Final draft of published Appendix 1 in Charlton, B.
Psychiatry and the human condition
Radcliffe Medical Press: Oxford, UK, 2000.

Evolution and the cognitive neuroscience of
awareness, consciousness and language

Bruce G Charlton MD

Reader in Evolutionary Psychiatry
Department of Psychology
University of Newcastle upon Tyne

Editor-in-Chief, Medical Hypotheses

Evolution and the cognitive neuroscience of awareness, consciousness and language
Consciousness is sometimes regarded as intrinsically mysterious - something probably beyond human comprehension, maybe even impossible to define. On the other hand consciousness is an ordinary fact of life - babies are born without it and develop it over the first few years of life. And whatever it is, it presumably evolved - like other complex biological phenomena. Even if we regard consciousness as a curse, then that makes it even more plausible that it has a biological benefit to counterbalance its obvious disadvantages - or else natural selection would have gotten rid of it long ago (saving a lot of hungry brain tissue in the process). We experience the dawn of consciousness every morning when we awaken.

So consciousness cannot be any more biologically mysterious than any of the other extremely hard to understand and explain abilities we and many other animals enjoy. Indeed, consciousness is almost certainly a great deal simpler than human vision - which requires enormous amounts of brain tissue and hundreds of millions of years of evolution to perfect. Consciousness seems to have taken only tens of millions of years to develop, and involves only a relatively small amount of the cerebral cortex.

Evolution of 'awareness'

People often find 'consciousness' mysterious, but the real mystery is awareness - and many other animals are aware, so this is not a specifically human mystery.

The question of consciousness can therefore be approached by considering the general phenomenon of awareness, of which consciousness is one particular example. The mystery usually consists in puzzlement over why humans are apparently 'conscious' (ie. aware) of at least some of their own 'thinking' (ie. cognitive processing), rather than cognitive processing simply generating behavior without awareness. Why do we know that we know, instead of just knowing it? Since most of mental life (including some of the most computationally difficult tasks, such as those concerned with vision) proceeds perfectly well without awareness of processing, the question arises: why should awareness exist at all?

Function of awareness
But this question is badly framed. The proper questions should be as follows. Firstly, is awareness an evolved adaptation? In other words did awareness evolve to solve a problem with reproductive consequences for the animal, or is awareness it perhaps an epiphenomenon? And if we assume that awareness is an adaptation with a biologically useful function, the second question concerns the biological nature of this adaptation - what is the mechanism by which awareness solves the problem it evolved to solve?

So, the first difficulty with the 'mystery' of consciousness is that the question as commonly stated bundles together several questions that require separate consideration. There is the general phenomenon of awareness, and consciousness is 'merely' a specific type of awareness. In fact, consciousness is awareness of inner body states, as will be described later.

Awareness is not distinctive to humans, many animals display the characteristics of awareness, it is something that seems to have evolved many times in many lineages. And awareness has a quite exact definition: it is the ability selectively to direct attention to specific aspects of the environment, and to be able cognitively to manipulate these aspects over a more prolonged timescale than usual cognitive processing will allow. To hold in mind selected aspects of the perceptual landscape.

Technically, awareness is attention plus working memory - ie. the ability to attend selectively among a range of perceived stimuli and a short term memory store into which several of these attended items can be 'loaded', held simultaneously, and combined. Awareness is a standard variable in psychological research, unproblematically measured in, for example, animal vision. It is studied by means such as measuring performance at memory tasks while monitoring gaze direction, delaying responses, and recording brain activity. When brain activity correlates exactly with performance of tasks then it can be assumed that that bit of brain is involved in that particular task. And the length of time which brain activity is sustained corresponds to an animals ability to 'hold in mind' information for immediate use. Researchers are therefore recording the operation of a temporary store.

Awareness is not therefore an aspect of social intelligence. Instead, awareness is a mechanism of integration. Awareness is a way of converging and combining information, and it is a functional ability that is found in complex animals living in complex environments. Awareness therefore relates to the ability to cope with complexity or perception and behavior, and it is found not only in social animals, but also in solitary animals. While awareness is found in animals right across the animal kingdom; consciousness is of much more limited distribution. I suggest that consciousness is probably confined to a small number of recently-evolved social animals such as the great ape lineage - especially common chimpanzees and bonobos - and perhaps a few other recently-evolved social mammals such as elephants and dolphins.

Awareness is in located in working memory
Awareness comprises attention and working memory (WM). To be aware of an perception it must be selectively attended to, and the representation of that entity must be kept active and held in the brain for a length of time adequate to allow other cognitive representations to interact with it and in a place where other cognitive representations can be projected. Working memory is such a place, a place where information converges and is kept active for longer than usual periods. Hence working memory is the anatomical site of awareness.

The nature of working memory can be understood using concepts derived from cognitive neuroscience. Working memory is a three-dimensional space filled with neurons that can activate in patterns. Cognition is conceptualized as the processing of information in the form of topographically-organized (3-dimensional) patterns of neural activity called representations - because each specific pattern 'represents' a perceptual input. So that seeing a particular shape produces a pattern of cell activation on the retina, and this shape is reproduced, summarized, transformed, combined etc in patterns of cell activation in the visual system of the brain - and each pattern of brain cell activation in each visual region retains a formal relationship to the original retinal activation.

Representations are the units of thinking. In the visual system there may be representations of the colour, movement and shading of an object, each of these having been constructed from information extracted from the original pattern of cell activation in the retina (using many built-in and learned assumptions about the nature of the visual world). The propagation and combination of representations is the process of cognition.

Cognitive representations in most parts of the brain typically stay active and persist for a time scale of the order of several tens of milliseconds. But in working memory cognitive representations may be maintained over a much longer time scale - perhaps hundreds or thousands of milliseconds - and probably by the action of specialized 'delay' neurons which maintain firing over longer periods. So WM is a 3-D space which contains patterns of nerve firing that are sustained long enough that they can interact with other 'incoming' patterns. This sustaining of cognitive representations means that working memory is also a 'convergence' region which brings together and integrates highly processed data from several separate information streams.

Any animal that is able selectively to attend-to and sustain cognitive representations could be said to possess a WM and to be 'aware' - although the content of that awareness and the length of time it can be sustained may be simple and short. The capacity of WM will certainly vary between species, and the structures that perform the function of WM will vary substantially according to the design of the central nervous system. In other words working memory is a function which is performed by structures that have arisen by convergent evolution, WM is not homologous between all animals that possess it - presumably the large and effective WM of an octopus is performed by quite different brain structures from WM in a sheep dog, structure that have no common ancestor and evolved down a quite a different path. The mechanism and connectivity of the human WM allows cognitive representations from different perceptual modalities or from different attended parts of the environment to be kept active simultaneously, to interact, and to undergo integration in order that appropriate whole-organism behavioral responses may be produced.

Working memory is reciprocally-linked to long term memory (LTM), such that representations formed in WM can be stored in LTM as patterns of enhanced or impaired transmission between nerve cells (the mechanism by which this occurs is uncertain but probably involves a structure called the hippocampus). So temporary patterns of active nerves are converted to much more lasting patterns of easier or harder transmission between nerves. The patterns in LTM may be later recalled and re-evoked in WM for further cycles of processing and elaboration.

This is how complex thinking gets done - a certain, maximum number of representations can interact in WM in the time available (maybe a couple of seconds). So there is a limit to what can be done in WM during the span of activation of its representations. To do more requires storing the intermediate steps in reasoning. The products of an interaction in WM can be summarized ('chunked') and 'posted' to LTM where they wait until they are need again. When recalled and reactivated these complex packaged representations from LTM can undergo further cycles of interaction and modification, each building up the complexity of representations and of conceptual thought.

WM is therefore conceptualized as a site for integration of attended perceptual information deriving from a range of sensory inputs. Awareness seems to be used to select and integrate relevant inputs from a complex environment to enable animals to choose between a large repertoire of behavioral responses. There is a selective pressure to evolve WM in any animal capable of complex behavioural responses to a complexly variable environment. So the cognitive representations in WM in non-conscious animals are derived from external sensory inputs (eg. vision, hearing, smell, taste and touch).

The critical point for this current argument is that non-conscious animals may be aware of their surroundings, but they lack the capacity to be aware of their own body states. Awareness of outer environment is common, but awareness of inner body states is unique to conscious animals.

Evolution of 'consciousness'

Awareness of body states
So the starting point for evolution of consciousness is an aware animal with an integration centre called working memory which is able to maintain the activity of attended perceptual representations. The evolutionary breakthrough to consciousness occurs when working memory receives not just external perceptual information from senses, but also projections of inner body states. In other words, in a conscious animal WM memory contains body state representations as well as perceptions of the external environment.

Consciousness arises when body state information becomes accessible to awareness. And consciousness depends upon the animal evolving the ability to feed information on its internal physiological state into WM, so that it can be integrated with sensory perceptual information.

First, some terminology. The physiological state of the body states constitutes what is more commonly called an emotion. To put it another way, emotions are body states as they are represented in the brain. Damasio points out that although we think of the brain as being concerned mainly with processing information derived from inputs by the five senses, in fact controlling internal body states is the primary evolutionary process of the brain. The primitive brain in lower animals is mainly a devices for monitoring what is going on in the body - brains (or rather a central nervous system) evolved when bodies got too large to allow communication to occur purely by diffusion of chemicals. So the main business of the brain is to monitor and interpret body states (including emotions), and to modulates these states.

Feelings is the term for emotions of which we are aware. For instance, 'fear' is activation of the sympathetic nervous system and preparation of the body for 'fight or flight' - so fear is the effects of the sympathetic nervous system on the disposition of internal organs ('viscera') such as muscle tension, heart rate, sweat glands and so on. When these body states register in the brain and affect behavior without awareness this can be termed the emotion of fear, when we become aware that we are frightened then this is termed the feeling of fear.

By this account emotions may be non-conscious, while feelings are conscious. And emotions are found in many animals, most of which are not conscious and not aware of their emotions. A cow can experience the emotion of fear (and react appropriately) but it will not have the feeling of fear - it will not know it is frightened, it will just be frightened. In other words, fear does not have a representation in the WM of a cow. By contrast a person can experience fear without or with awareness. Fear is present when the body state of fear is present, but humans may or may not have awareness of this - a person can be frightened without knowing, just as a cow can. For example when watching a horror film and absorbed in the action a person might experience the physiological state of fear (thumping heart, hair standing on end - a preparation for action). But the person may lack the awareness that they are frightened until such a point as they are interrupted and asked whether we are frightened - at which point an awareness of fear is produced, an awareness of tense posture, creeping skin, hair on end….

So that fear can be an emotion, leading to an adaptive behaviour such as fight or flight - and the behavior may occur without a person having any awareness of their inner state. But in conscious animals such as humans there may also be awareness of an inner state: the person may know that they are scared. The question is, what is the use of knowing that one is frightened, what is the adaptive function of consciousness - especially given that most animals function perfectly well without such knowledge? What function did consciousness evolve to perform?

The answer is that consciousness is an aspect of social intelligence, and the adaptive function of consciousness is to enable the cognitive modeling of social situations. Animals evolved the ability to project body state representations into working memory in order that emotions could interact with perceptions. And awareness of inner body states is an accidental by-product of bringing together cognitive representations of emotions and cognitive representations of social perceptions in working memory.

Working memory as a convergence zone for emotions and perceptions
Consciousness therefore exists because WM is the location - the only location - where the streams of internal and external information converge, where information on the environment is juxtaposed with information on the body, where emotional representations can interact with, modify and evaluate representations of social perceptions.

The point is critical: once consciousness had evolved, representations of socially-relevant perceptions (eg. a particular person) could be 'evaluated' by correlating a perception with subsequent changes in body state (eg. fear after seeing that particular person). So that a particular person would be evaluated as 'fear-provoking'. Social events often lead to emotional responses and adaptive behavior - and fear of a particular person may lead to avoidance without awareness of the process. But in a conscious animal, cognitive representations of both the social event and the resulting emotion can interact in working memory, and we can become aware that we fear a particular person because the juxtaposition has occurred in working memory.

This perceptual-emotional interaction creates the possibility for new kinds of cognitive representation: representations comprising information from both the senses and the body. It is an accidental by product of convergence in working memory that these new kinds of perceptual-emotional representation are able to become the subject of awareness. Awareness of inner body states is not the primary role of consciousness, rather it is an epiphenomenon of the fact that convergence is attained in working memory - that is just the way that things happened to evolve. If, in an alternative history, internal and external information converged in another part of he brain than MW, then presumably we would not be aware of body states - and we would not therefore be 'conscious'.

It is probable that consciousness is crucially dependent upon neural circuits located in the pre-frontal cerebral cortex of humans - this is the most recently evolved part of the human brain. The dorso-lateral (DL) prefrontal cortex - the upper-outer lobes of the front of the brain - are probably the site of working memory in humans. Indeed, it is perhaps specifically the DL frontal cortex of the dominant (language-containing) hemisphere. And working memory probably functions by having arrays of 'delay neurons' capable of remaining active for longer than most neurons, and by arranging these in an hierarchical pattern. Information from different perceptual inputs which is fed into working memory at the posterior part of the DL prefrontal cortex can become integrated as it converges towards the upper levels of the hierarchy. So, in most people, working memory is probably located in the large dome of brain above the left eye and extending about one third of the way back - and the further forward one goes, the more integrated the information becomes and the 'higher' the level of processing.

I am suggesting that body state representations are constructed in the parietal lobe of the non-dominant (usually right) cortex - so that information on the state of the body converges on the non-dominant parietal lobe, is interpreted for its emotional importance, and behaviours are initiated that are appropriate to this information - the whole process happening with out any need for conscious awareness. The parietal region seems to be necessary to interpret the biological meaning of body state feedback in terms of relevance to behavior.

In other words, the parietal region seems to conceptualize feedback in terms of a continuously updated body image, and the continual updating of this image is necessary to the experience of emotions. Without the relevant parietal region body states would not be interpretable. Destruction of the right sided parietal (for example in a stroke or a traumatic brain injury) will destroy the body image representation and the ability to determine what is body and what is not will be lost. In the common phenomenon of 'neglect' or anosognosia, a person with a non-dominant parietal lesion may lose their awareness of the opposite side of their body. While they are actually looking at the left hand, they may be able to comprehend that it is indeed a part of their body. But when they are not observing the hand and rely on internal information this awareness is lost - presumably the hand is omitted from their body image, and they may neglect to move or care for the left hand, may even deny that the hand belongs to them, or feel it is an alien hand. This emphasizes the extent to which we depend upon an internal representation of our body state I order to monitor and control the body.

Continually-updated body state representations are projected from the right parietal lobe, across the fibres of the corpus callosum that link the two sides of the brain, and to the left dorsolateral prefrontal lobe. Just as destruction to the non-dominant parietal prevents the body state information from being constructed into a body image, so any lesion to the fibres as the cross the corpus callosum or penetrate the prefrontal lobe will prevent body image representations reaching WM. Since the SMM loses emotional input, and perceptions cannot be evaluated by reference to the body states (emotions) that they evoke, lesions to the non-dominant parietal or to the corpus callosum are associated with impaired social intelligence (eg. severing the corpus callosum as a treatment for epilepsy or severing the horizontal connections between the prefrontal region of the cortex and the rest of the brain (as in certain types of 'leucotomy') will both severely impair social intelligence - consistent with the assumption that social intelligence needs emotional information (ie. body states) in order to perform its function. A patient with a right sided stroke will often deny they have any disability, and their social judgment is very poor. Similarly 'split brain' patients with a surgically severed corpus callosum apparently cannot cope with employment that requires the exercise of planning and judgment, and their social interactions are impaired.

Why can we be aware of body states?
Why should animals such as humans have evolved to become aware of body states? One answer is that awareness of body states was adaptive, it enabled evaluation of social information by emotions, and this gave the conscious animal competitive advantages in the social arena by enabling strategic social intelligence. But it is not the awareness of body states that is adaptive in itself - rather we are aware of our body states an accidental by product of the fact that they are juxtaposed with perceptions in working memory. The ability to 'introspect' and become aware of our internal milieu (heart beat, abdominal sensations, tiredness etc.) does not itself have an adaptive function.

It is often stated that the things that consciousness does could not be done equally well or better without consciousness - for example by a 'zombie'. Theorists arguing along this line either suggest that consciousness is non-adaptive, an accidental by product of something else. Or they argue that consciousness is a mechanism necessary to the solution of some particular adaptive problem that can only be solved by consciousness, or at least a problem for which consciousness provides the simplest or most efficient, engineering solution.

However, this is a non-biological and potentially misleading approach to understanding adaptive function. There are many theoretically potential solutions to any specific behavioral problem. The actual solution reached by natural selection is seldom the simplest or most efficient engineering solution. this arises because contingent historical factors constrain the possible directions natural selection can take, each evolutionary step must be an incremental improvement on what went before, and furthermore the genetic mutations upon which adaptations are built are random and undirected. In the case of consciousness, constraints such as the previously existing structure of the brain are critical in determining the range of possibilities for subsequent evolution.

Consciousness is sufficient to perform its adaptive function, but not necessary
The fact that a cognitive task could in principle be performed without consciousness is irrelevant to the adaptive argument. Even if the task could be performed more simply and efficiently by other methods the only thing that matters is what, as a matter of historical fact, actually was the solution arrived at by natural selection.

Consciousness is therefore required to be sufficient to, but not necessary for, the performance of the task which it is evoked to explain - just as legs are sufficient to, but not necessary for, locomotion. Wheels would also do the job. The 'ultimate' reason why humans locomote by legs rather than wheels is a matter of contingent historical constraints rather than, say, mechanical effectiveness or engineering simplicity. Whatever the relative functional pros and cons of wheels versus legs, humans just happened to have evolved from ancestors with legs - wheels were not an option. Similarly consciousness working by awareness of body states was not the only way of integrating emotions and social perceptions - the task could in principle (ie. under different constraints) have been does without awareness.

To recapitulate. In principle, humans would not need to be aware of body states in order to integrate body state information with perceptual information. If our evolutionary history had taken a different path, then integration might have been achieved in brain regions where cognitive processing did not reach awareness. But, by the 'accidents' of evolutionary history, WM happened to be the place in which emotions and perceptions were brought together. Hence, human awareness of emotions is a consequence of contingent evolutionary factors, it is not a formally necessary aspect of strategic social intelligence and in this sense its mechanism is accidental. And at the same time the awareness of emotions does not in its own right confer an adaptive advantage - it is only the convergence of emotions with social perceptions in working memory that is adaptive. Its mechanism will now be explored.

The somatic marker mechanism

Emotions and feelings
According to Damasio the primary evolutionary function of the animal brain was to serve as an integrative centre to monitor, coordinate and regulate the 'inner world' of a complex organism. Therefore the human brain, like the brains of other complex animals, receives on-line, continuously updated representations of the state of the body. These representations are mostly derived from sensory autonomic nerves from the inner organs and somatic nerves from muscles and skin, modified by hormonal chemical messages. They comprise the afferent or feedback arm of a feedback and control mechanism for monitoring, integrating and modulating the current 'state of the organism': internal viscera, skin, muscle, connective tissue, joints, blood chemistry and so on.

It seems likely that there is a region of the parietal lobe in the 'non-dominant' hemisphere of the cerebral cortex (ie. the side that does not have the language specialization - usually the right hand side) that is responsible for integrating information from body state feedback to create a continually updated representation of the body state. If this region is destroyed (for example when someone has a stroke affecting the right parietal lobe) they exhibit a phenomenon termed 'neglect' or anosognosia. The person becomes unaware of all or part of the left side of the body and visual field - that part controlled by the right cerebral hemisphere.

'Feelings' occur when body state representations in working memory indicate a change in body state in response to change in the environment. Hence, consciousness uses feedback concerning body states in order to evaluate perceptual inputs, by juxtaposing feelings with the perceptions that have preceded them. In other words, changes in the soma (body) are used to mark perceptions in WM. What is formed are perceptual-emotional representations - representations which encode information on both perceptual information and the body state that occurred in response. We might imagine a visual perception of an aggressive male as one representation and the emotion of terror as another representation - both active in WM at the same time. The SMM will combine these two representations to create a single representation (aggressive male-fear) that when it is activated in WM will elicit both recognition of the perception, and replaying of the emotion.

Perceptual-emotional representations evolved in order to deal particularly with social situations: consciousness is adapted to function as an aspect of social intelligence. In summery, the Somatic Marker Mechanism (SMM) evolved in order to evaluate social information and enable strategic social intelligence.

Theory of mind and the somatic marker mechanism
The somatic marker mechanism is a vital aspect of the so-called Theory of Mind Mechanism (ToMM) - although the full theory of mind ability requires (I will argue) language, as well as the SMM.

Theory of mind has been proposed as a cognitive mechanism by which overt behaviour is interpreted in the light of inferred mental attributes. In other words, an animal with the ToMM is able to make a 'theory' about the contents of another animal's mind - this is the ability that Baron-Cohen has termed 'mind reading'. The ability is termed a 'theory' of mind mechanism, because the attribution of mental contents of another animal is based on inference - obviously animals do not have direct access to the contents of each other's minds and every inference is in this sense a 'theory'. From observation of behavior and context I may draw the conclusion that someone is angry, I don't know for sure that the person feels anger - this is a theory designed to account for the situation and predict the future outcomes, and humans are good enough at this mind-reading that loss of the ability (as in autism) is a severe handicap.

In an animal with ToM, the primary interpretative inference is 'mentalistic', and overt behaviour is understood in the context of ascribed motivations, dispositions and intentions. For example, we infer the meaning of a smile by reference to a person's state of mind - the smile could be understood as one of sympathy, of shared delight, of ingratiating deception, or maybe a superior sneer - according to our understanding of the smiler's state of mind. By contrast, it is assumed that most animals - lacking a ToM - infer the meaning of behaviour directly from overt behavioral cues - to such an animal a smile is unambiguously a smile, an expression having a single meaning.

The selection pressure which led to the evolution of ToM was probably the potential ambiguity of social cues when overt behavior is ambiguous (eg. when behavior is complex, rapidly-changing, or deceptive). When behavioral cues are ambiguous, interpretation of a given cue becomes dependent upon inferences concerning intentions, dispositions and relationships.

For example, the approach of another human is ambiguous - it may have several meanings, some hostile some friendly. In the interpretative sequence 'That man is angry, and approaching me - therefore I must get ready to fight'; the mentalistic ascription of anger is logically prior to the interpretation of overt behavioral cues. If the ascription of disposition were to be changed from 'angry' to 'happy', then - even when the immediately perceived cues are identical - the inferred meaning of the overt behaviour 'approaching me' (and the implications for an adaptive response) would also change.

As a plausible example in chimpanzees, the approach of a male stranger might evoke fear - the physiological state of arousal in preparedness for 'fight or flight'. This response comprises a characteristic physiological state. Cognitive representations of changing body state are continually constructed in the brain from feedback from the afferent nerves, chemo-receptors and other inputs converging and being integrated probably in the right sided parietal lobe. So emotions are created in this way - using feedback from the body.

In an animal lacking an SMM, such cognitive representations of the changing body state may affect behaviour - so that the emotion of fear might provoke involuntary flight. But in an animal with an SSM, a cognitive representation of this changing body state may be projected forward into the prefrontal cortex and the region which performs the working memory (WM) function. In WM the body state representation that is the emotion of fear becomes accessible to awareness as a conscious feeling of fear. The cognitive representation of 'fear' may then be used as a somatic marker when sustained in WM in temporal juxtaposition to the perceptual representation of the male stranger's identity.

The juxtaposition of the somatic marker for fear with the stranger's identity that evoked it, creates a novel cognitive representation incorporating what is, in effect, the disposition of that individual - this stranger is disposed to be violent. Although it does not specifically make reference to the mind - this is a theory of mind inference, an inference that the stranger is of aggressive intent. The combined perceptual-emotional representation is implicitly one of 'that fear-evoking stranger': ie. aggression and hostility are attributed as a 'theory' of the stranger's mental contents. The combined representation can be stored in long term memory, and when recalled to WM it will be capable of re-evoking individual identity (perception) and simultaneously re-enacting the linked body state of fear (emotion) as a change in body state.

The ventro-medial (lower middle) prefrontal cortex appears to be necessary for the interaction of body state-representations with WM that enables consciousness. What possibly happens is that working memory in the upper-outer (DL) frontal lobe sends a message down through the inner-middle (VM) frontal lobe to deeper structures called the basal ganglia which then evoke the appropriate emotional state in the rest of the body. In patients who have suffered damage to either the ventro-medial prefrontal cortex or the basal ganglia, this damage can prevent the expression of secondary emotional states in response to cognitive modeling - the assumption is that upper-outer frontal, inner-middle frontal and basal ganglia form links in a chain that produce emotion in response to cognitive representations in working memory. If this chain is broken, then we would no longer be able to experience fear as a result of imagining frightening events although we would still experience fear in response to actual frightening events. We would no longer enact the emotion of fear when thinking about a tiger attack - although a real life tiger attack would still produce arousal and flight.

Internal-modeling of behavior by the SMM
The somatic marker mechanism is a system for internally modeling social behavior, and its emotional consequences. The possibility of creating a combined perceptual-emotional representations means that social relationships can be variously combined and sequenced in working memory, and the consequences of this deployment can be evaluated by re-experiencing the enacted emotional body state as gratifying or aversive.

For instance, in an ancestral situation perhaps the representation of my aggressive and lustful male cousin recalled from long term memory along with a representation of my beautiful but gullible daughter, and these two representations are juxtaposed in working memory. Their interaction will. perhaps, lead to the enactment of an aversive emotion of anxiety which would suggest that bringing these two people together in real life should only be done with caution. Or, modeling the possible outcome of my having a fight with this male cousin might lead to the enactment of a pleasurable sensation, which would encourage me to challenge to this potentially dangerous character.

These examples are simplified, but perhaps communicate the idea that because of the SMM, secondary emotions will accompany the interaction of representations in working memory, and these emotions serve as a guide to interpreting the past and planning future social behaviors. These emotions enacted in response to imagined scenarios are presumed to influence our choice of future action. At the simplest level we are more likely to pursue a course of action which, when played-out by the SMM, leads to a gratifying outcome than we are to pursue a course of action that leads to an aversive outcome. And this is what I mean by strategic social intelligence: it involves the ability to decide between which alternative strategies to pursue on the basis of modeling social interactions and evaluating them by the brain sensing how the predicted outcome feels to us as emotions are enacted in our bodies.

Somatic marking is therefore the actual mechanism of ToM, and in this sense the SMM is the basis of 'mind-reading': the SSM is a mechanism for inferring what are de facto intentions, motivations and dispositions. Nonetheless the SMM could be considered almost the reciprocal of the common cognitive conceptualization of the ToMM. For example, hostility would not be represented directly as the hostile contents of another's mind, but instead as the reciprocal attribution of the feeling of 'fear'. A 'hostile' male stranger would actually be represented by the SMM as a 'fear-evoking' male stranger - an identity 'marked' by an emotion - with the inference that if we feel fear, then he is probably hostile.

Tactical and strategic social intelligence
The behavior of Damasio's patients with damage to the system that enacts secondary emotions is characterized by exactly this kind of poor judgment in interpreting and planning social behaviours, and other complex behaviors such as business decisions and gambling. The concept of strategic social intelligence requires further elucidation, both here and elsewhere. Strategy can be contrasted with tactics - strategy being long-term and tactics concerned with the here-and-now, strategy to do with general plans and tactics a matter of immediate responses.

It is presumed that the somatic marker mechanism is not used in tactical social interactions; since the demand for high speed responsivity dictates that behaviours are elicited by directly reading-off the meaning of overt behavioral cues. Modeling in working memory occurs over a timescale of hundreds or thousands of milliseconds, and deploying emotions in the body by the autonomic nervous system and hormonal regulation occurs over an even slower timescale of seconds or minutes. When engaged in face to face argument or flirtation, facial expression, gesture and language must all be minutely and rapidly responsive to the situation (over a timescale that is tens or hundreds of times more rapid than working memory) - and this kind of tactical social intelligence occurs by 'instinctive' and unconscious mechanisms. But planned and reflective social interactions, strategic social intelligence, depends upon mental modeling.

Of course strategic and tactical social intelligence will interact. In conscious animals, mentalistic ascriptions of the ToM type form a 'mind-set; in-place in advance of tactical interactions. Each mind set establishes a tendency for interpretation of cues. So that if we fear an individual on the basis of strategic modeling of their intentions, then we will tend to interpret tactical behavioral cues in the light of them being hostile. If, for instance, you have decided that our lustful cousin has designs to seduce your daughter, then his 'tactical' here and now behavior will be interpreted in the light of that assumption - we will interact in a suspicious and cautious manner, and perhaps with a greater tendency to aggression.

Many behaviors are intrinsically ambiguous, having a variety of possible meanings or being capable of being deployed in a dishonestly manipulative fashion. A gift from the lustful male cousin may be generosity or seduction, and which is decided on the basis of strategic modeling of his motivations, intentions and dispositions. This is one plausible interpretation of the chain of events leading to persecutory delusions - strategic social intelligence creates a mind set in which we falsely assume hostile intentions in a person, then ambiguous behavioral cues associated with that person are consistently misinterpreted during tactical interactions, and these misinterpretations serve to reinforce the false belief in hostility.

The evolution of consciousness probably occurred some time before the divergence of the human and chimpanzee lineages, so that modern humans and chimpanzees are both conscious; although human consciousness differs from chimpanzee consciousness due (mainly) to the addition of abstract symbolic language in humans. Whether consciousness and strategic social intelligence extend further throughout primate species, or to other social mammals (eg. elephants, dolphins), is a question which would need to be explored in the light of an understanding of the SMM. My hunch is that elephants and dolphins are both conscious, and capable of strategic social intelligence. Time will tell.

Humans are essentially social creatures
The nature of the SMM and the functioning of consciousness means that the conscious human world is essentially social. This is a matter of common observation. We are aware of people rather than things, most of our conversation is gossip about the doings of other people, our aspirations are usually related to love and lust while our worst fears usually take the form of threats from other humans. Aside from the times when problems of ecological survival are immediate and urgent - extreme hunger, discomfort or danger - we see the world through social lenses and pursue social goals.

Kummer has commented on the fact that high status, power and wealth are usually achieved for success in the social realm of human versus human competition; and a person's performance at ecological survival tasks is given much less prestige except insofar as it impinges on this social realm. So that many of the lowest status, poorest and most powerless individuals are those doing the 'most important' work of growing and preparing food, sanitation, rearing children, building and so on. At the same time the rich and famous are often people like politicians, managers, entertainers or (in other societies) soldiers - people whose relationship to the world of survival is at best indirect.

This is a consequence of the success of social animals in solving problems of survival. When group living animals have succeeded in developing effective strategies for obtaining food and shelter and repelling predators, then their main source of competition becomes the other members of the group. So, although human cooperation is what made us such as successful social animal, at the same time intra-species, between-person competition characterizes the human condition. To use the biological jargon, ee are both intrinsically altruistic and agonistic.


In considering the human condition there is much that humans share with other animals, many causes of pleasure and pain, survival or death. But there is also much human that is unique, and of these feature probably the most obvious is language.

It is clear that in some sense, 'language' is indeed unique to humans - although exactly what it is about language that is unique requires further definition. Many other animals communicate, a few have extremely sophisticated systems of communication. It seems that only humans have evolved a complex, abstract symbolic language which also has that feature that is biologically crucial about human communication - human communication is both complex, and capable of displacement.

Defining 'language' as displacement-communication
Displacement refers to the capacity of language to refer to entities and events that are 'displaced' in time or space - ie. not you or me, not here, or not now. This is the subject matter of much human language - we do not just talk of the here and now, but conversation ranges widely over reminiscence of the past, hope and fears for the future. But although necessary to allow this, displacement is not necessarily a part of complex communications systems. For example the bee dance is an abstract symbolic communication of spatially displaced information but occurring in an extremely simple and specific communications system. The dance is capable of transmitting information about where to find nectar in relation to the place of the dance (ie. displacement to another place) but that is pretty much all the dance can do.

But displacement in human language is built on top of an already extremely sophisticated social communication system we inherited from our ape and primate ancestors - a system based on facial expression, gesture, and a range of sounds. For example, even without language we could talk about 'my brother' and what he is doing at the waterhole. So long as my brother and the waterhole are both present they can be indicated by gesture, and so long as we restrict ourselves to what is actually happening her and now then the subject matter can be indicated by further gestures and body language, and our feelings about them indicated by facial expressions and vocalization. So if he was swimming, this could be indicated by pointing at him and the waterhole and miming the swimming. This is how many social animals communicate, and how humans may communicate with other animals such as dogs and with people who do not have language (such as children), or people with whom we do not share language. But without some system of displacement we cannot refer to my brother and waterhole in relation to another time or place, and we could not refer to any person who was not present here and now.

Displacement would inevitably involve some way of symbolizing my brother and the waterhole by creating an abstract referent (eg. this stone is my brother and the leaf is the waterhole; or this gesture, or this word), and by indicating the nature of the relationship between bother and waterhole (ie. swimming - which might be done by facial expression, gesture and vocalization for simple concepts). Brother, waterhole and swimming are a scenario. The act of displacement involves indicating that the brother and waterhole and the relationship are to be understood as being at another time or place - probably by linking the described scenario with indicators of another time or place.

Using displacement we might talk of my brother (even though my brother is not here at present), we might talk of my brother at the waterhole (although the waterhole is not in sight), and the fact that he is swimming. The scenario is one of my brother swimming at the waterhole, and to perform the displacement we might talk of my brother swimming at the waterhole last full moon (past), or the conjecture that he may swim at the waterhole next full moon (future). The scenario is displaced by associating the scenario (brother-waterhole-swimming) with an indicator of different time or place. Displacement requires merely that we can symbolize the full moon and indicate whether we are referring to the last one or the next one, and that this can be linked to the scenario. My suggestion is that displacement works by establishing such linkages to displace scenarios to other times, or places.

Displacement is necessary, but not sufficient, for the definition of language
Displacement is here taken to be the defining feature of language as contrasted with communication. Displacement is defining since displacement is an aspect of communication that could not - even in principle - be replaced by gesture, grunts, facial expression, body language and other non-linguistic communications. Displacement requires symbols, and symbolic communication must already have existed before displacement could have evolved. This implies that symbols can occur without displacement, and that we have inherited a symbolizing ability from our primate ancestors. Common chimpanzees and bonobos who have been trained to use large vocabularies of symbols in communicating with humans seem to have considerable abstract symbolic ability. So human ancestors already had an ability to symbolize and 'only' required to evolve the ability to perform displacement.

Displacement is therefore proposed as necessary, although not sufficient, for a communication system to be termed a language. In other words, there is a great deal more to language than just displacement; but without displacement, communication does not count as language. Broadly speaking, chimpanzee ability communication plus the capability of displacement equals what most people would term a full language - human-type language.

There should be a clear distinction between speech as a system of communication, and the existence of and displacement to constitute language. The highly restrictive definition of only displacement-communication as language proper means that most of verbal communication (even in humans) is not language, since most communication is potentially, in principle, replaceable by non-language communication. By this account, most of 'linguistics' is not about language, but about speech. And chimpanzee communication - although it may be very sophisticated, capable of complex instructions, and perhaps even have its own 'grammar' - would only be considered a full language if turns-out to be true that chimpanzee communication of social information is indeed displaceable. And the speech of children, or adults with mental handicap, however subtle a form of here and now speech-based communication, would only be considered to constitute a full language if the individuals were able to make functional use of displacement.

But it is important to point out that the forms of displacement by themselves, such as the use of words indicative of past and future tense, are not by themselves evidence of displacement. For example it is possible to 'parrot' grammatical forms such as past or future tense without any understanding of how to use them in practice. A demented person may use phrases indicative of false memories or imagined fears for the future - but these are socially (and biologically) non-functional. The test is that full language uses the forms of displacement in a functional way with real world applicability.

The idea that displacement is a distinctive and defining quality of human language is certainly not currently accepted. For instance, at present, many of the formal 'tests' of language ability (e.g. the test 'batteries' used by speech and language researchers) do not function as tests of displacement. For example, when a doctor asks a patient who has had a stroke to name a watch or a pen being presented to them, this does not count as a test of language. Indeed, there are currently no language tests designed specifically to measure the ability to perform displacement. Displacement is not a recognized key variable in language function.

Furthermore, current 'linguistic analysis' is essentially the study of speech or written communication. The discipline of linguistics does not make a distinction between the displacement functions that are unique to language and the use of speech that could, in principle, be replaced by expression and gesture. This emphasizes the distance that the discipline of linguistics need to develop in order to become properly integrated as a biological science.

Why displacement evolved - role and adaptive benefits
The benefits of displacement may seem obvious, since it gives access to knowledge of social events that are remote in space of time, but since displacement does not seem to have evolved in other social primate species such as common chimpanzees, bonobos, gorillas, orang utans and baboons - it needs further explanation. If displacement is useful for humans, why has it not (or not obviously) arisen in other primate species?

The assumption that there is an adaptive reason why displacement evolved, and that language is not an accidental by product of some other adaptation, must be justified. The adaptive role of displacement is strongly suggested by the social intelligence perspective which sees language as primarily concerned with the communication of information about human beings and their doings. It should also be borne on mind that human brain is difficult to grow and develop, and is metabolically very expensive to maintain, in other words brain is a very costly tissue. This implies that there must be considerable benefits to offset the costs of substantial neural construction such as was required to support the advanced functions of language.

Almost any level of brain damage to the more recently evolved parts of the cerebral cortex will impair language function. Even when the actual production of grammatical speech is apparently unimpaired (as in non-dominant lobe lesions or frontal lobe lesions) so that the brain damaged subject can perform purely linguistic tests at a normal level; the actual applicability of language to social situations is almost always impaired. Close study of most patients with any significant degree of brain damage will usually reveal that factors such as appropriateness, prosody (ie. the rise and fall of intonation), or use of metaphors are impaired - in other words the social function of language is impaired. Language function certainly appears to depend on brain function, and indeed on sustained function of most of the brain. The first assumption is that whatever the reason for the evolution of displacement - it is social. This is in line with the social intelligence assumption that recent human evolution has been driven by social selection pressures. The social assumption reshapes the question about displacement into asking what is was about ancestral human social organization that made displacement so useful, and how these features differed from other related ape species which did not evolve displacement as part of their communication systems. This argument is, of course, based upon only a few species of primates, and (like almost all scientific theorizing) contains some elements of post hoc circularity, but this does not mean that the theory is untestable.

Even when methods of testability are not immediately obvious, once a scientific theory has been described in a clear and explicit fashion, ways can usually be found by which its novel consequences may be put to the test of observation and experiment.

Constraints on the evolution of displacement-language
In a conscious social animal without language (such as a chimpanzee) the SMM enables the modeling of differential social identity together with somatic markers to represent disposition, motivation and intention. The SSM can form combined perceptual-emotional representations (implicitly symbolic) such as 'that angry, aggressive male who hates me'. The SMM therefore allows abstract, symbolic thought - thought in which representations can interact and be manipulated. The addition of language to consciousness augments this combined social-emotional representation with further displacement-markers indicative of other times, other places, other persons.

The specifically socially-adaptive nature of language is supported by evidence from spontaneous language usage (most of which constitutes 'gossip' concerning the doings of other people), neuroanatomical correlations between regions concerned with language and social intelligence, and by temporal and genetic informational constraints on human evolution. Conservatively estimated, there has been only around 5-6 million years since divergence of the human lineage from that of chimpanzees, and less than 2000 genes (ie. under 2% of the genome) differ between humans and chimpanzees. This amount of DNA has been estimated to contain c.35 000 (35K) bits of useful design information - which is not much: certainly not enough to code for something as complex as the whole of the human communication systems.

And although the frontal lobe of the brain has expanded substantially since our lineage diverged from that of chimpanzees, the brain substance itself does not appear to changed qualitatively in its structure. There is no obviously different new cortical region which has been added to the chimpanzee brain in order to perform the function of language - the human brain just looks like more of the same. It seems as if humans have merely evolved more of the same kind of brain stuff as was already present in the ancestors we share with chimpanzees - a relatively quick and easy thing to evolve, since it merely requires a few genetic mutations to instruct the body to 'make more of this', and to 'wire-it -up' in such a fashion.

These constraints make it likely that the evolution of human linguistic capacity was largely dependent upon pre-established neuroanatomical circuitry, and the evidence on the specifically socially-adaptive nature of language means that the neuroanatomical circuitry of language is very probably the systems which evolved to subserve social intelligence - in other words, working memory and the somatic marker mechanism. The constraints of limited evolutionary time also imply that the extra computations required for language processing were relatively simple - the computations are probably of the same kind as those performed in the primate frontal cortex of other species - the difference is in the connectivity between the computational areas, particularly the addition of extra levels to the hierarchy of convergence and integration.

The extra power of the human brain may be a matter of greater integration. For example the visual system seems to have evolved by new brain regions sampling more different aspects of the visual information generated by the retina, and bringing together these different aspects in new syntheses to extract more and more information from the same initial signal. In other words the human brain is pretty much a bigger chimpanzee brain, with most of the extra brain at the front.

Displacement, group size, and the sexual division of labour
So why did displacement evolve? If the common chimpanzee is taken to be closely similar to the human ancestor of five million years ago, and if we assume that the bonobo also evolved from something very like a common chimpanzee, then a plausible scenario can be constructed. Both chimpanzee species exhibit extremely sophisticated 'here and now' tactical social communication by facial expression, gesture and verbal signaling - but apparently this communication relates only to individuals and circumstances that are currently present. The question is: under what ecological circumstances might a great ape benefit substantially from an extra ability, the ability to do something more than this and to communicate about individuals not present and events not now?

My suggestion is that displacement-language evolved for two reasons. The first reason is that social groups became sufficiently large that unique identities were required to keep track of and refer to individuals. Humans inhabit large social groups, compared with chimpanzees, and perhaps this led to an enhanced ability to use abstract symbols to refer to individuals since it would not always be possible to indicate individuals by gesture. Bonobos seem to have abstract symbolic ability to a higher degree than chimpanzees - which fits with the fact that their social groups are much larger. As well as the symbolic ability of chimpanzees and bonobos trained to use abstract geometric shapes to communicate with humans, examples of symbol use have been recorded in natural conditions - for example the young bonobo that used a log of wood as a 'doll' to play with exactly as if it were a baby; or the way in which bonobos communicate the need to move on to a new camp by dragging a large branch around the troop and showing it to each member - the branch seems to serve as a symbol (presumably learned) of the need to move.

But for displacement to be useful, as well as a large group size it may also be necessary that the group splits up for significant periods before being reunited. So the second selection pressure favoring displacement would be division of labour - tasks dividing the group for significant but temporary periods. Given that language (it is assumed) evolved for the communication of information on other people, splitting of the larger group into smaller groups for significant periods would mean that communication of information about other members of the group who were not present would require displacement. The advantages of this are that individuals and their behaviors can be evaluated in their absence - for example information could be gathered concerning the suitability of a potential mate, or rival.

Humans under ancestral conditions exhibited exactly such a sex-based division of labour. Ancestral groups were nomadic foragers, and these groups would have split-up frequently and for hours or even days at a time, due to the sexual division of labour - men going off to hunt while women remained near the camp, gathering vegetable food and looking after the children. My idea is that this splitting up of the troop would have provided a strong selection pressure to favour those people who could talk about others even while they were not present; to 'gossip' in order to understand their personality, interpret their past behaviors and predict their next moves.

For instance, one plausible scenario is that displacement-language may have evolved initially among women for exchanging information about the absent men, in order to evaluate potential mates and discover more about the behavior of the males of the family. Females could exchange knowledge (knowledge which might, of course, be biased or deceptive) concerning the absent males. Men are often the topic of conversation among women under such circumstances today. By contrast while displacement would certainly enhance the planning of hunting in principle, such planning is clearly not essential to the activity of hunting. Many, many animals including common chimpanzees hunt effectively in groups without the need for displacement-language, and hunting does not appear to be a selection pressure for displacement ability. And men who are hunting often speak little with one another.

So this scenario suggest that displacement-language evolved initially primarily to benefit women in exchanging information about men who were away hunting, but this language ability was also inherited by men since most inherited traits are shared and displacement would also have benefited men albeit is a secondary fashion. This speculation fits contemporary evidence of higher level linguistic ability ('verbal intelligence') in women, a high frequency of spontaneous language use among groups of women, and the observation that the subject matter of private women-to-women conversations if often focused on the subject of men.

I find the story both plausible and attractive, but inconclusive. It could be tested by further study of communication and language use in chimpanzees, and especially bonobos. Although bonobos do not have sexual division of labour, their groups are large (many dozens of individuals) and since bonobos live in jungle they would not be able to keep all group members under observation. This would be a selection pressure for some degree of displacement, and might explain the greater linguistic ability of bonobos relative to common chimpanzees.

Displacement in WM depends on sufficient spatial capacity for complex representations
The radical view put forward here is that there are no recently evolved specialized 'language centres' in the brain, but that instead displacement-language has been made possible by a quantitative expansion of the functional capability of working memory, on top of the already-evolved and pre-existing somatic marker mechanism.

This view is in contradiction to a vast amount of linguistic, neurological and evolutionary speculation which is based on a different conceptualization of language (a conceptualization which does not, for example, define displacement as the crux of language, and which does not clearly differentiate language from speech). But when the concept of language is built up in this step-wise and evolutionary fashion, by considering the somatic marker mechanism assumed to be present in chimpanzees and bonobos and adding the capacity of displacement, then it becomes plausible that the evolution of language may be a much simpler and straightforward matter than usually believed.

The basic, underlying principle of displacement of social relevant information could be the interaction of representations created by the somatic marker mechanism with further markers for displacement. What is envisaged is that - for example - a representation such as 'aggressive-male cousin' is a combined perceptual-emotional representation, comprising (at least) two representations that have been combined in WM. There is a representation of the perception of a specific person (the male cousin) and a representation of the associated emotional state from which infer that the male cousin is aggressive. To displace the representation of aggressive-male cousin requires nothing more than to create an association with a further symbolic marker which represents another time or another place.

So, this entity of aggressive-male cousin contains both perceptual information on individual identity and the information to trigger a specific emotional response. This perceptual-emotional representation may be associated with a symbolic marker that represents (for example) a temporal displacement such as 'tomorrow morning', or a spatial displacement such as 'water hole', or a symbol for another person (including the emotional response) such as 'my younger sister'. The process is one of incremental expansion and association of information in working memory, whereby representations are 'superimposed' and loaded with more and more information - both perceptual and emotional.

A large capacity working memory can therefore create very complex representations, and these representations can enable displacement - or indeed some other extra complexity of information. Incremental expansion of working memory over an evolutionary timescale is biologically plausible, and would enable progressively more and more of these iterative associations to be accumulated within the capacity of WM.

Limits of working memory
What are the limits of working memory, what defines how much it can carry? As working memory is essentially an anatomical space where representations (in the form of three dimensional patterns of neural activity) are sustained, then the major constraints on the capacity of WM are set by the complexity of representations that may be sustained, and by the duration that representations may be sustained. In other words, working memory size is defined by the capacity of WM (how complex a representation it can accommodate) and by the timespan of WM (how long the representations can be sustained).

Capacity is probably constrained by the size of the brain devoted to WM (the more neurons in WM, the greater the complexity of representation it can enact). Timespan probably constrains the number of representations that may be kept active simultaneously - along the lines of 'Miller's magic number' (known to all Psychology undergraduates) which suggests that a maximum of from 5-9 items can be sequentially loaded and retained in working memory at one time. The exact number is less important than the fact that there is some such temporal limit on the timepan an item can be kept active in WM.

In evolutionary history the capacity and timespan of WM would presumably have varied between species, according to evolutionary constraints - some species being able to retain items over a greater timespan, and other to enact larger and more complex representations than others. Some evidence from maze tasks suggests that humans and rats do not differ substantially in the timespan of working memory - and it is likely that the very great size of the dorsolateral prefrontal cortex implies that the special thing about human working memory is the size and complexity of representations that it is capable of sustaining (rather than a particularly large capacity for sequential loading with items). This is speculative, but it seems plausible that the large volume of the anatomical substrate of human memory evolved I order to allow complex (hence large volume) cognitive representations to be enacted.

According to current evidence, it seems that the WM of a common chimpanzee is insufficient to support the displacement of social information; although it is possible that specific individual chimpanzees with exceptionally large WM capacity may be able to perform displacement. It might also be that chimpanzees can perform displacement on tasks which are computationally simpler than social intelligence. As suggested from the work of Sue Savage Rumbaugh, bonobos may be much nearer to possessing a human-like language ability - which implies that they should have a larger capacity WM than the common chimpanzee. Clearly these questions would require specific exploration.

But the assumption here is that the chimpanzee working memory has the same nature and functional capacity as the human WM - the only important structural difference relates to size. The idea is that the evolution of the very large human prefrontal cortex was driven by the advantages of expanding working memory, and the primary function that was served by the expansion of WM was displacement-language - the ability to form complex associations not just of social identity and emotion, but also markers of other times and places. Such representations containing information on persons, emotional reactions to these persons and also markers of other places and/ or times would, presumably, be highly complex and large, requiring a greater capacity from WM.

It should also be emphasized that the neural substrate for displacement is not specific to displacement. The expansion of WM over human evolutionary history has a very wide range of other consequences, since it enables greatly increased complexity of all types of cognitive modeling - an increase in what many people would consider to be 'general intelligence'. Expanded WM also enables other types of complex grammatical construction such as representing contingency, and performing many layers of embedding of clauses.

In summary, the assertion is that the selection pressure for the expansion of WM capacity occured in order to enable displacement of language - that was its selective driving force - and therefore that enabling displacement is the adaptive consequence of expanded WM. The other consequences of expanded WM capacity such as many aspects of 'general intelligence', although perhaps more obvious under modern conditions, are epiphenomenal by-products when viewed from the evolutionary perspective.

Some consequences and predictions
To put it crudely but reasonably accurately, a human brain may be pretty much a chimpanzee brain with a larger working memory - (plus some motor and auditory specializations to enable speech). The larger size of human working memory means that bigger and more complex patterns of nerve activation (cognitive representations) can be accommodated, and these representations can become so complex as to include information on social identity, emotion and displacement. General-purpose human intelligence, as applied to the vast range of cultural activity, is an accidental consequence of the adaptive benefits of social intelligence - especially the somatic marker mechanism and the need to displace social information. The enhanced human ability to symbolize may also be a consequence of expanded WM - so that very complex representations can then be further linked to abstract perceptual entities (such as words).

Presumably symbolization is made possible by association of representations occurring in working memory. A perceptual representation interacts with an emotional representation to form a bridging representation that links both. So the bonobo links the perceptual representation of a baby with a particular baby-sized piece of wood, or the need to move camp with the dragging of a branch. The symbol also links the relevant emotions (evoked by a real baby, or a need to move) to the symbolic use of a particular shape and size of wood. With the doll, a specific piece of wood acts to trigger a mental representation which both recalls a baby, and also evokes the emotions appropriate to a baby. The wood is an effective symbol of the baby because it stands both for identity and emotion.

1. Enhancing chimpanzee working memory
This relatively simple scheme seems to include - in outline - many of the features a biological description of language would require. One way to test it would be somehow to increase the working memory of a chimpanzee, and see if this brought its communication ability towards that of humans. It is possible that this amplification of working memory is exactly what has been achieved by teaching chimpanzees the use of visual symbols on a board - each symbol is a convenient 'chunk' of the perceptual features of that to which it refers. Each symbol is 'stored' on a board for reference which leaves more of the chimpanzee working memory free to manipulate and integrate other representations. Symbol boards can be seen as an indirect method for amplifying chimpanzees working memory rather as humans do by writing or using counters to calculate.

Certainly it seems that the better a common chimpanzee or bonobo can master a symbol board the more human-like their language becomes, and it does not look too unlikely that Doctor Dolittle's desire to 'talk to the animals' may have been achieved by Sue Savage Rumbaugh and her bonobos.

2. Humans with low capacity WM may lack language (ie. lack displacement)
If displacement-language depends on WM in a quantitative fashion, then this has several other testable consequences. Humans who have WM capacity below a critical threshold, perhaps due to brain trauma, disease, congenital brain damage or other forms of intellectual handicap, would be expected to display redued WM capacity and also lack displacement language - even when they have speech and the ability to communicate theory of mind information. Once again it is important to emphasize the difference between speech and language.

So that - for example - the prediction is that mentally handicapped individuals with Williams Syndrome, who are supposed to have remarkable social and 'linguistic' abilities, would be found on specific examination to lack the ability to perform functionally efrfective displacement with language. Displacement is not - of course - merely a matter of being able to use the appropriate grammatical forms such as 'tense' - that could be achieved by merely 'parroting' (ie. repeating without understanding forms heard elsewhere). But the displacement-language must also be adaptive, must be appropriate to the social situation, and refer to real world events.

3. The social structure of language
The somatic marker mechanism is not restricted to social information, because working memory is not restricted to social information (which is not a distinct module - but is composed of projections from a wide variety of processed perceptual data as well as information on body states). This means that these associations are not confined to social, temporal and spatial information; but may be used to relate any kind of associations. Hence although the adaptive function of consciousness is specific to social intelligence, the mechanisms are general-purpose; and non-social information can use the SMM processes as an accidental by-product of brain connectivity.

Language, by its location in working memory, serves as a 'translation' device by which non-social domains of knowledge can gain access to the cognitive apparatus that evolved to deal with social intelligence. Working memory is an association mechanism, and any entities that can be deployed in WM can be associated with one onether. Presumably this is why humans can use their social intelligence to reason about non-social matters such as technology and natural history, by using the SMM in an expanded-capacity working memory.

In effect humans use social intelligence as a system for generating analogies, so that different classes of proposition are processed as if they were social problems. Much of high level human intelligence can be considered as analogical; a system in which SMM works by 'anthropomorphizing' non-social topics as if they were stories about intentional agents. I discuss this further in the chapter on creativity.

This capacity to 'over-learn' new topics onto the somatic marker evaluation-system of social intelligence has proved to be the crucial factor in the development of 'symbolic' human culture.

Language and the human condition
This book is only about language insofar as language is a major element in human nature - language is an aspect of what Bronowski called 'human specificity'. Language is a big part of what makes humans distinctive.

The current prevailing view of the nature and structure of language is dominated by the assumption that language is a general purpose specialization, by contrast the social theory of language assumes that language evolved for the purpose of communicating social information ('gossiping' as Dunbar terms it). The concept of displacement-communication is an incomplete evolutionary and neuroscientific account of full language. In particular, the above scheme distinguishes language from speech, and says nothing of the anatomical, motor and perceptual specializations necessary for verbal communication of language. However, the above view assumes that most of what linguists call 'language' is not biologically distinct from ,other forms of verbal and gestural communication. Only displacement-language is biologically distinct from other forms of here and now communication.

This view also overturns the idea of traditionally defined 'language areas'. Brain areas such as Broca's and Wernicke's areas are actually concerned with speech - with verbal communication rather than language. For example Broca's area is concerned with fine control of motor systems (including those involved in the articulation of speech), while Wernicke's are is probably concerned with specializations to the sense of hearing and verbal monitoring that evolved along with the evolution of verbal communication.

The remarkable sureness and rapidity of human language acquisition is seen as a consequence of the human drive to communicate combined with the gradual maturation of the central nervous system. There is no evolved 'language acquisition device' which makes us learn language. The drive to communicate is based upon our fundamental nature as social animals, and the immediate advantages a child gets from its ability to communicate social information. Normal children communicate as soon as the maturity of the nervous system allows them to do so - and it is the maturation of the nervous system which times the stages of linguistic development. Common observation shows that in a social milieu a child wants to communicate social information, and tries to talk because it is so useful to the social environment. Learning to talk happens when the physical apparatus of speech is mature, and when the working memory capacity has grown to a sufficient size. Displacement is the last major attribute of language to occur in a normal developing child (at about the time when myelinisation of the central nervous system is complete) and displacement presumably happens when working memory is large enough to deploy the complex representations which displacement requires.

The social theory of consciousness and language also makes some predictions about the structure of language - its grammar. It predicts that much of the structure of language derive from social intelligence, in other words from the structure and operation of the somatic marker mechanism. It is possible that social entities (intentional agents such as persons) and the nature of social interactions (as represented by the SMM) might constitute some of the fundamental categories of language. These speculations lie beyond the scope of this book, but will be pursued elsewhere.

The crucial point about the dependence of distinctively human intelligence upon the somatic marker mechanism is that, because it is based upon social categories and driven by social motivations, even our abstract thought world is saturated with emotions, preferences and aversions, pleasures and pain. At a deep level, this is why humans are able to care about that accidental and artificial product of human invention that we call culture.

Further Reading and References

Evolution of awareness
The nature and role of awareness depends upon Francis Crick's The astonishing hypotheses - although what Crick terms 'consciousness' I would term awareness. I found Zeki's A vision of the brain extremely useful in understanding the general organization of the brain as revealed by vision researchers. I have also drawn heavily upon the work and conversation of my colleagues in the Psychology Department at Newcastle University who are active in visual research and cortical connectivity - especially Malcolm Young, Jack Scannell, Martin Tovee and Piers Cornelissen.

Evolution of consciousness
The key to understanding the nature of consciousness is the somatic marker mechanism, as described by Antonio Damasio in Descarte's Error. But while Damasio describes many of the things that consciousness does and what happens when consciousness is absent (in patients with brain damage), he does not say why consciousness evolved. This is my concern here - to describe what was the social selection pressure for evolving consciousness, and what was the specific nature of the social task performed by this general mechanism. So, the neural nature of the SMM needs to be combined with an understanding of the nature of strategic social intelligence.

The somatic marker mechanism
I would consider Damasio's concept of the somatic marker mechanism to be one of he most fruitful ideas I have encountered. It is, however, a difficult concept to grasp on paper (although easier to explain verbally) which may explain its otherwise astonishing lack of major impact to date. One difficulty is that the SMM overlaps conceptually with 'consciousness' and 'theory of mind' - which are also slippery and elusive ideas. I have striven to bring some terminological clarity and precision to these matters, since the concepts themselves are not so difficult as the terminological tangle makes them. I have equated the SMM with consciousness, suggested that the function of consciousness is strategic social intelligence, and that theory of mind is based upon the SMM - but with the addition of abstract symbolic language in humans which introduces a new set of possibilities for what is represented.

Language is another terminological and conceptual minefield. This chapter was immeasurably helped by discussions with Tina Fry over about two years. Tina is trained in linguistics, and we are working together on the evolution of language. Most of the new ideas relating to the concept of displacement - the crucial aspect of displacement, the mechanism of displacement by association in WM, the putative ecological conditions for language evolution and so on - were developed in the course of discussions with Tina, and as a consequence of her knowledge. She has taught me a tremendous amount, and by-and-large eliminated some of my grossest errors - but there are a few points at which I have defied her advice, and for which I take full responsibility.

Adolphs R, Damasio H, Tranel D. & Damasio, A.R. 1996. Cortical systems for the recognition of emotion in facial expressions. Journal of Neuroscience 16, 7678-7687.

Baron-Cohen, S. 1992. Autism: a specific cognitive disorder of 'mind-blindness'. International Review of Psychiatry. 1990, 81-90

Baron-Cohen, S. 1995. Mindblindness: an essay on autism and theory of mind. Cambridge MA: MIT Press.

Barton R.A. & Dunbar R.I.M. 1997. Evolution of the social brain. In Machiavellian intelligence II: extensions and evaluations. (ed A. Whiten, R.W. Byrne). Cambridge University Press.

Bechara A, Damasio A.R., Damasio H. & Anderson S.W. 1994. Insensitivity to future consequence following damage to human prefrontal cortex. Cognition 50, 7-15.

Bechara A, Tranel, D. Damasio, H. & Damasio, A.R. 1996. Failure to respond automatically to anticipated future outcomes following damage to prefrontal cortex. Cerebral Cortex 6, 215-225.

Bechara A, Damasio, H. Tranel, D. & Damasio, A.R. 1997. Deciding advantageously before knowing the advantageous strategy. Science. 275, 1293-1295.

Bradshaw JL. 1997. Human evolution: a neuropsychological perspective. London: Psychology Press.

Byrne, R.W. 1995. The thinking ape: evolutionary origins of intelligence. Oxford University Press.

Byrne R.W. & Whiten A. (ed.). 1988. Machiavellian intelligence social expertise and the evolution of intellect in monkeys, apes and humans. Oxford: Clarendon Press.

Carruthers, P. & Smith, P.K. (ed.). 1996. Theories of theories of mind. Cambridge University Press.

Charlton, B.G. 1995. Cognitive neuropsychiatry and the future of diagnosis: a 'PC' model of the mind. British Journal of Psychiatry 167: 149-153.

Charlton, B.G. 1997. The inequity of inequality: egalitarian instincts and evolutionary psychology. Journal of Health Psychology, 2: 413-425.

Crick, F.C. & Jones E. 1993. Backwardness of human neuroanatomy. Nature. 361, 109-110.

Crick, F. 1994. The astonishing hypothesis: the scientific search for the soul. London: Touchstone.

Crick, F. & Koch C. 1995. Are we aware of neural activity in primary visual cortex? Nature 375, 121-123.

Damasio, A.R. 1994. Descartes' error: emotion, reason and the human brain London, Macmillan.

Damasio A.R. 7 Damasio, H. 1994. Cortical systems for retrieval of concrete knowledge: the convergence zone framework. In Large scale neuronal theories of the brain (Ed C Koch). MIT Press, pp 61-74.

Damasio, A.R. 1995. Towards a neurobiology of emotion and feeling: operational concepts and hypotheses. The Neuroscientist 1, 19-25.

Damasio, A.R. 1996. The somatic marker hypothesis and the possible functions of the prefrontal cortex. Philosophical Transaction of the Royal Society of London. B. 351, 1413-1420.

Dunbar, R. 1996. Grooming, Gossip and the Evolution of Language. London: Faber.

Foley, R. 1993. Causes and consequences in human evolution. Journal of the Royal Anthropological Institute (new series) 1, 67-86

Foley R. 1995. Humans before humanity Oxford: Blackwell.

Frith, U. (ed.) 1991. "Asperger and his syndrome." Autism and Asperger Syndrome (ed U. Frith). Cambridge University Press. Gazzaniga, M.R. (ed.) 1995. The cognitive Neurosciences. Cambridge, MA: MIT Press.

Goldman-Rakic, PS. 1995. Cellular basis of working memory. Neuron 14, 477-485.

Humphrey, N.K. 1976. The social function of intellect. In Growing points in ethology (ed. PPG. Bateson & RA. Hinde). Cambridge University Press.

Mithen S. 1996. The prehistory of Mind. London: Thames and Hudson

Pinker, S. 1994. The language instinct. New York: William Morrow.

Seyfarth, RM & Cheney DL. 1994. The evolution of social cognition in primates. In Behavioural mechanisms in evolutionary ecology (ed LA Real). University of Chicago Press, pp 371-389.

Shallice T. (1988). From neuropsychology to mental structure. Cambridge, Cambridge University Press.

Tovee, M.J. How fast is the speed of thought? Current Biology. 4, 1125-1127.

Whiten A. & Byrne R.W. (ed.) 1997. Machiavellian intelligence II: extensions and evaluations. Cambridge: Cambridge University Press.

Wierzbicka, A. Cognitive domains and the structure of the lexicon: the case of the emotions. In Mapping the mind: domain specificity in cognition and culture (ed. L.A. Hirschfield & S.A. Gelman.) Cambridge University Press, pp431-452

Worden, R. 1995. A speed limit for evolution. Journal of Theoretical Biology 176, 137-152.

Worden, R. In the press. The evolution of language from social intelligence. In The evolution of phonology and syntax (ed Kennedy & Knight). Cambridge University Press.

Young, M.P. 1992. Objective analysis of the topological organization of the primate cortical visual system. Nature 358, 152-154.

Young, M.P., Scannell, J.W., O'Neill, M.A., Hilgetag, C.C., Burns G. & Blakemore, C. (1995). Non-metric multi-dimensional scaling in the analysis of neuroanatomical connection data and the organization of the primate cortical visual system. Philosophical Transactions of the Royal Society of London. B. 348, 281-308.

Zeki S. (1993). A vision of the brain. Oxford, Blackwell.

Bruce G Charlton MD
Department of Psychology
University of Newcastle upon Tyne



also by Bruce Charlton
Peak Experiences
Cargo Cult Science
The Malaise Theory of Depression
Delirium and Psychotic Symptoms
Public Health and Personal Freedom
Psychiatry and the Human Condition
Psychopharmacology and The Human Condition
Injustice, Inequality and Evolutionary Psychology