Margery Lucas (1999) Selection Pressures for Language Prerequisites: Constraints and Limitations. Psycoloquy: 10(050) Language Prerequisites (2)

Volume: 10 (next, prev) Issue: 050 (next, prev) Article: 2 (next prev first) Alternate versions: ASCII Summary
PSYCOLOQUY (ISSN 1055-0143) is sponsored by the American Psychological Association (APA).
Psycoloquy 10(050): Selection Pressures for Language Prerequisites: Constraints and Limitations

Commentary on Burling on Language-Prerequisites

Margery Lucas
Department of Psychology
Wellesley College
Wellesley, MA 02481


Burling's proposal to find the selection pressures for naming and its cognitive prerequisites is critically evaluated, and it is argued that speculation along the lines he proposes is likely to be premature and limited in scope. The possibilities for making an adaptationist case for natural selection of the cognitive capacities underlying naming are considered in light of methodological constraints particular to the study of language and its related cognitive supports.


language; cognition; natural selection
1. Burling (1999) proposes four prerequisites for naming ability: a rich understanding of the external world, a theory of mind, use of icons and indices, and the ability to imitate. Before the question of selection pressure for these abilities is addressed, however, it would be useful to highlight some of Burling's assumptions which are potentially problematic. The first involves the four prerequisites that he proposes. These seem to be reasonable, but are they as self-evident as Burling suggests? There are some reasons to think otherwise. First, Burling finds his prerequisites by examining the capacities which children need in order to learn words. This suggests that Burling believes we can assume that ontogeny in some way recapitulates phylogeny, a strategy long-discredited in the biological sciences (Gould, 1977). For example, Burling claims that the first words in prehistory can never have been used alone because children convey a variety of meanings by using a single word embedded in contexts and accompanied by intonational and gestural cues. Although it is true that children do this, it is not clear that our hominid ancestors invariably relied on such contextual supports to supplement words across a variety of situations. Maybe there was a stage in the evolution of word use wherein words were used within a narrow range of situations and with limited contextual support, in order to signal a more productive (than animal calls) but still very limited range of meanings. This stage could have been lost in the progression from animal calls to full-blown naming just as the earlier stage of avian teeth is no longer seen in the development of the modern avian beak (from Kollar & Fisher's study, as cited in Lock & Peters, 1996). Such dissociations between phylogeny and ontogeny suggest that simple models of the relationship between child language acquisition and hominid language evolution, although seductive, are inadequate.

2. Furthermore, recent work in this area by other cognitive scientists suggests that Burling's list of cognitive prerequisites is either erroneous or incomplete. Donald (1991) has presented extensive evidence which suggests that mimesis, not imitation, was the essential precursor to naming. Although he acknowledges the importance of the use of icons and indices, Deacon (1997) argues that sophisticated mnemonic strategies and the ability to solve combinatorial problems are necessary before any system of symbols can be developed. Maybe some convergence can be found between Burlingr's list and other proposals, but the point is that before we can fruitfully discuss selection pressures, there needs to be agreement on the cognitive prerequisites. At present, that agreement appears to be lacking. It is possible that a comparative strategy may foster consensus on a list of necessary cognitive prerequisites if the study of nonhuman primates can shed light on the cognitive differences between primates that can use symbols and those that cannot. But the field of ape language is particularly contentious and the achievements of most primates learning language are overestimated (Seidenberg & Petitto, 1979; Terrace, 1979). The most ambitious claims have been made for the bonobo Kanzi (Savage-Rumbaugh, 1991), but even Kanzi's naming abilities are questionable (Pinker, 1994). The only and best hope for a comparative strategy for assessing cognitive prerequisites rests on acquiring stronger evidence for symbol use in the bonobo, along with stringent assessments of the cognitive abilities in that species line, compared to other species that have shown little talent in acquiring a symbol system.

3. Another of Burling's debatable assumptions is that language was an adaptation. However, there are least two respected scholars who have doubts about this (Chomsky, 1972, 1988;Gould, 1987). Although many accept Pinker and Bloom's (1990) arguments that language, like the eye, is too complex to have occurred by successive random mutations or genetic drift, language is not like the eye. The physiological structures supporting the eye are easily observable, but the same is not true for language and its cognitive supports. Even for words, there is disagreement and uncertainty about the brain structures involved (Pulvermuller, 1999). Many assume a neurological complexity that is, as yet, not evident. It is not impossible to imagine that relatively minor tweaks in the developmental program for the brain led to major changes in the cognitive abilities that permitted naming. Moreover, disputes in the arena of language innateness and development (e.g., the well-known controversy over learning the past tense of English verbs) indicate that it would be premature to take the adaptationist view of language for granted. The existence of a thriving industry in connectionist modelling of language abilities demonstrates that there is a viable alternative to the adaptationist account of language, the view that language phenomena, complex cognitions that they are, are the result of more domain-general cognitive mechanisms. On this view, the abilities that Burling lists (theory of mind, iconicity, imitation) may be, in fact, not the prerequisites for language but constitutive of language ability in and of themselves.

4. The lack of consensus on the cognitive prerequisites for language and doubts about the adaptationist position suggest that Burling's questions regarding selection pressures are premature. A less problematic version of his questions might be the following: if the adaptationists are correct, what kind of case, if any, could be made for the evolutionary origins of naming and its cognitive prerequisites, for any given list of prerequisites? Even without a consensus list of prerequisites, attempting to answer this question would not just be an exercise in idle speculation. It is important to know in advance of serious investigation into any scientific topic what methodologies, if any, are available to answer questions.

5. There have been several serious attempts to investigate the selection pressures for language and its cognitive prerequisites (Pinker and Bloom. 1990; Donald, 1991; Deacon, 1997; Dunbar, 1993 among others). Some of these have focused on syntax, others on word use. Burling wants to examine the selection pressures on the cognitive prerequisites, as well as on naming. Whatever the focus, the danger inherent in all these attempts is that compelling speculations will be elevated to the status of scientific theory. In order to have a viable scientific theory, there must be the possibility of testable hypotheses. Traditionally, fossil evidence supports evolutionary hypotheses, but there is, arguably, no reliable fossil record for language abilities or its precursors, beyond evidence for changes in brain size and laterality (Holloway, 1996). These limited findings present two problems. First, as Holloway points out, it is impossible to tell whether the changes in asymmetries indicate changes related to tool use and handedness, or whether they also signal changes related to naming and its cognitive prerequisites. Although the evidence for laryngeal mechanisms for speech occurred much later than the changes in brain size and laterality, we cannot rule out the use of gestural symbol systems in early hominids (Hewes, 1996). For lack of a written record, we will never know if homo habilis or homo erectus had naming capabilities or just the prerequisites for naming or unrelated cognitive and motor abilities. Secondly, the fossil record can only reveal very gross changes in neuroanatomy (Holloway, 1996). More detailed features and brain development below the surface of the cortex would not leave their imprint on an endocranial cast, even if we knew all the relevant neural areas to look for. This means that the fossil record will always be silent with regard to the evolutionary progression of many of the neural structures underlying language ability. Deacon (1997) has tried to compensate for these constraints on hypothesising about the evolution of language and the brain by using a comparative strategy to look at our nearest primate relatives (a sort of living fossil record). He has done a brilliant job of using comparative neuroanatomy to trace the changes in the neocortex that accompanied the ability to use language. Without an accompanying fossil and archaeological record that indicates when and under what conditions those changes occurred, however, there is no way to know for sure what selection pressures might have led to those changes.

6. Another strategy employed in determining the selection pressures for a process is used in evolutionary psychology (see Barkow, Cosmides, and Tooby, 1992, for a representative body of such work). For example, Cosmides and Tooby (1992) describe an elegant series of experiments designed to demonstrate that solving certain types of reasoning problems involves the use of a special purpose reasoning mechanism; the cheater detection module that would have facilitated social exchange in the Pleistocene era. This strategy relies upon establishing that there was a special design of the cognitive process in question. This is done by showing that the mechanism operates with precision and efficiency with regard to the function the mechanism serves (from Williams, 1966). Then, the mechanism is linked to conditions in the environment of our hominid ancestors in the Pleistocene, the age during which certain hominid lines began to exhibit the changes that eventually led to the emergence of homo sapiens.

7. The problem with the first part of this strategy as it relates to language lies in establishing the functions which naming and its prerequisites would serve. It seems obvious that naming and related cognitive abilities such as those suggested by Burling would have greatly increased survival rates for any organism living on the African savannah 1.5 million years ago. However, teasing out the relationships among the functions for naming and its prerequisites, other components of language ability, and possible environmental pressures is complicated by the possibility that these relationships may not be straightforward. Implicit in Burling's article is the idea that naming can be isolated from other aspects of language and its prerequisites, and that the selection pressures on naming can be found independently of those for its prerequisites or for other aspects of language ability. This strategy assumes accepting modularity and modularity, which, although popular in many quarters, is by no means well-established fact. Recently, Pulvermuller (1999) has persuasively argued that words and their meanings are distributed across several different brain systems, with Broca's and Wernicke's areas just being particularly intense areas of processing. According to this view, it becomes possible to conceive that whatever neural changes were required to allow for certain cognitive prerequisites (theory of mind, say) these were also sufficient to allow for naming. That is, it may be that once certain cognitive abilities were in place, the neural structures that supported them allowed naming and perhaps even syntax to develop. This possibility is suggested by Gould's hypothesis that language is a spandrel (as cited in Pinker & Bloom, 1990). Although some, most famously Pinker and Bloom, find this hypothesis implausible, there is no hard evidence against Gould's thesis. Until we know more about the neural structures serving language and its prerequisites, that is, whether there are domain-specific modules along the lines suggested by Fodor (1983) or Pinker (1997), it is prudent to be open to all possibilities.

8. Determining the selection pressures for language prerequisites is not made easier by considering the environment of our hominid ancestors, which is the other part of the strategy employed by evolutionary psychologists. Although it seems likely that the selection pressures for naming and its cognitive prerequisites facilitated social communication, cognitive abilities such as the use of icons or theory of mind would have given early hominids such an advantage over other organisms in so many situations encountered in the Pleistocene (not just in social situations but in enhanced internal planning and problem-solving), that it is difficult if not impossible to know which of early environmental pressures led to the improvements in cognitive abilities listed by Burling. The possible adaptive consequences of naming and its immediate precursors are not limited or obvious. Many of the probable functions are tied to social life and there is no archaeological record of social interaction during the relevant period of time. More importantly, there is little possibility of ever finding such a record.

9. Speculation about the evolutionary origins of language is fun, and satisfies the intuitions of some that language has an innate basis in a brain mechanism or set of mechanisms developed specifically for that purpose. But although evolutionary speculation is possible and useful in some areas of cognitive science, we must be especially wary of the pitfalls involved in speculating too freely in advance of evidence about the origins of language, for the reasons given in this paper. It is all too easy to accept reasonable and compelling stories about language origins as valid scientific theories, instead of the entertaining myths that they are. They may never be more than myths because the relevant aspects of the past that would help to elevate them to the status of scientific theory are inaccessible to us, and may always be so.

ACKNOWLEDGEMENTS: I would like to thank Pat Hayes for comments and suggestions.


Barkow, J.H., Cosmides, L., & Tooby, J. (1992). The Adapted Mind: Evolutionary Psychology and the Generation of Culture. New York: Oxford University Press

Burling, R. (1999) The cognitive prerequisites for language: Target article on language prerequisites. PSYCOLOQUY 10(32). psyc.99.10.032.language-prerequisites.1.burling

Chomsky, N. (1972). Language and Mind (extended edition). New York: Harcourt, Brace, & Jovanovich.

Chomsky, N. (1988). Language and Problems of Knowledge: The Managua Lectures. Cambridge, MA: MIT Press.

Cosmides, L. & Tooby, J. (1992). Cognitive adaptations for social exchange. In J.H. Barkow, L. Cosmides, & J. Tooby (Eds.), The Adapted Mind: Evolutionary Psychology and the Generation of Culture. New York: Oxford University Press.

Deacon, T. (1997). The Symbolic Species: The Coevolution of Language and the Brain. New York: W.W. Norton.

Donald, M. (1991). Origins of the Modern Mind: Three Stages in the Evolution of Culture and Cognition. Cambridge, MA: Harvard University Press

Dunbar, R. I. M. (1993). Coevolution of neocortical size, group size, and language in humans. Behavioural & Brain Sciences, 16:681-735.

Fodor, J.A. (1983). The Modularity of Mind. Cambridge, MA: MIT Press.

Gould, S.J. (1977). Ontogeny and Phylogeny. Cambridge, MA: Harvard University Press.

Hewes, G.W. (1996). A history of the study of language origins and the gestural primacy hypothesis. In A. Lock & C.R. Peters (Eds.), Handbook of Human Symbolic Evolution, New York: Oxford University Press.

Lock, A. & Peters, C.R. (1996). Ontogeny: Symbolic Development and Symbolic Evolution. In A. Lock & C.R. Peters (Eds.), Handbook of Human Symbolic Evolution, New York: Oxford University Press.

Pinker, S. (1994). The Language Instinct. New York: William Morrow & Co., Inc.

Pinker, S. (1997). How the Mind Works. New York: W.W. Norton.

Pinker, S. & Bloom, P. (1990). Natural language and natural selection. Behavioural & Brain Sciences, 13:707-784.

Pulvermuller, F. (1999). Words in the brain's language. Behavioural & Brain Sciences. 22, 253-336.

Savage-Rumbaugh, E.S. (1991). Language learning in the bonobo: How and why they learn. In N.A. Karasnegor, D.M., Rumbaugh, R.L. Schiefelbusch, & M. Studdert-Kennedy (Eds.). Biological and Behavioural Determinants of Language Development. Hillsdale, N.J.: Erlbaum.

Terrace, H. S. (1979). Nim. New York: Knopf.

Williams, G.C. (1966). Adaptation and Natural Selection. Princeton: Princeton University Press.

Volume: 10 (next, prev) Issue: 050 (next, prev) Article: 2 (next prev first) Alternate versions: ASCII Summary