today. Samuels looks for evidence in the “fossil record” of the modern human language faculty, and animal and human cognition, and this thesis is one that merits further research in evolutionary psychology. Also, the articles on language development and evolution (e.g. the Samuels article on phonology, the Reiss article on underspecification, the Pietroski article on lexical concepts) raise interesting questions on how phonology, syntax, the mental lexicon, and semantics might have co-evolved.
Other than a few articles, a thorough discussion of language evolution seems lacking in this volume. Jackendoff (2002), for example, describes the evolutionary hurdles to be overcome in language evolution, which deserve more attention here (these hurdles include the use of linguistic symbols such as words and segmental phonology, to allow for a larger and more productive lexicon, linear position of words for syntax and semantics, phrase structure, grammatical categories, and morphology). The theory of group selection from sociobiology also deserves serious attention in discussing the biological foundations of language and language architecture. Group selection posits that genes are selected because of the advantages they confer upon a whole group, not just individuals (e.g. Wilson 2006). For a cognitive faculty such as language, the role of group selection (if truly a mechanism of evolution) in language development merits serious research and discussion, as it could help us understand how language arose as a communicative tool in social contests. For such a volume, more on the evolution behind syntactic structure and interface would also be desirable. More cross-disciplinary work is needed in language evolution, such as evolutionary genetics (e.g. regarding the expression of the FOXP2 “language gene” in humans and our ancestors) and evolutionary psychology to better address questions of how language arose in humans.
Overall, the various articles in this volume will be of interest to those working within the Minimalist framework, Minimalist approaches to interface, and computational linguistics, and to those interested in language evolution, grammaticalization, and related issues of language development.
Elman, Jeffrey; Karmiloff-Smith, Annette; Bates, Elizabeth; & Johnson, Mark. 1996. Rethinking Innateness: A Connectionist Perspective on Development. Cambridge, MA: MIT Press.
Jackendoff, Ray. 2002. Foundations of Language. Oxford: Oxford University Press.
Jackendoff, Ray. 1997. The Architecture of the Language Faculty. Cambridge, MA: MIT Press.
Kintsch, Walter. 1993. Information accretion and reduction in text processing: Inferences. Discourse Processes, 16:1-2, 193-202
Wilson, D. S. 2006. Human groups as adaptive units: toward a permanent consensus. In P. Carruthers, S. Laurence & S. Stich (Eds.), The Innate Mind: Culture and Cognition. Oxford: Oxford University Press.
| ABOUT THE REVIEWER:
Kent Lee is a Research Professor at the Center for Teaching and Learning at Korea University, Seoul, Korea. His research interests include cognitive foundations of language, language education, psycholinguistics, writing studies, and phonology.