10 So how did it happen?

IN AN EARLIER chapter I suggested that we should view the past as if it were a drama. The interest in such a play is not so much the action, but what is going on in the minds of the actors when various events occur and actions are undertaken. I have concluded that the diverse range of new behaviours that appear in Act 4 of the play derive from a fundamental change in mental architecture. Thoughts and knowledge which had been previously trapped within chapels of specialized intelligence could now flow freely around the cathedral of the mind – or at least a part of it – harmonizing with each other to create new types of thoughts as part of an almost limitless imagination: a cognitively fluid mentality.

Explaining the rise of the flexible mind

My argument remains incomplete, because I have yet to explain how the new cognitive fluidity arose. I believe the explanation relates to changes in the nature of language and consciousness within the mind. Let me start my explanation with a simple proposition: once Early Humans started talking, they just couldn’t stop.

To understand how this led to cognitive fluidity we must first recall that in previous chapters I have followed the proposals of Robin Dunbar that the language of Early Humans was a ‘social language’ – they used language as a means to send and receive social information. This contrasts with our language today which is a general-purpose language, playing a critical role in the transmission of information about the non-social world, although a social bias remains. Now although the language of Early Humans can be characterized as a social language – and for the Early Humans after 250,000 years ago, as a language with an extensive lexicon and grammatical complexity – I believe there would nevertheless have been ‘snippets’ of language about the non-social world, such as about animal behaviour and toolmaking.

These would have arisen from two sources. The first is general intelligence. As I argued in Chapter 7, general intelligence was extremely important in the Early Human mind as it conditioned behaviour at the domain inter faces, such as the use of tools for hunting and use of food for establishing social relationships. As a result, behaviour at these domain interfaces remained extremely simple, because general intelligence could not access the cognitive processes located within each of the specialized intelligences. General intelligence is also likely to have enabled Early Humans to associate particular vocalizations with non-social entities and consequently produced ‘snippets of conversation’ about the non-social world – which would have been few in number and lacking in grammatical complexity. Indeed these snippets are likely to have been similar in complexity to the use of symbols by chimpanzees when trained in laboratories which, as I argued in Chapter 5, arises simply from possessing a general intelligence, rather than any linguistic capacity. The non-social ‘language’ of Early Humans may thus have amounted to a small range of ‘words’, used predominantly as demands, and with no more than two or three being strung together in a single utterance. They would have contrasted with the grammatically complex and diverse flow of utterances relating to the social world produced by Early Humans arising from their specialized social and linguistic intelligences. Yet the non-social vocalizations may have been embedded within this social language.

A second way for snippets of non-social conversation to arise may have been that the specialized intelligences were never totally isolated from each other, although the degree of isolation was sufficient to prevent them working together. I gave an example of this in Chapter 8 when I suggested that although Neanderthals may have lacked reflexive consciousness about their toolmaking and foraging activities, they may have had a fleeting, ephemeral, rolling consciousness about these – a ‘snippet of consciousness’, insufficient to have provided any introspection about their thoughts and knowledge in these domains. I explained why this may have been the case by using my analogy of the mind as a cathedral. The ‘sounds’ of reflexive consciousness at work may have seeped through the chapel walls of social intelligence, and then seeped into the chapels of technical and natural history intelligence, arriving in a heavily muffled or watered down form. I gave another example in Chapter 7 when I noted that in those very rare instances when Early Humans did work bone, they chipped it as if it were stone. This implies that if technical intelligence was indeed being used, it was not working effectively, since chipping is an inappropriate method for working bone. So we may also imagine that seeping in through the walls of social and linguistic intelligence were the muffled thoughts and knowledge coming from the chapels of technical and natural history intelligence. Consequently these were also available for use by linguistic intelligence when generating utterances.

What would have happened to these snippets of language about the non-social world? They must have entered the minds of other individuals as part of the flow of social language and have been decoded by linguistic intelligence and interpreted by social intelligence. In other words, the chapel of social intelligence began to be invaded by non-social information. Those individuals who could exploit these invasions to increase their own knowledge about the non-social world would have been at a selective advantage. They would have been able to make more informed decisions about hunting and toolmaking, enabling them to compete more successfully for mates and provide better care for offspring.

Further selective advantage would have been attained by those individuals who could add more non-social linguistic snippets into conversation, such as by introducing questions about animal behaviour or toolmaking methods. Perhaps these were individuals who, due to random changes made in the architectural plans they inherited, had particularly permeable walls between their specialized intelligences. These talkative individuals were gaining their selective advantage by exploiting the non-social knowledge of other individuals by using language, as opposed to relying on behavioural observations alone. As a consequence, social language would very rapidly (in evolutionary time) have moved to a general-purpose language; my guess would be in the time period between 150,000 and 50,000 years ago. Natural selection, the most important architect of the mind, simply would not have allowed this opportunity to improve the exchange of non-social information, and hence increase reproductive success, to pass by.¹

There is evidence of this switch from a social to a general-purpose language surviving in our conversation today. As Robin Dunbar described, we still predominantly talk about social issues – we have a love of gossip. Moreover, when we talk about physical objects we often appear to ascribe to them an intrinsic tendency towards motion and imply that they possess ‘minds’ as if they are living, social beings. This has been explained by the linguist Leonard Talmy.² He argues that sentences such as ‘the book toppled off the shelf’ and ‘the ball sailed through the window’ imply that these objects move under their own power, since they are equivalent in their structure to sentences such as ‘a man entered the room’. More generally, utterances appear to use the same range of concepts and structures whether they are referring to mental states, social beings or inert objects – which linguists refer to as the ‘thematic relations hypothesis’.³ They assume that the original use of language was for the last of these, and those concepts became transferred into utterances about the social/mental world by ‘metaphorical extension’. Yet it makes more sense to see it the other way round: the structure of language arose when talking about the social world and was metaphorically extended for talking about physical objects.

The superchapel of the mind

Returning to our evolutionary scenario of a switch from social to general-purpose language, we must ask what happened to the chapel of social intelligence as it began to be invaded by non-social ideas and information? The cognitive scientist Dan Sperber has provided the answer: it became a type of superchapel in the cathedral of the mind. As we saw in Chapter 3, he described this superchapel as the ‘module of metarepresentation’ (MMR). He suggested that the MMR is an expanded version of the theory of mind module, although my position conceives of it as an expanded – perhaps even exploded – version of a more general domain of social intelligence. Sperber states: ‘As a result of the development of communication, and particularly of linguistic communication, the actual domain of the meta-representational module is teeming with representations made manifest by communicative behaviours…. An organism endowed with … a meta-representational module … may form representations of concepts and beliefs pertaining to all conceptual domains, of a kind that the modules in those domains might be unable to form on their own.’⁴ (see Figure 27)

The critically important point that Sperber is trying to express is that knowledge about the world comes to be represented in two different locations within the mind – within the specialized cognitive domain where it ‘belongs’, and within what had been the domain of social intelligence but which now contains knowledge about both the social and the non-social world. Indeed, the multiple representations of knowledge within the mind is a critical feature of Annette Karmiloff-Smith’s idea for how cognitive fluidity arises during development.

This idea helps us understand what often appear to us to be contradictory views held by living hunter-gatherers, and indeed any Modern Human, about their world. Recall, for instance, the attitude to the polar bear held by the Inuit that I described in Chapter 3. This animal is thought of as a fellow kinsman, but it is also killed and eaten with delight. This combination of a deep respect for the animals they hunt, often expressed in terms of social relationships, and the lack of any qualms about actually killing them appears to be universal among hunter-gatherers. Such a combination of attitudes appears contradictory to us, until we appreciate that knowledge about these animals may be contained in two different cognitive domains – one where it relates to natural history and the problems of securing food, and one where it is mixed up with social intelligence. Another example is the attitude of the Australian Aborigines to their landscapes. To exploit these they rely upon a profound understanding of ecology. They are expert natural historians with detailed knowledge about the cycles of life and death. Yet they also understand their landscape as continuously created by Ancestral Beings, who have no respect for any laws of ecology. There is no contradiction or confusion in the Aboriginal mind: they simply have two mental representations of their environment, located in different cognitive domains.

27 The role of language in creating cognitive fluidity.

Sperber suggested that the invasion of social intelligence by non-social information would trigger a ‘cultural explosion’.⁵ We do, of course, see precisely that cultural explosion at the start of the Upper Palaeolithic, and indeed see a rumbling anticipation of it after the first Modern Humans enter the play of our past 100,000 years ago. And as part of that cultural explosion we see the appearance of concepts and beliefs which no single domain could create by itself, concepts such as art and religion.

A new role for consciousness

A critical feature of the change to a cognitively fluid mind was a change in the nature of consciousness. Throughout this book I have followed Nicholas Humphrey’s arguments that (reflexive) consciousness evolved as a critical feature of social intelligence: it enabled our ancestors to predict the behaviour of other individuals. But just like any other microdomain of social intelligence, consciousness was not accessible to thought in other cognitive domains – there is no reason to expect Early Humans to have had an awareness about their own knowledge and thought processes concerning the non-social world (other than the ephemeral rolling consciousness I described in Chapter 8). But if, via the mechanism of language, social intelligence starts being invaded by non-social information, the non-social world becomes available for reflexive consciousness to explore. This is, in essence, the argument that Paul Rozin made in 1976 regarding the evolution of advanced intelligence. The critical feature of his notion of accessibility was the ‘bringing to consciousness’ of the knowledge which was already in the human mind but located within the ‘cognitive unconsciousness’.⁶

Quite how much knowledge was brought to a level of conscious awareness is unclear. As I discussed in Chapter 8, a large proportion of our mental activity is likely to remain closed to us in our unconscious mind. Craftspeople, for instance, often appear unaware of the technical knowledge and skills they are using. When asked how they undertake tasks such as throwing a pot they often have difficulty explaining what they do unless they can provide a demonstration. Actions do indeed speak louder than words when technical knowledge is trapped within a specialized cognitive domain. This emphasizes the importance of verbal teaching of technical skills, which only began at the start of the Upper Palaeolithic as is implied by the spatial proximity of knapping debris produced by skilled and unskilled knappers at sites such as Etiolles in France and Trollesgave in Denmark.⁷ When knowledge is acquired by verbal instruction it is by definition passed into what had once been the chapels of social and linguistic intelligences, where it becomes available for reflexive consciousness.⁸

The new role for consciousness in the human mind is likely to have been the one identified by the psychologist Daniel Schacter. In an article written in 1989 he argued that, in addition to creating the subjective feelings of ‘knowing’, ‘remembering’ and ‘perceiving’, consciousness should be viewed as ‘a global data base that integrates the output of modular processes.’ He goes on to argue that such an ‘integrative mechanism is crucial in any modular system in which processing and representations of different types of information are handled in parallel by separate modules.’⁹ In the Early Human mind, general intelligence was the only device available to play this integrating role, and it hardly played it at all. But because language acted as the vehicle for delivering non-social thoughts and knowledge into the chapel of social intelligence, consciousness could start to play this new integrating role within the cathedral of the mind.

We have seen the consequence of integrating knowledge from separate domains in the previous chapter – a vast increase in human creativity. A final argument that consciousness is playing a fundamental role in achieving this integration and resulting creativity comes from the philosopher John Searle. In his 1992 book The Rediscovery of the Mind, he considered those sufferers of petit mal seizures that I referred to in Chapter 8. Recall that during their seizures they were able to continue with their goal-directed behaviour but without any consciousness. Referring to the change in the manner in which they under took their activities, such as piano playing, Searle writes: ‘the patients were performing types of actions that were habitual, routine and memorized … normal human conscious behaviour has a degree of flexibility and creativity that is absent from … [these] … cases of unconscious drivers and unconscious pianists…. One of the evolutionary advantages conferred on us by consciousness is the much greater flexibility, sensitivity and creativity we derive from being conscious.’¹⁰

Early Humans did not lack consciousness altogether; it was simply restricted within their domain of social intelligence. And consequently their social interactions showed considerable flexibility, sensitivity and creativity. But this was markedly absent from their non-social activity – as anyone who has had the task of describing handaxe, after handaxe, after handaxe will know. But as soon as language started acting as the vehicle for delivering non-social information and ideas into the domain of social intelligence, reflexive consciousness could also get to grips with the non-social world. Individuals could now become introspective about their non-social thought processes and knowledge. As a result, the whole of human behaviour was pervaded with the flexibility and creativity that is characteristic of Modern Humans.

Nursing females, cognitive fluidity and extended childhood

The scenario I have offered for the evolution of cognitive fluidity suggests that by 150,000 years ago the Swiss-army-knife mentality was beginning to break down. Those individuals who were able to exploit snippets of non-social conversation were at a selective advantage as they could integrate knowledge which had been ‘trapped’ within specialized intelligences. We can, I think, identify one particular class of individual within these societies who would have been under particular selective pressure to achieve cognitive fluidity: sexually mature females.

Females at any time during human evolution were only able to give birth to relatively small-brained infants. This is due to the anatomy of the pelvis which needs to be narrow to allow efficient walking on two legs.¹¹ Consequently the offspring of Modern Humans have a brain size no larger than that of a newborn chimpanzee – about 350 cc. Yet unlike the chimpanzee, in the immediate period after birth the human brain continues to grow at the same rate as that of a foetus. By the age of four a human brain has tripled in size, and when maturity is reached it is around 1400 cc, four times the size at birth. In contrast the chimpanzee brain has only a small postnatal increase in size to around 450 cc.¹² During the period of brain growth after birth, human infants have a very high degree of dependency on adults. There are substantial demands on the mothers to supply the energy to fuel the growth of the infant brain, and indeed anatomy in general. These demands would have become particularly strong during the second period of rapid brain expansion that began after 500,000 years ago.

The social anthropologist Chris Knight and his colleagues have argued that the Early Modern Human females solved the problem of how to ‘fuel’ the production of increasingly large-brained infants by extracting ‘unprecedented levels of male energetic investment’.¹³ They suggest that coordinated behaviour by females forced males to provide them with high-quality food from hunting. An important element of the female action was a ‘sex strike’ and the use of red ochre as ‘sham menstruation’. They describe this as the first use of symbolism and find evidence for it in the increase of red ochre after 100,000 years ago associated with the Early Modern Humans of southern Africa.

While I am sceptical about their ideas of coordinated female action, they have identified a social context in which food becomes critical in negotiating social relationships between the sexes. In this context ‘snippets’ of language about food and hunting may have been especially valuable in the social language between males and females. Females, in particular, may have needed to exploit this information when developing their social relationships with males. This may indeed explain why the first step towards cognitive fluidity, as seen in the behaviour of the Early Modern Humans of the Near East, was an integration of social and natural history intelligence.

The increase in the time between birth and maturity that arose as brain size enlarged during the course of human evolution¹⁴ has another consequence for the switch from a Swiss-army-knife to a cognitively fluid mentality. This is simply that it provides the time for connections between specialized intelligences to be formed within the mind. As I described in Chapter 3, the developmental psychologist Annette Karmiloff-Smith has argued that the mind of a modern child passes through a phase during which cognition is essentially domain-specific, after which knowledge becomes applicable beyond the special-purpose goals for which it is normally used. In Chapter 7 I argued that the cognitive development for young Early Humans effectively ceased after the specialized domains of thought had arisen and before any connections had been built. Consequently, with regard to development, the source of cognitive fluidity must lie in a further extension of the period of cognitive development.

There is indeed evidence in the fossil record that child development of Modern Humans is considerably longer than that of Early Humans. This comes in the form of the skeletal remains of the few Neanderthal children that exist. These show that Neanderthal children grew up rather quickly, developing robust limbs and a large brain at an early age compared with Modern Humans. A particularly important specimen comes from the site of Devil’s Tower on Gibraltar and dates to around 50,000 years ago. This consisted of no more than five fragments, but reconstructions have shown it to be of a three- or four-year-old child. The teeth of this child demonstrate that dental eruption occurred earlier than in Modern Humans. Of more interest, however, is that at this young age the brain size of this Neanderthal, at 1400 cc, was approaching that of a mature adult. Such a rapid rate of brain expansion appears to be a general feature of Neanderthal children, being found in several other specimens.¹⁵ The most recently discovered and best-preserved Neanderthal child is a two-year-old from Dederiyeh Cave in Syria. This appears to have possessed a brain size equivalent to that of a six-year-old Modern Human.¹⁶

In essence, there was no time for cognitive fluidity to arise before the development of the Neanderthal mind – and I assume the Early Human mind in general – had ceased. Unfortunately we lack any child skulls of the 100,000-year-old Early Modern Humans from the Near East, or those of the first Upper Palaeolithic hunter-gatherers. But my guess would be for a gradual extension of the period of development between 100,000 and 50,000 years ago.

The rise of the modern mind: an overview

Let me conclude this chapter by summarizing my explanation for the evolution of cognitive fluidity. The seeds were sown with the increase of brain size that began 500,000 years ago. This was related to the evolution of a grammatically complex social language. The utterances of this language, however, carried snippets of non-social information as well. Those individuals who were able to exploit such non-social information gained a reproductive advantage. In particular, females who were nursing infants for prolonged periods – and therefore unable to feed themselves adequately – would have come under selective pressure to adapt in this way, because their patterns of social interaction with males had become bound up with a need for food. As social language switched to a general-purpose language, individuals acquired an increasing awareness about their own knowledge of the non-social world. Consciousness adopted the role of an integrating mechanism for knowledge that had previously been ‘trapped’ in separate specialized intelligences.

The first step towards cognitive fluidity appears to have been an integration between social and natural history intelligence that is apparent from the Early Modern Humans of the Near East, 100,000 years ago. This is before Modern Humans dispersed into Asia and Europe where they either replaced or inter bred with existing Early Human populations. The final step to a full cognitive fluidity occurred at slightly different times in different populations between 60,000 and 30,000 years ago. This involved an integration of technical intelligence, and led to the changes in behaviour that we refer to as the Middle/Upper Palaeolithic transition. In other words, it created a cultural explosion: the appearance of the modern mind.