What is Language?

Tuesday, October 4, 2011

Robert McCauley and Susan Tamasi

Last spring, Dr. Susan Tamasi (Linguistics) and the CMBC’s own Dr. Robert McCauley co-taught the CMBC-sponsored undergraduate course Language, Mind, and Society (LING 301), a required course for the Linguistics major and the joint major in Psychology and Linguistics. On the first day of class, students were asked to complete a short writing assignment, tackling the question “What is language?” Fittingly, this question also served as the launching point for discussion when Tamasi and McCauley reunited for the CMBC’s first lunch seminar of the fall semester on Thursday, September 22. Following opening remarks from each speaker, student and faculty attendees across a range of disciplines offered their own insights on the nature of language. In this commentary, I’ll highlight some of the key issues raised in the speakers’ presentations and in the lively group discussion that ensued, while also weaving in my own thoughts and reactions.

In her opening remarks, Tamasi noted the difficulty of coming up with any single definition of language. Various definitions offered by students in LING 301 focused on language’s role as a biological construct, a reflection of thought, a transmitter of information, and a marker of social identity, with some arguing that these characteristics render language a uniquely human trait. Tamasi suggested that defining language may not simply be a matter of combining all of these characteristics; rather, there may be some additional, perhaps intangible, property yet to be identified that captures what language is at its core. The fundamental nature of language, Tamasi went on to suggest, cannot be identified merely by examining individual languages, nor by focusing exclusively on the structure of language apart from how it is used by individual speakers and communities.

McCauley’s opening remarks centered on defining language as an abstract concept, or system, akin to the study of religion, as opposed to individual religions. Drawing on his perspective from the philosophy of science, McCauley suggested that explanations of systems are most fruitful when they are mechanistic; that is, when they define the parts, their organization, and their contributions to the operation of the system. Chomsky’s conception of language as primarily a system of thought, rather than a medium of communication, lends itself to such mechanistic analysis (even if Chomsky himself has not pursued that end). According to McCauley, psychological accounts of language have the benefit of localizing the mechanisms underlying language to a physical substrate, namely the brain, with advances in the study of cognition in turn contributing to our understanding of language. McCauley also characterized language as one of several maturationally natural systems, defined in part by their significance in addressing the basic problems in life, their comparatively early appearance in development (within the first two decades of life), the automaticity with which they are engaged, and their lack of dependence on culturally distinctive support.

One issue that came up in discussion was how these criteria apply in cases of so-called feral children (e.g., “Genie”), those who grow up isolated from human contact from a very young age and who consequently are unable to develop normal language abilities, among other cognitive skills. If language does not depend on culturally distinctive support, it might seem that such children should be able to acquire language, certainly once discovered and exposed to the social world. According to McCauley, however, invoking “culture” in such cases may be overly grandiose; what feral children lack may be reduced simply to the opportunity to interact with one other conspecific. Although interactions between individuals must be critical to the typical development of language, they constitute culture in only its most minimal sense. McCauley’s point is that there is nothing special about the social interaction necessary for acquiring language, only that it must be present in some form. While parsimonious, this line of reasoning does not address the question, what is culture? At what point does a particular type of experience become distinctively cultural rather than merely providing the “bare bones” of culture? For ethical reasons, it has often been regarded as impossible to isolate the social ingredients essential for acquiring language, but perhaps greater understanding of the cognitive mechanisms supporting various social phenomena may shed light on what types of social experience are necessary versus merely ornamental.

Another discussion point concerned the extent to which language is unique to humans. Non-human animals certainly have communication systems (and some have even been successful at acquiring sizeable vocabularies), but Tamasi suggested that equating such systems with human language assumes that the principal function of language is to communicate information. Instead, Tamasi stressed, we must take seriously the structural complexity of language at multiple levels (e.g., phonological, morphological, syntactic, etc.), which allows for the coordination of thought and a level of generativity unseen in other communication systems. But do such systems possess a hidden complexity that we cannot recognize simply from observable behaviors? Perhaps, but while we cannot zoom in on the structure of such systems directly, there is one place we can look: the brain. We might posit that the “waggle dance” of the bee, which serves to recruit other bees to forage in the same area, is incredibly complex, but it is unclear how the neural system of the bee could support such complexity, particularly given that we now have some understanding of how properties of human language like recursion are neurally instantiated. As McCauley noted, knowing where to look for the mechanisms underlying language offers a methodological opportunity to understand the nature of linguistic complexity. This opportunity may be wasted if we insist on parity across species.

Interestingly, however, many linguists assume parity across individual languages. Although one language might be regarded as more complex than another in some aspect of its structure (with consequences for how easily such structure is learned), languages are assumed to be equally complex overall because they all serve the same purpose. But if languages are defined by their complexity, this argument seems circular. If some recently discovered language was found to be less complex overall, we would have to conclude that it was not in fact a language (or, alternatively, that its complexity had yet to be discovered). I wonder if it would benefit the field to abandon the notion that there are necessary and sufficient features for language, given the observation from cognitive science that it is virtually impossible to define necessary and sufficient features foranything, whether it be chairs, games, or ideas. Acknowledging a continuum of complexity in language would not render the construct “language” meaningless; it would merely suggest that there is no clear distinction between what counts as a language and what doesn’t. Of course, given that language has often been viewed as a window into the mind, abandoning the parity assumption might open another can of worms by implying an extreme form of linguistic relativity, or cognitive differences among speakers of different languages. Nevertheless, recent research indicates that there is no simple one-to-one mapping between language and the conceptual system (e.g., Malt et al., 2011; for a review, see Wolff & Holmes, 2011), suggesting that differences in linguistic complexity do not necessarily entail analogous cognitive differences.

One important aspect of what defines language seemed to be missing from the lunch discussion: the world. As cognitive scientists have noted, the world is richly structured, with inherent discontinuities in how properties are distributed. For example, concrete objects like dogs and tables form more coherent perceptual bundles than relational notions typically encoded in verbs and prepositions (e.g.,throw and in). This structure constrains which components of meaning are encoded in language more generally and also serves as a standard of comparison when considering semantic variation across languages. Recent evidence suggests that universal properties of human perceptual experience may lead to a conceptual space largely shared across languages, with different languages partitioning the space differently. Our understanding of language may be enriched by considering the complex interactions among mind, world, and society that give rise to this fundamental human capacity.


Malt, B. C., Ameel, E., Gennari, S., Imai, M., Saji, N., & Majid, A. (2011). Do words reveal concepts? In L. Carlson, C. Hölscher, & T. Shipley (Eds.),Proceedings of the 33rd Annual Conference of the Cognitive Science Society (pp. 2884-2889). Austin, TX: Cognitive Science Society.

Wolff, P., & Holmes, K. J. (2011). Linguistic relativity. Wiley Interdisciplinary Reviews: Cognitive Science2, 253-265.

About Kevin Holmes

Kevin Holmes earned B.A. and M.A. degrees from Stanford University and a Ph.D. in Psychology (Cognition & Development) from Emory University. He subsequently completed a postdoctoral fellowship in Cognitive Science at UC Berkeley. Since 2014, he has served as an Assistant Professor of Psychology at Colorado College. Kevin's research investigates the structure of human thought, exploring how the mental categories we rely on to think, perceive, and act upon the world are related to the languages we speak, as well as how people think and reason about concepts of space, time, and number.
This entry was posted in 2011 Archives. Bookmark the permalink.

Leave a Reply

Your email address will not be published. Required fields are marked *