On the recent origin of symbolically-mediated language and its implications for psychological science

Second draft chapter prepared for
S. Lea and M. Corballis (Eds) Evolution of the Hominid Mind.
Oxford: Oxford University Press.

Andrew Lock
Massey University


When might humans have begun to use language? The recent literature that deals with this question provides hypotheses of all shades. For example, Pinker (1994, pp. 352-3) claims that 'the first traces of language could have appeared as early as Australopithecus africanus. ... It is easy to imagine some form of language contributing to [the success of Homo erectus].' Holloway (1969) and Tobias (1987, 1991) also argue for an early emergence of language, not with australopithecines, but perhaps with Homo habilis. Bickerton (1990) is more conservative. He draws a distinction between protolanguage and language, and argues that 'protolanguage did not develop until erectus emerged. This would give protolanguage a time depth of around a million and a half years' (p. 141). In his view, real language, as distinct from protolanguage, arose more recently, with modern humans, giving it a time-depth of at least 100 000 years. Bickerton is thus in agreement with Lieberman (1991) who is confident that 'we can date language as we know it back to at least ... 100 000 years ago at the edge of Africa and Asia' (p.172). Noble and Davidson (1996) are more cautious still, concluding that it was 'sometime between about 100 000 and 70 000 years before the present [that] the behaviour emerged which has become identified as linguistic' (p. 217).

Yet there is precious little evidence for language even at that late stage: for, as Halverson (1993) notes, 'we do not know whether the Cro-Magnons [at 35,000 bp] even had a fully developed, grammaticized language' (p. 762). And if we were to appeal to direct data, then the earliest incontrovertible evidence that humans could speak is not much more than 100 years old (i.e., Edison's recordings). Beyond that, we must work by inference from less direct evidence. What do these sources tell us?

Setting the boundaries

The earliest dates for the appearance of the species to which we belong, Homo sapiens, are unclear. The earliest specimens which may be 'like us', such as those from Petralona and Heidelberg at around 300 000 years ago, are theoretically indistinguishable from the later members of Homo erectus. Current evidence points to anatomically modern Homo sapiens as being a very recent species, less than 200 000 years old. The earliest dates for anatomically-modern human fossil remains are around 100 000 years ago (eg., those from Klasies River Mouth; South Africa: Skhül and Qafzeh, both Near East---see Campbell 1996 and Waddell and Penny 1996 for details). Mitochondrial and nuclear DNA studies are consistent with the conclusion that the human species had an African origin within the past 200 000 years, and began to move out of Africa from around 100 000 years ago or later (though not necessarily in one wave). What do we know about these early modern humans?

Human species characteristics

We are faced today with an organism that would appear to have possessed a human-like brain for at least 1.5 million years; that has an upper respiratory tract modification dating back around 0.75 million years that is most parsimoniously explained as an adaptation for human-like vocalisation capabilities; and whose technological skills appear to have been based on modern intellectual capacities for around 0.3 million years. The evidence for these clams is reviewed below.

Brain organization

The first member of the human genus, Homo habilis, is dated to a little over 2 million years ago (Wood 1992). Homo habilis, or its conspecific Homo rudolfensis, provides the first evidence for a human-like specialization of the brain. There is a consensus that the endocast of the specimen KNM-ER 1470 shows a somewhat more complex and modern-human-like third inferior frontal convolution than is typical of modern apes, suggesting that by this point the brain was becoming organised into a more human rather than ape form (see Holloway 1996 for a review). Also evident from this skull is a reorganization of the posterior cortex, which Wilkins and Wakefield (1995, p.171) take as indicative of a shift in cognitive functioning from that of apes to a 'uniquely human modality-free sensory representation'. Thus, from very early in our genus, with respect to the inferior frontal and inferior parieto-temporal cortices, there is evidence of a more human-like structure than previously seen. However, no data are available from this time as to the configuration of the basicranium, and thus there is no way of telling whether the evidence on the development of Broca's area provides a basis for enhanced right-hand motor control or the voluntary control of vocalisation (see Peters 1996). Later endocasts through Homo erectus and archaic sapiens specimens are devoid of clear convolutional detail (Holloway 1996), thus making it impossible to be certain about any further possible re-organizational changes within our own genus. Evidence as to the appearance of modern brain organization is scarce. But, a reasonable assumption would be that the brain of modern humans is a species characteristic, and as such, the brains of early examples of our species at around 100 000 years ago would be the same kinds of organs as those we find in individuals who have existed since then.

Vocal capabilities

It is not until Homo erectus sensu stricto, particularly specimen KNM-ER 3733 (Homo ergaster sensu Wood 1992), that an incipient flexion of the basicranium is found. This partial flexion may well correspond to a partial descent of the larynx, which would have increased this specimen's ability to make more human-like sounds (Laitman and Heimbuch 1984; Laitman et al. 1992). The configuration of the basicranium of the skull of earlier hominid fossils, as well as later Neanderthal ones, indicates that their upper respiratory anatomy was not fully modern, but was more akin to the general primate configuration (which allows for simultaneous swallowing and breathing without the danger of choking that modern humans risk). Data are not available for later H. erectus specimens other than the Salé specimen - possibly early sapiens rather than late erectus (Hublin 1985) - which appears to show the modern configuration. Other specimens of middle to late Pleistocene hominids assigned as archaic H. sapiens, such as those found at Petralona and Broken Hill, have basicranial configurations that lie within the modern range. On the basis of these analyses it can be inferred that their upper respiratory tracts were also similar to our own (e.g., Laitman et al. 1992). This would suggest sound production (or speech) capability on a par with that of current humans (see Arensburg et al. 1989; Laitman and Reidenberg 1988), but this point has perhaps been over-emphasised in the literature. Thus it would appear that there have been changes in the structure of the hominid upper respiratory system over a period of at least 750 000 years which can only be explained as an adaptation to vocal communication. The full adaptive anatomical suite is a species characteristic of anatomically modern Homo sapiens.


When might modern cognitive abilities have come into being? How far back in the record can we go before we fail to find evidence of capabilities that would fall within the range we can infer for modern humans? The only source of evidence is the material record.

H. habilis is regarded as having made those stone tools referred to as the Oldowan industry, consisting of simple core and flake tools of basalt, quartz and quartzite. Wynn (1996) notes of these tools that:

There is considerable question as to the reality of the tool types in the Oldowan. If by type we mean a well-defined category of tool that existed in the minds of the tool-makers, then we would be unable to argue for their existence. ... It is doubtful whether there were any design criteria whatsoever, beyond perhaps big and little (p. 267).

Similarly, Toth and Schick (1993) note that the so-called 'core tools' are 'not necessarily tools, nor do they necessarily correspond to 'mental templates' held by their early makers'. Toth and Schick (1993) and Fagan (1989) agree that this is an ad hoc technology, and Wynn (1996) concludes that 'Oldowan tools did not require a particularly sophisticated intelligence' ( p. 267). Nevertheless, these writers are of the view that the cognitive capabilities of these tool-makers may show some advance over that of modern apes in two ways: first in the planning of the sequences of motor actions needed to produce such tools; and second in judging the angles and overhangs of the 'blank' cores with respect to the angle and force of the blow required to effectively flake the stone.

The technology associated with early forms of H. erectus, by contrast, appears to be much more sophisticated, and the typical Acheulean hand axe or biface is 'the first type tool that is clearly outside the range of an ape technology' (Wynn 1996, p. 269). The symmetrical nature of the early bifaces could be produced without their makers having any mental concept of symmetry (Wynn 1996; see also Davidson and Noble 1993), and there is in fact no consensus that these bi-faces were actually tools, as distinct from the cores or debris left from making flake tools (e.g., Bradley and Sampson 1986; Potts 1989). But the later bifaces show such a high degree of three-dimensional symmetry that to regard them as being only core remains and not the intended finished artefact seems to be stretching it. For example, both of the cores in Fig. 1 appear to have been worked on in order to modify their final shape, and the flakes that have been removed appear to be waste rather than cutting flakes that were intended as such. With respect to Fig. 2a, Wynn (1979) comments that 'here the symmetry has been created by four short sections of retouch (A,B,C and D) which are not contiguous. In order to do this the maker had to have a competence in the relation of whole to parts' (p. 377): thus, the end result appears to have been anticipated prior to the observable retouching. And with respect to Fig. 2b he notes that

Both faces of the original flake have been retouched to yield a remarkably straight edge. ... in order to have produced such artificial straightness the knapper had to have related each flake removal to all the others and also to a stable point of view (p. 378).

That stable point of view had to be 'conserved in the knapper's imagination during the actual flaking process' (p. 380). On this basis, Wynn concludes that

these artifacts required the organizational abilities of operational intelligence and ... therefore, the hominid knappers were not significantly less intelligent than modern adults (p. 371).

His overall conclusion is that 'essentially modern intelligence was achieved 300 000 years ago' (p. 371).

It would appear, then, that early modern humans had an essentially modern intelligence, and on that basis also had the capacity to act in the ways that later modern humans do. What evidence is there that they used this capacity?

Indices of intelligent behaviour

The material artifacts associated with these early moderns actually lack many of the characteristics of the modern tool-kit, and overall, early anatomically-modern human remains are associated with industries of greater antiquity and non-modern characteristics, leading Stringer (1989), for example, to conclude that

if the postulated dispersal of anatomically modern humans from Africa was associated with new forms of cultural or behavioural expression, this was not reflected in any simple or direct way in the character of the associated lithic industries (p. 7).

Thus, modern human behaviour is not a species characteristic. In fact, there is an emerging consensus in palaeoanthropology of a marked temporal disjunction between the appearance of anatomically-modern human forms---the species Homo sapiens---and modern species-typical behaviour (for a marshalling of the evidence, see Noble and Davidson 1996). For, quite remarkably, the period from the emergence of our species through to around 40 000 years ago shows little substantive change in the archaeological record associated with modern human forms as compared to more archaic ones. Throughout this period the artefacts of modern humans suggest remarkably un-modern activities. The complexity of the technology associated with both premodern and modern humans stays fairly constant with respect to technique, raw materials, number of components combined together, and the number of stages involved in the construction of tools (see Figs 2 3, and 4). And perhaps most tellingly, there is no evidence of any symbolic practices (e.g., Chase and Dibble 1987; Lindly and Clark 1990). As Lindly and Clark (1990) conclude from their review, 'neither archaic H. sapiens nor morphologically modern humans demonstrate symbolic behaviour prior to the Upper Palaeolithic' (p. 233).

'Modern' behaviour

What characteristics does behaviour need to have in order to count as modern? Contemporary human activities rely on the social and cultural maintenance of symbolic resources. The means of both the maintenance and transmission of these symbols are conservative; the symbols themselves are volatile. The physical instantiations of contemporary human symbolic activities show temporal volatility, differential spatial distribution, and spatial relocation. Temporal volatility is not necessarily associated with functional utility. Clothing, for example, is functional, but fashion is volatile in entirely nonfunctional ways; hem lines on dresses go up and down for reasons unrelated to the non-symbolic function of wearing a dress, just as accepted male and female garb differs according to custom rather than function. Differential spatial distribution means that the pool of cultural products of human activities that would be found within five kilometres of where you are reading this chapter will have a different constitution from those within five kilometres of another person in Tibet or New Guinea. If an inventory were available for your actual present location five years ago, then the temporal volatility of those products would again be apparent. Spatial relocation is demonstrable in the same way: The natural availability of the products that make up cultural materials within five kilometres of any particular point is vastly different from the proportions in which they have been assembled by that culture. The resources available to us move over large distances; even the food on one's plate can be a geography lesson.

Taking these properties as criteria for characterising modern human cultural practices, we can ask when, in the archaeological record, is there evidence that humans acted in characteristically modern ways? That is, when did the material artifacts associated with humans show patterns of temporal volatility, differential spatial distribution, and spatial relocation?

1. The evidence is almost completely negative before 40 000 years ago. Prior to that time there is abundant evidence, in the form of artifacts, that cultural activities were a characteristic of human life, way back to non-modern Homo species approximately 2 million years ago: but there is very little unequivocal evidence of symbolic mediation, as indexed by the above characteristics.

2. The evidence is conclusively positive after 20 000 years ago (a point in time often referred to as the 'symbolic explosion'); the evidence of the intervening 20 000 years is less clear cut, but strongly supportive of human life being a symbolically-mediated activity, even if the spatial and temporal distribution of cultural artifacts were not fully modern in their characteristics.

Thus, there is a temporal gap of around 60 000 years between the appearance of anatomically-modern humans, and their giving evidence of acting in characteristically modern ways (for reviews, see Conkey 1996; Gamble 1993; Lindly and Clark 1990; Noble and Davidson 1996; White 1996; Wynn 1996).

Note that here the distinction between modern and not-modern is being drawn on grounds of artifact style and its volatility rather than inventory. The transition point noted above at about 40 000 years ago has been referred to as the 'Mid/Upper Palaeolithic Transition'. Prior to that time, there is little in the archaeological inventory that is comparable to the typical subsequent inventory. It is generally accepted that certain lithic techniques (blades as opposed to flakes), tool forms (e.g. hafted spears), and materials (e.g., no bone or antler-based tools) distinguished the two periods. Such a distinction is probably a conceptual oversimplification, and if held rigidly both raises unnecessarily the issue of the continuity or discontinuity of 'traditions' and draws attention away from the interactive assemblage of mosaically-patterned developments that feed into each other to provide bases for new possibilities. In addition, the distinction is always open to empirical refutation, such as evidence of otherwise modern bone tool technologies reported by Yellen et al. (1995) at around 90 000 years ago in Zaire.

But, it is not the particularities of individual artifacts that is the dividing line. Rather, the distinction being drawn here is based on the suite of recovered indices. Thus, Gamble (1990) has speculated that

elements of the full Upper Palaeolithic package such as modern skulls, blade technologies, hearths, even, I would suggest, undecorated bone tools, will be found over wide areas and throughout periods from 200 000 BP to 40 000 BP (p. 170).

What is presently lacking is evidence for the assembled suite being in place before the Upper Palaeolithic. Before 40 000 years ago there was no art, and no spatial or temporal volatility of style in the tool-kit record. Foraging patterns were apparently based more on scavenging than hunter-gathering, and there was less geographical movement of tool materials. After the transition, a tool, even if still made of stone rather than materials such as ivory or bone, might be found 900 or more kilometres from the nearest known natural source of that stone. After the transition, too, the archaeological record contains evidence of typically modern human practices of a symbolic nature. (The one piece of present evidence that is problematic for this transitional date of 40 000 years ago is the date for the human colonization of Australia. Whenever this was - and it could have been around 60 000 years ago - it is unlikely it could have been accomplished accidentally, but required the deliberate building of a boat. Whether a boat could be built by a human group that lacked language is a moot point (see Davidson and Noble 1992).)

Social life

The above change in 'technological' assemblages is correlated with changes in the social organization of human groups. Social life prior to this transition may be inferred as still not fully modern. There is some early evidence for the use of fire, and perhaps cooking, as early as 700 000 years ago at Zhoukoudian (Stringer 1985), but there is no substantive evidence for hearths, storage pits, or architecture. Again, there is some early evidence for true hunting, if only of smaller mammals (Binford 1985; Shipman and Rose 1983; Thieme 1997), even if scavenging were still the major source of animal remains. Particular sites appear to have been used for particular activities (e.g., de Lumley 1975; Freeman 1975; Keller 1973), suggesting a spatial division of labour. But the overall picture here is complex and difficult to interpret, and the evidence is scanty. Gamble (1993) concludes his review of this technologically pre-modern period by noting that, with respect to social life:

It leaves an overwhelming impression of spontaneous, highly episodic behaviour where stone tools were made to do the job in hand before being dropped and their makers moving on. ... What is lacking ... is any indication for such modern practices as detailed planning, widespread contracts (sic), or elaborate social display. There is no physical evidence of storage, raw materials all come from within a radius of 50 km, and usually less than 5 km of the sites where they were used and any form of art, ornament, jewelry, or decoration is entirely absent. ... [But] the fifteen minute culture as revealed in the manufacture and use of stone tools is a poor guide to the length of time over which social information could be retained. The occupation of seasonal environments provides a clear indication that such memory was now substantial (pp. 138-9, 143).

After the transition to modernism, the evidence indicates changes in social organization in two directions. First, there is an increasing spatial and temporal extension that elaborates and sustains

extended kinship networks, communication beyond face-to-face encounters and exchange of information beyond the here-and-now, the organization of logistical economic strategies, and the extension of the time depth of adaptation to environmental fluctuations (Whallon 1989: 451).

Second, there is an intensification in the organization of the immediate social environment. Built shelters and semi-permanent 'villages' are found after 40 000 years ago, but not before (see, for example, Gamble 1986). At first sight it might seem paradoxical, but these two changes can reinforce each other with respect to their effects in elaborating a linguistically-mediated awareness of the world. Both increasingly break the commonalities of shared knowledge between an individual and others: on the one hand in the meeting of strangers, and on the other in the creating of strangers through the implicit demarcation of the public and private within the domestic society (see, for example, Wilson 1988, for a fuller discussion). I will address this issue of the structuring social relations as providing a zone of proximal development in a later section.

What might we make of all this? The conclusion to be drawn from all of the sources of evidence just reviewed is that modern human languages were established after the emergence of our biological species. The biology that supports these languages therefore evolved in the absence of such languages, and was subsequently co-opted to support present abilities as a result of quasi-historical changes in human social and technological practices through a long process of 'bootstrapping'. What follows is, then, a socially-driven account, similar to that recently put forward by Noble and Davidson (1996), who argue that

'understanding' describes an interpersonal achievement, full stop. Brains are needed for this, but they are not where understanding occurs. ... Far from 'mind' as a personal possession, it is better characterised as socially distributed (p. 105).

This social distribution of mind is part-and-parcel of the communicative use of symbolic signs, which they regard as the defining feature of language. It is these symbolically-mediated social practices 'which happen to be unique to humans, [that] recruit the structures of the brain [to support them], rather than being determined by them. ... Practices interact with structures' (p. 18).

The central point to be made in what follows is that social structures put varying pressures on the communication systems that sustain them, through the different levels of presuppositionality a society's members share with each other. With low social differentiation and interchangeable roles, people communicate with each other against a background of a shared or common orientation, in a context of interpersonal relations founded on common perceptions, values, interpretive competence and so on. As social structure differentiates, presuppositionality decreases. Thus, the communicative practices within an increasingly elaborated society must develop ways of creating common contexts within the medium of communication itself. The contexts which make interpersonal communication possible have to be lifted out of the everyday milieu and created within the symbolic system of the lingua franca. Essentially, contexts will be 'pressured' into symbolic codes, and become conceptualized as objects of knowledge rather than remaining solely as implicit processes which sustain knowledge. In the process of this translation from context to code, grammar can be seen as emerging from the co-option of hierarchical ordering abilities that were already pregnant within the human mental apparatus. Grammar functions to 'freeze' the implications of otherwise immediate contexts into explicit forms. Grammar keeps track of these implications by structuring them into a symbolic tool kit that mediates our relation with the world we live in. The evidence broadly supports a view of communicative symbol systems as the providers of 'cognitive technologies' or 'tool-kits' that variously afford analytic, context-independent thought. This is not to claim that language determines thought, but rather that particular symbol systems can make some universal mental operations more or less difficult to carry out.

Culture, biology, and development

The internalization of cultural forms of behaviour involves the reconstruction of psychological activity on the basis of sign operations. ... The internalization of socially rooted and historically developed activities is the distinguishing feature of human psychology, the basis of the qualitative leap from animal to human psychology. As yet, the barest outline of this process is known (Vygotsky 1978, p. 57).

The evidence on human evolution reviewed above indicates that the biology required to support modern behaviours was present well before those behaviours were put in place. Thus, when we ask questions about the evolution of language, which I include as a prime component of modern behaviours, we need to jettison the notion of language having an innate basis; that there is some module in the human brain that is specifically given and evolved to support grammar. We need to do this because there is no evidence that modern, grammatically-structured languages were in existence during the period when the biological substrate of the modern brain was being selected for. Thus, it would appear that whatever hardware we bring to support current-day tasks was put in place before the behaviours they now support were elaborated.

This is not as outlandish as it might at first sound. An example is suggested by the practices of reading and writing. The claim here is that all of the biological machinery that evolution had to construct to enable people to write was in place 100 000 years ago, yet reading and writing were not accomplished until about 95 000 years later. In technical terms, writing systems are extrasomatic achievements. In the process of learning to read and write, it is quite likely that the structural and functional characteristics of our simultaneously developing and maturing brains are modified from the final patterns they might have ended with in the absence of this extrasomatic environment. However, these changes are not directly related to any changes in the phenotypic information that evolutionary processes are typically conceived to operate on. In this way, evolutionary questions with respect to modern behaviours, such as speaking a language, are shifted to the extrasomatic domain. The biology that enables them was there. The real questions are, firstly, how were linguistic systems put together by the possessors of this biology? And secondly; how do these systems, once they are established, interact with the ontogenetic elaboration of the biology that supports them? But before I can tackle these questions, there is some groundwork to be done.

My emphasis on the extrasomatic domain needs some clarification. I am not looking to present arguments that cultures 'evolve' in a fashion analogous to the way organisms evolve. Rather, it is quite clear that evolution is a far more complex process than proponents of the neo-Darwinian Modern Synthesis have made out. That view is exceptionally individualistic in locating the site of evolution as the individual animal (or even the individual gene). It is characterized by a rigid separation of organisms from their environments. But evolution is not just based in the differential culling of variations between individuals by autonomous and natural selection forces emanating from a fixed environment. Environments evolve along with the organisms that inhabit them. And as environments evolve, they can be used by the evolutionary process itself, to allow biological structures to be put to new functions.

Such an expanded view both legitimizes and makes orthodox our approaching human evolution in terms of an extrasomatic emphasis, since there are good grounds for assuming that this is a viable way for an evolutionary process to do things. In the present context, if we eject the so-called information required to sustain language 'out of the brain', we do not need to account for how a language module evolved. This is a useful move, for two reasons. First, it is extremely difficult to elaborate an evolutionary account of a 'mental organ' which is by its nature discontinuous with anything that has gone before it (cf. Chomsky 1968, p. 60). Second, how else might we account for the temporal gap between the completion of the biology and the emergence of the behaviour? Brain mechanisms cannot have been selected to better adapt people to something that was not present in their environment.

Instead, we can look to processes of co-option or exaptation (Gould 1991). Hence, if we now, as modern adults, can be shown to have something that can be described as a 'language module', then this can be seen as developing as a result of the interaction between phenotypic information and extrasomatic information. Biological processes are co-opted by environmentally-conserved information to support new ends and outcomes. What I will be arguing is that, for the elaboration of human languages and linguistically-mediated behaviour, we are becoming aware of how these extrasomatic factors have their influence. They work by changing the structuring of the attention of those who grow up within them, and thereby enter into determining the phenotypic outcome of the developmental process.

There is evidence that different social structures facilitate the development of different mental ways of dealing with the world (see Lock and Symes 1996, for a review). How we attend to what we see is mediated in our developing years (a period when 60% of brain development occurs) by the actions of others. And part of our development is taking over these patterns of attending into our own conduct. As what we 'see' becomes increasingly symbolic, so our patterns of thinking get differently founded. This point is implicit in the central concepts of Vygotsky's (1978) developmental account, that:

Every function in the child's cultural development appears twice: first, on the social level, and later on the individual level; first, between people (interpsychological), and then inside the child (intrapsychological). This applies equally to voluntary attention, to logical memory, and to the formation of concepts. All the higher functions originate as actual relations between individuals (p. 57).

The argument here is that structuring principles and information for developmental processes are conserved extrasomatically, both in durable artefacts and durable cultural practices. These provide new substrates for the developmental experiences of later generations. Hence, as patterns of social interaction change, they enable new ways for attention to be structured (because the individual is immersed in a perceptual world increasingly structured by social interaction and its products). We therefore need to look to the changing patterns of human social interaction for an explanation of our ability to have elaborated pre-existing communicative abilities into linguistic abilities that are symbolically mediated and constituted. Emerging brains of an already-modern sort handle these culturally-conserved sources of structuring information as best they can, by the use of pre-existing potentials rather than the deployment of already-given, task-specific modules. Those language-related modules, when (or if) they exist, are most likely the result of development under particular conditions of environmental input: they are not inevitable, in the way that modules for recognizing faces, say, are likely to be.

A central question in this exploration asks how what Vygotsky has called the zone of proximal development can be generated sui generis, in the absence of more enculturated others. A lot of developmental processes have come to be seen as creative ones. For example, from a Piagetian perspective we conceive of the child as an active problem-solver; and from a Chomskyan perspective we see the child as creatively elaborating language. But in each case there is a model, something there to be mastered, an agreed end-point. The Vygotskyan perspective has been read as one of 'apprenticeship' or 'implicit tutoring': Through the actions of someone who has already mastered the end point, the child is assisted in getting there. Development is scaffolded by its embeddedness within social interactions that point to a known outcome. But when we turn to evolution, we have the problem of the outcomes not being there already. The development of the ability to count, for example, can be seen to be scaffolded by those who can already count, but what enables counting to be constructed when no-one can count? Fortunately, we have some clues as to how to refine and pursue this question.


Implicit properties

Popper (1972) has pointed out very clearly how particular symbol systems contain the seeds of their future elaboration. He points to the number system as one example. Many four-year-olds can count from 1 to 10, at least after a fashion. But they do so in complete ignorance of the properties of the numbers they are using. Among those properties are the distinction between odd and even numbers, the nature of prime numbers, and the nature of perfect numbers, for example. Hence, there is built into the system a whole new order of distinctions that could be made explicit. Making these things explicit is not necessarily easy. First, the way in which numbers come to be represented or symbolised can affect the likelihood of these implicit properties being discovered. Basic mathematical operations such as addition, subtraction, division and multiplication are made more or less easy depending on how the number system is notated: the Arabic system is better suited to this than the Roman one; and an abacus is better still.

Second, similar views have been put forward for alphabetic literacy. One such view (e.g., Goody 1968) is that writing functions to abstract the events it deals with from their ongoing context, and so helps foster abstract thinking. But the empirical evidence suggests that coming to exploit this possibility of writing, to use the written medium as a way of inspecting different accounts for their logical consistency, for example, is not a guaranteed outcome of being able to read and write. Rather, the social practices that writing is used for differentially predispose readers and writers to discover or not discover the possibilities made available by the system they have (e.g., Scribner and Cole 1981). In addition, the nature of the system for visually representing language---logographic, syllabic, alphabetic, for example---may also make this discovery more or less easy or difficult, in the same way that different systems of mathematical notation do (although the empirical evidence to warrant this claim has been very difficult to establish).

Third, environments can be changed by the actions of organisms so as to bring into being entirely new objects that were never intended and would not have otherwise existed. Popper (1972) uses the creation of a path through vegetation as an example of this. Paths come into existence through animals initially moving along the lines of least resistance, which are determined jointly by the contours of the terrain and its vegetation along with the predispositions of the animals that move across that terrain. He note of a path that:

It is not planned - it is an unintended consequence of the need for easy or swift movement. This is how a path is originally made - perhaps even by men - and how language and any other institutions which are useful may arise .... In this way, a whole new universe of possibilities or potentialities may arise... (p. 119).

What I want to emphasise here is that, unlike the possibilities that are implicit in counting systems, what has been constructed in this instance is immediately available to the organism: The path is as real in the organism's experience as the terrain and vegetation through which it runs.

At this point, then, the issue becomes clearer. Can we find ways in which social practices can act to create the analogues of paths in human perception, such that those paths not only establish new possibilities and potentialities for a physiological system (the brain) to use, but simultaneously provide the notational system that eases the discovery of those things that are implicit in them? If we can, then we have delineated the self-constituted zone of proximal development that can bootstrap the elaboration of those symbolic cognitive technologies that it is our task to elucidate in accounting for the evolution of human abilities.

Developmental clues

Another set of clues comes from the study of language development by present-day human infants. To see such information as helpful here is not to imply a crass recapitulationism (see, for example, Lock and Peters 1996). Rather, I just want to make the simple point that development is an interactive process. Both what an infant learns and how this is learned occur in close contact with other human beings. And the actions of these others contribute a massive structuring to the world as-attended-to by the infant. Vygotsky has captured one facet of this structuring process in his conceptualization of the zone of proximal development. Intrinsic to these actions are the 'signs' (words) available for linguistically-mediated behaviour. Within the broad Vygotskyan perspective can be found the seeds of a theory of the relation between the signs used in interpersonal communication and the character of the social relations within the group who use those signs. Vygotsky's contemporary, Volosinov, puts it thus

Every sign, as we know, is a construct between socially organized persons in the process of their interaction. Therefore, the forms of signs are conditioned above all by the social organization of the participants involved and also by the immediate conditions of their interaction. .... This is the order that the actual generative process of language follows: social intercourse is generated ...; in it verbal communication and interaction are generated; and in the latter, forms of speech performances are generated; finally, this generative process is reflected in the change of language forms. (Volosinov 1973, pp. 21, 37).

This way of looking at the issue helps refine it by seeing the elaboration of social structures as providing the substrate that constitutes, sui generis, a zone of proximal development that leads language-users on. What I will go on to show, later in this chapter, is how changing relations between people can act to structure each other's perception and attention in new ways, simultaneously bringing into being new things to talk about. In this way, social relations scaffold the elaboration of a linguistic system by providing the requisite structuring principles to recruit the available biology, and, at the same time, also provide the cognitive technology or 'signs' that those biologically-based processes can use (and thereby simultaneously elaborate themselves within the constraints of the developing phenotype).

Comparative clues

A third set of clues comes from attempts to establish the language capabilities of apes (see, for a review, Ristau 1996). There has been a tendency for this work to be seen as establishing the limits of ape abilities. In this way, such work can highlight particular abilities that human individuals must possess that enable them, as a species, to have transcended the limitations apparent in apes. The dominant cognitive paradigm of contemporary Western psychology tends to favour casting these differences within a model of internal representations and the devices by which these representations are handled. These devices include things like our having a larger working memory capacity (Russell in press), a greater imitative facility (Tomasello et al. 1994), or better cross-modal transfer of information (Ettlinger and Wilson 1990), etc.

But this perspective can be reversed, and we can begin to look at what is going on outside of the organism, and the role this plays in scaffolding what our brains provide the substrate for. When we do this (see Lock and Colombo 1996, pp. 631-3), then the inherent power of the extrasomatic structures provided by particular human cultural practices becomes apparent. That is, certain modern practices, when conducted with apes as the conversational or interactive partners, can produce psychological alchemy: They transmute apes into something beyond what any ape has ever achieved on its own. This is clear evidence of the power of cultural practice. We get apes with what Premack (1980) has called 'upgraded minds'. What these human cultural practices can do with human biology is evidenced in the study of language ontogeny. The difference is that the ways in which we upgrade our children's minds have been created by us, ourselves, and were not handed down to us by some god-like being in any experimental situation (well, as far as we know).

Returning language development to the social

[The human infant] cannot, even theoretically, live an isolated existence;...he is not an independent individual. He lives a common life as one term in a personal relation. Only in the process of development does he learn to achieve a relative independence, and that only by appropriating the techniques of a rational social tradition (Macmurray, 1961, p. 50).

In the majority of studies of language acquisition to date, the focus has been predominantly on the individual infant as providing the skills that enable language development to occur. Here I want to redress this by turning to an emphasis on the social nature of these developments. The line I will be following is one that I have elaborated elsewhere (e.g., Lock 1980): that the communicative acts that infants are engaged in have only an implicit content to begin with, and that, at least in the early stages, development occurs as these contents are realised explicitly. This implicit context exists 'in'---although in a sense also 'beyond'---the ongoing world in which the infant is jointly situated interactively with another who is more skilled. What an infant pays attention to and finds of significance in this world is in large part constructed through the actions of that other. Development is thus a guided reinvention that is negotiated interactively. Some examples are useful here.

1. Here is a generalisation about infants that I have interactive data on (see Lock et al 1989; Lock 1991): At around 6 months of age the majority of things adults do in their interactions with infants serve to set up conditions that enable themselves and their infants to focus their attention on particular things and events in their immediate perceptual world. They do this by, for example, expressing exaggerated surprise at particular noises and events, by emphasizing objects as 'worth looking at' by making them interesting, by jiggling them up-and-down, pointing at them, putting them in piles, etc. The object of the exercise appears to be to co-ordinate the attention of both parties so that they could sustain a topic as a focus for interaction

Once routines are established that enable the infant to pay sustained attention, around 12-13 months, then adults shift to attempts to specify what is being attended to more precisely, and to being able to direct and re-direct the infant's attention more precisely. They do this by increasing their emphasis on the names of things as well as their properties, and by asking questions such as 'can you show me the x?', 'where's the y?'; etc. The suggestion is that some aspects of the infant's perceptual world are being made more perceptually salient as compared to others in this way, and this can act to provide the conditions whereby implications in the world are transformed into explicit components of it.

Once infants begin to 'take over' language from their caretakers, around 18 months, then the other shifts to elaborating on the properties and characteristics of the shared focus of attention, including those beyond the here-and-now, thus accreting further significance to the objects the child perceives. Some examples of this, taken from Lock et al. (1989), are concerned with the relationships that can exist between objects, that is, how objects can be regarded as parts of larger wholes (as reflected here by just the mother's speech from particular interactional episodes):

    (a)    Show it to Teddy.

    (b)    Put it on the saucer.

    (c)    Put that bug into the bucket.

    (d)    Are you going to get the snake to catch this bubble?

2. Further, this particular mother juggles objects and contexts in even more sophisticated ways, with the result that symbolic play is engineered:

    (e)    Would you like a cup of tea?

        Yes, go on, it's a nice little cup.

        You have some. (Child pretends to drink)

        Give it to Teddy.

        Go on, give Teddy a drink. (Child holds cup to Teddy's lips).

Here she gets the infant not merely to put two objects in conjunction within the present perceptual world, but to do so in a way which only makes sense through their relation to something that is not immediately present: the 'tea' is imaginary, and must be brought in from absent contexts to inform present action. It is quite a complicated activity: one object (the cup) is to be related to another (tea), which has to be lifted out of absent contexts (and lifted out of absent objects to be put in this one), and then both the cup and its imaginary contents have to be regarded as a whole and placed in relation to a third, and not just Teddy, but the right part of Teddy---his mouth.

My suggestion is that, by working in the infant's ever-shifting zone of proximal development, adults are acting to foreground the implicates of the infant's present abilities and thus acts to bootstrap their explication. At root, these interactions are ones in which a co-ordination is being attempted between two different Umwelten, and it is the nature of the attempts at co-ordination that act to bring the Umwelt of the developing infant more towards a commonality with that of the adult. The view being argued here is that the implications of the adult's knowledge of the world are constitutive of her Umwelt, such that she perceives, say, a book as a book, and this in turn implies for her a number of ways of going forward with the interaction. By contrast, a six-month-old infant sees the same object, but perceives it as something to suck, bang, etc., and not something to look at as a book. In her Umwelt, the same physical object is constituted differently from the way her mother constitutes it, both as a perceptual object and in terms of its implications for action.

The point I want to draw out here is that the contributing partners to such interactions are pursuing different projects because they bring different interests and abilities to bear on their supposedly shared arena of interaction. The notion of both interactants being attuned to a common, objective world is not useful here. Rather, the issue is

how the 'objective outside world' in which we observe ourselves and other living things arises out of our subjective (private) universe (our Umwelt), that is to say how the former [the objective] is derived from the latter [the subjective] as an abstraction (T. von Uexkull, 1982, p.10).

And this is achieved through, first, the co-ordination of attention, and its structuring, in interactive situations, and second, by negotiating an increasingly explicit symbolic marking of those co-ordinates. The fact is that, for humans at least, private universes are structured, from the outset, by their intersubjective situation in joint action.

Similarly, interaction can act to give a 'constitutional workout' to abilities that have previously been accorded solely an individual, cognitive status. An example is the ability to analyse constituents into their component parts. This skill has been highlighted as integral to the early development of symbolic abilities in the work of Bates et al. (e.g., 1979, 1988). They offer an explanation of the different styles of early language use, the 'expressive' and 'referential' as noted by Nelson (1973), on the basis of a differential reliance by individual children on either arbitrary or analysed learning skills:

the 'expressive' child tends to employ acquisition through perception of contiguity, imitating and using unanalysed phrases as means to ends prior to analysis of vehicle-referent relations. The 'referential' child is faster at analysis, so that the use of imitated whole forms is short-lived and infrequent. When he does imitate ... he may do so in parallel with and/or following a rapid breakdown into its components and the relationships into which that form can enter. ... Both arbitrary and analysed learning are necessary for rapid and efficient acquisition of language. Nevertheless, individuals may vary in the 'relative' use of one process over the other, and in the timing of analysis with imitation (Bates et al. 1979, p. 361).

The point is that if one wishes to attribute differential facilities with imitation vs. analysis to infants, then one should conceive of those skills as, at least partially, socially constituted and not individually given. These differential abilities are not based solely in individually-given cognitive operations, but are embedded in the perceptual world of interaction itself. Thus, while from one perspective one might choose, for some purposes, to conceive of infants as having differential amounts of analytic skills which they employ on the world, it is equally the case that they have come to live in different socially-constituted perceptual/conceptual worlds, some of which contain more perceptible parts than others (i.e., the examples of interaction given earlier are not typical of all dyads, and part-whole relations are not always so distinctively marked out as in the examples given (see Lock et al. 1989). In sum, it is not that infants develop differently by the exercise and practice of their individual skills; those skills, and the infants themselves, are constructed differently via socially-constitutive interaction (see Lock et al 1989, for further details).

The relevance of this developmental view for language evolution

The suggestion with respect to language evolution that I am putting forward here is that the initiating communication system used by anatomically modern humans was at least as complex as that found in current one- to two-year-olds and those apes that have been immersed in language programmes (and this is not a new suggestion; see, for example, Bickerton, 1990). However, it is not a straightforward task to take these extant exemplars as indicative of the functional characteristics of a primordial system. The major reason is that these modern exemplars function within an already-mature communicative environment. This adds the complication that there is a developmental inequality in place between the participants of contemporary communicative acts in which this possibly primordial system is encountered: the human adult participant actually possesses the fully-elaborated system. What this means in practice for the infant is that the immediately apparent benefits of possessing the primordial system are extended beyond the infant's mental horizon. This happens through the adult partner to the communicative act being able, in interaction, to explicitly 'point out' what it is the infant might mean. A concrete example is useful here in clarification.

At a very general level of analysis, the language abilities of one- to two-year-old infants have a quite limited spatial and temporal extension beyond the infant's immediate 'here-and-now' projects. I have previously described a conversational tactic used by some mothers of young infants as 'context hopping' (e.g., Lock 1991, p. 289). It is this tactic that, I am suggesting, provides the infant with the resources for exploiting the inherent properties of the words they are using to go beyond the here-and-now. An example of context hopping that occurred between a mother exploring a picture book with her 18-month-old son is entailed in the mother's following remark: 'Oh look...he's got boots on just like Daddy's'.

It seems unlikely that the infant, on his own, would bring Daddy to mind at this point without the word being spoken. Similarly, it is unlikely that the infant, unaided, would abstract out of the picture the similar features that link the pictured boots to Daddy's boots. But the conversational resources the mother is able to bring into their shared here-and-now can very effectively mediate the infant's perception for him. That is, she can 'see' similarities without much effort: what she directly sees in the picture is a pair of boots 'like Daddy's'. The shared words are amplified in their signification such that the features of similarity can be interactively foregrounded as a topic. It is this tactic, I am suggesting, that provides the infant with the resources for exploiting the inherent properties of the words they are using to go beyond the 'here-and-now'. What was otherwise implicit in the here-and-now is made explicit: similarly, what was a possible meaning of the indicated object boot is assisted in its transition to being an actual meaning. Adopting a different terminology, one of the participants in such conversations is able to act in the zone of proximal development of the other participant (Vygotsky 1962, 1978). Now, this is clearly not the case for two participants who have abilities at the same level.

For the issue at hand---the initial elaboration of language systems---the problem becomes this: How are participants who are at the same level of functioning with a communicative system able to go beyond its constraints? The essential claim being put forward here as a possible solution is two-fold. First, the functional elements of the presumed communicative system have immanent potentialities beyond those exploited by the early stages of their use. Second, under particular circumstances, the social characteristics of the communicative process can create their own zone of proximal development. These two facets of the communicative process at hand are being conceived here as the providers of bootstrapping conditions for the elaboration of new abilities which represent the explication of properties implicit in the potentialities of the initiating system. This re-orientation is such that we can get a handle on how cultural practices can, of themselves and sui generis, act as a 'zone of proximal development' for bootstrapping what can be thought about; that is, how implications are constructed as 'explications-in-waiting' that cognition can grab onto.

The same point emerges from the literature on creative problem solving, and how the chance of finding a successful solution can be 'primed' in the way the problem is structured. Thus, for example, the task of joining two hanging ropes that are simultaneously beyond the grasp of an individual with the assistance of a hammer is facilitated if the problem solver sees a pendulum operating beforehand. This alerts the solver to the possibility of using the hammer as a pendulum weight, tying it to one rope so as to set it in a swinging motion that brings it in reach when the other rope is being held. That such possibilities are difficult to grasp without this kind of predisposing zone of proximal development is captured in Piaget's observation on how Darwin developed the concept of evolution:

the two results which seem the most interesting to me are, first, the time that Darwin needed to become aware of ideas which were already implicit in his thought, and, second, the mysterious passage from the implicit to the explicit in the creation of new ideas ... One might have believed that this passage concerned only the relationship between thought and action, and that, on the level of thought itself, the passage from 'implicit' schemas ... to their reflective explication would be much more rapid. [But] ... even in a creator of the greatness of Darwin the passage is far from immediate. This delay establishes ... that making things explicit leads to the construction of a structure which is partially new, even though contained virtually in those structures which preceded it (Piaget, 1974: x).

The role of a zone of proximal development is to enable what Piaget calls here 'reflective explication'.

At this point, recall that we know something of the power of modern social practices in reconfiguring the demonstrable cognitive abilities of modern apes. We see in the current world the analogous reconfiguration of the cognitive abilities of morphologically modern infants, through their transformation and amplification by the 'cognitive technologies' constituted by symbolically-based communicative systems. A paradigm example that elucidates the interplay of social relations with language in the constitution of abilities is provided, at first sight tangentially, in Elias's outline (e.g., 1978, 1982) of the establishment of modern western practices of politeness.

Elias's account

One interpersonal activity that Elias elucidates concerns blowing one's nose. Elias establishes the historical course of elaborating western practices for dealing with the assorted accumulations of material that periodically inhabit the human nasal passages. At the root of the changes Elias documents in what is considered polite is a hierarchy of actions: blowing the nose; hiding the blowing of it by using a handkerchief; hiding the blowing of it into a handkerchief. But, most importantly, embarrassment is being invented. Embarrassment thus comes to be seen as an emotional state created by the explication into discourse of this hierarchy: For it to be realized, a self-censorious ability has to be established. People have to become able to reflect on their own behaviour---that is, on how they act in company---where previously they had not done so.

In Elias's view, this kind of change in interpersonal behaviour is not just one of fashion; further,

it does not involve solely changes of 'knowledge' or transformations of 'ideologies', in short, alterations of the content of consciousness, but changes in the whole human make-up, within which ideas and habits of thought are only a single sector (Elias 1982, p. 284, emphasis added).

These changes reflect a reorganization and transformation of

the whole personality throughout all its zones, from the steering of the individual by himself at the more flexible level of consciousness and reflection to that at the more automatic and rigid level of drives and affects (Elias 1982, p. 284).

Elias considers that it is the relationship between the psychological functions controlling an individual's actions that changes during historical time; that it is

these relationships within man between the drives and affects controlled and the built-in controlling agencies, whose structure changes in the course of a civilizing process, in accordance with the changing structure of the relationships between individual human beings, in society at large (Elias 1982, p. 286).

Interaction, discourse and cognition

How are these changes in the basic structuring of Western psychological structure related to the issues at hand here?

First, we need to remember that the changes in conduct that Elias charts are occurring against a background of a diversifying society. Travel was becoming easier, leading to more and different people meeting each other. Cities were emerging in a contrastive role to that of the countryside, resulting in trade and trade specialization, the establishment of a merchant class, and so on. In general, people were becoming less socially homogeneous, roles were becoming specialized, and so people were sharing less common knowledge among themselves; they had fewer common presuppositions. This had obvious repercussions on the process of communication between fellows. Essentially, it became much harder to make oneself understood. It is likely that this same process is at the root of Armstrong and Katz's (1981) finding of a correlation between the number of basic colour terms in a society's language and its social complexity, and work by Fischer (1966, 1973) on the different structuring of two closely-related Micronesian languages in relation to the social structures of the cultures that use them. In both cases, the more differentiated the social structure, the more differentiated the referential domains marked by words in the language. Thus, below, in outlining Elias's views with respect to selves and personalities, we may take his specific focus as indicative of changes occurring in other referential domains, changes that are related to changes in social organization. Recall, this is one of the major changes occurring at the point at which the human record begins to show evidence of symbol use.

Second, we need to think about the consequences of these changes with respect to the sort of knowledge a person would need in order to act effectively in such a changing world. Because it was much more difficult to get a message across to another, communication failure could occur much more frequently, making the individual aware of the communication process itself. Further, it would provide the individual with many more perspectives on the presentation of his or her self.

Third, language began to code new concepts and to be used more explicitly. On the one hand, the loss of presuppositionality in discourse will force an increasingly elaborated and explicit linguistic coding of communication: People would have been required to make explicit information they had previously left implicit. On the other hand, an increasingly complex society can create all sorts of new situations and experiences among the people whose actions bring it into being: these may come to be expressed in language. Hence, information concerning the presentation of the self would have been available directly to an individual. A society with differentiated roles might well force an awareness upon an individual of the aspects of individuality that those roles are simultaneously constructing. It may eventually provide linguistic concepts for rendering these explicitly. The main point here is that the socially constructed facets of personality will increase, through the explicit realisation in discourse, of the number of different perspectives an individual can formulate of his or her self, while at the same time increasing the individual's ability to transcend his or her presupposed, unreflective, non-meta-awareness of his or her self through a richer fabric of communication and concept.

Fourth, Luria and Yudovich (1971) have noted that words have a profound affect on individual psychological functioning:

When he acquires a word, which isolates a particular thing and serves as a signal to a particular action, the child, as he carries out an adult's verbal instruction, is subordinated to this word.... By subordinating himself to the adult's verbal orders the child acquires a system of these verbal instructions and gradually begins to utilize them for the regulation of his own behaviour (pp. 13-4).

In this view, through the medium of language, one of the major transitions noted by Elias---from external control of behaviour by threat of punishment to internal control via the self-reflexive censors of shame, guilt and embarrassment---can be effected; and through these particularly social forms of emotional cognition 'people become ... sensitive to distinctions which previously scarcely entered consciousness' (Elias, 1982: 298).

The point is that our personalities, our conceptions of ourselves and others, our emotional experiences, and our views of the world are all explicated from phenomena whose existence is created beyond us in our social worlds. In this view, human conceptual systems are explicated renderings into mental form of human social discourses. These explications have become, initially, salient aspects of the human Umwelt through changes in social relations. They become preserved in cultural practices and the linguistic resources that enable the discourses that sustain these practices. Thus, they are able to act as a zone of proximal development sui generis. And:

I know of no way in which intelligence or mind could arise or could have arisen, other than through the internalization by the individual of social processes of experience and behaviour .... And if mind or thought has arisen in this way, then there neither can be nor could have been any mind or thought without language; and the early stages of the development of language must have been prior to the development of mind and thought (Mead 1934, pp. 191-2).

Back to evolution

The prospective summary, then, is that the process of explicating the implicit possibilities of cognitive endowments is contributed to by the potentialities of the various symbol systems that humans use, as well as the nature and structure of the human practices within which these are sustained. It is becoming clearer that (1) the properties of particular symbol systems and (2) the conditions under which they are employed affect the ease with which humans can use them for particular purposes. Thus changes in social relations have provided the major resources for the establishment of modern abilities, and these are not solely founded on preadaptive changes in human biology. The structure of relations between people and the properties of the discourses and practices they engaged in constituted the conditions that enabled their discovery of the unintended properties of the systems of symbols, social relations, and discourse practices their modes of life constituted. It is these discoveries that underlie the evolution of human linguistic abilities: these discoveries, not the biological kit alone.

The position argued for here is that interaction constructs contexts that language can come to symbolize, thereby providing a cognitive technology that bootstraps the increasing discovery of those 'things' that are implied by what has already been symbolised (e.g., Donald 1991, 1993; Lock and Peters 1996; Olson 1996). The crucial missing link in the pre-Upper Palaeolithic was the lack of the cultural support system of socially-constituted 'pregnant implications' in the perceptual Umwelten of these ancestral humans, such that there was an insufficient basis to exaptively co-opt those biologically-based functions that have come to be regarded as the 'Language Acquisition Device'. There was no naturally constituted 'zone of proximal development' to bootstrap the explicit construction of a symbolic lexicon that benefited from a neurology that predisposed, on the basis of its functional architecture, the duality of patterning found in modern language systems, from the phonetic through to the syntactic. The evidence for the increasing elaboration and complexity of social life in the Upper Palaeolithic is compelling (see, for example, Gamble 1993, pp. 179-202, for a recent review, and earlier this chapter). Rather, the process of elaborating the symbolic support for, and amplification of, cognitive abilities, is embedded in a nexus of influences that feed forward and back to each other. These hypothesised links are portrayed in Fig. 5.

The scenario suggested here is thus close to that argued for by Bickerton (1990) and Whallon (1989, drawing on Bickerton 1981). It would credit hominids from 40 000 or so years ago back to at least mid-H. erectus specimens represented by KNM-ER 3733 at around 700 000 years ago with a vocal communication system that Bickerton characterises as protolanguage, a 'mode of linguistic expression that is quite separate from normal human language and is shared by four cases of speakers: trained apes, children under two; adults who have been deprived of language in their early years, and speakers of pidgin' (Bickerton 1990, p. 122). In Whallon's opinion, such protolanguages would facilitate communication, but could well lack the means for explicitly expressing and reflecting on the temporally and spatially absent world, the imaginary world enabled at the truly explicit symbolic level, that is required for elaborating and sustaining the characters of modernity. Thus, communication existed, as did the biological base for its elaboration, and 'the social context provides the exaptive process to upgrade this to language. The link between, on the one hand, exploration and the use and knowledge of space, and on the other the scale, depth and complexity of social life is irrefutable' (Gamble 1993, p. 174).

What that social context does is as follows:

1. It provides for the structuring of attention by embedding objects within cultural practices such that they became different objects in our perception, making available new implications and courses for action directly. For example, given the similarity in the visual systems of chimpanzees and present-day western humans, it seems a reasonable assumption that both species 'see' the same physical object when they look at a flat surface supported on 'legs'. But the two species 'perceive' different objects: humans perceive a 'table', and know what and what not to do with it; chimpanzees perceive something to walk and sit on, not a table.

2. It conserves previously elaborated ways of keeping track of implications that have been made explicit: that is, particular symbol systems that have been arrived at. These systems can variously constrain modes of thinking. As an example of what I mean by this, I earlier indicated an analogy with the differences between the Roman and Arabic number systems. The mental operations of multiplication and division are available to users of both systems, but the symbolic tool-kits affect the ease with which these operations can be performed: it is easier to divide 63 by 9 to get 7 than LXIII by IX to get VII. It is even easier to do it on an abacus, but here we are at a further remove in that a technology for manipulating symbols has been introduced. In addition the device---the abacus---can be 'internalised' and used as fast and accurately as using it physically, and often faster. In addition, skilled abacus users show very specific enhancements to their mathematical skills: digit memory---forward and backward---increases to 15 digit strings, while for other item lists it remains around the magic number 7 +/- 2 (Hatano 1982; Stigler et al. 1982).

3. It conserves 'patterns for handling' implications that have been made explicit: what we term 'grammar'. Bickerton (1990), for example, has pointed out that:

Syntax is, to a large extent, a projection of the lexicon, wherein all the sub-categorization frames of verbs and all the grammatical items (and much else) are stored. If the initial lexicon was dramatically limited, then not all of the structures implicit in syntactic principles could immediately be realized. To take an extreme example, if there were no verbs of reporting, and no 'psychological' or 'intentional' verbs, then biclausal structures of the type he said/thought/believed/ hoped that X had happened/would happen would be impossible to construct ... (pp. 180-1)

In other words, biologically-instantiated possible abilities require a substrate for their actualisation. And the emphasis in this chapter shifts that substrate from being some abstract object that is analysed by some 'Language Acquisition Device', as the situation is captured from a cognitive perspective, to one in which the learner is the generator of, and participator in, shared acts of 'making sense'. That is, forms of social being are in process of elaboration, and it is within these social practices and the child's Umwelt that this substrate needs to be located.

Two empirical examples are useful here. First, in a microanalytic study of the emerging use of self words, Budwig (1995) finds a clear relation between reference, grammatical form and caregiver's language for children in the age-range 20-32 months in their use of words such as I, me and my in American English, and ich and man in German. The use of these words by the children studied was tied in with the notions of self and agency that were established in the course of interacting with caretakers, indicating that the children were picking up subtle cues about agency, and thereby articulating different forms of selfhood that were then reflected in the grammatical uses of self referential words. The caretaker-infant interaction is thus revealed as functioning to delineate a 'problem space' in which 'referential objects'---forms of selfhood and agency---are constructed simultaneously with the vocabulary items that allow these objects to be structurally marked in the child's emerging speech.

Similarly, Bamberg (1996), drawing on Capps and Ochs (1995a, 1995b), takes this tack further in his study of emotional socialisation with respect to children's discourses about anger and sadness. What he terms the 'grammars' of being angry and sad, in talking from both the first-person and third-person perspectives, are quite complex in English, and work so as to disentangle the different discursive requirements of handling empathy and blame. Bamberg sets his particular findings into a general framework that sees the development of a linguistic ability to handle 'emotion talk' as 'a process of appropriating the tools necessary to talk meaningfully about the social relationships in which emotions are embedded'. Thus, grammar is set up as 'a set of linguistic constructions ... directly tied to the discursive purpose to which the particular constructions are put to use':

grammar, if understood correctly, i.e. not as abstract principles of a universalist nature, but as social know-how relevant for the construction of social meaning, plays an integral role in coming to grips with what emotions do and what they are used for in social communicative practices. As such, learning to use the linguistic construction procedures for socially appropriate purposes is deeply embedded in cultural practices (Bamberg 1996, p. 22 ).

In this view, again, the social and the cognitive are intimately intertwined, and grammar becomes a system used in the service of keeping track of the explications of what otherwise remains implicit in the child's Umwelt.

Now, we need to remember that, developmentally, the grammatical means for differentially handling anger and sadness talk are pregiven in the child's language community---'they have their social existence before they are put to use in social practices' (Bamberg 1996, p. 22 ). Phylogenetically, this is not the case. But the point here is that the close relation between social practice and grammatical device established by Bamberg is one where the latter are instantiated in the former. That is, in line with Bickerton's point earlier about the relation between lexical entries and syntax---that the latter is wheeled on to the mental stage to meet the needs of the former---the driving force is the social structure of obligations captured constitutively by words. These words come with a particular set of implications that are unpacked and pointed to by a set of grammatical structures, perhaps structures that are new, but more likely reworked in a bricolage fashion from something older. Thus, as social structures become more complicated, vocabularies and grammars will change. Vocabularies will get larger as there are more things to talk about. Grammars should become more complex in an attempt to signpost explicitly the more complex implications of having, at least, more things to talk about.

Rolfe (1996) has put forward a detailed argument as to how this elaboration is also motivated by the demands of different communicative purposes, thus leading to the bootstrapping of a propositional grammatical system from earlier structural principles. Similarly, the point is reinforced by Bloom's studies (e.g., 1981) on the differential ease of solving problems that are expressed through grammatical devices that are more or less familiar as tactics to speakers of different languages. Like literacy (e.g., Scribner and Cole 1981), neither grammar, symbols, nor social structures are magic bullets, but are predisposers of certain outcomes in interaction with the cultural practices they constitute and support.


...it is precisely the conceptualizing of existing social relations and the formulating of them as symbolic 'tools', e.g. words, that mark the true transition from prehuman ancestor to man [sic]. It was when our ancestors began to gain awareness of the relationships they already recognized in terms of differential behaviour that man emerged (Reynolds 1976, p. 64).

In putting forward the above formulation, the focus on the late emergence of language under the push coming from the changes in social relations that provide, sui generis, a zone of proximal development, a different way of looking at the biology of human evolution is constructed. Language is now a species characteristic, but it is not one that was initially selected for. The biological substrates that support it have not been arrived at through their having evolved, but are constructed and reconstructed through an ontogeny that is embedded in a set of social relations that continuously bootstrap themselves and the resources they conserve, thereby enabling the exaptation or co-option of an existing biology to ever more complex cognitive technologies. But this is not meant to completely de-biologize our views of human nature altogether. Certainly I am socialising human nature, but it is also certainly the case that many aspects of those social relations that contribute to and enable this humanising process have their bases in biological constraints. Examples that come to mind here concern incest taboos, mate choice (Buss 1994), and social problem solving (Cosmides 1989). It is precisely such phenomena that we might expect evolutionary processes to constrain and select biases for. But we were not designed for language use.

The consequences of this view are many. Here I draw out two points. First, the study of human discourse practices is moved from the fringe of psychological science to its centre, for the processes fundamental to the project of the so-called 'cognitive revolution' are no longer encapsulated within the head of an individual, but distributed in the symbolically-mediated practices that comprise human cultures, distributed between the individual and the social. Bruner (1990) has also argued for this view, though from a different starting point. He reflects analytically on the cognitive science that has been spawned by the cognitive revolution that he played such a role in forming. He deals with the consequences of what he sees as a 'neglect for meaning' in this cognitive science.

There is no question that cognitive science has made a contribution to our understanding of how information is moved about and processed. Nor can there be much doubt on reflection that it has left largely unexplained and even somewhat obscured the very large issues that inspired the cognitive revolution in the first place. So let us return to the question of how to construct a mental science around the concept of meaning and the processes by which meanings are created and negotiated within a community (p. 10).

He accomplishes in this work what Harré (1992) has called 'the second cognitive revolution'. Taking this line does not mean a retreat into subjectivism; an 'Umwelt approach' (Harré 1990) is computable (Bechtel 1993; see also Bechtel and Abrahamsen 1991). Nor does it mean attempting the impossible of comprehending what the world might look like to a bat, for example (Nagel 1974). What it does mean is that a set of approaches to human psychology that have often been argued by their own proponents to be antithetical to a cognitivist stance, and that have been marginalised by the mainstream of cognitive science, are central to the project.

Second, the analysis offered here credits discourse with tremendous powers. Discourse practices are being credited with a central role in the creation of the psychological abilities that underpin discursive practices themselves. Cognitive mechanisms, if they exist, are constructed out of an interaction between developmental, maturational, and socio-cultural processes, a constructive process in which the driving engine is a consequence of the changing structures of discourse and social relations. That is, it is no longer sensible to assume a biologically-given individual with a set of pre-given cognitive abilities that allow information 'to be moved about and processed', nor, for that matter, to use this assumption to legitimate the currently dominant psychological paradigms. Meaning-full second-generation cognitive science needs to pursue the Umwelt project; and in doing so it needs to bring a whole new paradigm of work into the mainstream.


        Arensburg, B., Tillier, A. M., Vandermeersch, B., Duday, H., Schepartz, L., and Rak, Y. (1989). A Middle Palaeolithic human hyoid bone. Nature, 338, 758-60.

        Armstrong, D. F. and Katz, S. H. (1981). Brain laterality in signed and spoken language: a synthetic theory of language use. Sign Language Studies, 33, 319-50.

        Bamberg, M. (1996). Language, concepts, and emotions: The role of language in the construction of emotions. Visiting Scholar Series 6. Department of Psychology, Massey University, Palmerston North, New Zealand.

        Bates, E., Benigni, L., Bretherton, I., Camaioni, L., and Volterra, V. (1979). The emergence of symbols: Cognition and communication in infancy. Academic Press, New York.

        Bates, E., Bretherton, I and Snyder, L. (1988). From first words to grammar. Cambridge University Press.

Bechtel, W. (1993) The case for connectionism. Philosophical Studies 71: 119-54.

        Bechtel, W. and Abrahamsen, A. (1991) Connectionism and the mind: An introduction to parallel processing in networks. Oxford: Blackwell.

        Bickerton, D. (1981). Roots of language. Karoma, Ann Arbor, MI.

        Bickerton, D. (1990). Language and species. University of Chicago Press.

        Binford, L. R. (1985). Human ancestors: Changing views of their behavior. Journal of Anthropological Archaeology, 4, 292-327.

        Bloom, A. H. (1981). The linguistic shaping of thought: A study in the impact of language on thinking in China and the West. Erlbaum, Hillsdale, NJ.

        Bradley, B. and Sampson, C. G. (1986). Analysis by replication of two Acheulian artefact assemblages. In Stone age prehistory (ed. G. N. Bailey and P. Callow) , pp. 29-45. Cambridge University Press.

        Bruner, J. S. (1990). Acts of meaning. Harvard University Press, Cambridge, MA.

        Budwig, N. (1995). A developmental-functionalist approach to child language. Erlbaum, Hillsdale, NJ

        Buss, D. M. (1994). The evolution of desire: Strategies of human mating. Basic Books, New York.

        Campbell, B. G. (1996). An outline of human phylogeny. In Handbook of human symbolic evolution (ed. A. J. Lock and C. R. Peters), pp. 31-52. Clarendon Press, Oxford

        Capps, L. and Ochs, E. (1995a). Constructing panic: The discourse of agoraphobia. Harvard University Press, Cambridge, MA.

        Capps, L. and Ochs, E. (1995b). Out of place: Narrative insights into agoraphobia. Discourse Processes, 19, 407-39.

        Chase, P. and Dibble, H. (1987). Middle Paleolithic symbolism: A review of current evidence and interpretations. Journal of Anthropological Archaeology, 6, 263-96.

        Chomsky, N. (1968). Language and mind. Harcourt, New York.

        Conkey, M. W. (1996). A history of the interpretation of European 'palaeolithic art': magic, mythogram, and metaphors for modernity. In Handbook of human symbolic evolution (ed. A. J. Lock and C. R. Peters), pp. 288-349. Clarendon Press, Oxford.

        Cosmides, L. (1989).The logic of social exchange: Has natural selection shaped how humans reason? Studies with the Wason reasoning task. Cognition, 31, 187-296.

        Davidson, I. and Noble, W. (1989). The archaeology of perception. Current Anthropology, 30, 125-55.

        Davidson, I and Noble, W. (1992) Why the first colonisation of the Australian region is the earliest evidence of modern human behaviour. Archaeology in Oceania, 27, 135-42.

        Davidson, I and Noble, W. (1993) Tools and language in human evolution. In Tools, language, and cognition in human evolution (ed. K. R. Gibson and T. Ingold), pp. 363-388. Cambridge University Press.

        Dennell, R. (1983). European economic prehistory: A new approach. Academic Press, London.

        de Lumley, H. (1975). Cultural evolution in France in its paleoecological setting during the Middle Pleistocene. In After the australopithecines (ed. K. Butzer and G. Isaac), pp. 745-808 Mouton, The Hague.

        Donald, M. (1991). Origins of the modern mind: Three stages in the evolution of culture and cognition. Harvard University Press, Cambridge, MA.

        Donald, M. (1993). Precis of Origins of the modern mind: Three stages in the evolution of culture and cognition. Behavioral and Brain Sciences, 16, 737-91.

        Elias, N. (1978). The civilizing process: Vol I: The history of manners. Blackwell, Oxford.

        Elias, N. (1982). The civilizing process: Vol II: Power and civility. Blackwell, Oxford.

        Ettlinger, G. and Wilson, W. A. (1990). Cross-modal performance: Behavioral processes, phylogenetic considerations and neural mechanisms. Behavioral Brain Research, 40, 169-92.

        Fagan, B. (1989). People of the earth: An introduction to world prehistory (6th edition). : Scott, Foresman and Co., Glenview, IL.

        Fischer, J. L. (1966). Syntax and social structure. Truk and Ponape. In Sociolinguistics: Proceedings of the UCLA sociolinguistics conference, 1964 (ed. W. Bright), pp. 168-87. Mouton, The Hague.

        Fischer, J. L. (1973). Communication in primitive systems. In Handbook of communication (ed. W. Schramm, I de Sola Pool, N. Maccoby, E. B. Parker and F. W. Frey), pp. 313-36. Rand McNally, Chicago.

        Freeman, L. (1975). Acheulian sites and stratigraphy in Iberia and the Maghreb. In After the australopithecines (ed. K. Butzer and G. Isaac), pp. 661-744. Mouton, The Hague.

        Gamble, C. (1986). The Palaeolithic settlement of Europe. Cambridge University Press.

        Gamble, C. (1993). Timewalkers: The prehistory of global colonization. Penguin, London.

Goody, J. (1968) Literacy in traditional societies. Cambridge University Press.

Gould, S. J. (1991). Exaptation. Journal of Social Issues, 47, 43-65.

        Halverson, J. (1993). Mythos and logos. Behavioral and Brain Sciences, 16, 762.

        Harre, R. (1990). Exploring the human Umwelt. In R. Bhaskar (ed.) Harre and his critics: Essays in honour of Rom Harre with his commentary on them. Oxford: Blackwell.

        Harre, R. (1992). Introduction: The second cognitive revolution. American Behavioral Scientist, 36, 5-7.

        Hatano, G. (1982). Cognitive consequences of practice in culture specific procedural skills. Quarterly Newsletter of the Laboratory of Comparative Human Cognition, 4, 15-7.

        Holloway, R. (1969). Culture: A human domain. Current Anthropology, 10, 395-412.

        Holloway, R. (1996). Evolution of the human brain. In Handbook of human symbolic evolution (ed. A. J. Lock and C. R. Peters), pp. 74-108. Clarendon Press, Oxford.

        Hublin, J. J. (1985). Human fossils from the north African Middle Pleistocene and the origin of Homo sapiens. In Ancestors: the hard evidence (ed. E. Delson), pp.283-8. Alan Liss, New York.

        Keller, C. M. (1973). Montagu Cave in prehistory. University of California Press, Berkeley, CA.

        Laitman, J. and Heimbuch, R. C. (1984). The basicranium and upper respiratory system of African Homo erectus and early H. sapiens. American Journal of Physical Anthropology, 63, 180.

        Laitman, J. and Reidenberg, J. S. (1988). Advances in understanding the relationship between the skull base and larynx with comments on the origins of speech. Human Evolution, 3, 99-109.

        Laitman, J., Redenberg, J. S., Friedland, D. R., Reidenberg, B. E, and Gannon, P. J. (1992). Neandertal upper respiratory specializations and their effect upon respiration and speech. American Journal of Physical Anthropology, 6, 129.

        Lieberman, P. (1991). Uniquely human: The evolution of speech, thought, and selfless behavior. Harvard University Press, Cambridge, MA.

        Lindly, J. M. and Clark, G. A. (1990). Symbolism and modern human origins. Current Anthropology, 31, 233-261.

        Lock, A. J. (1980). The guided reinvention of language. Academic Press, London.

        Lock, A. J. (1991). The role of social interaction in early language development. In Biological and behavioral determinants of language development (ed. N. Krasnegor, D. Rumbaugh, R. Schiefelbusch, and M. Studdert-Kennedy), pp. 287-300. Erlbaum, Hillsdale, NJ:

        Lock, A. J. and Colombo, M. (1996). Cognitive abilities in a comparative perpective. In Handbook of human symbolic evolution (ed. A. J. Lock and C. R. Peters), pp. 596-643. Clarendon Press, Oxford:

        Lock, A. J. and Peters, C. R. (1996). Ontogeny: symbolic development and symbolic evolution. In Handbook of human symbolic evolution (ed. A. J. Lock and C. R. Peters), pp. 371-99. Clarendon Press, Oxford

        Lock, A. J. And Symes, K. (1996). Social relations, communication, and cognition. In Handbook of human symbolic evolution (ed. A. J. Lock and C. R. Peters), pp. 204-35. Clarendon Press, Oxford.

        Lock, A. J., Service, V., Brito, A. and Chandler, P. (1989). The social structuring of infant cognition. In Infant development (ed. G. Bremner and A. Slater), pp. Erlbaum, London.

        Lucy, J. A. (1992). Linguistic diversity and thought: A reformulation of the linguistic relativity hypothesis. Cambridge University Press.

        Luria, A. R. and Yudovich, F. A. (1971). Speech and the development of mental processes in the child. Penguin, London.

Macmurray, J. (1961). Persons in relation. Faber and Faber, London.

        Mead, G. H. (1934). Mind, self and society. Chicago University Press.

Nagel, T. (1974)

        Nelson, K. (1973). Structure and strategy in learning to talk. Monographs of the Society for Research in Child Development, 38 (1 and 2), Serial No. 149.

        Noble, W. and Davidson, I. (1996). Human evolution, language and mind. Cambridge University Press

        Olson, D. R. (1996). Towards a psychology of literacy: on the relations between speech and writing. Cognition, 60, 83-104.

        Peters, C. R. (1996) Tempo and mode of change in the evolution of symbolism (a partial overview). In A. J. Lock and C. R. Peters (eds.) Handbook of Human Symbolic Evolution. Oxford: Clarendon Press

        Piaget, J. (1974). Foreword. In Darwin on man: A psychological study of creativity (au. H. E. Gruber) [pages?] E. P. Dutton, New York.

Pinker, S. (1994). The language instinct. Morrow, New York.

        Popper, K. (1972) Objective knowledge: An evolutionary approach. Oxford: Clarendon Press.

        Potts, R. (1989). Olorgesailie: new excavations and findings in Early and Middle Pleistocene contexts, southern Kenya rift valley. Journal of Human Evolution, 18, 477-84.

        Premack, D. (1980). Characteristics of an upgraded mind In Bericht uber den 32. Kongress der Deutschen Gesellschaft fur Psychologie in Zurich, Vol. 1 (ed. W. Michaelis), pp. 49-70.

Reynolds, V. (1976) The biology of human action. San Fransisco: W.H. Freeman.

                Ristau, C. (1996) Animal language and cognition projects. In Handbook of human symbolic evolution (ed. A. J. Lock and C. R. Peters), pp. 644-85. Clarendon Press, Oxford.

        Rolfe, L. (1996). Theoretical stages in the prehistory of grammar. In Handbook of human symbolic evolution (ed. A. J. Lock and C. R. Peters), pp. 776-792. Clarendon Press, Oxford.

        Russell, J. (in press). Development and evolution of the symbolic function: The role of working memory. In Modelling the early human mind (ed. P. A. Mellars and K. Gibson) . McDonald Institute for Archaeological Research, Cambridge.

        Scribner, S. and Cole, M. (1981). The psychology of literacy. Harvard University Press, Cambridge, MA.

        Shipman, P. and Rose, J. (1983). Evidence of butchery and hominid activities at Torralba and Ambrona: An evaluation using microscopic techniques. Journal of Archaeological Science, 10, 465-74.

        Stigler, J. W., Barclay, C., and Aiello, P. (1982). Motor and mental abacus skills: a preliminary look. Quarterly Newsletter of the Laboratory of Comparative Human Cognition, 4, 12-4.

        Stringer, C. (1989). The origin of early modern humans: A comparison of the European and non-European evidence. In The human revolution: Behavioural and biological perspectives on the origins of modern humans (ed. P. Mellars and C. Stringer), pp. Princeton University Press.

        Thieme, H. (1997) Lower Palaeolithic hunting spears from Germany. Nature 385 807-10.

        Tobias, P. V. (1987). The brain of Homo habilis. Journal of Human Evolution, 16, 741-61.

        Tobias, P. V. (1991). Olduvai Gorge. Cambridge University Press.

        Tomasello, M., Savage-Rumbaugh, E. S., and Kruger, A. (1994). Imitative learning of actions on objects by children, chimpanzees, and enculturated chimpanzees. Child Development, 64, 1688-705.

        Toth, N. and Schick, K. (1993). Early stone industries and inferences regarding language and cognition. In Tools, language, and cognition in human evolution (ed. K. R. Gibson and T. Ingold) [pages?] Cambridge University Press.

                Volosinov, V. N. (1973). Marxism and the philosophy of language. Harvard University Press, Cambridge, MA.

        von Uexkull, T. (1982). Introduction: Meaning and science in Jacob von Uexkull's concept of biology. Semiotica, 42, 1-24.

        Vygotsky, L. S. (1962). Thought and language. MIT Press, Cambridge, MA.

        Vygotsky, L. S. (1978). Mind in society.Harvard University Press, Cambridge, MA.

        Waddell, P. J. and Penny, D. (1996). Evolutionary trees of apes and humans from DNA sequences. In Handbook of human symbolic evolution (ed. A. J. Lock and C. R. Peters), pp. 53-73. Clarendon Press, Oxford.

        Whallon, R. (1989). Elements of cultural change in the later Palaeolithic. In The human revolution: Behavioural and biological perspectives on the origins of modern humans (ed. P. Mellars and C. Stringer), 433-54. Edinburgh University Press.

        White, R. (1985). Thoughts on social relationships and language in hominid evolution. Journal of Social and Personal Relationships, 2, 95-115.

        White, R. (1996) On the evolution of human socio-cultural patterns. In Handbook of human symbolic evolution (ed. A. J. Lock and C. R. Peters), 239-62. Clarendon Press, Oxford.

        Wilkins, W. K. and Wakefield, J. (1995). Brain evolution and neurolinguistic preconditions. Behavioral and Brain Sciences, 18, 161-226.

        Wilson, P. J. (1988). The domestication of the human species. Yale University Press, New Haven.

        Wood, B. (1992). Origin and evolution of the genus Homo. Nature, 355, 783-90.

        Wynn, T. (1979). The intelligence of later Acheulean hominids. Man, 14, 379-91.

        Wynn, T. (1996). The evolution of tools and symbolic behaviour. In Handbook of human symbolic evolution (ed. A. J. Lock and C. R. Peters), pp. 263-87. Clarendon Press, Oxford.

        Yellen, J. E., Brooks, A. S., Cornelissen, E., Mehlman, M. J., and Stewart, K. (1995). A Middle Stone Age bone industry from Katanda, Upper Semliki Valley, Zaire. Science, 268, 553-6.