To appear in Consciousness edited by Max Velmans, Goldsmith's College, London.


Social Construction and Consciousness

Rom Harré


There does not seem to be the 'problem of consciousness', only various conundrums that come to the fore when we reflect on the special character of higher organisms. Roughly one might say that common to many puzzles about consciousness is the thought that human beings, at least, are aware that they are aware of material things, some of the states of their own bodies, and can monitor themselves thinking about these and other matters. In this paper I propose to use the analysis of certain relevant kinds of talk as a way of investigating how far some of the phenomena comprehended under the broad term 'consciousness' are socially constructed. By that I mean the question as to whether they are brought into being, or if already existing, are they modified in various ways, by social, that is interpersonal processes and structures. I shall work with the widely shared principle that the main instrument through which such creations and modifications come to be is language.

Analyzing consciousness talk

It is a useful, indeed indispensable part of the opening up of a research programme in psychology to do a little lexicography. Phenomena are made available to people, picked out from the 'hurly burly' of life (as Wittgenstein put it) by perceptual skills that are in part made what they are in the learning of a language. The rules for the use of a word can give us an insight into the domain of its application. Thus the word 'conscious' and its cognates, though late entrants into English, are now well established, so well established that a group of people can be brought from all over the world to study what ever it is that the word 'consciousness' picks out. In doing a lexicographical exercise as a preliminary to defining a research programme we must bear in mind Wittgenstein's reminder that the diverse phenomena that are found within the domain of the words 'conscious' and 'consciousness' may not be unified by a common essence. He thought that many ways that a common word is used is best pictured as field of family resemblances, structured by an ever shifting network of similarities and differences in the way the words at the focus of an investigation are used. While bearing this point in mind must be careful not to slip into the post-modernist exaggeration, that there is no firm ground of usage. On the contrary symbol driven human life is possible only because among the shifting patterns of usage of words there are persisting invariants that serve generations of the users of a language to conduct their everyday lives. We must agree, as Wittgenstein remarked, in a grammar, though we must acknowledge that no grammar lasts for ever.

    Several of the everyday uses of the word 'conscious' can be replicated by phrases making use of the words 'aware' and 'attend'. Thus the content of 'I was conscious of a disapproving atmosphere at the meeting' can be rendered by 'I was aware of a disapproving atmosphere at the meeting' while the content of 'I suddenly became conscious of a ticking sound' could be expressed with the phrase 'paid attention to'. This group of words are characteristically used to express a relation between a person and an intentional object. Their use presupposes an ontology in which the basic particulars are material things some of which are animated as persons. Examination of a larger slice of 'aware of'/'attend to' discourse would disclose a structure in which a person served as the 'origin' of an array of entities of various sorts, patterned by the 'conscious of' relation. Such structures are typically continuously being transformed around fairly stable invariants, as one takes a walk, has lunch and so on.

    But this field of family resemblances includes the abstract noun 'consciousness'. I am of the opinion that the appearance of nouns in the setting up of a research programme in psychology is a danger sign. If psychology is the science of what people do then verbs and adverbs ought to carry the burden of descriptive psychology rather than nouns, and especially abstract nouns. Once a word like 'consciousness' is given a central role, then we are strongly tempted to begin to look for a referent for it. 'What is consciousness?' then takes on the air of a serious question, answered by examining it closely, like a question put to a physicist, 'What is heavy water?'. The fact that we cannot answer it readily makes it look as if the question is a hard one which could be answered with little more effort. But perhaps there is something wrong with the formulation of the question, drawing our thoughts off in a fruitless direction.

    Let us return to our linguistic preliminaries. What do we use the word 'consciousness' for? Two uses stand out. One is simply a further candidate for replacement by one of the 'aware of'/'attend to' vocabulary. It is exemplified by such expressions 'rose to consciousness' or 'drifted into consciousness', roughly equivalent to 'became aware of' or 'noticed'. Then there are uses such as 'lost [regained] consciousness'. Here they seem to refer to the existence of an indefinite range or domain of perception, numerous intentional objects of which someone is aware. Closely allied to this is the use of words like 'awake', said of one who is at that moment aware of the state of the material environment both inside and outside the person's body. To wake up or to regain consciousness is to come to be in such a condition as to be able to become aware of all sorts of things. But it is not just to be aware of this and that state of affairs, but it also suggests that a certain enabling condition for being aware of intentional objects, now or once again obtains in the organism itself.

    Where we were confronted with an abstract noun, used very broadly, we now have something complex but quite concrete. A person as an embodied being, is in a certain state a state which is a necessary condition for coming to perceive things, that is to stand in a certain relation to them.

Prescribing consciousness research

To whom should further explorations of consciousness be entrusted? I shall argue that the study of fields of intentional objects should be the domain of discursive psychologists, while the study of the enabling conditions for the existence of such fields is the province of neuropsychology, exploring the state of the brain and nervous system characteristic of those who are awake, paying attention and so on. The first of these suggestions might look very much like a take-over bid by phenomenologists, and to a certain extent that is just what it is. The legitimacy of trying to convey to another person the nature of one's private experience by describing it needs to be defended. The second of these suggestions might look very much like a capitulation to a reductionist programme in psychology, in which seemingly mental state terms are given a new meaning as having a neurophysiological reference. Since that is not my intention I owe the reader at least a sketch of how a phenomenological vocabulary and a neurophysiological vocabulary could be related so that the one does not displace the other.

    Let us begin in the phenomenological mode. On what are perceptual domains centred? They seem to be laid out in different orientations to a person. This person is a singularity around which what is perceived is arrayed. To explore the nature of that singularity we enter into the study of selfhood, or at least of one aspect of this multivocal notion. That aspect is the sense each person has of the uniqueness of their location and trajectory through space and time, of being one and only one embodied person.

    Here is a problem: do we perceive an ordered world, each from his or her own point of view because we each have an original or native sense of our own singularities? Or do we each have such a sense because we perceive the world in an ordered and centred way? To encounter the structure of the perceptual domain of another person, which will include many of the same objects that one can oneself perceive, one can place one's body where the other person has been very shortly beforehand. Soon it is enough to imagine what the view from somewhere else would be like. But what guarantees the authenticity of the experience as that of the perceptual domain of someone else? It can be from what they tell us about what they saw, or felt, or heard. But by what devices do we distinguish a description from the point of view of the speaker from a bland, neutral description? The key, I argue, is in our grasp of the grammar of the first person, explicit as in 'I saw the postman go by', or implicit as in 'Can't you hear the phone?'

The expression of consciousness as awareness of something

What sort of display would be an expression characteristic of someone or some creature that was conscious? It might, for instance, be a report of how things looked from the point of view of the speaker. This is done by using the first person singular in the report. The point of view from which the world is being seen, felt, heard etc. is implicitly centred on the body of the speaker.

    Using the first person in such a report also accomplishes a more complex social act, namely the taking of responsibility for what has been reported. In saying 'I can see the taxi at last' an anxious member of a travelling party not only indexes the content of the statement with his or her position in space as an embodied being, but also puts his or her character 'on the line', standing behind the report, so to say. If the plume of dust turns out to have been made by a herd of antelope and not a taxi then the reporter is liable to censure, not just for getting it wrong but for raising everyone's hopes falsely. A parrot learns to imitate the sounds of speech and croaks out 'I can see the taxi now'. There isn't one. Do we censure the parrot? Not usually. A parrot does not partake of enough of the human form of life to be liable to praise and blame for its perceptual reports.

    That of which a person is conscious, that is, is aware of and/or attends to, has a structure, centred for each person, on his or her body. The fine tuning of the perceptual centring of consciousness, in this sense, will depend on many factors, including whether the person is blind when the field of awareness is tactile and auditory, or perhaps is a patient of Oliver Sachs and suffers from some curious psycho-anaesthesia. People speak and write of the 'structure of consciousness' and most of the time they simply mean the structure of the local environment, as revealed to the visual, auditory, tactile etc. perceptual systems. The discursive expression of that structure, as contrasted say with its expression in a landscape by the use of perspective, is usually revealed in the use of first person constructions, rather than neutral descriptions. Compare 'Look! There goes Sadie' with 'I can see Sadie!' Agentive activities, such as making a shot in billiards, could also reveal the person centred structure of the environment as perceived by the player.

    Radically different uses of the first person might persuade us that the structure of consciousness was different for some person, even perhaps for some tribe if the use of an exotic grammar (from our point of view) was widespread. Where does madness end and cultural diversity begin? Only some radically different uses of first person indices would count. And just how far these uses could diverge from the customary ways of speaking is limited. In the case of people displaying multiple personality syndrome we find no evidence in their off-beat grammars of disembodied or diverse locations for perceiving the ordinary furniture of the locality. For example there is no reason at all to think that Miss Beauchamp's or Eve White's perceptual experiences were ordered in any abnormal way. As Mary Watkins expressed the matter:

It is paradoxical that the illness (sic) of multiple personality is precisely because of its singleness of voice at any one moment, not because of its multiplicity. Improvement starts when dialogue and reflection between the selves begins to happen, when there is a multiplicity in a single moment of time, rather than multiplicity over time (Watkins, 1996: 12)

Of course the centredness of consciousness, as a field of perceived objects related to the perceiving person, cannot be multiple, even before improvement starts.

Consciousness and other kinds of awareness

So far I have discussing the grammar of 'aware of' in cases in which that awareness is perceptual. There are other uses of the phrase, more cognitive, such as understanding a meaning and drawing an inference. 'Consciousness' is so multivocal an expression that it turns up in all sorts of contexts, where it is actually doing duty for something other than the perceptual. We find feminists writing of the 'divided consciousness' of women. Examining this usage closely reveals that it does not mean anything thrilling such as 'each and every woman has two centres of awareness somehow located in the one body'. It usually means something quite mundane, that women often find themselves with two sets of (sometimes) incompatible duties; or that women often find themselves having to pay attention to more than one thing at the same time; their take home work on the family computer and the grizzling of a damp and hungry infant. But all these objects competing for attention are located in a single material environment which the person perceives from her singular location as an embodied being. Why use expressions like 'divided consciousness'? One can only surmise that such a word is chosen for its rhetorical impact. It is a striking metaphor, but a far from a transparent terminology for a psychology of gender role differences!

Psychological research programmes

The material embodiment of human beings as persons appears in two different ways in our analysis. A place in the body, somewhere just behind the eyes, serves as the centre of the material environment including one's own body as part of it. However, I have so far not begun to develop an account of other fields of consciousness, in particular memory and imagination. The organisation of recollections and anticipations as the span of a life is patterned around a temporal origin, the 'now' of bodily existence. This too is an indexical, like the first person, and the source of the possibility of tensed statements as the public expression of thoughts about the past and the future. 'Was' means 'before now', while 'will be' means 'after now'. 'Now' is the moment contemporaneous with an act of speaking. Further research will take us deeper into the study of indexical expressions, locatives of both places and times.

    Some state or pattern of states of the body seems to be a necessary requirement for all the modes of awareness that I have touched on. These are the enabling conditions as neural structures and processes. It is important that the right neural phenomena could not be identified without the use of phenomenological criteria, a point to which I will return.

    From these two considerations it follows that the form of a psychological investigation of some aspect of human life, such as consciousness (perception) or the emotions or reasoning or attitudes or anything else, takes the following form:

  1. Since every psychological process or phenomenon is made possible by virtue of a certain condition or state of the brain and nervous systems of those engaged in the activity there is a well defined though ever changing set of questions that can be put to neuroscientists. Following well established tradition I shall call the neurological foundation the 'enabling condition' for the activity. Some enabling conditions exist by nature, so to say, while others are established by training and practice.
  2. Every psychological process or phenomenon is a skilled performance by a person or persons, including perception. A performance is skilled if it is intentional, that is directed to some end and is subject to criteria of correctness, propriety, in short is normatively constrained. Deeper research into the structures of these processes is facilitated by the fact that human beings have various ways of publicly expressing what is directly known privately to each individual. These ways, since they are public, can be studied, analyzed and the principles of their construction discovered. It has been said that perception is not a skilled activity, but there are good reasons for treating it as such. For example, one must learn to identify the type or category to which an object belongs, as well as learning how to discriminate a figure from a ground. For example playing a musical instrument is a psychological phenomenon, made possible by the changes that have occurred in the brain and nervous system of someone who has acquired the skill.

The action of playing is indeed psychological, since it is intentional, the player is trying to play something, and it is normative, that is the performance is subject to a variety of standards in reproducing the score, coordinating one's performance with those of others,phrasing 'musically' and so on. Just in the same way performing an arithmetical calculation, making a decision, displaying an emotion, forming a friendship, seeing a constellation, tasting a fine claret, and so on are all to be construed within the basic framework of enabling conditions and skilled actions In so far as consciousness is a catch-all term for perceptions of various modes and kinds, the study of 'consciousness' too falls within this generic hybrid methodology.

    Enabling conditions are somewhat more complex than the brief introductory sketch above would suggest. Not only do they include the neural state of the human organisms involved, but also there are environmental conditions that must also obtain for a skilled performance to be possible. One cannot display one's skills as a mountaineer unless there is some rock face (or artificial surrogate) to climb. One cannot be jealous unless there is someone whose rights to a good are thought by you to be inadequate relative to yours, and so on. One cannot read without some text to follow. In what follows, that is a discussion of consciousness in the discursive frame, the environmental conditions will play a minor though essential part, since it is only in a certain kind of discursive environment that consciousness as we know it, can come into existence.

The grammars of everyday life

Contemporary Anglo-American life, conceived as the joint production of meaningful and normatively constrained patterns of action, of which verbal intercourse is one species, seems to be shaped by four main 'grammars', that is more or less coherent systems of norms and semantic rules.

    There is the Soul or S-grammar, which makes use of concepts like 'God','soul', 'sin', 'redemption' and so on. This grammar, once universal in the 'West', is now confined to certain rather restricted groups, tribes and regions where it was once dominant. It is still part of working practices of Mormon Utah, for instance. A main feature of S-grammars is the independence of the prime source of activity, the soul, from any necessary relation to the body it 'inhabits'.

    There is the Person or P-grammar, in which persons are the basic particulars and originating sources of activity. It comprises the tribal dialects and idiolects of everyday life in all those regions of the world where some measure of individualism prevails and where the peculiar blend of Roman and Anglo-Saxon legal traditions are in place. Among some of the specialised dialects of this generic grammar are the idioms of the courtroom, of the shrill claims for rights of the 'entitlement society, and so on. A main feature of P-grammars is the conceptual tie between singular embodiment and personal uniqueness and individuality.

    There is the Organism or O-grammar: Current Western discourses make use of a third grammar, that in which the basic powerful particulars or active beings are organisms. While it has, so to say, its natural domain of application in discussions about animals it has some important uses in discourse about human beings. The use of this grammar is becoming very widespread in reports of and discussions of the role of a human being's genetic endowment in the pattern of one's life. In so far as one's actions are the product of neural mechanisms of genetic origin one's life is embedded in a discourse in which the concept of 'person' is replaced by that 'organism' in discussions of human beings. Animals are agentive and act teleologically, but the person grammar is extended to them only with difficulty and usually only metaphorically. Animals do not act intentionally in the full sense that would bring into play the grammar of responsibility attributions except in rare cases. Responsibility talk addressed to family pets is surely metaphorical. When addressed to certain primates, such as domesticated chimpanzees it may have a deeper significance, widening the scope of the domain of moral agents. Though babies act for an end they surely do not act for a purpose. If Vygotsky is right the use a P-grammar with its load of 'responsibility' concepts, for talking about, and usually to, neonates, is part of the necessary conditions for a human organism to mature into a person.

    There is the Molecular or M-grammar, in which molecules and molecular clusters are the basic particulars and originating sources of activity. Among the dialects shaped by M-grammar is human physiology and molecular biology. Discourse framed in this grammar includes such attributions of agency to molecules as the power (alleged) of melatonin to put one to sleep in the sense of a change in the brain rhythms, and excess stomach acid to cause heartburn, in the sense of discharges in the pain receptors. Unlike the P-grammar the M-grammar is strongly hierarchical and displays emergent properties at every level. Why did top tennis players eat a banana between sets? This query requires an answer couched in M-grammar. But why they no longer seem to do so may need the P-grammar. And so too does an account of the use of cortisone to reduce the inflammation in a cartilage, and so on.

    Thus we have a loose cluster of grammars that set the standards of proper discourse for the human domain, the S-, the P-, the O- and the M-grammars. Each has variants, and in certain circumstances they fit together into hierarchies, and, in other circumstances, they complement one another.

    Not only does each of these grammars have its associated ontology, or catalogue of sources of activity, but each also comprises taxonomies of dependent particulars, such as action-types in the P-grammar. These give us criteria for partitioning the flow of the activities of persons into meaningful acts. In the O-grammar we have a classificatory resource for identifying behaviour-types as partitions of the activities of pets and some of the wild animals well known to us. The M-grammar allows for the identification of chemical reactions in organ systems, for example the interaction between carotene and cholesterol. Organs, in the M-grammar, are treated as functional partitions of the hierarchical clustering of molecules.

    Each grammar has its own distinctive principles of sequence and order among basic and dependent particulars. In P-grammar these include semantic and syntactic rules, moral imperatives and story-lines. Thanks to the work of the ethologists we now see the lives of animals teleologically. This is expressed in the O-grammar in terms of repertoires of actions directed towards maintaining their forms of life, leading to a complex repertoire of means/end principles. But in M-grammar there are only causal laws.

    S- and P-grammars differ from O- and M-grammars in the way that responsibility is dealt with. This is particularly important for a philosophy of psychology, since the transition from infancy to maturity of a being that has native agentive powers and acts teleologically, occurs along the dimension of growing responsibility for what it does. Shaver (1985) has proposed an analysis of responsibility dimensions that will do very well as a working grammar for much of the P-grammar of current English language folk psychologySee footnote 1. The attribution of responsibility according to Shaver runs as follows:

A judgement made about the moral accountability of a person of normal capacities, which judgement usually but not always involves a causal connection between the person being judged and some morally disapproved action or event (Shaver, 1985: 66)

The 'causal connection' presupposes the agentive powers of a person, not a Humean regularity of stimulus-type and response-type. Hart (1963), analyzing that variant of the P-grammar that is to be found in English law, cites three necessary conditions for attributions of responsibility:

  1.     That the person understand what is required.
  2.     That the person has deliberated on the matter in hand.
  3.     That the person conforms to the result of the deliberation.

One can see that in the S-grammar the concept of 'sin' depends on very similar conditions obtaining.

    Another striking way in which the S- and P-grammars differ from the O- and M-grammars, is in everyday discourse of remembering. Only people (or souls) remember, not their brains nor the molecular structures that ground our recollective capacities. Why is this? To say 'I remember .. ' is to claim some kind of authority, to commit myself to what I assert about the past. How my claims to remember are taken depends on my moral standing in the community, my reliability. It engages my personal qualities. It does not refer to me as an organism.

    Playing tennis is another example that requires the P-grammar. The exchange of shots is constrained by conventions of meaning: 'On the line is out'; and of procedure: 'Change ends after four games'. Scores accrue to people and it is people who play shots, good and bad, for which they are responsible, not their bodies nor their racquets.

    Finally I draw attention to the grammatical fact that the experiences of falling asleep and of waking, of being in pain and so on, are described in the P-grammar, since it is persons who notice what is going on, and its persons who suffer.

How is a phenomenology of experience possible?

The upshot of the discussion so far is that the 'study of consciousness' is a hybrid project, requiring intensive work in the phenomenology of perception, proprioception in all the sense modalities together with a study of acts of recollection, and requiring intensive work on the neurophysiology of the conditions which enable acts of perception and recollection. The former are done by people, the latter happen in brains. But if the phenomenological project is to be pursued an age old problem must be solved: how is it possible for there to be a common, public language by the use of which people can discuss how the world appears, how their bodies are, and so on, each from his or her own point of view?

The P-grammar of descriptions and avowals

Many experiences are not shared with others. While we both see the same tennis ball only I feel the pain of my 'tennis elbow'. While we both see the same slice of Tin Roof Fudge Pie only I enjoy the exquisite sensations since you foolishly chose carrot cake. Yet I can discuss the pain of my tennis elbow with my doctor, and I can mortify you with my panegyric on the gustatory delights of a famous 'desert'. But there seems to be an insoluble impasse in philosophy in which the authority of experience clashes with the conditions for sharing a meaning with someone else. In the course of criticising the idea that a language could be established on the basis only of private and personal referents for the teaching the meaning of words, Wittgenstein also provided the means for resolving the impasse.

Wittgenstein's Private Language Argument (PLA)

The shape of the argument

The 'argument' involves three interwoven and mutually supporting strands that are relevant to the distinction between descriptions of states of affairs and avowals of experience. One strand is a demonstration that the general principle that learning the meaning of a word is a achieved by a teacher pointing to an exemplar cannot account for how words for private experiences can be learned. There is no public exemplar to which the teacher can point and learner can attend. Meaningful words for sensations and bodily feelings could not have been learned by 'ostension'. Furthermore our experiences, both private and public, are ordered around spatial and temporal 'I-poles', to use Husserl's useful phrase, a feature of experienvce that contributes massively to the human sense of self. Again there is no public exemplar for learning the words, such as first person pronouns, that we use to express the structured or centred 'shape' of our personal experience. Wittgenstein's demonstration that there must be some other way to learn these important classes of words than by pointing to exemplars, depends upon a commonsense but fundamental observation. The words for discussing publicly the nature and struture of private and personal experience are learned with ease and used without a qualm.

    The assumption we have just criticised, about the role of exemplars in learning words, would make no sense in the case of private experiences, without the further assumption that all meaningful words denote objects. The second strand of the PLA is an argument to show that feelings are not objects in the sense that the objects that are pointed to in ordinary situations of ostensive learning, say 'potato', are objects. This argument is based on a reflection on the criteria by which judgements of 'same feeling' are made in intrapersonal and in interpersonal comparisons. According to Wittgenstein the concept of 'sameness' for feelings cannot be analyzed along the same lines as we analyze concepts of sameness for such things as potatoes.

    Judgements of 'sameness' in material contexts are based on criteria of qualitative and numerical identity. Two things are qualitatively identical if they have more or less similar properties. Something is numerically identical, 'still the same' in different circumstances, if its integrity and continuity of existence is uninterrupted. In some of later paragraphs in the exposition of the PLA Wittgenstein shows in some detail that these criteria of identity have no place in language games in which we make such judgements as 'I feel just as bad as I did yesterday' or 'I know just what you are feeling. I felt the same when I had finished my exams'. Yet everyone makes lots of everyday unproblematic judgements of sameness and difference of their own and other people's feelings. Since they do not meet the criteria for 'sameness' and 'difference' of material things, bodily feelings are not objects in the relevant sense. But even if they were their privacy precludes their use as exemplars for denotational learning.

    How to resolve this seemingly intractable bouquet of problems? We use feeling words perfectly well, yet it seems that we are never in a position to learn them. We talk about feelings yet they are not at all like the entities of the material world about which we also talk. The impasse comes about because we took two things for granted. We assumed that words are all learned by the teaching pointing to an exemplar of the kind of object talked about while the pupil attends to it. We also assumed that feelings were a special sort of object or entity, for which the usual, thing-related criteria of identity, perhaps adjusted a little, were appropriate. But feelings are not thing-like. So we have two gaps to fill in understanding how it is possible to discuss our private feelings with others and to reflect on what we felt in the past. How is it that words for feelings do get their proper meanings? And what do we mean by 'same feeling' in both intra and interpersonal contexts?

    In a quite similar way an intractable problem arises about that to which the word 'I' refers when I use it of myself in reporting how things are for me. How could such an expression ever get established in public language if it could be learned only from a public exemplar? There is no public exemplar of the centredness of experience, nor is there any thing to which attention could be drawn. My 'self' is no more a thing-like entity than is my pain. And again we can see this by comparing my sense of my own personal identity with my sense of yours. I think you are the same person I saw yesterday mostly by noticing the similarity of certain salient features of your appearance today as yesterday. Of course I might be mistaken. I do not have to use any criterion to judge if I am still the same person I was yesterday, and certainly not one based on bodily appearances. It does not make sense to declare: 'All along I was wrong about the person I am. I have discovered I am you.'

    There are subtly different cases of uniqueness of point of view here. I can't have your pains in the way I can't sneeze your sneezes, to borrow another of Wittgenstein's images. To be that instance of pain it has to be your feeling. Thus only you can have it. I can see a herd of cows from your point of view by simply standing where you were a moment ago, but the second term of the relational structure of the seeing of the material environment is uniquely you or uniquely me. But in both cases we slip into seemingly intractable puzzles by taking for granted that it is objects that are in question, where 'the feeling' or 'the self' are assumed to obey the same grammars as 'the colour' and 'the driver'.

Wittgenstein's suggestion that solves the linguistic problem

The suggestion that allows us to transcend the impasse is to be found in the following remarks:

    How do words refer to sensations? - There does not seem to be any problem here; don't we talk about sensations every day, and give them names? This question is the same as: how does a human being learn the meaning of the names for sensations? - of the word 'pain' for example. Here is one possibility: words are connected with the primitive, the natural, expressions of the sensation and used in their place. A child has hurt himself and he cries; and then adults talk to him and teach him exclamations and, later, sentences. They teach the children new pain behaviour.

        'So you are saying that the word "pain" really means crying?' -On the contrary: the verbal expression of pain replaces crying and does not describe it. (Wittgenstein, 1953: 244).

There are plenty of language games which have the right combination of natural expressions and alternative verbal expressions. A child picks up a toy and chortles. A footballer writhes in the aftermath of a vigourous tackle. Gaza weeps with disappointment and frustration when he misses the vital shoot out against Argentina. These are natural expressions of how one is feeling. Words for feelings are learned as alternatives to natural ways of expressing feelings. Instead of chortling we learn to say 'I'm delighted'. Instead of weeping some footballers will tell you that they are 'as sick as a parrot'. So here we have the essential move in a resolution of the problem of how it is possible to learn how to use a word for a private feeling in a public context. We learn to use a verbal formula as a substitute for a natural expression. Just the same holds for the centredness of 'consciousness' and the use of the first person. In the same language games we can find the groundings of judgements of 'same and different feeling'. They are not made by a comparison between your feelings and mine, or between my feeling of yesterday and my feeling of today. They are made by attention to parallel patterns in the language games in which our feelings are publicly expressed.

    Wittgenstein goes on to develop a thoroughgoing distinction between descriptions of public objects and states of affairs, and avowals of 'how it is with me', the form and content of private experiences. In the former case there is room for evidence, and error. In the latter there is no epistemological gap between the groan and the feeling, nor between the centring of my personal environment on my body and the use I make if 'I' in reporting on that environment from my point of view. The groan expresses the feeling. It does not describe it. Feeling and the disposition or tendency to groan, sigh, rub the spot, weep and so on are integral parts of the same phenomenon. Abstract any of them and the phenomenon disappears. If we have no tendency to groan then whatever the feeling is, it cannot be pain. Expressive ways of using language depend on a pre-existing ethological repertoire of natural expressions of the ways we are feeling, sad, in pain, happy, and so on. This is part of the natural history of humankind (Robinson and Harré, 1997).

    The message of this paper is that the same treatment should be given to reflexive expressions like the first person pronouns and verb inflexions as Wittgenstein gave to words for feelings. They are expressions of the centred layout of what we can perceive, of what we can recollect and so on rather than devices used to refer to a mysterious entity within the person. In the Cartesian account of personal identity the inner 'ego' is not only the centre of expereince but also a substance of which our experiences are properties. The word 'experience' springs easily to the pen in discussions like this, but it too is fraught with the possibility of bad grammatical analogies, as if the experience of something were a different phenomenon than the thing which we see, feel and so on.

    We owe to Wittgenstein an all important insight for a methodology for psychological research, that the ontological distinction between thought (ineffable and meaningful) and language (audible, visible and tangible and meaningful) does not matter in certain key cases, because in the case of avowals, there is a holistic unity. Part of the art of psychological method is to be right in distinguishing those psychological phenomena for which the holistic principle holds and those for which it does not. There is no epistemological gap between a feeling and the expressions of a feeling, in those cases in which there is a natural expression, a part of human ethology, from which the substitution of words for non-verbal modes of expression can take its start. In the case of perception and proprioception the natural expression is to be seen in the way actions are taken -- returning a serve, picking up a spoon, scratching an itch and so on. The P-grammar includes the rules for the use of first person indexicals and these are the basis of verbal substitutes for the natural expression of the structure of the environment as we perceive it.

The grammar of the expression of the environment as I perceive it

What is the role of the first person singular in everyday talk? At first glance it might look as if 'I' was a queer kind of name, one that each person could use to refer to him or herself. But already we can see that there is something odd about that suggestion. Unless the person addressed is aware of who is speaking the self-referential function of 'I' is ineffective. So who is 'I' for the moment is discovered, so to say, by knowing who is speaking (or writing). In this way it is quite unlike a proper name. I know who is being referred to when someone uses the words 'John Lennon', even though he is long since dead and gone and only his songs live on. I am addressed by my proper name, but never by 'I'. The pronoun of address is 'you', but that is not a name either. Neither pronoun is used to refer to anybody in the way a proper name can be used to refer to someone.

    One common use for the first person singular, in English, is to take responsibility. 'I'll look after the children' is so close to a promise that it would usually be taken to be just another way of promising. However, in order to know where the responsibility lies the interlocutors must know who is speaking. Immediate presence is not necessary. One can make a commitment that is activated only in the future (a will) and one can make promises over the phone. Embodiment is not a necessary condition for this use of 'I'. Commitments could be entered into at a séance.

    However there is another use of 'I', often going along with the one just described. 'I' indexes the perceptual content of a report with the spatial position of the speaker. This is common in the everyday life of positioning oneself in space. It is also the core of autobiography. Telling about an event in the first person places the speaker at or near the event in question. 'I watched the planes circle and come in to drop their bombs on the bridge' carries at least the conversational implicature that 'I was there!'. The archetypal autobiographical anecdote has both place and time indexes: 'When I was in Nevada ... ' Authority for myself as the author of the story is claimed just by the use of the first person. Compare the fairy tale opening 'Once upon a time there lived a King who had three sons' with the autobiographical 'I was one of three brothers'.

    Places are locations in physical space. They are defined by relations between material things. In so far as I have a place in space I have it by virtue of the fact of my embodiment. In so far as the tenses of the verbs in my utterances relate what I am saying to the times of events as past or future, my utterances themselves must be events in the same time frame as those events. In so far as those events are happenings in the material world, so must my utterances be. At least one aspect of myself must be a trajectory of connected locations in space and time.

    But there is more. To be a person is to be a singularity, to have just one trajectory in space and time, to be embodied in just one material organism. This is not a ubiquitous fact that students of human life have discovered. It is part of the grammar of the 'person' concept. Strawson (1964) pointed out that it is only as embodied beings that people are routinely identified and reidentified as the singularities they are. They share this metaphysical property with the material things among which they are embodied. As a grammatical remark about people it fits well with the observations I have made above about the indexical force of the first person singular in reports how the environment strikes me and autobiographical narratives. Personhood is so bounded by the singularity of each human being's embodiment that neither claims to more nor to less than one person per body are permitted to stand. The former exploits the distinction between the use of 'I' to take responsibility and its use to index reports with their spatial and temporal location. If there is a second 'I' in use by this body, then it may be taking responsibility for actions performed by the first 'I' who is then no longer to be blamed for them. The latter is so transparently a ploy if taken contemporaneously that it seems to turn, not on the question of material location as the centre point of perception, but on issues of responsibility. Fascinating though the issue is it is not central to my analysis in here.

    In summary then, in order to understand a perceptual report I have to know where the speaker is in space. But that spatial position is fixed by the relations that the speaker bears to the things in the material environment. To have a place in such an array a speaker must be embodied, as a thing amongst things. The acts performed by the active person with whom we are engaged are embedded within sequences of events in the material environment, such as the risings and settings of the sun, the ticking of clocks and so on. So a speaker's acts must be carried by events in the material world. To understand a perceptual report fully I must know how the event of making the report and the event reported are related in time. Both must be material events, and this again requires that the speaker be embodied. The general principle that acts must be sustained materially has long been accorded a place in social psychology and ethology in the act-action-behaviour distinction. Behaviours are material phenomena, actions are material phenomena seen as intended by the actor, and acts are the social meanings of actions as they are seen with unfolding social episodes.

    The traditional 'mind/body' distinction is no longer required to make sense of the distinction between private and public activities. Instead we are using a distinction between personal and impersonal points of view on the immediate and the distant environment. Singular embodiment, one person per body and one body per person, focuses our attention on that which is localised in space and time. People are many and history is rich in events. The emphasis I have placed in embodiment ought not to obscure the importance of studying cognitive phenomena that are generated in the symbolic interactions of crowds and in the historical unfolding of long term 'conversations'.

    If consciousness talk is nothing other than a way of talking about what one perceives from one's own unique point of view, then the 'structure of consciousness' is nothing other than the structure of the field of perceived entities from the point of view of the perceiver. There is a singular but imperceptible 'origin' and a 'pencil' of relations to an ordered array of objects, both within and without the body. A public expression of this private and personal structure is to be found in the indexicals of the P-grammar with which we publicly make available to others 'the environment for me'.

Is consciousness socially contructed?

This sounds like a thrilling question. However, looking more closely at what might be involved it loses its exciting character. The mark of a socially constructed aspect of mind is variation in different cultures, in particular in the different ways languages perform the same or similar functions. Given that everyone's perceptual system, say as described by Gibson (1966), is inherited with certain built-in capacities in what ways could our uses of that system differ?

Fine structure

In addition to the indexical pronouns the grammar of demonstratives also reveals ways that the 'environment-for-me' is structured. English makes do with 'this' and 'that', for things seen, heard and so on near and far from the speaker. The language of the Maya of Yucatan is richer in demonstratives. For instance 'that-seeable-by-you-but-not-by-me' is lexically distinguished from 'that-seeable-by-both-of-us'. In short Maya demonstratives facilitate joint attention to the finer structure of an array of seen objects than does English. In this mundane sense we could say that the difference between Maya consciousness and English consciousness is socially constructed, in that the language game of referring to distant objects is learned in social practices, and what is learned during the acquisition of the English demonstrates is different from what is learned when acquiring Mayan.


Even less thrilling is the commonplace that our capacities to differentiate objects in our environment is fine tuned by learning whatever discriminatory criteria are necessary to manage successfully in some local form of life. For instance the ethnobotany of one tribe will differ from that another, so what members of each will pay attention to in their environments will differ.

Further steps in the analysis

There is at least one major issue raised by the treatment offered in this paper which must remain 'unfinished business'. Eliminating 'consciousness' talk in favour of more mundane ways of expressing and commenting upon what people can perceive both inside and outside the envelopes of their bodies, and interpreting the first person as indexing avowals with the location of the speaker, presumes that the distinction 'speaker-non-speaker' is in place. This might be the 'I/thou' distinction of some theologians, or it might be the Kantian double synthesis of 'self-as-perceiver' over against 'object-as-perceived'. Is the maturing infant built to make this distinction or is it created in Vygotskian psychological symbiosis with beings who have already acquired it? Infants and chimpanzees, though at first unable to realise that a spot of paint on their foreheads, visible in a mirror, is on their own faces, soon do realise this, and try to pick it off. Is the transition due solely to maturation of the perceptual system or is it, in some measure, socially constructed? I believe the answer is yet to be found.

    Since the P- and O-grammars are both required to delineate a research programme in the psychology of perception, the question of their relationship needs to be tackled. I have taken for granted that the criteria by which organs and organ systems relevant to perception require distinctions drawn in the P-grammar. Unless the criteria for seeing had been established it would be impossible to identify eyes as the organs of sight. P-grammar criteria are taxonomically prior to O-grammar criteria. In short no account of enabling conditions is possible without an account of the phenomena that they enable. This point has been made often enough in different terms, for instance by the use of the phrase 'top/down'. However a robust defence is still required in the face of reductionist arguments from Davidson (1980) and reductionist obiter dicta from Crick (1994).



Crick, F. 1994. The astonishing hypothesis: the scientific search for the soul. New York: Scribner.

Davidson, D. 1980. Essays on actions and events. Oxford: Clarendon Press.

Gibson, J. J. 1966. The senses considered as perceptual systems. New York: Houghton Miflin.

Hart, H. L. A. 1963. Law, liberty and morality. Stanford, CA: Stanford University Press.

Robinson, D. N. and Harré, R. 1995. "On the primacy of duties". Philosophy 70: 513-532.

Shaver, K. G. 1985. The attribution of blame. New York: Springer Verlag.

Strawson, P. F. 1964. Individuals. London: Methuen.

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

Watkins, M. 1996.

Wittgenstein, L. 1953. Philosophical investigations. Trans. G. E. M. Anscombe and G. H. von Wright, Oxford: Blackwell.

Footnote: 1 I owe notice of Shaver's work to M. A. Spackman.

Converted by Andrew Scriven