The Growth of Language: Universal Grammar, Experience, and Principles of Computation
Total Page:16
File Type:pdf, Size:1020Kb
Neuroscience and Biobehavioral Reviews 81 (2017) 103–119 Contents lists available at ScienceDirect Neuroscience and Biobehavioral Reviews jou rnal homepage: www.elsevier.com/locate/neubiorev Review article The growth of language: Universal Grammar, experience, and principles of computation a,∗ b c d e,f Charles Yang , Stephen Crain , Robert C. Berwick , Noam Chomsky , Johan J. Bolhuis a Department of Linguistics and Department of Computer and Information Science, University of Pennsylvania, 619 Williams Hall, Philadelphia, PA 19081, USA b Department of Linguistics, Macquarie University, and ARC Centre of Excellence in Cognition and its Disorders, Sydney, Australia c Department of Electrical Engineering and Computer Science and Department of Brain and Cognitive Sciences, MIT, Cambridge, MA 02139, USA d Department of Linguistics and Philosophy, MIT, Cambridge MA, USA e Cognitive Neurobiology and Helmholtz Institute, Departments of Psychology and Biology, Utrecht University, Utrecht, The Netherlands f Department of Zoology and St. Catharine’s College, University of Cambridge, UK a r a t i b s c t l e i n f o r a c t Article history: Human infants develop language remarkably rapidly and without overt instruction. We argue that the Received 13 September 2016 distinctive ontogenesis of child language arises from the interplay of three factors: domain-specific prin- Received in revised form ciples of language (Universal Grammar), external experience, and properties of non-linguistic domains 10 November 2016 of cognition including general learning mechanisms and principles of efficient computation. We review Accepted 16 December 2016 developmental evidence that children make use of hierarchically composed structures (‘Merge’) from the Available online 7 January 2017 earliest stages and at all levels of linguistic organization. At the same time, longitudinal trajectories of development show sensitivity to the quantity of specific patterns in the input, which suggests the use of Keywords: probabilistic processes as well as inductive learning mechanisms that are suitable for the psychological Language acquisition constraints on language acquisition. By considering the place of language in human biology and evolu- Generative grammar Computational linguistics tion, we propose an approach that integrates principles from Universal Grammar and constraints from Speech perception other domains of cognition. We outline some initial results of this approach as well as challenges for Inductive inference future research. Evolution of language © 2017 Elsevier Ltd. All rights reserved. Contents 1. Internal and external factors in language acquisition . 104 2. The emergence of linguistic structures . 104 2.1. Hierarchy and merge . 104 2.2. Merge in early child language . 105 2.3. Structure and interpretation . 107 3. Experience, induction, and language development . 108 3.1. Sparsity and the necessity of generalization . 109 3.2. The trajectories of parameters . 110 3.3. Negative evidence and inductive learning . 112 4. Language acquisition in the light of biological evolution . 114 4.1. Merge, cognition, and evolution . 114 4.2. The mechanisms of the language acquisition device . 115 Acknowledgments . 116 References . 116 ∗ Corresponding author. E-mail address: [email protected] (C. Yang). http://dx.doi.org/10.1016/j.neubiorev.2016.12.023 0149-7634/© 2017 Elsevier Ltd. All rights reserved. 104 C. Yang et al. / Neuroscience and Biobehavioral Reviews 81 (2017) 103–119 1. Internal and external factors in language acquisition to the design of language and therefore crucial for the acquisition of language (Chomsky, 2005): Where does language come from? Legend has it that the Egyp- tian Pharaoh Psamtik I ordered the ultimate experiment to be a Universal Grammar: The initial state of language development conducted on two newborn children. The story was first recorded in is determined by our genetic endowment, which appears to be Herodotus’s Histories. At the Pharaoh’s instruction, the story goes, nearly uniform for the species. At the initial state, infants inter- the children were raised in isolation and thus devoid of any linguis- pret parts of the environment as linguistic experience; this is tic input. Evidently pursuing a recapitulationist line, the Pharaoh a nontrivial task which infants carry out reflexively and which believed that child language would reveal the historical root of regulates the growth of the language faculty. all human languages. The first word the newborns produced was b Experience: This is the source of language variation, within a bekos, meaning “bread” in the now extinct language, Phrygian. This fairly narrow range, as in the case of other subsystems of human identified Phrygian, to the Pharaoh’s satisfaction, as the original cognition and the formation of the organism more generally. tongue. c Third factors: These factors include principles that are not spe- Needless to say, this story must be taken with a large pinch of cific to the language faculty, including principles of hypothesis salt. For starters, we now know that consonants such as /s/ in bekos formation and data analysis, which are used both in language are very rare in early speech: the motor control for its articula- acquisition and in the acquisition of knowledge in other cog- tion requires forcing airflow through partial opening of the vocal nitive domains. Among the third factors are also principles of tract (hence the hissing sound), and this takes a long time for chil- efficient computation as well as external constraints that regulate dren to master and is often replaced by other consonants (Locke, development in all biological organisms. 1995; Vihman, 2013). Furthermore, the very first words children produce tend to follow a consonant-vowel (CV) template: even if In the following Section we discuss several general principles /s/ were properly articulated, it would almost surely be followed of language and the remarkably early emergence of these princi- by an epenthetic vowel to maintain the integrity of a CV alterna- ples in children’s biological development. Section 3 focuses on the tion. But the Pharaoh did get two important matters right. First, he role of language-specific experience and how the latitude in chil- was correct to assume that language is a biological necessity—in dren’s experience shapes language variation during the course of Darwin’s words, “an instinctive tendency to acquire an art”. He acquisition. Section 4 reviews some recent effort to reconceptu- was also correct in supposing that, Phrygian or otherwise, a lan- alize the problem of language acquisition, focusing specifically on guage can emerge simply by putting children together, even in the how domain-general learning mechanisms and the principles of absence of an external model. Indeed, we now know that sign lan- efficient computation figure into the Language Acquisition Device. guages are often spontaneous creations by deaf communities, as recently documented in Nicaragua and in Israel (Kegl et al., 1999; 2. The emergence of linguistic structures Sandler et al., 2005; Goldin-Meadow and Yang this issue). Second, the Pharaoh was correct in supposing that child language develop- As observed long ago by Wilhelm von Humboldt, human lan- ment can reveal a great deal about the nature of language, its place guage makes “infinite use of finite means”. It is clear that to acquire in the mind, and how it emerged as a biological capacity that is a language is to implement a combinatorial system of rules that unique to our species. generates an unbounded range of meaningful expressions, relying The nature of language acquisition has always been a central on the biological endowment that determines the range of possi- concern in the modern theory of linguistics known as generative ble systems and a specific linguistic environment to select among grammar (Chomsky, 1957; this issue). The notion of a Language them. von Humboldt’s adage underscores a central feature of lan- Acquisition Device (Chomsky, 1965), a dedicated module of the guage; namely, linguistic representations are always hierarchically mind, still features prominently in the general literature on lan- organized structures, at all levels of the linguistic system (Everaert guage and cognitive development. In this article, we highlight some et al., 2015). These structural representations are formed by the major empirical findings that, together with additional case stud- combinatorial operation known as ‘Merge.’ ies, form the foundation of future research in language acquisition. At the same time, we present some leading ideas from theoretical investigations of language acquisition, including promising themes 2.1. Hierarchy and merge developed in recent years. For several of the additional case studies, see Crain et al. (this issue). The hierarchical structure of language was recognized at the Like the Crain et al. paper, our perspective on the acquisition very beginning of generative grammar. The complexity of human of language is shaped by the biolinguistic approach (Berwick and language lies beyond the descriptive power of finite state Markov Chomsky, 2011). Language is fundamentally a biological system processes (Chomsky, 1956), as exemplified in sentences that and should be studied using the methodologies of the natural sci- involve recursive nonlocal dependencies: (1) a. If . Then . ences. Although we are still quite distant from fully understanding b. Either . Or . the biological basis of language, we contend that the develop- c.