On the Human: Rethinking the natural selection of human language

Introduction

Since Darwin’s time, the human language capacity has been a perennially cited paragon of extreme complexity that defies the explanatory powers of natural selection. And it is not just critics of Darwinism who have argued that this most distinctive human capacity is problematic. Alfred Russel Wallace—the co-discoverer of natural selection theory and in many ways more of an ultra-Darwinian than Darwin himself—famously argued that the human intellectual capacity which makes language possible, is developed to a level of complexity that far exceeds what is achievable through natural selection alone. While fiercely defending natural selection theory with respect to the traits of other species, he argued that in the case of humans, “… natural selection could only have endowed the savage with a brain a little superior to that of an ape.” (p. 392) And Charles Lyell—who personally promoted Darwin’s work and generally supported the evolutionary perspective—also worried that language was just too complex to have evolved by natural means. Not only are the vast vocabulary and baroquely structured grammar and syntax of even the most simple of natural languages orders of magnitude more complex than any other species’ communication system, but the capacity this all provides for expressing esoteric concepts and conveying aesthetic experiences seems far removed from anything with direct adaptive consequence.

Darwin himself fretted over the possibility that natural selection alone might be incapable of accounting for exaggerated functional complexity in nature. In a letter he wrote to Asa Gray shortly after the publication of On the Origin of Species, he admits that “The sight of a feather in a peacock’s tail, whenever I gaze at it, makes me feel sick!” Despite the spectacular and elaborately formed details of this adornment, it was a burden that negatively impacted health and survival and so could not have been the subject to natural selection with respect to the environment. But it was the extravagance of traits such as this, despite their lack of utility, that suggested to Darwin an approach to the challenge of explaining human mental capacities.

In the case of the peacock tail, and other similar traits, Darwin realized that, indeed, something other than natural selection with respect to environmental conditions was responsible. Recognizing that reproduction rather than individual survival was the critical factor in evolution, he argued that competition with respect to reproductive access (sexual selection) could result in runaway selection on certain traits, independent of their environmental suitability. Darwin argued that a display feature or fighting ability that led an individual to out-compete others in gaining access to mates would also favor proliferation and evolutionary exaggeration of these traits, even at some cost to individual health and survival. Analogously, he postulated that selection with respect to sex might also explain such extravagant and highly divergent traits as human language. In his book The Descent of Man and Selection in Relation to Sex—which is typically referred to by only the first half of its title—he argues that language and other human traits that appear exaggerated beyond survival value, can be explained as consequences of sexual selection. So, for example, he imagines that language might have evolved from something akin to bird song, used as a means to attract mates, and that the ability to produce highly elaborate vocal behaviors was progressively exaggerated by a kind of arms-race competition for the most complex vocal display.

Unfortunately, there are strong reasons for doubting the relevance of sexual selection to this most distinctive of human traits. This is because sexual selection inevitably produces complementary divergence of male and female traits, as is exemplified by peacock tails and moose antlers, which are exhibited only by males. While there are indeed a few highly divergent traits distinguishing women from men (e.g. patterns of fat deposition in breasts and hips, etc.), the sexes differ only very subtly in their intellectual and language abilities. Thus accounting for the extravagant complexity of language in terms of sexual selection requires explaining why it lacks this otherwise ubiquitous mark of extreme sexual dimorphism. To explain the origin of the highly structured human-unique adaptation inevitably requires addressing Wallace’s challenge concerning the complexity and apparent non-adaptive aspects of these features.

Long evolution in an artificial niche

In my work I use the phrase, symbolic species, quite literally, to argue that symbols have literally changed the kind of biological organism we are. I believe that we think and behave in many ways that are quite odd compared to other species because of the way that language has changed us. In many respects symbolic language has become a major part of the environment to which we have had to adapt in order to flourish. In the same way that our ancestors’ bodies evolved in the context of the demands posed by bipedal foraging with stone tools and incorporating meat into the diet, their brains evolved in the context of a rich fabric of symbolic cultural communication. As it became increasingly important to be able to enter into the social web of protolinguistic and other early forms of symbolic social communication in order to survive and reproduce, the demands imposed by this artificial niche would have selectively favored mental capacities that guaranteed successful access to this essential resource. So rather than merely intelligent or wise (sapient) creatures, we are creatures whose social and mental capacities have been quite literally shaped by the special demands of communicating with symbols. And this doesn’t just mean that we are adapted for language use, but also for all the many ancillary mental biases that support reliable access and use of this social resource.

But this claim depends on language-like communication being a long-time feature of hominid evolution. Theories suggesting that human language is a very recent and suddenly evolved phenomenon would not make this prediction. To them language is almost epiphenomenal. This is particularly true if the claim is that language appeared suddenly due to some marvelous accidental mutation that transformed dumb (but large brained) brutes into articulate speakers. This sort of scenario has become commonplace in recent years, though the evidence supporting it is mostly very indirect (e.g. archeological evidence of representational forms and objects for adornment, appearing in the Upper Paleolithic). I think that it is mostly a reflection of a caricatured view of the human/animal distinction and a sort of hero metaphor imposed upon the fossil evidence. The way that modern human brains accommodate language can be used as a clue to how old language is.

If language is a comparatively recent feature of human social interaction, that is if it is only, say, a hundred thousand years old or so, then we should expect that it had little effect on human brains. Any structural tweaks of brain architecture that evolved to support it would have had to be either minimal or else major but dependent on comparatively few genetic changes. A recent origin of language would give it little opportunity to impose selection pressure on human brains, so language function would not be supported by any widespread and well integrated neurological changes. This would predict that language abilities are essentially an evolutionary after-thought, inserted unsystematically into an otherwise typical (if enlarged) ape brain. With little time for the genetic fixation of many supportive traits to occur, this adaptation would likely depend on only a few key genetic and neurological changes. As a consequence, language function should be poorly integrated with other cognitive functions, relatively fragile if faced with impoverished learning contexts, susceptible to catastrophic breakdown as a result of certain small but critical genetic defects, and severely affected by congenital mental impairment.

None of these seems to be the case.

On the other hand, if language has been around for a good deal of our evolutionary past, say a million years or so, that amount of time would have been adequate for the demands of language to have affected brain evolution more broadly. A large network of subtle gene changes and neurological adjustments would be involved, and as a result it should be a remarkably well integrated and robust neurological function. Indeed, there is ample evidence to suggest that language is both well-integrated into almost every aspect of our cognitive and social lives, that it utilizes a significant fraction of the forebrain, and is acquired robustly under even quite difficult social circumstances and neurological impairment. It is far from fragile.

The co-evolutionary interaction goes both ways. Languages also have to adapt to brains. Since the language one learns has to be passed from generation to generation, the more learnable its structures, and fitted to human limitations, the more effective its reproduction in each generation. Languages and brains will evolve in tandem, converging towards each other, though not symmetrically. But brain evolution is a ponderously slow and unyielding process in comparison to the more facile evolution of languages. So we should expect that languages are more modified for brains than brains are for language. Nevertheless, if we have been evolving in a symbolic niche for a million years or more, we should expect that human brains will have been tweaked in many different ways to aid life in this virtual world.

The world of symbols is an artificial niche. Its ecology is radically different than the biological niche we also find ourselves in (or at least our ancestors found themselves in). In the same way that beaver dam building has created an aquatic niche to which beaver bodies have adapted over their evolutionary history, our cognitive capacities have adapted to our self-constructed niche: a symbolic niche. This is not a new idea. Indeed the anthropologist Clifford Geertz suggested something like this many decades ago. I think that today we may be at a point in our evolutionary theorizing and our understanding of brains to begin to explore exactly what this might mean.

The most intense and unusual demands of this niche should be reflected in the ways that human cognition diverges from patterns more typical of other species. Although it has long been popular to think of the human difference in terms of general intelligence, I think this bias may have misled us into ignoring what may be a more important constellation of more subtle differences. These likely included differences in social cognition (e.g. joint attention, empathy, the ability to anticipate another’s intended actions), differences in how we learn (e.g. superior transfer learning, a predisposition to assume that associations are bidirectional—known as stimulus equivalence, a comparative ease at mimicking) or even just unusual motor capacities (e.g. unprecedented articulatory and vocal control). These are members of a widely distributed and diverse set of adaptations that fractionally and collectively contribute to our language abilities.

With respect to the brain, we need to confront another mystery. How could these many diverse brain traits have become so functionally intertwined and interdependent as to provide such a novel means of communication? This is particularly challenging to explain because language is in effect an emergent function, not some prior function just requiring fine-tuning. Our various inherited vocalizations, such as laughter, shrieks of fright, and cries of anguish, are comparatively localized in their neurological control (mostly subcortical) as are other modes of communication in animals. In comparison, language depends on a widely dispersed constellation of cortical systems, each of which can be found in other primate brains, but evolved for very different functions. These brain systems have become collectively recruited for language only because their previously evolved functions overlapped significantly with some processing demand necessitated by language, though evolved for quite different functions altogether. Indeed, the neural structures and circuits involved in the production and comprehension of language are homologous to structures found ubiquitously in most monkey and ape brains: old structures performing unprecedented new tricks.

A related mystery concerns the extent to which this dominant form of communication depends on information maintained by social transmission. Even for theories postulating an innate universal grammar, the vast quantity and high fidelity of the information constituting even a typical vocabulary stands out as exceedingly anomalous from a biological point of view. How did such a large fraction of our communicative capacity wind up offloaded onto social transmission? And what explains the remarkable reliability of this process?

Relaxed selection and complexity

Perhaps the most surprising and controversial point to be made follows from the realization of the importance of relaxed selection. The higher-order synergy of systems that contribute to language requires the cooperative functioning of component brain systems. But it appears to paradoxically require that this synergy among diverse systems must already be in place in order for selection to have honed it for language.

The co-evolutionary niche construction scenario sketched above still does not account for the generation of the novel functional synergy between neural systems that language processing requires. The discontinuities between call control systems and speech and language control systems of the brain suggest that a co-evolutionary logic alone is insufficient to explain the shift in substrate. Recent investigation of a parallel shift in both complexity and neural substrate in birdsong may be able to shed some light on this.

In a comparative study of a long-domesticated bird, the Bengalese Finch, and its feral cousin, the White-Rump Munia, it was discovered that the domesticated lineage was a far more facile song-learner with a much more complex and flexible song than its wild cousin. This was despite the fact that the Bengalese Finch was bred in captivity for coloration, not singing (Okanoya, 2004). The domestic/feral difference of song complexity and song learning in these close finch breeds parallels what is found in comparisons between species that are song-learners and non-learners. This difference also correlates with a much more extensive neural control of song in birds that learn a complex and variable song.

The fact that this behavioral and neural complexity can arise spontaneously without specific breeding for singing is a surprising finding since it is generally assumed that song complexity evolves under the influence of intense sexual selection. This was, however, blocked by domestication. One intriguing interpretation is that the relaxation of natural and sexual selection on singing paradoxically was responsible for its elaboration in this example. In brief, with song becoming irrelevant to species identification, territorial defense, mate attraction, predator avoidance, and so on, degrading mutations and existing deleterious alleles affecting the specification of the stereotypic song would not have been weeded out. The result appears to have been the reduction of innate biases controlling song production. The domestic song could thus be described as both less constrained and more variable because it is subject to more kinds of perturbations. But with the specification of song structure no longer strictly controlled by the primary forebrain motor center (called nucleus RA), other linked brain systems can begin to play a biasing role. With innate motor biases weakened, auditory experience, social context, learning biases, and attentional factors could all begin to influence singing. The result is that the domestic song became more variable, more complicated, and more influenced by social experience. The usual consequence of relaxed selection is genetic drift—increasing the genetic and phenotypic variety of a population by allowing random reassortment of alleles—but neurologically, drift in the genetic control of neural functions should cause constraints to become less specific, generating increased behavioral flexibility and greater conditional sensitivity to other neurological and contextual factors.

This is relevant to the human case, because a number of features of the human language adaptation also appear to involve a relaxation of innate constraints allowing multiple other influences besides fixed links to emotion and immediate context to affect vocalization. Probably the clearest evidence for this is infant babbling. This unprecedented tendency to freely play with vocal sound production occurs with minimal innate constraint on what sound can follow what (except for physical constraints on vocal sound generation). Babbling occurs also in contexts of comparatively low arousal state, whereas laughter, crying, or shrieking are each produced in comparatively specific high arousal states and with specific contextual associations. This reduction of innate arousal and contextual constraint on sound production, opens the door for numerous other influences to begin to play a role. Like the domesticated bird, this allows many more brain systems to influence vocal behavior, including socially acquired auditory experience. In fact, this freedom from constraint is an essential precondition for being able to correlate learned vocal behaviors with the wide diversity of objects, events, properties, and relationships language is capable of referring to. It is also a plausible answer to the combinatorial synergy problem (above) because it demonstrates an evolutionary mechanism that would spontaneously result in the emergence of multi-system coordination of neural control over vocal behavior.

But although an evolutionary de-differentiation process may be a part of the story for human language adaptation, it is clearly not the whole story. This increased flexibility and conditionality likely exposed many previously irrelevant interrelationships between brain systems to selection for the new functional associations that have emerged. Most of these adaptations remain to be identified. However, if such a dedifferentiation effect has been involved in our evolution, then scenarios hypothesizing selection for increased innateness or extrapolation from innate referential calls to words become less plausible.

Some concluding speculations

In closing, I would like to reflect on some of the more esoteric features of humanness that may be illuminated by the paired processes of symbolic niche construction effects and relaxed selection.

For example, I think it makes sense to think of ourselves as symbolic savants, unable to suppress the many predispositions evolved to aid in symbol acquisition, use, and transmission. In order to be so accomplished at this strange cognitive task, we almost certainly have evolved a predisposition to see things as symbols, whether they are or not. This is probably manifest in the make-believe of young children, the way we find meaning in coincidental events, see faces in clouds, are fascinated by art, charmed by music, and run our lives with respect to dictates presumed to originate from an invisible spirit world. Like the flight play of birds, the manipulation of objects by monkeys, the attraction of cats to small feathered toys, our special adaptation is the lens through which we see the world. With it comes an irrepressible predisposition to seek for a cryptic meaning hiding beneath the surface of appearances. Almost certainly many of our most distinctive social capacities and biases—e.g. tendencies to conformity and interest in copying the speech we hear as infants—are also reflections of this adaptation to an ecosystem of symbolic relationships. And of course there is literature and theater. How effortlessly we project ourselves into the experiences of someone else, feeling the joys and sorrows almost as intensely as our own.

Relaxation of selection, on the other hand, may have contributed to another suite of distinctively human traits. Widely distributed dedifferentiation at the genetic and epigenetic level would have increased flexibility of a variety of once phylogenetically constrained cognitive and motivational systems. Perhaps the most striking feature of humans is their flexibility and cultural variety. Consider the incredible diversity of marital and kinship organizations. Most species have fairly predictable patterns of sexual association, kin association, and offspring care, and although they are somewhat flexible, this variety is mediated almost entirely by individual motivational systems. In contrast, despite the evolutionary importance of reproduction, human mating and reproduction are largely controlled by symbolically mediated social negotiations. This offloading of one of the most fundamental biological functions onto social-symbolic mechanisms is perhaps the signature feature of being a symbolic species. Thus, because of symbols and with the aid of symbols, Homo sapiens has been self-domesticated and adapted to a niche unlike any other that ever has existed. We have been made in the image of the word.

References

Darwin C (1859) On the Origin of Species by Means of Natural Selection, or the Preservation of Favoured Races in the Struggle for Life (John Murray, London), 1st Ed.

Darwin C (1860) Letter 2743 Darwin, C. R. to Gray, Asa, 3 Apr 1860. Source: http://www.darwinproject.ac.uk/entry-2743.

Darwin C (1871) The Descent of Man and Selection in Relation to Sex (John Murray, London).

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

Deacon TW (2009) Relaxed selection and the role of epigenesis in the evolution of language. Oxford Handbook of Developmental Behavioral Neuroscience eds Blumberg MS, Freeman JH, Robinson SR (Oxford University Press; New York) pp 730-752.

Lyell C (1863) Geological Evidences of the Antiquity of Man (John Murray, London).

Okanoya K (2004) The Bengalese Finch: A window on the behavioral neurobiology of birdsong syntax. Annals NY Acad Sci 1016:724735.

Wallace AR (1869) Sir Charles Lyell on Geological Climates and the Origin of Species. Quarterly Review, April, cxxvi: 359-94.

30 comments to On the Human: Rethinking the natural selection of human language

  • Terry Deacon hits the nail on the head in urging us to reconsider the interaction between natural selection and human language. His emphasis on “long evolution” is apt. His focus on “relaxed selection” “in an artificial niche” is original, potentially seminal. He offers here a major extension of his proposals in The Symbolic Species—which I reviewed favorably when it was first published (see http://ssrn.com/author=1058129).

    My disagreement with Deacon is captured in his title: “The Natural Selection of Human Language.” There was no natural selection of human language per se, independent of other advanced abilities. Rather, the basic mental capacity for conceptual integration, otherwise known as blending, has been evolving since the mammalian line. Human beings evolved a more advanced ability for blending. This advance made it possible for several related behaviors to arise as a suite of coordinated and mutually-supporting products of blending. Human language was one of them. Gilles Fauconnier and I made this proposal in The Way We Think (Basic Books, 2002; see http://markturner.org/wwt.html).

    Evolutionary accounts often take their cue from a rough-and-ready list of advanced human behaviors we see all about us—advanced music, advanced social cognition, culture, religion, sign systems, gesture, language, art, science, advanced dance, fashion, advanced tool use, math, . . .—and then, considering them metaphorically as independent traits, seek an evolutionary story for each, regardless of the others. What most needs to be reconsidered is that impulse. To the extent that we are able to interrogate the archeological record, these behaviors seem to be recent and to have come up in concert. To the extent that we are able to study these behaviors in living human beings, they appear to develop in concert. Blending theory proposes that these behaviors cooperate because they are siblings, drawing on the same mental source.

    Consider surfing. It’s an amazing behavior, highly complicated, creative, demanding. Its Polynesian invention is, evolutionarily speaking, very recent, and its development and proliferation have occurred mostly in my lifetime. (Fins are a recent invention; the leash was invented when I was in college.) No one imagines that there has been natural selection for this behavior per se. Pushing the timeframe back a little, consider writing. Writing is at most 8000 years old. Writing as we know it is at most only a few thousand years old, and writing as a common activity is at most only a few hundred years old. No one imagines that there has been natural selection for this behavior per se. Push this back to fifty, sixty, seventy thousand years or so. I propose that instead of remarkably simultaneous independent evolution or coevolution of all these different advanced behaviors, there was long evolution of the enabling mental ability, which reached a crucial tipping point. From that tipping point, all these products arose and coevolved. The suite was highly adaptive, and the result was cognitively modern human beings.

  • Derek Bickerton

    Both Terry Deacon and Mark Taylor focus on how language evolved once it had appeared, but they have little to say about its initial emergence. (Mark’s remark that “Human beings evolved a more advanced ability for blending” leaves a host of questions unanswered–how did it evolve; why did it evolve in the way it did in humans; why, if it was so common, didn’t it evolve in the same way in other species; and on and on.) Relaxed selection is a good idea–I already pointed out in Adam’s Tongue that language could not emerge until communication had broken its link with fitness–but it can only apply AFTER a trait is present; after all, Bengal finches could sing even before domestication. What nobody in the field of language evolution (or in evolution generally) seems to realize is the extent to which the very notion of being able to use an arbitrary sign to symbolize entities or actions in the real world is totally alien to the way other animals think. No explanation of how language evolved can be viable unless it first faces this problem and then explains how it was overcome.

  • Deacon’s reasoning beautifully transcends the dichotomy of human versus natural sciences by accounting for both the multiplicity of human experience and its biological conditions. I am in complete sympathy with the argument as a whole, but would like to add a few suggestions on how it could be expanded. First, the distinction between artificial and biological niche may not be as significant as Deacon proposes, if we accept Uexküll’s conclusion that all species inhabit subjective worlds defined by their particular capacities to process signs. Not only do different species navigate environments defined by their specific perspectives (a circumstance recognized by many indigenous peoples and reported by anthropologists from Hallowell to Viveiros de Castro), but the same of course applies to human cultures. As Sahlins and others have shown, there is no unmediated, objective world accessible to humans. Even for our ancestors, any biological niche would have been ethno-biological, whether linguistically codified or not. Second, the proposal that the relaxation of natural and sexual selection has been conducive to linguistic and cultural diversity wonderfully harmonizes with ecological theories on the creative proliferation associated with biological colonization of new and less competitive habitats. The relaxation of constraints appears to encourage creativity in both contexts. Finally, having spent years trying to grasp the significance of kin terminologies, I think Deacon’s argument could more explicitly make the point that the operation of such terminologies, in defining potential spouses, is not only a reflection of cultural diversity unleashed by the relaxation of selective constraints but recursively also a source of such relaxation. In the co-evolutionary process that he has identified, language has relaxed selective pressures by delegating reproductive success to social conventions. If the relaxation of selection is an explanation of language, the opposite is equally true.

  • Derek Bickerton

    Mark, none of your chapters seem to be available on the web, so I can’t comment specifically on your proposal within the time-frame of this discussion. For now I’ll just say that my position is diametrically opposed to yours, since I believe that enhanced cognition comes from language, rather than vice versa. I take it that your position excludes any task-specific mental mechanisms for language, a position which my own work on creole languages renders implausible (although please note I do NOT endorse Chomsky’s version, or versions). However, I would really welcome a chance to pursue this discussion (I am not among those who automatically rule out any opinion contrary to their own), so if you have available electronic versions of any of the three articles you cite, could you send them to me at ?

  • Lorenzo Magnani

    I would like to add a few suggestions on how to extend this wonderful analysis of language as a cognitive niche given by Deacon. It would be interesting – taking advantage of this framework – to address the problem of the intertwining between language and cooperation, to increase knowledge about an important topic: violence. Here the example of fallacies. The hypotheses generated by the so-called fallacies in a dialectic interplay (but also when addressed to a non-interactive audience) are certainly conflict-makers but they do not have to be conceived absolutely as a priori “deal-breakers” and “dialogue-breakers”. I would contend that the potential deceptive and uncooperative aim of fallacies can be intertwined with pieces of both information and disinformation, logical valid and invalid inferences, other typical mistakes of reasoning like perceptual errors, faulty memories and misinterpreted or wrongly transmitted evidence, but fallacious argumentations still can be – at first sight paradoxically – “civil” ways/conventions for negotiating. That is, sending a so-called deceptive fallacy to a listener is much less violent than sending a bullet, even if it “can” enter – as violent linguistic behavior – a further
    causal chain of possibly “more” violent results. Also in the case of potentially deceptive/uncooperative fallacious argumentation addressed to a non-interactive audience, the listener is “in principle” in the condition to disagree with and reject what is proposed (as in the case of deceptive and fallacious advertising or political and religious propaganda). In summary, language is possibly “a tool exactly like a knife”, as René Thom already contended some decades ago.

  • Terrence Deacon

    On the Mark Turner / Derek Bickerton debate:

    I am in sympathy with Derek Bickerton in treating the capacity to use conventional symbolic reference as a more fundamental (and simpler) threshold distinguishing recent hominid evolution. I have been arguing this point since the early 1990s. Moreover, it should be obvious that our ability to utilize symbolic tools has greatly facilitated conceptual blending. But I also recognize that the notion of conceptual blending may also extend to domains beyond symbolic processes, though this fact doesn’t make this process more basic. But I worry that conceptual blend theory needs more unpacking. It is too abstract in its current cognitive science incarnation, and needs to be decomposed into component neural processing dynamics so that it isn’t treated like some monolithic “function” that is present or absent. In service of this goal I think we need to consider it along side its earlier close cousins: i.e. the related notions of Arthur Koestler’s “bisociation” – Gregory Bateson’s “double description” – Charles Pierce’s “abduction” and “extensive abstraction”. I actually think that there could be a useful collective unpacking of all five concepts to get at some of the underlying process assumptions, on the way to a unified cognitive neuroscience theory. But the either/or framing of this issue is counter-productive. These capacities co-evolved. As cognitive supports for each evolved and were socially developed they reciprocally amplified each other. We are good double-scope blenders because we can tap into vast symbolic supports and analogues, and we can develop remarkably rich symbolic systems because we can create complex conceptual synergies via conceptual blending. Again we need to approach this issue from a synergistic co-evolutionary perspective, which is a core assumption of the target piece. This view implies that both symbolic-linguistic and blending capacities evolved in tandem. At an early phase of this evolutionary ratchet I suspect that the two capacities are not differentiable from one another and so a chicken and egg debate may be pointless. More to the point, I think that both capacities require similar neurological augmentation, particularly involving prefrontal cortical functions (e.g. enhancement of transfer learning, combinatorial working memory, displacement of intrinsic salience/reward effects, stimulus equivalence bias). These are phylogenetically atypical cognitive capacities. The question I address here is how this suite of interdependent capacities relevant to language and symbolic cognition evolved to its present state. Relaxed selection effects are particularly relevant to this debate. De-differentiation is a prerequisite to both the interaction of the relevant neurological systems and the ability to juxtapose divergent cognitive processes as required by both symbolic reference construction and cognitive blending.

  • Derek Bickerton

    I am grateful to Mark Turner for kindly making available to me electronic versions of two of the items he cited in his second post on this thread. They made for fascinating reading, and I certainly wouldn’t dream of questioning the centrality of conceptual blending in all aspects of modern human cognition, including language. But the centrality of something now does not automatically entail its centrality then, and from my perspective, conceptual blending does not explain how language evolved.

    Before I start, however, a sincere word of apology to Terry. Much of what I shall say has only an oblique bearing on Terry’s case (although I am here in a sense arguing against him, since I do believe that natural selection, albeit in a bizarre and indirect way, did cause language to evolve). Be it noted however, that my thinking on this is and has been for the last few years heavily influenced by my (somewhat tardy) acceptance of Terry’s contention that symbolism, not syntax, is the major defining mark of the human species. And I do accept that, once language had begun to evolve, “long evolution in an artificial niche” is about as good a capsule description of what took place as you’re likely to get. The question I want to zero in on is, “Was natural selection responsible for starting language?”

    Mark’s case in many ways resembles Terry’s, but to my great surprise, it is also in many ways similar to Chomsky’s. I know this sounds counterintuitive, but note the following:

    1) Neither account makes substantive reference to pre-human evolution..

    2) Both accounts root human-prehuman discontinuity in thought rather than behavior.

    3) Both accounts implicitly assume that the cognitive/conceptual system of human predecessors was similar to that of humans.

    4) Both accounts assume that the internal revolution that made language possible was fully completed before the utterance of a single word.

    5) Both accounts assume that language emerged quite abruptly and substantially in the form in which we find it today; in other words, they assume that there was nothing intermediate between an alingual state and modern human language–no kind of “protolanguage” that did not already have all the major properties of modern human language.

    In fact, the only significant difference between the two models seems to be that the engines driving them were in the one case syntactic (the process “Merge”) and in the other semantic (conceptual blending).

    Reverting to (1), above, although Mark claimed in his first post to this thread that “the basic mental capacity for conceptual integration, otherwise known as blending, has been evolving since the mammalian line”, at the beginning of Chapter 9 of The Way We Think he states that “the considerable efforts of animal psychologists have uncovered no evidence that other species can reach very far in conceiving of counterfactual scenarios (like those underlying pretense), metaphors, analogies, or category extension.” These statements seem to me flatly contradictory, and if the second is correct, what caused the explosion of blending capacity in a single mammal?

    He asserts (without giving details) that “each step of the [blending] capacity would have been adaptive because each step gives existing cognitive ability to compress, remember, reason, categorize and analogize.” But he gives no evidence of any evolutionary results from this increased capacity. I agree that the fact that conceptual blending IS so complex and far-reaching suggests that it must have had some evolutionary history. But this fact is equally consistent with the view that conceptual blending began (and could only have begun) once the first overt symbolic units (call them words, proto-words, word-like objects or whatever you want) came into existence perhaps a couple of million years ago. And the latter view is indeed more consistent with the large cognitive gap that, as Mark admits, exists between humans and other animals.

  • Terry Deacon certainly opened a beautiful can of worms, elegantly and provocatively, but it is still a can of worms. For what he says to make sense, considerable filling-in is required. Since brevity is demanded by the venue, I apologize for the extreme compression of what follows.

    1. Complexity and relaxed selection: There is an unfortunate tho no-doubt unintended inference here that a highly complex system, with a lot of cross-modular interaction and co-evolution, necessarily transcends adaptive selection. Such an inference is dubious in biology, where its most ardent proponents are the ID crowd. It is equally dubious when it comes to language. Plenty of run-of-the-mill cellular, somatic, and neurological systems are highly complex but still unimpeachably adaptive, selected and genetically coded. By itself, systemic complexity does not translate into relaxed selection. Rather, it points out to more a complex, interactive, multi-factored selection. ‘Epiphenomena’ or ‘spandrels’ may be one consequence of complex selection, but they are hardly counter-evidence to it. Human language is a prime example of a highly complex multi-modular system that is nonetheless adaptive and selected. And the presumed arbitrariness and extravagance of its ‘spandrels’ is vastly over-stated.

    2. Innateness, variability and selection: In general, the older a system is in bio-evolution, the more rigidly-programmed (‘innate’) it tends to be in the genome. Older evolved systems exhibit more within-species uniformity, are more fully automated at birth, and are less sensitive to epigenesis, maturation and learning. In contrast, more recently evolved systems tend to be less rigidly programmed in the genome, exhibit more phenotypic and behavioral variability, and are more susceptible to maturation and learning effects. But this flexibility and variability of recently evolved systems is hardly evidence for relaxed selectional pressures. On the contrary, such variability and behavioral flexibility are the innovative vanguard of adaptive selection (Mayr 1976, 1982; Fernald & White 2000, West-Eberhard 2003). Through the interaction of niche construction and genetic assimilation, flexible behavioral experimentation eventually becomes assimilated into the genome (Baldwin 1898; Waddington 1942, 1953; West-Eberhard 2003). In principle, phenotypic and behavioral variability in the early stages of evolutionary change is a highly adaptive phenomenon. It allows for multiple adaptive solutions to compete head on, before selection pares the multiplicity down to the few that have proven themselves over multiple generations and contexts. These success stories are then coded more rigidly in the genome. The progression from old (‘reptile’) brain to mid-brain to cortex is a terrific example of this evolutionary trend, with decreased genetic specification, increased phenotypic and behavioral variability, and increased maturation-and-learning sensitivity.

    3. Cognition vs. communication: I must confess I find the Turner-Bickerton exchange here rather unenlightening. On the one hand, the neurological systems that underlie pre-human and pre-linguistic cognition–of objects-and-events, of somatic sense and motor control, of space-and-time, of action and agency–are ancient mammalian and pre-mammalian systems (Ungerleider and Miskin 1982; inter alia). Not only do they predate human communication by a wide margin, they also serve as the platform upon which the more elaborate cognitive representation system that underlies human language was erected. True, language has prompted a great elaboration, complexity, and increased abstraction of its precursor representational systems. But the core neurological sub-systems that underlie language processing arose much earlier and still perform many of their pre-linguistic functions (Amunts et al. 1999; Bookheimer 2002; Badre and Wagner 2007; Kaan 2009; Hagoort 2009; Fiederichi 2009; Friederici & Brauer 2009; Fernández-Duque 2009; inter alia).

    On the other hand, to suggest–as so-called ‘cognitive linguists’ are fond of doing– that somehow the communicative apparatus of human language and in particular grammar ‘falls out’ of the cognitive representational system is to ignore both what grammar does as an adaptive system and how uniquely it is structured; as well as to ignore the distinct, complex, multi-modular neurology of grammar (Friederici 2009, inter alia). Communication is not mental representation. Rather, it is the transfer of mental representations from one mind to another. It certainly presupposes mental representation, but does not ‘fall out’ of it.

    4. Diversity vs. universality: Human language, as a relatively recent and highly complex adaptation (from ca. 1.8 million to ca. 100,00 years ago) bears all the marks of a recently-evolved biological system: Flexibility, maturation-and-learning dependence, and considerable phenotypic and behavioral variability. It is too tempting, however, to over-state linguistic variability and ignore the great amount of universality of language functions, structures and controlling principles. Just as it is too tempting from the opposite perspective to exaggerate universality and pooh-pooh variability (Chomsky 1992; Hauser et al. 2002). From my middle-ground vantage point, I see plenty of universals of both phonological and grammatical structure and function. Like all good species-specific universals, they surely are genetically coded. Still, such universal allow considerably phenotypic and behavioral variation, and such variation is biologically natural functional–given the complexity and evolutionary recency of human language.

    5. The adaptive niche of human language: To understand the difference between pre-human and human communication, one must begins with two core features of non-human communication: (a) non-displaced reference; and (b) strictly manipulative speech-acts.

    The first feature indeed predicts the second: When communication is only about here-and-now, you-and-I, this-and-that visible in the shared speech-situation, there is no need for declarative speech-acts. All relevant information (epistemics) is fully shared. Only manipulation (deontics) needs to be communicated. Humans are the only species that, overwhelmingly, communicates about displaced reference by transacting declarative and interrogative speech-acts. The adaptive niche for human communication must have, therefore, entailed some vital information that is not available in the here-and-now but rather pertains to there-and-then, and is therefore not shared equally by speaker and hearer. Several possible niches qualify for this condition: (a) reporting on and recruitment for cooperative foraging of remote food-sources (big-game hunting, Washburn & Lancaster 1968; big-game scavenging, Bickerton 2009). (b) reporting on and planning for migration to a remote new territory; (c) reporting on and planning for hostile engagement elsewhere. All three ecological niches are unique to the hominid line. All three demand declarative/interrogative communication about remote referents. Is it an accident that the two adaptive niches in which honey bees developed their symbolic communication are: (a) reporting on and recruitment for foraging of a remote food-source; and (b) reporting on and recruitment for migration to a remote new hive-site (von Frisch 1967; Gould and Towne 1987)?

    6. Human language as a complex adaptation: Saying that human communication is ‘symbolic’ is saying relatively little. The devil is in the extremely complex detail. Human language has two distinct codes–phonology and grammar–that perform different functions and evolved at different times. Phonology–the sound code–is much more concrete, arbitrary (‘symbolic’) and easier to get a grip on. It functions primarily to code the lexicon of entities (nouns), events (verbs) and states (adjectives); that is, to code semantic memory. All the available evidence suggests that phonology evolved much earlier, giving rise first to pre-grammatical pidgin (Givón 1979, 2009) or ‘proto-language’ (Bickerton 1981, 2009). But while seemingly a lexical code, single-word symbols in actual communication do not code mere lexical information, but rather propositional speech-acts. Thus the vervet monkeys’ predator calls (Cheney and Seyfarth 1990) do not code the merely-lexical ‘eagle, ‘leopard’, ‘snake’, but rather the propositional speech-acts ‘hide from the eagle’, ‘watch out for the leopard’, ‘run from the snake’.

    Grammar evolved much later and gradually, and is much more complex and abstract. Its adaptive motivation is much harder to understand, due primarily to out-of-context contemplation of structures detached from their adaptive contexts (Chomsky 1965). Once grammar’s adaptive-communicative function is understood, it turns out to be much less arbitrary than the sound code. Grammar encodes multiple, highly specific communicative functions, most of them having to do with the speaker’s assessment of the hearer’s deontic (intention) and epistemic (belief) mental states during ongoing communication. It is a highly-automated instrument for representing other minds (Givón 2005), much more efficient and less error-prone than pre-grammatical pidgin. Not only is grammar as a whole adaptive, but specific grammatical construction perform specific communicative functions, in a way that is highly constrained and far from arbitrary (Haiman ed. 1985; Givón 1995, 2001, 2005). The familiar story of grammar as an extravagant ‘spandrel’ (e.g. Hauser et al. 2002) thus needs to be tempered with the understanding of grammar’s adaptive function.

    References:

    Amunts, K., A. Schleicer, U. Burgel, H. Mohlberg, H.B.M. Uyling and K. Zilles (1999) “Broca’s region re-visited: Cytoarchitecture and intersubject variability”, J. of Comparative Neurology, 421.1
    Badre, D. and A.D. Wagner (2007) “The left ventrolateral prefrontal cortex and the cognitive control of memory”, Neuropsychologia, 45
    Baldwin, J.M. (1896) “A new factor in evolution”, The American Naturalist, 30
    Bickerton, D. (1981) Roots of Language, Ann Arbor, MI: Karoma
    Bickerton, D. (2009) Adam’s Tongue, NY: Hill and Wang
    Bookheimer, S. (2002) “”Functional MRI of language: New approaches to understanding the cortical organization of semantic processing”, Annual Review of Neuroscience, 25
    Cheney, D. and R. Seyfarth (1990) How Monkeys See the World, Chicago: University of Chicago Press
    Chomsky, N. (1965) Aspects of the Theory of Syntax, Cambridge, MA: MIT Press
    Chomsky, N. (1992) “A minimalist program for linguistic theory”, MIT Occasional Papers in Linguistics, , Cambridge, MA: MIT Press
    Fernald, R.D. and S.A. White (2000) “Social control of brains: From behavior to genes” in M. Gazzaniga (ed. ) The New Cognitive Science, 2nd edition, Cambridge, MA: MIT Press
    Fernández-Duque, D. (2009) “Cognitive and neural underpinnings of syntactic complexity”, in T. Givón and M. Shibatani (eds) Syntactic Complexity, TSL #85, Amsterdam: J. Benjamins
    Friederici, A. (2009) “Brain circuits of syntax”, in D. Bickerton and E. Szathmáry (eds) Biological Foundations and Origin of Syntax, Cambridge, MA: MIT Press
    Friederici, A. and J. Brauer (2009) “Syntactic complexity in the brain”, in T. Givón and M. Shibatani (eds) Syntactic Complexity, TSL #85, Amsterdam: J. Benjamins
    Givón, T. (1979) On Understanding Grammar, NY: Academic Press
    Givón, T. (1995) Functionalism and Grammar, Amsterdam: J. Benjamins
    Givón, T. (2001) Syntax: An Introduction (2 vols), Amsterdam: J. Benjamins
    Givón, T. (2005) Context as Other Minds, Amsterdam: J. Benjamins
    Givón, T. (2009) The Genesis of Syntactic Complexity, Amsterdam: J. Benjamins
    Gould, J.L. and W. Towne (1987) “Evolution of the dance language”, The American Naturalist, 130.3
    Hagoort, P. (2009) “Reflections on the neurobiology of syntax”, in D. Bickerton and E. Szathmáry (eds) Biological Foundations and Origin of Syntax, Cambridge, MA: MIT Press
    Haiman, J. (ed. 1985) Iconicity in Syntax, TSL #6, Amsterdam: J. Benjamins
    Hauser, M., N. Chomsky and W.T. Fitch (2002) “The faculty of language: What it is, who has it, how did it evolve?” Science, 298
    Kaan, E. (2009) “Fundamental syntactic phenomena and their putative relation to the brain”, in D. Bickerton and E. Szathmáry (eds) Biological Foundations and Origin of Syntax, Cambridge, MA: MIT Press
    Mayr, E. (1976) Evolution and the Diversity of Life, Cambridge, MA: Harvard University Press
    Mayr, E. (1982) The Growth of Biological Thought, Cambridge, MA: Harvard University Press
    Ungerleider, L.A. and M. Mishkin (1982) “Two cortical visual systems”, in D.G. Inge, M.A, Goodale and R.J.Q,. Mansfield (eds) Analysis of Visual Behavior, Cambridge, MA: MIT Press
    von Frisch, K. (1967) The Dance Language and Orientation of Bees, Cambridge, MA: Harvard University Press
    Waddington, C.H. (1942) “Canalization of development and the inheritance of acquired characters”, Nature, 150
    Waddington, C.H. (1953) “Genetic assimilation of an acquire character”, Evolution, 7
    Washburn, S.L. and C. Lancaster (1968) “The Evolution of Hunting”, in R.B. Lee and I. DeVore, Man the Hunter, Chicago: Aldine
    West-Eberhard, M.J. (2003) Developmental Plasticity and Evolution, Oxford: Oxford University Press

  • Derek Bickerton

    Terry, when I hear words like “bisociation”, “double description”, or “abduction”, I reach for my Occam’s razor. Aside from that, I think we’re (mostly) on the same page. “Both symbolic-linguistic and blending capacities evolved in tandem.” Right on! After that our difference is merely one of focus. You are, as you say, concerned with their subsequent evolution, while I am concerned with how they got started in the first place. Which was hen and which was egg, or was it a third party? Something had to get the whole thing started, but people don’t seem to see that this is the real problem–just about anything can evolve further once it’s gotten started–so they either ignore it or give the kind of answer that shows they haven’t really thought it through.

  • Terrence Deacon

    At this point, there is far too much in Tom Givon’s response for me to even contemplate responding to, except to say to all involved that you shouldn’t assume that I have said everything that I wanted to say about this issue in the 2500 word essay that I have contributed. In order to stay within the size limitations of this venue I only opened a narrow window on the relationship between natural selection and language, and even at that I could only hint at the details. This inevitably has led Tom to conclude that just because I have emphasized one previously unappreciated aspect of the problem (self-organization effects intrinsic to epigenetic processes and the way that they are exposed via niche-construction) that I am denying the importance of the other (natural selection on biological substrates to generate functional correspondences). Far from it.

    For the most part we are in agreement on many points, and perhaps in a later post I can address a few of the ways we diverge. Here I only comment on the general points made about natural selection. There is a much larger biological background behind my approach which of necessity has had to go unmentioned. It traces to my work on brain development and evolution, and more broadly it borrows from work that currently runs under the banner of “evodevo” and which has begun to illuminate once problematic issues in evolutionary genetics, molecular cellular biology, and epigenesis. My point is not to discount the contributions of natural selection, which I agree is the final arbiter of functional adaptation, but to bring attention to another unnoticed facet of the evolutionary process. Natural selection is explicitly NOT the generator of the biological phenomena that it prunes in the process that leads to increased adaptation. Not only are the variants of existing organismic subsystems generated irrespective of function (e.g. by genetic “damage”) but the expression of these varieties of structure and dynamics depends on generative processes whose details we tend to hide in generic concepts like epigenesis and reproduction. New stuff, new structures, and new processes need to be generated so that there is raw material fed to the engine of natural selection. The second law of thermodynamics has to be locally tamed in order for this to be possible. And natural selection theory is so widely applicable precisely because it can be agnostic as to how any of this is achieved, so long as it is. Surprisingly, despite our many disagreements about innateness, I find some resonance in Noam Chomsky’s periodic suggestion that some of the complexity of grammar may have emerged from general laws of physics analogous to the way that the Fibonacci regularities exemplified in the spirals of sunflower seed and pine cone facets emerge. Natural selection has “found a way” to stabilize the conditions that support the generation of this marvelous regularity of growth because it has important functional advantages. But natural selection didn’t generate it in the first place, geometric regularities that can become amplified due to center-out growth process are the ultimate source (as has now been demonstrated also in growth-like inorganic processes). I also agree that flexibility CAN be an adaptive response to a variable and demanding habitat, but not necessarily. And I hope I have shown that there is another mechanism potentially available to explain some of the complexity (both neurologically and functionally) and some of the flexibility, besides natural selection and innate algorithms. In this essay, I am only interested in pointing out the importance of processes that have tended to be overlooked by those assuming that only natural selection in its most basic sense is at work, and that any universals must be assumed to arise from genetic origin. Just appealing to genetic accident is little better than invoking a miracle.

    Indeed, this is a can of worms, and like Pandora’s box, even just opening it a crack runs the risk of unleashing explosion of linked conceptual problems. Derek (and Mark), don’t be too sure that you can ignore the various theories related to conceptual blending. I do think that buried in this mess is an important insight into the crucial transition that we both find so interesting. I am happy to welcome Occam into the game but I am wary of trying to simplify too soon, and as a biologist I am constantly confronted with the fact that mother nature doesn’t respect this razor. This goes for restricting our explanations to simplifying assumptions about natural selection, mechanisms contributing to biological complexity, theories attempting to explain the source of language universals, and how we account for the neurological teaks that provide us with such a distinctive capacity.

    In response to my remarks suggesting that I am only interested in the “subsequent evolution” of language capacities, I meant this only in the context of the present essay. Indeed, Derek and I agree that explaining the transition from non-symbolic to symbolic communication is ground zero.

  • Derek Bickerton

    Your last paragraph is well said, Terry, but when it comes to ground zero, what seems to you the role of natural selection there? After all, any reader faced with your bare title might have expected to hear something about that.

  • Terrence Deacon

    Ground zero:
    I still stand by the general scenario as sketched in my book The Symbolic Species, in which I argue that the shift in foraging ecology precipitated by the development of the first stone tool technology at 2.4 million years ago was the first spark. This was a first niche construction activity, that embedded these early hominids in a radically different ecology and demanded a fundamental restructuring of social-sexual relationships that only symbolic communication (because of its displacement of reference) could stabilize. Indeed, I don’t believe that stone tool supported foraging could have been maintained across generations without it. The meager capacity of great apes to acquire a very small repertoire of symbolic signs (with highly structured support) suggests to me that the cognitive flexibility of these australopithecine (A. garhi) tool users was just barely sufficient to achieve this level of symbolization (probably mostly in the form of ritualized performances). But this created a complex synergy between tool technology and communicative “technology” that could only be sustained by social transmission. This semiotic niche is the hominid equivalent to a beaver dam. It both relaxed selection on many traits no longer critical, and shifted selection onto others, specifically those that would support the stability of this virtual niche. Cognitive flexibility – niche construction – shifted selection and relaxed selection – shifted cognitive predispositions and increased cognitive flexibility … My answer to your question, then, is that neither natural selection nor some lucky genetic accident precipitated this transition. It was cognitive flexibility and the social capacity to acquire novel behaviors by observation learning.

  • Derek Bickerton

    Terry, I agree with everything you say, but you still aren’t getting to grips with the real crux of the matter, which is, how and why were the first symbolic signs created? I’ve said it before, I’ll say it again, if it comes to that, YOU’VE said it–the idea of using a sound or gesture to symbolize a concept of some real-world entity is something totally alien to non-human minds. Some kind of ritual doesn’t begin to explain things, because nothing in a ritual necessitates having a concept of anything. For instance a chimp’s rain dance could develop into some kind of ritual, getting more complicated over time, without it symbolizing anything–it’s just what they do when raindrops keep falling on their heads. For all the many virtues of The Symbolic Species, it did not, unless memory fools me, contain any explicit account of how the first words, manual signs or whatever came to be.

  • Full twenty years after the publication of the Pinker-Bloom BBS paper that many see as foundational for the entire field of evolution of language, it is very interesting to see a text by Prof. Deacon on the same general topic – the role of natural selection – but reaching a novel and provocative conclusion. When Pinker and Bloom argued for a specific role of natural selection in designing subsystems of the human language faculty, Deacon points to the possible role of the relaxation of selection pressures in (phylogenetic) language development. However, the text starts with alluding to many important points regarding the evolution of language – let me start from enumerating these points (of, I hope, consensus) because they are well worth re-emphasizing:

    1. Language itself must have been a selection pressure – humans as a species (thus, their cognition) show clear signs of having been selected for efficient language use.
    2. The ‘language-brain coevolution’, as suggested by the name, is CO-evolution, so it works in both directions: language evolves to adapt to its environment, i.e. the human brain/mind. This idea has been especially prominent in the works of M. Christiansen.
    3. Before there was ‘full language’ there must have been a simpler communication system, but already distinguished by the use of arbitrary and conventional signs, –> symbols. This of course is a development on D. Bickerton’s seminal idea of protolanguage.
    4. Language (or protolanguage) couldn’t have developed from nothing – as it were, ‘bare ape cognition’ – but instead must have been built upon a rich collection of cognitive preadaptations for language, such as Theory of Mind (further subdivisible into joint attention, empathy, etc.), mimesis, and many others. Prof. Bickerton comments on this in discussion.
    5. Language, or at least protolanguage, is something phylogenetically old. ’50kY [50,000 year] theories of language origin’ are deeply problematic. It is symptomatic that the idea of a very recent emergence of language seems to be popular with very many researchers with a passing interest in this field, but only very few researchers with more that a passing knowledge (e.g. seasoned Evolang delegates). For many years Deacon has been consistent in his criticisms of the ’50kY theories’.

    I can offer little in the way of commenting on the novel argument concerning the possible effect of the relaxation of selection pressures, but I would like to venture two questions.

    1. One important (indeed, crucial, but often neglected) feature is missing from the explanation and that is the transfer of honest information (already alluded to by Prof. Givón). What language accomplishes is essentially *altruistic* transfer of honest information to nonrelatives – how could an altruistic form of behaviour have evolved certainly is non-obvious and requires an explanation.
    Could your account perhaps be made compatible with some particular explanation of the emergence of altruism/cooperation as manifested in linguistic communication?

    2. As it stands, ‘relaxation’ is a rather general concept: the removal or subsiding of what *specific* selection pressures should be subsumed under this term in the present context?

    Finally, this is probably a good opportunity for thanking Prof. Deacon for his presentation at the Protolang conference in Torun last year – a big thank you once again!

    Slawomir Wacewicz

  • Terrence Deacon has written a thought-provoking article on the co-evolution of language and the brain, from the perspective of niche construction. He submits an evolutionary account that conjures up an adaptive/exaptive perspective that in some ways conflicts with the language that he uses to explain it, for instance when he says that “we are creatures whose social and mental capacities have been quite literally shaped by the special demands of communicating with symbols” and “modern brains accommodate language,” as well as by his characterization of the social niche in which language evolved as “artificial.” It all sounds as if “symbolic and cultural communication” had origins separate from and/or independent of the brain and hominin social interactions, worst of all an existence that is anterior to the state of the hominin brain that could use it. Therefore the brain would have been adapting to language rather than language reflecting what the brain is capable of. Selection would have operated through the survival, at different stages of the phylogeny of mankind, of those hominins whose brains could “accommodate” the complex structures and functions of then-emergent languages all the way to modern ones.

    The scenario as presented sounds contrary to what one would expect to have happened. A comparison with other animal species suggests that hominins had always been able to communicate, albeit in less complex ways at earlier stages of our phylogeny; evolution has led us to gradually develop more and more complex “systems” for communication (based on predominantly symbolic devices built upon more elemental phonetic/signed units and combinatorial conventions). Part of the complexity itself must have emerged from the interactions of modules that produced various strategies at different times. It is thus also plausible to conceive of languages as emergent phenomena (see below).

    I think it is relevant to consider the fact that parrots can imitate speech, though the anatomical structure they use is different from that of humans. Thus the particular shape of humans’ bucco-pharyngeal structure is secondary compared to the role of our minds. The reason why parrots cannot produce novel utterances that they have never heard before (which characterizes linguistic competence), thus cannot use speech creatively, is that they are not endowed with the kinds of minds that can behave linguistically in the human style. This appears to be the same reason why they have not capitalized on their potential to produce phonetic sounds to develop a human-like communicative system for themselves. It is likewise significant that great apes have not developed signed languages of their own (though they have forelimbs similar to humans’) nor any complex social cultures comparable to humans’, among other things that distinguish them from mankind. The 2% or so of genetic materials they do not share with humans include the absence of the kind of mental capacity that modern humans are endowed with.

    It seems to me that human languages (spoken and signed) reflect the complexity of the minds that produced and use them, and in fact have continued to reshape them to date, although perhaps in respects evolutionarily less significant than those associated with the protracted emergence of language. An account that sounds plausible would involve hominins’ adaptive quest for more and more suitable means of communication as their minds evolved toward increasingly more complex structures/activities and they produced more and more complex social life styles and organizations that called for more and more sophisticated modes of communication. Such a scenario would account for the emergence of “symbolic communication and culture.” Even if these evolved rather later in the phylogeny of mankind, they would still typify humans as a unique “symbolic species” unlike any other in the animal kingdom.

    I think it is important to underscore the fact that communication with any means, apparently in a form more holistic than combinatorial, occurred much earlier. A challenge we face is to explain the transition from holistic vocalizations and gestures to compositional speech and signed language, respectively, with hierarchically-structured discrete units. (Linguists should think here of “double articulation,” though reality suggests more levels!) None of this evolution would have been possible without the emergence, after the separation of the Homo line from the other primates, of mental structures capable of domesticating the human anatomy to meet new and more complex communicative needs. Language may have enhanced human cognition, in fact it facilitates its ontogenetic development (thanks to its world-creating capacity), but comparisons with other animals suggest that the hominin ability to think and solve problems differently from other primates seems to have preceded the emergence of “symbolic language” itself.

    The question of how diversity emerged would be more puzzling if the cradle of mankind had lain in one single camp or village in East Africa and language had been designed. Linguistic systems have been so resistant to reductions to a few regularities largely because they consist of phenomena that emerged at different times in response to specific communicative needs, with the innovations being only partly constrained by current strategies. In a nutshell, we must bear in mind that any kind of cultural evolution has involved innovators and copiers, with these roles varying with every feature. If, as I assume, there were different hominin colonies not bound to meet their communicative needs in identical ways, and if innovators in the same colonies could produce alternative responses to the same or similar challenges, then the question should be not why there is variation (both language-internal and cross-linguistic) but why there is not more typological variation than would be possible. The answers need not lie in the monogenesis of language. They probably lie in the kinds of species-specific evolutionary factors (physiological and mental) that constrained the languages (at any evolutionary stage) that different populations produced. Note that “Language” is the invention of the scholar, a useful generalization indeed, but not a phylogenetic singular suggested by the geographical distribution of the available paleontological evidence. These considerations lead us to the question of what the “language organ” or “Universal Grammar” proposed in theoretical linguistics means and what it really consists of, if such a device need be posited.

  • Leigh Van Valen

    Yes, interesting, and I look forward to a more fleshed-out version. There is also the possibility of selection specifically for a relaxation of constraints, not mit einem Schlag but a progressive loosening of a relatively more closed behavioral system in a prelinguistic or protolinguistic context of more flexible overall behavior. Ostensive identification or protodefinition would seem to require such capacity.

    Such a process, like relaxation of selection (from what initial state?), gives complexity, and perhaps a less disordered complexity. Constraints would be relaxed in a way that enhances function, an outcome that is less expected under the random dissociations that would be produced by drift under relaxed selection.

    Results of active selection along the open-closed continuum are well known in the context of learning and fixed action patterns. The active selection I suggest here would operate on the same connections that Deacon envisages.

    An aside: Language evolution, in recorded cases, often proceeds partly by structural simplification, making the language easier to learn and the overall trend thus advantageous. There is then the obvious question of how the earlier complexity arose. This complexity doesn’t follow from the relaxation of constraints, which merely permits it. Many languages with few or no speakers today are remarkably complex, in different ways. Might the size of isolated small groups somehow promote complexity? There may be a correlation here, but an actual causal process seems mysterious.

  • Derek Bickerton

    Slawomir Wacewicz has asked an excellent question, one which every language evolutionist should be prepared to answer: why should anyone transfer useful information to non-kin (and why should non-kin believe them–the “cheap tokens” problem!)? It seems to me that for these objections to be overcome, at least the following four conditions should hold:
    1) Speaker gains an immediate benefit from giving the information (not just some possible future reciprocity).
    2) Speaker can only obtain this benefit if the information is given.
    3) Speaker can derive no benefit from giving the information if it is false.
    4) Hearer can quickly verify whether the information is false or not.
    I know of only one situation in which all these conditions can be met. Around 2 million years ago, human ancestors began to engage in scavenging megafauna carcasses. In order to exploit such carcasses and cope with a fearsome array of competitors, those ancestors were obliged to gather large (> 50) groups. To do so they would have had to inform conspecifics of the nature and whereabouts of those carcasses–without adequate numbers, the operation could not be carried out and no-one would benefit. Note that with this scenario, you solve at one blow the evolution not just of language but of co-operation–the only other major dimension on which we differ from our primate kin. For a detailed account, see chapters 4-8 of Adam’s Tongue.

  • Derek Bickerton

    Leigh van Valen’s “aside”: Lee, I’m afraid you’re confusing language evolution with linguistic change. Since modern homo, languages have been cycling around within the relatively narrow search space provided by the mature language faculty. If a language is used by a small homogeneous community for a long time undisturbed, it will accrete complexities like a ship accretes barnacles. If a language undergoes severe social disturbances or gets to be used by large numbers of people, it will shed those complexities. I think that’s one of the few things few linguists would quarrel about.

  • Sara Waller

    Thanks so much for this enlightening discussion, and thanks to Mark Turner especially for letting me know about it. I think I’ve missed the deadline for a response — any chance for an extension of continuation? It would be fun to talk more about animal communication, language, and human thought.

  • Sue Savage-Rumbaugh

    The exchanges in this dialogue are fueled by a desire to understand how language could have evolved from something that appears to be ‘not language,’ and by a firm belief that language sets humanity apart from other living beings on the planet. Most of the participants agree that only human beings possess language, and most agree with Terry that symboling (not grammar) is the key event in hominid evolution – the ‘Ground Zero.’ Some continue to view the gestures and vocalizations of other animals as symbolic – but ‘holistic’ in nature. Some continue to assert that combining symbols is a truly unique human innovation and that Terry’s emphasis on symbols is misguided. I would like to set that debate aside for now, due to lack of space. For those who wish to know, my view is that true symboling and grammar are inextricably intertwined. It is fundamentally impossible to have one without the other and previous attempts to view them as separable has misled humanity in its search for a solution to the origins of language.

    More important for the current debate however, is the fact that this debate began before science achieved the ability to rapidly sequence the entire genome and to follow genes across generations. Such new data is challenging the ground rules of evolutionary theory at its core. We no longer need infer that genes have changed, because anatomy and behavior have changed. We can look and see if genes have changed and what genes have changed. We now know the following things.

    a) The biggest genetic differences between apes and human beings lie in genes that code for immunity not those that code for ‘intelligence’ or language in any simple sense, or anatomy. Apes and humans are true sibling species.

    b) The onset of FLUENT speech could indeed have been a sudden genetic event. This is different from the onset of speech itself, or the linkage of speech to a wide array of complex cognitive processes.

    c) Epigenetic markers are more likely than genes themselves to account for the behavioral changes we associate with language, thus making the ape to human transition more of a cultural transformation than a biological transformation.

    d) Genetic activation profiles during prenatal and postnatal development, as a function of culture, can manifest differently in different environments, with major life changing consequences.
    The basic question of whether evolutionary theory can account for language must now be understood on the basis of this new knowledge.

    Ape rearing requires clinging, human rearing does not. Thus the ape and human developmental trajectories that are built from birth forward, produce different kinds of beings and very different epigenetic profiles. This one difference, clinging, affects the tendency toward bipedalism, toward free vocalization, and toward object use and manufacture.

    When bonobo infants are reared in an environment which decreases or eliminates the constant need to cling to the mother, their hands and eyes become free from birth to engage in a variety of object manipulatory activities. This allows the brain to construct eye/hand/mouth/object manipulatory systems that are co-ordinated in ways that follow the pattern of human infants. The need to cling inhibits this and channels the neurological development of eye/hand coordination in a ‘non-relaxed’ developmental manner. This phenomenon is analogous to Deacon’s ‘relaxation’ hypothesis, but it is the culture which is the relaxing constraint, not the genes. Not only do patterns of eye/hand coordination change in striking ways with human rearing, the vocal system becomes relaxed as well. This is not because it becomes free of any genetic constraint. It is freed because of a decrease in the risk of predation. Humans infants living in stable dwellings are far less likely to be attacked than bonobo infants. When an bonobo infant, in an arboreal a world, gets very far away, makes very much noise or gets interested in playing with objects rather than watching its mother and rushing to her before she moves, its chances of survival decrease markedly. If a bonobo mother puts a very young baby down, it is immediately subject to predation. Human reared bonobo infants face no such pressure and are free, because of human culture, to follow the developmental trajectory of the human infant.

    As they do so the plasticity of their nervous systems allows them to process and comprehend human speech, both symbols and syntax. Thus it becomes clear that the ‘ground zero’ question does not apply to whether apes have language ability – they do. The question is what is the cultural trajectory required to facilitate a human language. The answer is that the required cultural trajectory is obviously a human one. Were a human child to be reared along an ape cultural trajectory, even though it had a human brain, it would not manifest a human language, even though it possessed a human brain and would, like free-ranging apes, manifest the ability to communicate symbolic meaning with intentionality.

    While some linguists continue to quibble about whether what Kanzi has demonstrated is really exactly human language, this debate is really no longer relevant. Kanzi clearly has acquired sufficient language ability to indicate that the human rearing trajectory is the key ingredient in the emergence of human language. In this vein, it is essential to note that Kanzi did not have a uniquely human rearing – he was reared in a Pan/Homo culture. Had Kanzi been reared with a uniquely human trajectory, his human language capacity would perhaps have become sufficiently well developed that not only would he be producing vocal approximations of human words and sentences, but quite clear ones. (For those readers interested in seeing what Kanzi and family have accomplished, go to Kanzi.bvu.edu)

    Kanzi is a ‘first generation’ attempt to determine what happens when scientific investigation pushes the cultural envelope toward the human trajectory. Because Kanzi was co-reared by bonobos and humans, his cultural envelope was also constantly pulled toward the bonobo norm by Matata. It is not possible to know how far Kanzi could have gone had this not been the case. The participants of this debate seem to assume that the data is in with regard to what apes can do. Such is not the case. Only a few apes have been reared on a truly human trajectory from birth (Gua and Vickie), and none of them stayed on an exclusively human trajectory, with an extended human family, into adulthood. Moreover, neither was provided an alternative symbol system to overcome the differences between ape phonemes and human phonemes. The so called ‘language failures’ of Nim and Washoe, represent attempts (unlike Gua and Vickie) to focus on raising apes as ‘subjects’. These projects did not plan to extend the true ‘expectancies of humanness’ to these apes. The apes were experimental entities, not family members. One would thus predicate that these apes would fail, because of a lack of sufficient cultural inclusion to allow the cultural bias to imprint itself upon their developing neurological system.

    The fact that Kanzi and family have acquired significant components of human language does not eliminate the ‘ground zero’ issue raised by Bickerton, it merely pushes the question further down the phylogenetic scale. Apes are sibling species with humanity. There are no real fossil apes between 5 million years ago and today. Chimpanzees have 48 chromosomes, humans 46. The human chromosome number two is formed by combining two of theirs — or alternatively, the human chromosome number two broke, resulting in their 48 chromosomes. This could have happened at anytime. It could even happen today. In vast areas of our genome, for example those genes dedicated to bodily odor, the difference between various human beings exceeds those between apes and humans. Where we really differ genetically is where it matters most. We differ in immunity. Because we live in radically different environments, we require different immunological profiles. Unless, of course, apes were to be reared in human cultures. We should then expect to begin to see changes in immunity profiles.

    But monkeys are not siblings to our species. Monkeys are not self-aware, nor are they self-reflective in the sense that apes are. There has been no indication that monkeys would acquire language if reared in a human world, though many have been so reared. Without a concept of the self as a knowing, free-willed causal agent, language makes no sense. Monkeys can solve symbolic tasks and monkeys can utilize sounds to inform others of predators, but they do not manifest self-cognizant, self-reflexive intentional actions. Apes and humans do.

    This self-reflexiveness, which becomes manifest in the grammar of our language, enabled humanity to become the symbolizing species to which Deacon refers. Apes have this potential as well, when reared on a human developmental trajectory. Monkeys appear to remain at ground zero. This behavioral discontinuity between living species of monkeys and ourselves aligns with the genetic data. Trying to uphold a similar distinction between humans and apes is contrary to all genetic evidence to date.

    It is probable that language of a different form than human language will be found in free-ranging apes. It is probable that monkeys will be found to have a far larger array of sounds than we now know and that many of them will have multiple meanings depending upon how they are embedded in the context. They may even combine sounds according to some rules. What does not seem likely is that monkeys will begin to talk and think in a self-reflexive manner.

    We now know that various familial lines seem to be biologically distinct at very basic levels. Recent work has revealed that during embryogenesis, epigenetic markers are stripped off at various points before implantation. These timing points differ across the basic mammalian lines. Therefore all mammalian lines are not transmitting epigenetic markers in the same way, as would be predicted by evolutionary theory. The new knowledge of embryogenetic transmission of epigenetic markers is already requiring many to rethink their basic views of how evolution operates. With the coming advances in these fields we will gain a clearer understanding of what has truly happened, at the genetic and epigenetic level, between humans and apes. As we do, the current era speculation and debate will fade under the weight of the genetic facts.

  • I am convinced Terrence Deacon raised a cluster of important questions in a very fertile way in The Symbolic Species. The need for explaining the vast use of human symbols, the Peircean framework taking symbols to be composite signs involving iconic and indexical aspects – and the framing of these issues in a co-evolutionary process of semiotic and brain evolution. The idea of relaxation of evolutionary constraints add a fascinating idea to this scenario.

    Derek Bickerton now pinpoints the issue by saying that “… the very notion of being able to use an arbitrary sign to symbolize entities or actions in the real world is totally alien to the way other animals think.” It is definitely true that the vast human system of symbolic communication is alien to animal semiotics. But that is not the same as claiming there is no semiotic arbitrarity (or convention, or habit) in the animal kingdom at all. Indeed, the admittance of germ-like forms of arbitrarity in animals makes the issue of the “semiotic missing link” between animal and man less steep and forbidding. Bickerton points to animal thinking rather than communication. But in the way higher animal think, a basic arbitrarity resides in their ability to learn. Take good old Pavlovian conditioning – this amounts to learning the arbitrary association between a bell sound and feeding. After such conditioning, the bell now functions as an arbitrary sign for food. In the Peircean terminology, such a sign is definitely a symbol. This is not, of course, to say those dogs by themselves use such symbols to communicate. But it shows that higher animals, capable of ontogenetic learing, has the ability of forming stable, arbitrary symbols by associating events in their surroundings (see the discussion in my Diagrammatology, 2007).

    Even further down in the animal kingdom, we find symbols on the species level – even on the interspecies level, such as the yellow-black stripe pattern as a sign for danger in and among insects and their predators. This symbol is, of course, prompted by the existence of insects displaying that pattern and really being poisonous, such as wasps – but during evolution, other, non-poisonous species have adapted by acquiring a similar pattern, protecting them, to some degree, against predators. This is not, of course, a symbol developed nor learned during ontogenetic life like the conditioned symbol – it is rather a phylogenetically evolved symbol emerging during the slow process of combined natural selection of several species in certain ecosystems. But still, the symbol is effective in the life and interaction of individual organisms. And still the symbol is arbitrary – there is no intrinsic or iconic connection between the yellow-black pattern and danger.

    This points to the issue that the categorical way of phrasing the issue presupposing arbitrarity is an all-or-none issue is misleading. Rather, arbitrarity and symbolicity come in degrees; many signs are arbitrary in some respects and motivated in others, symbols and icons at the same time. An important aspect of semiotic-cognitive co-evolution is the appearance of ever more arbitrary and interconnected symbols – and finally their appearance also in ontogenetically acquired communication between individuals. The famous vervet monkey warning calls seem to constitute an example of such a simple symbolic communication system.

    Thus, on such an account, there is no clear non-symbolic/symbolic borderline between animals and human beings. But this does not in any way invalidate Deacon’s approach – rather, it places the dramatic increase of symbol use in human beings on a basis on which it seems easier to understand how it might emerge in the first place.

    Professor Givón adds an important observation to such a scenario: the notion that the vervet monkey calls should not be seen as simple signs, but rather as full-blown propositional speech acts. My guess would be that all efficient animal signs are such holist propositions. This makes it more easily understandable how human propositions may evolve – namely by the internal differentiation of propositional structure, in turn making human propositional stance much more flexible because propositions may now be constructed from different simpler signs – which was not earlier beforehand.

    This finally ties in neatly with Deacon’s intuition that grammar – or, at least, central parts of grammar – may have an a priori status. Propositions, even if holist and not-yet differentiated, must at the very least do two things at once. They must 1) indicate the presence of an object, and 2) say something about it. Like “There is a leopard, run!”. This, of course, corresponds to the basic S-P or noun-verb structure of grammar which seems to be present as aspects of the symbol already in biosemiotic propositions – even if of course not present in the sense of explicit grammatical parts of the proposition.

    Such a scenario makes animal semiotics and cognition much simpler than human language and cognition – but not because animals use simpler signs which humans later compose into more complicated patterns. Rather because animals use signs which are later subdivided, differentiated, and made plastic and controllable by human beings. Such a scenario thereby has the force that the animal-human transition becomes not less dramatic – but more continuous.

  • Charles Wolverton

    First, a meta-comment: thanks to all the participants in this wonderful forum, especially, of course, Prof Deacon. Having little or no background in the main subject matter, with respect to it I can truthfully say “almost everything I know I learned from” … you folks.

    Next, an observation. Having been influenced in my youth by Marshall McLuhan’s observation that “the content of a medium is another medium” and later by working with layered network protocols, I instinctively view “communication” as involving multiple layers and try to be careful in tying any discussion to the relevant layer(s). In the present context, I see the bottom layer as being roughly the ability to make and detect audible sounds, the top layer as being natural language as we currently know it, and the intermediate layers as being some assortment of the results of evolutionary steps between the bottom and top layers. The point here isn’t to define the layers, only to suggest that focusing on specific layers might help in minimizing some misunderstandings that I think arose in the discussion. For example, I see Prof Deacon’s essay as mainly focused on higher layers, while some of the comments were more apropos lower layers. Explicitly identifying the layers under consideration might avoid some confusion.

    Although he didn’t make it explicit, I think Prof Givón’s comment (which was immensely helpful in getting at least a tenuous grip on the exchanges) essentially addressed this idea, especially the critical point that especially in considering evolutionary history, the timing and relevant considerations may vary considerably depending on the layer being addressed.

    Finally, somewhat of an aside. In Susan Blackmore’s interview with Vilayanur Ramachandran (in her book Conversations on Consciousness), he alludes (unfortunately, there are no details) to the more-or-less simultaneous emergence (in his view) of language, consciousness, and the sense of self – which seems reminiscent of Prof Turner’s conceptual blending. I see this as possibly relevant to the present discussion since I am leaning toward a view of consciousness as an emergent epiphenomenon attendant to the brain’s facility for creating an internal virtual “picture” of the neurological correlates to sensory inputs from the organism’s environment, a “picture” that is “viewable” by the virtual “self” that represents the organism in that environment. (The Cartesian Theater denied by Dennett, only here explicitly assumed not to actually exist, so no latent dualism emerges.) The ability to create this virtual “picture” would include the ability to associate the virtually “visible” symbols with objects in the environment, a step toward the required symbol-object association required for language, as noted by several commenters.

    Addendum: I see from Ms Savage-Rumbaugh’s just-posted and very interesting comment that the self-language tie may be even closer than that implicit in my “aside”.

  • As Terry knows, I resonate very much with the notion that relaxation of natural selection played a major role in human evolution. I have gone further in a recent review in

    Nature Reviews Genetics 9, 749-763 (October 2008) Human uniqueness: genome interactions with environment, behaviour and culture http://www.nature.com/nrg/journal/v9/n10/abs/nrg2428.html

    suggesting that aspects of human uniqueness arose because of a primate evolutionary trend towards increasing and irreversible dependence on learned behaviors and culture — perhaps relaxing allowable thresholds for large-scale genomic diversity. Examplars of this trend at the two extremes that are mentioned are:

    – the lack of genetic fixation of very long-standing hominin behaviors like control of fire

    – the loss of many aspects of pre-existing genetically fixed behaviors that are required for optimal mothering: significant in monkeys, worse in apes, worst in humans.

    Thus, one explanation for “Wallace’s Conundrum” is the human evolutionary path of offloading critical functions and behaviors from the genome to the phenome, and from the individual to the group.

    Not being an expert on language I will not specifically comment further, except to suggest that the evolution of human language abilities may have had a similar trajectory, with only core “innate” aspects now being genetically fixed, and the rest being dependent on input learned behaviors and cultural input from conspecifics.

    Ajit Varki, UCSD. http://cmm.ucsd.edu/varki/

  • Derek Bickerton

    It’s all too often assumed that arbitrariness is a distinguishing property of language, or at least some kind of “semiotic missing link”, in Frederik Stjernfelt’s words. It isn’t. Arbitrariness is a property of many animal calls, including all warning calls–think for a moment, you’ll see it could hardly be otherwise. Moreover, the assumption that vervet calls “mean” (or at least include) “leopard” is a back-projection of our own language-saturated view of the world. Vervet calls are intended to (a) make others aware of danger and its likely location (b) get others to take the appropriate evasive action. Thus the “eagle” and “leopard” calls are probably interpreted by vervets as things along the line of “Danger from the air!”, “Danger from the ground!”, “Get down from the trees”, “Get up into the trees”, etc., with the identitification of the predatory species being marginal if an issue at all. Vervet calls are indexical, not symbolic, as Terry’s book showed.

    Linguistic note: “There is a leopard, run” does NOT correspond to the Subject-Predicate structure of language, because “leopard” is not the subject of “run”.

  • Terry has raised some provocative and important questions, and sparked much stimulating commentary. I agree with most of what he says, but would add that as important to the evolution of higher cognition and language as relaxation of biological selection pressures was the introduction of a second form of evolutionary pressures, those associated with cultural evolution. I agree with Mark Turner about the relevance of ‘blending’. In the psychology literature this ability is generally referred to not as blending but as the ability to combine concepts and adapt them to new situations, and it is unfortunate that these two approaches to what is essentially the same phenomenon have not ‘blended’, for they have much to gain from each other. In any case, this ability surely played a pivotal role in the emergence of our ability to evolve culture. Terry aptly comments that blending “is too abstract in its current cognitive science incarnation, and needs to be decomposed into component neural processing dynamics”. I have a paper coming out this month in Creativity Research Journal that proposes a neural-level explanation of mechanisms underlying blending. The basic idea is that concept combination or blending is facilitated by being in an associative or divergent mode of thought, and the shift to a more associative mode of thought is accomplished by recruiting neurons that respond to abstract or atypical microfeatures of a task or situation. Since memory is distributed and content-addressable, this fosters the forging of associations to potentially relevant items previously encoded in those neurons. For example, in an associative mode of thought, someone who wanted to invent a comfortable chair might have a neuron activated that responds to anything that has the feature ‘conforms to shape’. Since beanbags have this feature, the person might combine the concept BEANBAG with the concept CHAIR and invent the BEANBAG CHAIR. The full paper can be found at:
    https://people.ok.ubc.ca/lgabora/papers/crj2010_neurds.pdf
    There is of course much left to be done here but I believe we are making significant headway toward understanding these complex issues right down to the neural level.

    A brief comment on ‘drift’. Terry says that “The usual consequence of relaxed selection is genetic drift—increasing the genetic and phenotypic variety of a population by allowing random reassortment of alleles.” However drift, as Wright (1969) defined it, generally decreases variety rather than increasing it. The term ‘drift’ is generally used to refer to refer to changes in the relative frequencies of alleles (forms of a gene) as a statistical byproduct of randomly sampling from a finite population. In other words, because the population is finite, some alleles are not sampled, and therefore eliminated, thereby reducing variety.

  • I could not resist grabbing this last chance before the final comment arrives from Terrence Deacon. To begin with, the notion of ‘symbol’ probably needs some refinement. As we’ve seen even from this discussion, arbitrariness alone is ‘too weak’ – true symbols are also conventional, which is a related but distinct feature (a point developed by e.g. Jordan Zlatev). Secondly, in Deacon’s formulation in TSS symbols were defined by their ability for entering combination, i.e. – to use Saussurian terminology – by their actual paradigmatic and potential syntagmatic relations. So on this account, symbols do presuppose a certain ‘grammar’ (an important caveat: as linguists including Derek Bickerton often point out, such basic combinatoriality is still very different from human syntax with its hierarchical structures and long distance dependencies).

    An extremely interesting comment came from Sue Savage-Rumbaugh. I feel that a vast majority of researchers in the evolution of language reject her conclusions. However, the idea of giving more prominence to the differences in the developmental trajectories rather than in the ‘raw genetic makeup’ between humans and (other) apes is indeed likely to become more and more influential as research progresses. I am particularly interested in Terrence Deacon’s reply to this post.

  • TERRENCE DEACON RESPONDS:

    In what follows I respond to some of the core issues raised by commentaries that I have not previously responded to. They are not in order of their appearance in this online discussion, but in an order that I think reflects a logical progression through an engagement with progressively deeper theoretical concerns.

    Salikoko Mufwene on the cart before the horse:

    I was surprised that Salikoko Mufwene didn’t get the gist of the co-evolution logic, and simply re-iterated the brain-changes-first mutation-leads-the-way logic that has for so long dominated and impeded progress in the study of human evolution. Hopefully few professional evolutionary biologists still cling to this simplification, even if the evodevo perspective hasn’t yet found its way into introductory textbooks and popular accounts of human evolution. Working from this perspective he appears to fear that this is “an adaptive/exaptive perspective that in some ways conflicts with the language that he uses to explain it” in which I postulate function that is present “anterior to the state of the hominin brain that could use it.” Not so. As I tried to make clear in my introductory essay, functional plasticity is the critical missing element. Ask yourself, did we need to have a special music-mutation before we developed musical ability and appreciation or a literacy mutation before reading and writing became possible. Indeed, there has not been brain specialization for these functions, and so from a Darwinian perspective, they are probably “value-added” side-effects of the flexibility of our evolved cognitive specializations.

    I obviously assume that “hominins had always been able to communicate”—indeed, communication is ubiquitous in the animal kingdom, it just isn’t symbolic or linguistic nor even a little bit, as Bickerton and other commentators are also quick to point out. In suggesting that “the hominin ability to think and solve problems differently from other primates seems to have preceded the emergence of “symbolic language” itself” we are left with two things to explain: what selected for that prior adaptation in our lineage alone, and how could those special abilities have been so well suited to this otherwise anomalous adaptation, language? The simple exaptation argument for was well criticized by Steve Pinker over a decade ago, pointing out that exapted functions are inevitably poorly suited to the new function and only a fraction of the prior adaptive complexity is likely to transfer. So to argue that this new function “emerged from the interactions of modules” begs the question that relaxation offers a partial answer to (without postulating modularity, only relative functional autonomy). The missing term here is synergy. And the mystery I address is precisely how a novel functional synergy could emerge, given that all contributing component processes must be reciprocally interdependent for that function to emerge. In case the logic of this mechanism is difficult to follow in such a short essay, I have provided a much more elaborate account with many additional examples of these sorts of processes in Deacon (2009). It’s time we stop with the 2% magic mutation and hopeful monster scenarios.

    Liane Gabora on creativity and drift:

    Liane Gabora’s efforts to dig beneath the superficial details of creative processes of all kinds follows a trail that should also be valuable for the cognitive difference questions pertaining to the neurological adaptations that support our unprecedented ease at communicating and thinking with symbols. We are on the same track here and ultimately there will probably be a convergence between the ideas of Turner, Gabora, and myself on this issue.

    Gabora is right to point out that the way I caricatured the concept of “drift” in my essay was incomplete and even somewhat misleading. Although drift is characterized by an increased influence of chance factors over selection, it is not necessarily, and indeed seldom, a relaxation effect. Wright was emphasizing a different source of non-selection effects. Thus, due to sexual reproduction and recombination in small genetically isolated populations, simply the random shuffling of alleles from generation to generation can cause some to go to fixation and others to go to extinction, by the luck of the draw, so to speak. Relaxation effects are also by definition non-selection effects but increase allelic variability irrespective of fixation or extinction effects of drift. So this is more an effect analogous to neutrality (though in a more-or-less sense), and is focused on mutational variety whereas drift is focused on the higher level of allele distributions within a population.

    Ajit Varki on anthropogeny:

    I am glad that Ajit Varki had an opportunity to offer access to his recently published paper. He presents precisely the sort of genetic approach that can detect relaxation effects in the human genome, and provides some interesting data that is consistent with this hypothesis. Specifically, it is suggestive of a general loosening of constraints on the genome, not only at the level of many loss-of-function mutations and deletions, but also at the level of chromosomal variability. More importantly, he has been the mastermind behind an invaluable resource on human evolution, gathering the theoretical and empirical contributions of a wide network of experts struggling to understand what he describes as “anthropogeny”—the processes that have shaped our uniqueness as a species. It is called CARTA, and many of its resources should be generally available on the web in the very near future.

    Sue Savage-Rumbaugh on clinging, epigenesis, and the gray divide:

    Sue makes an important point when she boldly states that “the ‘ground zero’ question does not apply to whether apes have language ability – they do.” I am often assumed to be arguing that the transition to symbol use that here has been dubbed ‘ground zero’ is an all-or-nothing discontinuity. In fact, my argument depends on Sue’s being right, though with some qualification. Bears can be trained to roller skate and parrots can be trained to produce appropriate English answers to questions about objects in their environment (as Irene Pepperberg demonstrated with Alex the parrot) even if these abilities never evolved, or ever could evolve, in the wild. With appropriate support, I suspect that many species can achieve some level of symbolic ability. The difference is not the presence or absence of some special symbolic neural module, but the capacity to assemble the complex mnemonic system upon which even simple symbolic reference requires, and this can be significantly aided by external support.

    Symbolic reference makes unusual demands on cognition, but as Sue has shown, when some of these demands are externally supported (as with a human symbolic cultural context) chimpanzees and bonobos appear to be able to join the game, at least at a minimal level (though Sue would probably argue that it is more than minimal). The evolutionary question is larger than this, however. I assume that our australopithecine and early hominin ancestors were at least as cognitively capable as our sister ape species in their ability to acquire symbolic abilities. But this is not sufficient. For symbolic communication to be a sustainable socially transmitted adaptation, this supportive context must also have been provided and maintained. For this reason I remain skeptical of Sue’s bet that symbolic communication is more widespread among apes in the wild than just in human laboratory contexts. The co-evolutionary ratchet that has honed both our symbol-learning adaptations and also the symbolic-culture transmission adaptations that guarantee the presence of a supportive context, doesn’t appear to have been initiated in any other animal lineage. If it had, I believe the social and behavioral consequences would be easily recognized because of their divergence from other forms of communication. Of course, if we don’t know what to look for, we may miss it, especially if it has only reached a very minimal stage of evolution.

    In conclusion, Sue is probably right that monkeys are really at ‘ground zero’ if that means incapable of symbolic cognition even with extensive support, but for me the hominid ‘ground zero’ was the point at which the co-evolutionary ratchet involving both the cognitive capacity, developmental flexibility (which Sue also points to), and the sufficiently reliable social context could take off. The error, which she highlights, is forgetting that adaptation is always to be understood in context, and the boundary between what is inside and outside is porous, especially during early development. This is the essence of the co-evolutionary perspective.

    Givón on evolution, genes, and flexibility:

    Let me preface my remarks by noting that Tom Givón is my model for a linguist who has taken seriously the entire range of topics necessary to understand language; including neurology, development, evolutionary biology, social psychology, and semiotic theory. I only wish more of our colleagues would recognize that the proper study of language is not language alone. That said, none of us involved in this project can be experts in all relevant areas, and I am always enlightened by what Tom says about linguistics. Nevertheless we do differ on several points, especially those regarding the evolution and genetics of language and the brain.

    I previously responded to certain of Tom Givón’s arguments against a role for relaxation of selection in the evolution of language, so I won’t reiterate those remarks here, except to point out that those features he cites as the “innovative vanguard of adaptive selection” are precisely those that will be most sensitive to relaxed selection. And this isn’t merely armchair speculation. Not only can the effects of relaxed selection be studied in the Finch case, but there are many other empirical exemplars available to study as well, especially involving domestication and founder effects. And as Varki (above) indicates, we are beginning to accumulate evidence for its genetic signature as well—even in humans.

    I was slightly more troubled by Tom’s repeating some evolutionary speculations (that still pass for established fact in certain texts) most of which reflect quasi-recapitulationist or progressivist conceptions of brain evolution. For example, he takes as given that “Older evolved systems exhibit more within-species uniformity, are more fully automated at birth, and are less sensitive to epigenesis, maturation and learning.” He adds that “The progression from old (‘reptile’) brain to mid-brain to cortex is a terrific example of this evolutionary trend, with decreased genetic specification, increased phenotypic and behavioral variability, and increased maturation-and-learning sensitivity.” The triune brain of Paul Mclean was a useful heuristic but it is no more than that. The telencephalon is as old as the mesencephalon and the homeotic genes that determine the major divisions of the brain have changed minimally in their structure or regional expression throughout vertebrate evolution. Descent with modification is the rule, not addition. Even cerebral cortex, which is distinctively mammalian, has now been identified with non-mammalian telencephalic homologues.

    There are indeed consistent differences in plasticity and structural-functional variability that distinguish, say, medulla from cerebral cortex; but these are not the result of relative recency of their evolution. Most of these claims are not critical to most of his points, and even when he says that “…the core neurological sub-systems that underlie language processing arose much earlier and still perform many of their pre-linguistic functions” he is actually consistent with a descent-with-modification view rather than an addition view (though some of his genetic arguments suggest a bias toward the latter; see below).

    Although in my introductory essay I did not present a detailed argument against strong genetic-based theories of grammar, I am indeed a critic of this view, and I can see why Tom might have responded with a modest defense of the nativist view, though among linguists he is often a critic of strong nativist claims. In defense of a genetic source for some features of universal grammar he says “Like all good species-specific universals, they surely are genetically coded.” The phrase “genetically coded” is highly loaded. As someone who has spent his career trying to understand the relationship between gene expression and species differences in brain structure, I am highly sensitive to simplistic claims about genes “coding” for certain traits. So while I completely agree that it is beyond doubt that there is a significant naturally selected genetic basis for human language abilities, I am far more wary of trying to draw a gene-to-grammar correlation, just because something is “universal.” There are other explanations for traits that are highly invariant—especially traits subject to so many levels of causal influence and functional constraint. Genes are only one.

    Tom’s characterization of humans as “the only species that, overwhelmingly, communicates about displaced reference by transacting declarative and interrogative speech-acts” is an insightful condensation of the functional features that are most distinctive of language functions, and implicitly also involves the issue of cooperative and honest exchange of information that both Wacawicz and Bickerton mention (discussed in more detail below). But while his comparison between hominid and honey bee foraging is worth pursuing in more detail, the range of selection scenarios that fulfill these criteria is not small. One would like additional constraints to help narrow this range.

    Givón, Bickerton, and Stjernfelt on arbitrarity of symbols:

    In response to a comment made by Frederik Stjernfelt, Derek Bickerton was quick to point out an important fact: that arbitrariness is not the critical defining feature of the transition to symbolic communication in our early ancestry. There is both a terminological dispute and a conceptual dispute implicit here. First, the terminological point. The term ‘symbol’ has come to be used differently in different traditions, and so first we need to be clear what we are talking about. If all that is meant is a mark that need not share any specific quality with its object of reference, then the term has trivial consequences. This gloss of the concept makes it easy to dismiss its importance for evolution, and indeed this simplification has been the motivation for many language origins researchers to imagine that it is syntax that demands explanation. This assumption about the concept of symbol is also reflected in Tom Givón’s comment that “Saying that human communication is ‘symbolic’ is saying relatively little.” The common usage of the “code” analogy also reflects this simplification, and for similar reasons leads to serious theoretical misunderstandings. A code does indeed involve an arbitrary mapping or correspondence relationship, but that is precisely why its reference is opaque and is the basis for encryption. This is because a code is a mapping of a parallel set of sign tokens to a language. So to describe language or any of its attributes, such as the basis for phonology, syntax, or semantics as a code, merely begs the question.

    Derek is entirely correct to point out that “Arbitrariness is a property of many animal calls” and that “the assumption that vervet calls “mean” (or at least include) “leopard” is a back-projection of our own language-saturated view of the world.”… “Vervet calls are indexical, not symbolic, as Terry’s book showed.” Thus arbitrariness is a red herring, and should indeed not be held up as a critical defining feature of language. But symbolic reference is not merely arbirariness and conventionality. As Charles Peirce pointed out over a century ago, we must distinguish properties of the sign vehicle (which he terms a representamen), which can include being an arbitrarily defined (i.e. conventional) type of sign vehicle, from properties taken to link it to its object of reference. Thus although current vernacular has habitually termed alphanumeric characters ‘symbols’ this usage ignores any referential relationship, and if not used carefully, in recognition of this shorthand, it can lead to all manner of theoretical confusion. Thus when your computer begins randomly spewing alphanumeric characters onto your screen they are indices of a malfunction, not symbols of anything. And likewise ;-) does not refer symbolically, even though it is composed of conventional tokens as is the term I-beam, though it mixes iconically and symbolically interpreted components. Peirce terms conventional sign vehicle types “legisigns,” and argues that symbols must also employ legisigns, but also that icons and indices can as well. A symbol is, in contrast, doubly conventional. It involves a conventional sign type that is additionally convention-mediated in the way it represents.

    Frederik knows this well, as does Tom, though it may not be clear from their remarks. We differ in our interpretations of Peirce, and how to define symbols. Because this may be confusing to readers, I will try to spell out in some detail the difference and explain why I think arbitrarity is a distraction.

    Arbitrariness is a negative way of defining symbols. It basically tells us that neither likeness nor correlation is necessary. But this is inadequate, even though it is a common shorthand way of characterizing symbolic reference. All sign relationships include some degree of arbitrarity, because those attributes that are taken as the ground for the sign-object linkage can be chosen from many dimensions. Thus, anything can be treated as iconic or indexical of almost anything else depending on the interpretive process.

    For example, with a bit of imagination a face can be discerned on the full moon, or in a cloud formation, and it might even remind you of someone you know. But iconism can also be highly abstract, as in Peirce’s demonstration that a mathematical equation refers iconically, once you know how to discern its symbol-mediated isometry (e.g. between the structure of the equation and a corresponding geometric or dynamical relationship). An equation can be interpreted to be iconic (e.g. of a parabolic trajectory) only, however, if you know how to discern the way that differences in the values or operations directly correspond to differences in the geometric object of reference. Frederik has recently written extensively on this point, so I found his remarks about arbitrarity a bit disappointing.

    Indices refer by contiguity in space, time, or substrate. A simple correlation can therefore be the ground for indexical reference. A lipstick smear on a man’s shirt collar can be a troublesome indication to his wife, a urine scent on a branch can be a sexual index to a female lemur, and the mobbing call of a small bird can indicate the present of a raptor. What gets correlated and how (accidental, cultural, evolutionary) is arbitrary, only the fact of correlation is not. Thus, a rat in a Skinner box pressing a bar in response to a bell in order to get a water reward has learned that the bell is an arbitrary index of the state of the apparatus. These states are arbitrarily paired in the experimental design, but that doesn’t make the one a symbol of the other. Repeated pairing—correlation—is, similarly, the basis for the indexicality of a vervet monkey alarm call sound and a type of predator and an appropriate defense activity. In this case, however, the pairing has occurred throughout the recent evolution of vervet monkeys. Other pairings that may have occurred did not also correlate with a sufficient level of survival. This evolutionary correlation is the basis for the innate indexicality of these calls. Elsewhere (Deacon 1997, 2003a) I have argued that innate iconicity and indexicality can evolve by natural selection, as in the above example, but that because of their displacement from intrinsic referential relationships, it is all but impossible for innate symbolic reference to evolve as a biological trait. This is part of the reason that I doubt the evolution of innate grammatical relationships, since they are largely symbolically derived.

    Arbitrarity is indeed aided by using convention to mediate the referential relationship, as opposed to some intrinsic sign attribute, since any similarity and correlation relationship involved is, in this case, produced external to the sign-object relationships. But arbitrarity is hardly diagnostic. So when I have used the term symbol I have adhered strictly to the notion of a doubly conventional referential relationship. And so it is the competence to construct this conventional system that mediates the interpretation of the referential relationship that matters and—to use Derek’s challenging image—this is ground zero.

    Though I have not had the context to make the argument here, I also agree with Sue Savage-Rumbaugh’s comment that “true symboling and grammar are inextricably intertwined. It is fundamentally impossible to have one without the other.” Once we overcome the tendency to treat symbolic reference as merely arbitrary correlation we can begin to discern the many contributions of the iconic and indexical supports of symbolic reference that have become incorporated into the constraints that define the grammar of language (for the general argument see Deacon 2003b).

    Wacewicz and Bickerton on cooperation and altruism:

    Consensus points? Slawomir Wacewicz begins his comments by listing four points of consensus that he believes should now be widely accepted, but which were highly contentious or even radical propositions at the time that my book The Symbolic Species was written. If he is correct, then considerable progress has been made in this field since each of these—language as niche construction, brain-language co-evolution, incremental evolution through a protolanguage phase, and a significant phylogenetic age for the origins of language—was first proposed. Indeed, each hypothesis has gained in acceptance due to the theoretical, simulative, and empirical results that have derived from efforts to test them.

    He pushes further, though, when he asks “Could your account perhaps be made compatible with some particular explanation of the emergence of altruism/cooperation as manifested in linguistic communication?” This same refrain is echoed by Derek Bickerton when he asks “why should anyone transfer useful information to non-kin?” Derek outlines his own recently articulated scenario, which loosely parallels the one I offered in The Symbolic Species: that foraging for animal remains with stone tools (which began at least 2.4 million years ago) required the evolution of a form of cooperative behavior that could only be maintained by the sharing of “honest” information. I agree with both Wacewicz and Bickerton that the evolution of symbolic communication and extensive cooperative prosocial behavior are necessarily linked. Just a brief reflecting on the astronomical quantity of socially inherited information that characterizes language and culture makes it obvious that it could not be reliably transmitted in a largely competitive selfish context. Its complexity and fidelity of transmission depend on a significant enhancement of prosociality in our species. But, beyond just recognizing this dependence, I believe that this linkage is the result of a co-evolutionary process that also involves a relaxed-selection plus niche-construction logic.

    In a paper soon to be published (Hui and Deacon, 2010), we offer a scenario for the evolution of what we describe as social addiction in which we argue that relaxed selection may be a crucial contributor to the evolution of altruistic and other prosocial behavioral tendencies. Recall that in the case of domestication, relaxation of selection can lead to degradation of functional autonomy and that this can increase the options for synergistic interactions. If cooperation is necessary to scavenge carcasses (some members fending off predators/scavengers while others cut off meat, and later share among all including mothers and infants who must stay apart because of the danger of predation), reliance on this strategy itself can relax selection on a variety of traits that would otherwise be necessary to success at non-cooperative foraging.

    One uncontroversial example of a relaxation effect was the significant reduction of the teeth and jaw muscles once critical to the ancestral australopithecine adaptation to a vegetarian diet. A half million years after the first appearance of stone tools, Homo habilis had lost this entire complex, including the loss of a specific myosin protein long evolved for the special demands of chewing in mammals. This degeneration effect is not unlike the degradation of the ability to endogenously produce vitamin C that occurred (because loss-of-function mutations of the GULO gene were not selected against) during early anthropoid evolution as a result of the regular incorporation of fruit in the diet. Loss of the ability to process the major food source available to non-cooperators would have selectively favored prosocial adaptations to keep a cooperative social group together because group dissolution would now be costly to all.

    Prosociality is more than altruism, however. It involves policing and means of promoting/enforcing egalitarian behavior (e.g. in mate access) as well as sharing of resources and information. The ratchet-like effect by which increased social addiction leads to increased offloading of inherited adaptive information onto social transmission processes leads to increased social addiction would have produced a deep interdependence between the cognitive adaptations for language and prosocial behavior, as both Wacewicz and Bickerton imply. This is what leads me to predict that the language adaptation complex is probably not centered around a built-in grammar so much as a diverse synergistic system of cognitive supports for social learning, social transmission, and reconstructing the thoughts and intentions of others.

    Again, let me re-emphasize (since some commentators, e.g. Mufwene, seem to miss the point, see below) that relaxation of selection is not an explanation for adaptation—that requires natural selection—it is rather an important and generally unrecognized precursor to functional reorganization and exaptation in evolution. Like genetic mutation, its consequences are generated irrespective of functional outcome. The point is that the dedifferentiation that inevitably follows from relaxation of selection can be a contributor to the emergence of unprecedented interactions between subsystems that were previously specialized for separate and independent functions. This simply increases the possibility that synergies will spontaneously emerge. Nevertheless, by enabling previously prohibited (by natural selection) interactions between separate adaptive mechanisms, this can expose these previously unavailable interaction-effects to natural selection which can now hone any synergistic consequences. Although I entirely agree with Liegh Van Valen that there can also be direct selection for flexibility, I would argue that this requires a more unusual and persistent context than do relaxation and degradation effects. To put it simply, it takes far more restrictive conditions to drive an adaptation to fixation in a species, than to degrade it.

    To conclude:

    The many thoughtful responses to this target essay offer no more than a tiny peek into a field of study that has seen an explosion of research interest in the past two decades. They also demonstrate both the diversity of approaches and of fields of study relevant to this topic. The study of the origins and evolution of language is still in its infancy. As a result, as Givón notes, discussing it inevitably opens a can of (speculative) worms. It poses one of the most difficult scientific challenges of our age, but our science has matured to the point where it is no longer just a playground for armchair speculation. To approach this complex mystery with the respect it deserves, however, we must be prepared to give up on simple one trick accounts, innate mentalese, miraculous mutations, increase in general intelligence, and so forth, and embrace its complexity as a semiotic-biological-epigenetic-social phenomenon whose structural features reflect the convergent co-evolutionary interactions of all these levels of causal process. Because of this, language defies evolutionary arguments that fail to account for the co-evolutionary mechanisms operating across these many levels and involving many diverse processes—and not merely natural selection. In this respect, I agree with Mufwene’s characterization of language as an “emergent” phenomenon, though for different reasons. In this respect, studying language origins offers an unprecedented view into the complex circuitous creative processes that have produced nature’s most spectacular masterpieces of design. I thank all who have responded for a lively and challenging dialogue.

    References:

    Deacon, T. (1990) Fallacies of progression in theories of brain size evolution. International Journal of Primatology 11: 193-236.

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

    Deacon, T. (2003a) Multilevel selection in a complex adaptive system: the problem of language origins. In B. Weber and D. Depew (eds.) Evolution and Learning: The Baldwin Effect Reconsidered. MIT Press, Cambrige, MA, pp. 81-106.

    Deacon, T. (2003b) Universal grammar and semiotic constraints. In Language Evolution, M. Christiansen & S. Kirby, eds., Oxford U. Press, pp. 111-139.

    Deacon, T. (2005) Language as an emergent function: some radical neurological and evolutionary implications. Theoria 54: 269-286.

    Deacon, T. (2009) Relaxed selection and the role of epigenesis in the evolution of language. In Mark S Blumberg, John H Freeman, Scott R Robinson (Eds.), Oxford Handbook of Development Behavioral Neuroscience, Oxford University Press.

    Hui, J. and Deacon, T. (2010) The evolution of altruism via social addiction. In Social Brain, Distributed Mind. R. Dunbar, C. Gamble, and J. Gowlett (eds.). Proceedings of the British Academy/Oxford University Press.

    Striedter, Georg F. (2004) Principles of Brain Evolution. Sinauer Associates.

  • This conversation, while ending here, continues on Facebook. Join us there by logging on to your Facebook account and proceeding to our group: http://bit.ly/OnTheHumanFacebook.