To appear in the proceedings of the 50th anniversary
conference of the Korean Psychological Association, edited by
Uichol Kim, Psychology dept, Chung-ang University, Seoul,
There are two main pictures of human beings and their modes of life that have been dominant in various branches of human studies since the nineteenth century. On the one hand there is the picture of human life as the sum of interactions of individual 'mechanisms' with each other and with the environment, the behaviour of each of which can be explained in cause-effect terms. On the other hand there is the picture of human life as a collective activity, in which individuals work with others to fulfil their intentions and achieve their projects according to local rules and norms. Despite more than a century and a half of criticism the causal assumptions of the first picture, manifested in the use of causal language, is to be found everywhere in contemporary 'main stream' psychology. The language of meanings, intentions, plans and rules, reflecting a belief in the agentive powers of people to act intentionally, is characteristic of discursive psychology and other related radical alternatives to mainstream psychology, such as cultural psychology (Shweder, 1991) and ethnomethdology (Garfinkel, 1957).
The contrast between the causal picture and the agentive picture shows up very clearly in the differing roles assigned to persons in each paradigm. In the causal picture the concept of a 'person' is secondary if it is invoked at all. Human beings are conceived as hierarchically organised clusters of cognitive mechanisms of most of the workings of which people are unaware. In some versions of mainstream psychology these mechanisms are presumed to be material and in others mental. In the agentive picture the concept of 'person' is fundamental. People are taken to be active beings using all sorts of tools, including their own brains, for carrying on their life projects according to local norms and standards.
In following out the causal picture psychologists have persisted with a methodology which involves trying to find correlations between treatments and behaviours (even now) and, in the metaphysical framework of the first cognitive revolution, to hypothesize cognitive mechanisms 'behind' these correlations. It is important to emphasise that this is, in general terms, the methodology of chemistry. Chemical reactions are observed and explanations sought in unobserved molecular processes. In following out the agentive picture psychologists practising the methodology of the alternative paradigm, try to make explicit the implicit norms, rules and meanings immanent in what people are doing, and to pick out from the turbulent flow of everyday life such patterns of individual and joint action as people themselves recognise as coherent episodes. This is the methodology of history, sociolinguistics, social anthropology and conversation analysis. The methodology of the current mainstream tends to be abstract and inductive, while that of the alternative psychologies tends to be concrete and analytical.
If we can call the construction of information processing models of cognition the realisation of a First Cognitive Revolution, the turn to discourse and the adoption of conversation as the generic model for psychology, should perhaps be called the Second Cognitive Revolution, even though the debate between those advocating a causal picture and the defenders of the conception of humans as agents goes back to Wundt (1917 - 1926), Dilthey (1985) and Stern (1925). According to proponents of the Second Cognitive Revolution the only mechanisms relevant to psychology are neural. For example see Johnson and Erneling (1997) and particularly pp. 17 - 18 of Chomsky's contribution (Chomsky, 1997).
The history of psychology displays a character quite
different from the histories of the natural sciences.
Associationism came and went leaving nothing behind. The
behaviourism went through the same cycle, at the same time as
psychodynamics rose and declined. Currently the same
phenomenon can be seen with cognitive 'science'. In physics
the Newtonian synthesis remains valid at a certain level of
analysis, while the two Relativities provide the means of a
deeper analysis of material processes by reforming the
theoretical basis of mechanics and electromahnetism.
Einstein's revolution was conceptual. Dilthey's conceptual
criticisms of the mechanistic trend in the human sciences have
been simply ignored, as have subsequent revivals of his
criticisms by others. This is quite unlike the situation in
physics. It strikes one as very odd for a field of endeavour
which aspires to the status of a 'science'. Perhaps one day
may get a historical-sociological explanation. There is
continuity and development in the alternative psychologies,
but radical discontinuities in the various fashions that have
overtaken the mainstream. Numerical disparities in the number
of psychologists of each persuasion has no more significance
than the numerical preponderance of alchemists in the
fifteenth century undercuts the truth of chemistry, practised
by a tiny minority.
In order to identify the current version of the tradition of agentive psychology, which I have called the Second Cognitive Revolution, more precisely, we must sketch the history and fate of the First. The context for the 'new cognitivism' was defined by the demise of behaviourism. Behaviourism involved both a theory of psychology and thus of human nature as well as a repertoire of investigative techniques, which, allied with the statistical analysis of the behaviours of aggregates of experimental subjects, bequeathed to psychology the main forms of contemporary American experimentalism. It has survived in methodology long after it died out as a general theory. Behaviourism as a general theory was based upon two major philosophical theses.
Positivists did not eschew theorising completely. Rather they regarded theory as an empirically empty machinery by means of which predictions and retrodictions of new relations between treatments and responses could be derived. McCorquodale and Mehl's (1948) distinction between hypothetical constructs and intervening variables opened up the possibility for a non-positivistic psychology, but the turn to realism, the adoption of a methodology that took unobserved processes seriously, had to await the developments in Harvard to be discussed below. See footnote 1
It is hard to believe from our current perspective that a version of this philosophical theory was also rampant amongst chemists in the 19th century. At that time there was a vigorous attempt to delete the concept of the chemical atom from chemical science on the grounds of the unobservability of atoms and molecules. Chemistry was to consist simply of the description of correlations between those chemical treatments and chemical responses which were observable in the laboratory. However, by the end of the 19th century chemists were fully convinced of the reality of chemical atoms, though much remained to be discovered about their nature. This attitude to hypothetical entities, that it makes sense to take the possibility of their existence seriously, is usually called 'scientific realism'. A similar debate emerged in physics in the 20th century as the problems of conceiving the underlying processes responsible for quantum phenomena proved to be very difficult to solve.
I do not know whether the impetus towards a realist psychology came from a diffusion into the ideas of the most forward thinking psychologists of the realist debates in physics but there is no doubt that the parallel between the growth of realism in chemistry and physics and the advent of the First Cognitive Revolution in psychology is very close.
That revolution was made by a group of psychologists who met
together in Harvard in the 1960s and forged a programme of
theory and research on the assumption that it was proper to
assume that there were unobservable cognitive processes, or at
least that models of such processes might be conceived, which
would play the same sort of role in psychology Both Miller (in
Miller, Galanter and Pribram, 1960) and Bruner (1983) took for
granted that such models were essential in the development of
the scientific psychology. For example Bruner (19 ) provided
powerful experimental support for the seemingly indispensable
role of tacit knowledge in the psychology of perception. To
understand the way that the First Cognitive Revolution
developed, we shall have to take a closer look at the way that
models are employed in the natural sciences.
The word 'model' is used widely by psychologists, but in a sense rather different from its use in the natural sciences. For the most part psychologists use the expression sometimes as a synonym for 'theory', but commonly simply to mean 'hypothesis' or even 'concept'. However, in the natural sciences a model is a representation or analogue of its subject. In the physical sciences, a typical example of a model would be the molecular theory of gases. A gas is pictured as a swarm of molecules, minute material particles, in continuous motion. It is important to distinguish between the claim that a gas is a swarm of molecules in continuous motion, and the claim that it can be pictured as such a swarm. It is the latter claim that is the basis of the molecular model. Of course, if one has sufficient ingenuity, any number of models can be constructed that will simulate the observable behaviour of gases and so serve the unifying function of the molecular model in bringing together all the particular laws of the behaviour of gases into one single unified system. If that were the whole story about models in the physical sciences, then they could be dismissed as little more than heuristic aids to thinking about the empirical regularities found in the laboratory. However, they are much more than that. The physical sciences have made such startling progress because, for the most part, it has been taken for granted that the question of the resemblance between a model, the behaviour of which simulates the behaviour of something in the real world, and the relevant structure and processes in the real world causing that behaviour is a sensible question to ask. Models in the physical sciences are constrained in two ways. They must be imagined as behaving in such a way as to simulate the observable behaviour of some feature or aspect of reality and they must be plausible in the sense that they are possible representations of whatever it is in reality that is causing that behaviour, that they so nicely simulate.
Model building and model testing are part of the methodology that is typical of the realist attitude to theory that is now dominant in physics and chemistry. The justification for adopting this attitude comes, at least in part, from the history of science. The practice of taking models seriously has paid off handsomely. But how is one to ensure that the models that one imagines are plausible, that is, could be representations of a possible reality? One obvious way to proceed would be to construct the model for the theory in accordance with the requirements for things and processes one knows already do exist. To understand the deep character of the First Cognitive Revolution, we must look closely at the source of the models of the unobservable cognitive processes that the first generation of cognitive psychologists took to lie behind such phenomena as differences in the times to recognise different classes of words, the tendency to see tilted circles as less elliptical than geometrical optics would predict, and so on, to choose some examples of Bruner's research.
The way that the architects of the First Cognitive Revolution constrained their model of the human mind quickly developed into the analogy of the computer and the running of its programmes. The basic idea was an analogy between the active human brain and an information processing device such as a computer running a programme. Many interesting proposals were derived from this idea. An early version of the model is to found in Miller, Gallanter and Pribram's famous 1960 book, Plans and the structure of behaviour, in the which the idea of a TOTE hierarchy was introduced. The computational 'theory of mind' has not only animated recent psychological investigations in the realm of Artificial Intelligence, for example in the field of 'knowledge representation'. It has also been influential in social psychology in a less formal way, in the script theory of Abelson (1976) and in the role-rule model proposed by Harré and Secord (1972).
As a psychology, this form of cognitivism had a number of disturbing aspects. It preserved a generally Cartesian picture of 'the mind' as some kind of diaphanous mechanism, a mechanism which operated upon such non-material stuff as 'information' See footnote 2. On this view, though the mind was assumed to be materially grounded in the neurophysiology of the brain, a number of concepts were subtly inserted into the conceptual resources of psychology that had a strong Cartesian ring, such as 'mental state' and 'mental process', and, more recently 'mental model'. In developments in the philosophy of mind, there also stemmed, though in many cases indirectly from the revolution of Bruner and Miller, versions of such seemingly everyday concepts as 'belief', 'attitude', 'memory' which were interpreted as generic terms for types of mental entities, such as mental states.
But the Cartesianism of the First Cognitive Revolution appeared in a stronger form still, since it led to the belief in a shadow cognitive reality behind the overt phenomena of ordinary human thinking and acting. In addition to the activities of conversing with other people, of displaying emotions, of talking to oneself, of remembering a past event, of telling that memory to someone else, of asserting a belief, of displaying an attitude and so on there were assumed to be covert cognitive processes.
The principle that we should not multiply entities (and processes) beyond necessity, Ockham's Razor, should be part of every scientists intellectual equipment. The most striking violation of Ockham's Razor among the theoreticians of cognitive psychology influenced by this general point of view can be found in the works of Fodor in particular in his book, The language of thought (1976). The idea of a shadow reality behind the overt phenomena of thinking, talking and acting is the central theme. As well as the language of everyday there is a language of thought.
The Cartesianism of this point of view was muted for many
of its adherents by the use of the brain:computer analogy to
transmute the nature of the hidden cognitive mechanisms from
immaterial substances to neuropsychological processes. It was
thought that the difficulties of Cartesian metaphysics were
somehow avoided if the processes which were the bearers of
cognition were located in a material mechanism, the brain. But
as we shall see, though this was surely a move in the right
direction it led to another error, an equally misguided
reductive materialism. this too was a mistake. It was
individualistic, assuming that cognitive processes, such as
remembering, could be confined to the workings of brains,
taken one by one. At the same time it led to a systematic
confusion of questions about the instruments of cognition with
questions about cognition as a psychological phenomenon. It
was as if the whole of tennis as a competitive game could be
explained by reference to the physics of elastic impacts and
The movement, currently called 'discursive psychology' (Edwards and Potter, 1992) and also, somewhat misleadingly 'social constructionism', has a very long history. In almost every decade since the mid-nineteenth century versions of essentially the same point of view have appeared. It is to be found in Wilhelm Wundt's (1917 - 1926) volkerspsychologie, in Wilhelm Dilthey's geisteswissenshaft (1985, new edition), but above all in the work of William Stern (1925) and Lev Vygotsky (1962).
The latest version of this point of view, which I am calling the Second Cognitive Revolution, developed from several different sources. Amongst these was a reaction against the metaphysical implications of the information processing analogy, that is, the idea that behind the overt phenomena of cognition were certain hidden cognitive processes, whether these were to be found in a psychologistic interpretation of the transformational grammar of Chomsky's theory of language or in the idea of information retrieval as the basis for a psychological theory of memory.
It is interesting to observe that a sufficiency of
observations and arguments against this mind-behind-the-mind
metaphysics were already available as early as 1953. But they
appeared in the writings of Ludwig Wittgenstein, a philosopher
particularly the Philosophical Investigations. This work is
not only profound but profoundly difficult to read. It has
been a long and difficult journey to reach an understanding of
the force of Wittgenstein's analyses and arguments for the
development of a science of psychology purged of extraneous
metaphysical elements. Amongst the many detailed and deep
arguments that appear in Wittgenstein's philosophical
investigations, two are of great importance in the
clarification of the foundations of psychology.
The first of these arguments is exemplified in Wittgenstein's observation that one is said to be calculating, working out a sum, in two quite different circumstances. One may be performing public arithmetical operations with a pencil and paper without there being any particular private conscious shadow version of the calculation accompanying what one is doing. Or one may be working 'in one's head', doing mental arithmetic. In this case too there need be no shadowy but unconscious version of the calculation accompanying one's performance. Calculating is the exercise of a typical cognitive skill. It is a skill because, like almost all those performances we call 'cognitive', it can be judged for whether it is exercised correctly or incorrectly. In the case of calculating judgements of correctness and error are made by reference to rules. In the case of other cognitive skills, such as remembering, which is of course not just recollecting the past but recollecting it correctly, the basis of normative judgements turns out to be very complex, and, as we shall see, could not be a matter of individual cognitive processes. Nor is it the case that in doing mathematics what seems right to an individual person is right.
There are two important points that emerge from these
But when we study habits psychologically we must be looking for the rules that were invoked in training and are immanent in the activity however habitual it becomes because the question whether it was done correctly or incorrectly can always arise.
Wittgenstein's second line of investigation into psychology involves the study of the major psychological concepts that are expressed in the vocabularies of ordinary languages and of their technical extensions. Wittgenstein believed that the royal road to error that beset human beings in their attempt to understand themselves, was to adopt too simple a way of characterising these fundamental concepts. It is quite wrong to think that Wittgenstein thought that the language of everyday life had some privileged place in the foundations of psychology. Rather it was a particularly rich breeding ground for language-driven mistakes. But so too are many of the neologisms of 'scientific psychology', such as 'affect', 'implicit memory' and so on (Danziger, 1997). Wittgenstein used the notion of an over generalised grammatical model to diagnose the sources of confusion in psychology (and elsewhere too). For instance, the word 'belief' is a noun and many nouns, for example 'sheep' and 'goat', refer to entities. So if we try to understand the way the noun 'belief' is used on the model of the way the noun 'sheep' is used we may be inclined to look for beliefs in somewhat the same manner as we might be inclined to look for sheep, though 'in the mind' rather than in the field. But if we survey all the ways that the word 'belief' and its cognates are used, we are soon made aware of a great diversity of usage, and of many different kinds of phenomena brought into some kind of complex structure and order via the use of a common term. In the case of 'belief', we find the word being used for, amongst many other possibilities, dispositions, declarations, tentative claims, and many others. In none of these uses do we find a justification for the assumption that beliefs are, in general or essentially, some kind of mental entity. Similarly in his analysis of thinking about the future, there is no good ground for supposed that there some one process that is going on when we are properly said to be expecting someone, to be hoping for something, to be wishing for something. The grammar of these verbs are different and complex, so too must be the cognitive processes for which they are the means. The literal and metaphorical use of the word 'tool' for the means by which intentional activities are carried out has been elegantly dissected by Vygotsky (1978: 54 - 57).
The assumption that Wittgenstein works with, and which was also the basis of Dilthey's conception of a human wissenschaft, is that ordinary language defines the broad topics of psychology. It delineates a vast range of psychological phenomena in which we can take an interest, since it is the means by which they are brought into being. Outside this range of topics lie other phenomena, for example pathologies of thought and feeling. Here too the trap of false grammatical analogies lies in wait to snare psychologists who neglect the study of the grammar of these 'dialects' into deep error. So the broad structure of the field to be investigated by psychologists is discoverable by studying the grammar of the psychological words of ordinary language, of psychopathology, of the various branches of experimental psychology, such as the study of perception.
We shall see, however, that the turn to language, the discursive turn, as it is sometimes called, has other consequences than the taxonomic. But it is worth noticing that I have put this matter in the singular, as if there was just one ordinary language, with one, however complex, system of psychological terminology, and that in its investigation would emerge a one and only grammar of the terminology for the psychological phenomena of all humankind. It is also widely assumed among mainstream psychologists, especially in The United States, that their technical vocabularies are first of all trouble-free and secondly unproblematically exportable to guide the investigation of psychological phenomena the world over, and across vast aeons of history.
Wittgenstein himself does not address this issue in any great detail, though he does refer to it from time to time in pointing out that there are diverse forms of life possible for human beings within the generic human form of life. These diverse ways of being human will, of course, be displayed in the grammars of the ordinary and technical languages of the various tribes. In recent years, one of the major developments in the Second Cognitive Revolution has come from the realisation that there are many ordinary languages, many human forms of life and many actually distinct human psychologies. A technical vocabulary that has developed as an extension of the everyday language of a culture that is rooted in radical individualism may be of little use, even as a technical vocabulary, when exported to a culture rooted in radical collectivism.
We owe to Wittgenstein a third and all important insight for a methodology for psychological research. It is the ontological distinction between thought (ineffable and meaningful) and language (audible, visible, tangible and meaningful) does not matter in certain key cases. These cases bridge the seeming gap between private experience and public discourse, because in the expressive mode of the use of gestures and words there is a holistic unity between the felt experience and the tendency to give it public expression.
This insight is part of his complex discussion of the
conception of a 'logically private language' and his
demonstration that the very idea makes no sense. In the course
of his analysis he points out that if it were the case that
all words have to be learned by attention to an example of
what they refer to, words for private sensations like 'pain'
could never be learned. But they are learned. So there must be
something wrong with the theory of meaning that exclusively
interprets meaningful words as names for things. In the case
of private sensations there are two things wrong with the
traditional puzzle of how someone can know about another
person's private and personal experiences. There must be some
other way that words get a meaning than by being to something
that be pointed out to by a teacher and noticed by a learner.
Wittgenstein develops an ancillary argument to show that
whatever private sensations are, they should not be taken to
be mental things. There are no mental objects to play the role
of exemplars even if we could overcome the impossibility of
using someone's else's private experience as an exemplar for
learning a word. Wittgenstein's solution is powerful. We learn
such words as 'pain' by being taught to substitute a verbal
expression for a natural expression of pain. The painful
feeling is not evidence for a (defeasible) claim that one is
in pain. Rather part of what it is to be in pain is to be
disposed to groan and express one's feeling in other ways.
Substituting a verbal expression for a behavioural one leaves
the basic distinction between expression and description
intact. The internal tie between expression and feeling could
be called the 'holistic principle'.
In looking closely at the methodology of traditional
psychology, one is struck by the way in which psychologists
seem to take for granted that there is some kind of common
mechanism which produces all the phenomena of our mental
lives. But if we turn our attention to the diverse ways in
which psychologies are realised in the various languages of
humankind and the discursive processes in which they figure
along with other symbolic devices, the assumption of a common
processing mechanism becomes more and more difficult to
sustain. The point has never been better put than by Shweder
The main force in general psychology is the idea of
that central processing device. The processor as it is
imagined stands over and above (or transcends) all the
stuff upon which it operates. It engages all the stuff of
culture, context, task and stimulus material as its
Given that image, the central processor itself must
be context and content independent. That means in effect
that the processor must be describable in terms of
properties that are either free of context/content
(abstract formal structural properties) or general to all
contexts/contents (invariant universal properties).
Ontologically speaking, knowledge in general psychology
is the attempt to imagine and characterise the form of
shape of an inherent central processing mechanism for
psychological functions, discrimination, categorisation,
memory, learning, motivation, inference and so on.
Epistemologically speaking, knowledge-seeking in general
psychology is the attempt to get a look at the central
processing mechanism untainted by content and context and
The main force in general psychology is the idea of that central processing device. The processor as it is imagined stands over and above (or transcends) all the stuff upon which it operates. It engages all the stuff of culture, context, task and stimulus material as its content.
Given that image, the central processor itself must be context and content independent. That means in effect that the processor must be describable in terms of properties that are either free of context/content (abstract formal structural properties) or general to all contexts/contents (invariant universal properties).
Discursive psychology, the psychology of the Second Cognitive Revolution, challenges the very idea of such a central processing mechanism. If there are no mechanisms other than the skilled use of symbolic systems maintained by reference to the normative constraints of a culture, then there is certainly no central processing mechanism unrestrained by the normal constraints of anything. Even the brain, which is 'the only mechanism in town', is structured and restructured as specific skills are learned (Luria, 1973).
In some current cognitive theories the mind is supposed to consist of a number of modules, each performing a well-defined task, the whole linked up into a cognitive system. The Shweder objection to universalistic claims is equally telling against this development of the original 'brain as computer' model. One of the strongest and most convincing cases against a universal cognitive system has come from the work of Michael Cole (1996). He describes, in detail, the failure of a research project in Africa. The research began by using assumptions as to what must be cognitive foundations of numerical competence based on what was known about these foundations from studies in the United States. The researchers arrived at the conclusion that the local children were seriously innumerate. but Cole describes his astonishment and chagrin when he saw those same children engaged in elaborate numerical operations in the local markets. This was his 'road to Damascus', recruiting a psychologist of the highest calibre to the discursive point of view!
In summary then, the Second Cognitive Revolution can be characterised as the abandonment of that form of the doctrine of scientific realism which animated the computer analogy initiated by Miller to fill out the cognitive hypotheses that were introduced within the framework of the hypothetico-deductive conception of theory.
Are we back to positivism? If we abandon the idea of hidden mechanisms, the processes which are responsible for the observed phenomena, the psychological phenomena so to speak, are we not confining psychology to the study of behaviour? The answer to this question is 'No'. This is because the Second Cognitive Revolution is as cognitive as the First, and, by virtue of the way that it is cognitive, brings the private and personal into the same arena as the public and collective. Though it is Vygotskian in that it is based upon the idea that the paradigm cases of cognition are linguistic or symbolic interactions among groups of people it is based on the principle that individual minds come into being just in so far as the skilful use of words and symbols acquired in the public domain are used in ways that are private and personal. Cf Vygotsky's (1978: 57) aphorism:
Every function in the child's cultural development
appears twice: first, on the social level, later, on the
There is a basic model to which discursive psychologists seem to adhere, implicitly or explicitly: namely, 'conversation', bearing the general force of an analogy for an ordered sequence of meaningful exchanges having, more often than not, some upshot or end, defined by local rules and customs. The use of this model is rooted in the Vygotskian belief that the way individuals think and act is to be explained by reference to the aspects of interpersonal conversation that they have appropriated. Overarching this idea is the principle that the distinction between public and private action and discourse is, effectively, the result of a human practice, reflecting the art of keeping things that one thinks, feels, plans and so on to oneself, an art not acquired until the second or third year of life.
To the question, 'What is the human mind?' the discursive psychologist answers, it is the private conversation, the private use of symbols, the private exercise of cognitive skills. But it is also the public conversation, the public exercise of cognitive skills in the joint creation of a conversation or some other kind of symbolic interaction, say a game of tennis.
So indeed, unlike the behaviourists, the cognitive psychologists of this second wave are insistent upon the reality and importance and accessibility of human minds. That which has been made private can become public. Reporting a private conversation is no more remarkable an epistemic feat that reporting a public one. It was the idea that there was something radically distinct, ontologically diverse, between the mind and its operations, and the body and its behaviour that led to the Cartesian impasse in the first place. But if mind is not a substance but a symbolic process, then there is as much mind in interpersonal conversations as there is in private soliloquies. There is as much mind in the architect's drawing office and the drafting of a plan by a team of designers, as there is in the private picturing of some scheme of urban renewal, yet to be realised, but thought about by an ambitious politician. From the discursive point of view, beliefs, attitudes, memories, emotions, ratiocination of all kinds are not mental states and processes, They are not entities of any sort, mental or otherwise. They are phenomena which have their being as attributes of public and private activities, in which people put local symbolic systems to work for all sorts of purposes. For example memories are products of remembering, nor sources of recollections.
How are we going to tie an individual to his or her practices if there is nothing behind language but language use, nothing behind reasoning but dialectical discourse (Billig, 1987) and so on? What must be true of an individual who can perform these wonderful feats? They must possess the relevant skills. On this view, the fundamental explanatory level that defines the limits of the science of psychology will be the identification of the basic skills the joint implementation of which results in those discursive processes in the which the psychological phenomena in which we take an interest are immanent. Wittgenstein remarked that it would be a disaster if having eliminated the Cartesian mind from playing the mythical role of the mechanism behind what people did, we were to introduce the brain instead. But brain damaged people lose their skills! Surely 'skill' cannot be the deepest level psychological concept? Are there not brain states and processes that account for skills and so must be admitted as the proper study of a psychological science? In a way this right, but the point must be made with great care.
How are human brains, human skills and human conversational interactions related? I have been arguing that from the point of view of discursive psychology the brain is one amongst the various bodily organs which can be put to use by people in carrying out the tasks which from time to time are required of them. It is more useful than the hand, while the hand is more useful than the knee, in most circumstances. My argument will be directed towards the idea that the brain has a place in human life comparable to that of the hand. It is something that we use, not something that uses us. Just as one uses one's hand to open a door, one uses one's brain to solve the problem of remembering the combination of the lock if one has forgotten to write it in some convenient but concealed place. It follows that just as tennis players take the Newtonian physics of the movements of racquet and ball for granted, though without these movements the instruments of play would not perform the way they do, so must psychologists recognise that in human affairs there are two quite different mutually irreducible sciences. There is the science of meanings, skills, strategies and rules (psychology) and there is the science of neural structures and processes and the genetic sources of the basic forms that these take (biology) with, as its domain of study, those aspects of the human form of life, such as the motor skills, natural expressions of bodily feelings and so on, without which the culturally various and historically contingent psychologies of human kind could neither develop nor be transferred to future generations. To ground a skill in a neural process is not to reduce the one to the other, but to identify the means by which the skill is exercised.
In the light of these observations, we can now formulate the main thesis of the Second Cognitive Revolution. It is that many psychological phenomena exist only as properties of discourses, that is, discourses considered rather generally as any kind of symbolically mediated interaction, including interactions with oneself. For example, conversation and soliloquy are the kinds of processes within which, and as properties of which, we would expect to find many of those phenomena which have been taken to be psychological, for example, attitudes, emotions, decision making, ad so on.
The model for the relation between the neurophysiological substrate of psychological performances and the psychological processes that are the subject matter of the discipline is emphatically not based on the Cartesian paradigm in any form. Mental phenomena are not distinguished from neural phenomena as attributes of different substances, the one immaterial and other material. That way lies 350 years of philosophical frustration and badly formed psychologies. Instead discursive psychology, very much in the spirit of Wundt's two human sciences, must frame psychology in a different metaphysics in which there is only one substance, the mysterious material stuff of the world. Intentionality is not an additional and mysterious property that symbols have and non-symbols lack, A symbol just is a material thing used by people for this or that purpose.
The adoption of the concept of 'person' for identifying the fundamental particulars of a possible scientific psychology distances the theories developed within that framework as far as it is possible to go from the Cartesian point of view. Persons are essentially embodied. We pass them in the street; they drive buses; man the checkouts of supermarkets; too many of them in an elevator car will trigger the mechanism that halts the system; they are routinely subject to moral praise and censure; they are not born as such but a produced by the work of other persons; and so on. The last thing they are is immaterial substances, contingently attached to otherwise inanimate bodies. The idea that persons are the basic particulars of the human world, the sources of cognitive activity and agentive in all sorts of other ways, is again nothing new. In recent times it served as the foundation for Stern's (19 ) Psychology from a Personalist Point of View and was the analytical conclusion of Strawson's (1959) famous study, Individuals. Just as the concept of 'elementary charge' grounds the science of physics, so 'person' ought to ground the science of psychology.
The duality between powerful particulars, basic agents, and the means by which their joint activities are realised, between tool-users and tools, allows us to draw a basic analogy between physics and psychology. Research in physics is controlled by the thesis that the observable dispositions of material things are to be explained by the powers of this things to act in the circumstances specified in the disposition. A material body falls under gravity, not because it has been given a kick from outside, but because its tendency to accelerate towards the centre of the relevant field has been released by removing some constraint that stopped it falling. Because material things as we know them in everyday life are complex aggregates of parts, the hierarchies of dispositions and powers that are deployed by physicists are elaborated until a level is reached in which there are only charges and their fields, that is beings which are possessed of powers which are not further grounded.
In psychology we reach a level at which observable skills
and dispositions are explained by primary powers ungrounded in
any underlying psychological processes or structures rather
quickly. Nevertheless these powers are grounded in neural
structures. My skills as a speaker are grounded in structures
and processes in my brain and nervous system but are not
reducible to them. The irreducibility comes simply from the
fact that distinct entities in each level are related to one
another in different and irreducible networks, meanings with
meanings, neural states with neural states. My words have the
meaning they do by virtue of the joint act of speaking and
listening in a highly particular cultural context. My
molecules, at least at a simple biochemical level of analysis,
require no other molecules to exist, and, at least their
atomic constituents are stable in a wide variety of
Why should one accept this proposal for yet another U-turn in
the long history of psychology's many volte-faces? What can be
done finally to realise the dreams of such as Dilthey and
Stern for a truly scientific psychology. It will not be
'scientific' in the superficial sense of taking numerical
measures and formulating laws in algebraic form. It will be
scientific in trying to realise the metaphysical foundations
of physics in a realm of active agents, and in following
ideals of precision of description in a mode appropriate to
the type of phenomena being studied, namely symbolic
practices. I shall try to justify the inception of the Second
Cognitive Revolution, a move from the invention of diaphanous
mental mechanisms to studies of the norms, rules and
conventions of concrete symbolic practices by demonstrating
the power of this approach in the study of some particularly
I suppose the most striking successes of the discursive point of view in enriching and opening up a field of research possibilities has been in the psychology of the emotions (Harré & Parrott, 1996). This development is now so well known that I shall give only a brief sketch of its several aspects.
The study of emotions was transformed by the realisation
that a bodily state supports an emotion only when certain
cognitive conditions are met:
As a consequence of noticing the role that locally valid interpretations played refining universal physiological reactions, it seemed likely that different cultures would have repertoires both a differing versions of some emotions and, in some cases, different emotions altogether (Wierzbicka, 1992).
In the light of these observations it is clear that emotion displays have proper (and improper) locations in the sequences of actions in which a group of people are engaged. The 'conversation' model seems appropriate in many cases. For example a display of embarrassment can be read as expressing something like 'I am sorry I fluffed it', and a display of anger as a protest and an accusation. These displays will have antecedents in the context of which they will make sense, and they will have consequents, which will depend on their meaning for those engaged, even if it is just the one person as sometimes happens. Of course this not to deny a role for neurophysiology in the study of emotions, and even of genetics, but they provide, at best, a substratum to emotions proper.
Part of the art of psychological method is to be right in distinguishing those psychological phenomena for which the holistic principle that we have seen as central to Wittgenstein's conception of how private experience and public manifestations are linked, holds and those for which it does not. There is no epistemological gap between a feeling and the expressions of a feeling. I shall be arguing that there is no epistemological gap between one important, since universal, aspect or component of a sense of personal uniqueness and the expression of that sense in one's use of pronouns and other indexical devices.
To develop this insight as a contribution to the complex topic of the psychology of personal identity, in analysing and accounting for one's sense of personal uniqueness, the next step will be to look closely at the grammar of the first and second person. I will try to show that it is not like the grammar of referenial expressions such as proper names. 'I' and 'you' are not names, and particularly they are 'variable names'.
Using words to refer to things is part of the language game of describing. If self-talk is expressive rather than descriptive then 'I' (and 'you') may have no referential function to inner entities, such as Cartesian ego or the soul. There need be no object, reference to which guarantees that the words of the first person and second person vocabulary are meaningful. Their grammar is related to the singularity of the publicly identifiable persons who use them.
Pronouns fall into two broad classes, from the point of view of the topic in hand. A third person pronoun is, literally a pro-noun, and can, with preservation of sense, be substituted for and by a proper name or other referring expression. Such pronouns are said to be anaphoric. In 'Jim fell off his bike and broke his collar bone' the successive occurrences of 'his' carry forward the same reference as does the proper name 'Jim'. Nothing about Jim's sense of personal uniqueness can be expressed by me in using the third person to refer to him. In so far as Jim is a unique being he appears as such in the public world where he is individuated by such criteria as bodily appearances and temporal continuity.
Third person expressions are sometimes called endophoric'. In 'Bill took up his saxophone and he played it to the surprise of his guests' the reference of 'he' and 'his' can be understood because we do not need to attend to anything external to the statement to fix their sense. First and second person pronouns are sometimes said to be exophoric, their sense being fixed in part by attributes of the person speaking and of the act of speaking itself, which are matters external to the utterance.
First and second person pronouns are indexicals, having
similar grammatical properties to 'this', 'here', and 'now'. I
shall try to show that they are used to express one's sense of
personal uniqueness, in several dimensions. And in accordance
with the generally Vygotskian tone of this paper, that one's
sense of uniqueness is, in part, brought into being through
the acquisition of the skill of using these pronouns in a
locally acceptable way.
For example in 'I can feel it now' which person is feeling the draft and when they are feeling it is fixed exophorically by speaker location and moment of utterance. The time of the occurrence reported is not expressed in the first person pronouns of Indo-European languages, but in the tenses of verbs.
The indexicality of a word can be interpreted as the fact that its meaning is completed on each occasion of use by some knowledge of the conditions of each particular utterance. It can also be interpreted as indexing the descriptive content and/or the moral force of an utterance with particular facts about the speaker and the utterance at the moment of the speaking. In what follows I shall be emphasizing the latter aspect of indexical expressions. For example the situated use of 'I' indexes the empirical content of a statement like 'I can hear a noise outside' with the spatial location of the embodied speaker, that is we know from which point the noise is audible.
The grammar I shall be describing as a general background to understand how the sense of personal uniqueness differs from culture to culture, is an amalgam of the grammars of many different languages, each of which includes some but not all the indexical properties we know from comparative studies to be available to speakers of particular languages.
An indexical expression marks the sense of an utterance with the relevant location of the speaker and his or her utterance in some array of entities within which an act of speaking occurs. There are two major arrays involved in first person grammar. There is the array of material things amongst which the body of the speaker has a place. And there is the array of events amongst which the event of the act of uttering is located. The structures of these arrays, as they underpin particular conversations are organised in all sorts of ways in different cultures. For example in Yucatan Maya the demonstratives, elaborations on 'this' and 'that', index the content of an utterance with the location of speaker, addressee and object spoken about in to 'line of sight'. 'That tree' is expressed differently when the tree is visible to both of us from when it is visible only to me.
The 'positions', the technical sense of Positioning Theory (Harré & van Langenhove, 1998) of speakers in locally relevant arrays of embodied persons is also reflected indexically. An array of persons can be ordered with respect to the speaker by such moral relations as duty and obligation and by multiple patterns of status relations.
Drawing on a wide range of languages we can assemble the following generic indexical grammar of the first person, reminding ourselves that the main first person indexical expression in some languages may not be a pronoun:
Where would one look for an expression of the seemingly elusive sense of personal uniqueness? Presumably, among other possibilities, in the uses of the first person, in pronouns and verb inflexions. Somewhere in the grammar of these devices the sense of self might be revealed. My sense of self, of being a singularity, a unique being then is a sense of having a location in space, a moral position in relation to other persons, a social standing in relation to other persons, and of having a life trajectory in time, of events that come before and will, perhaps, come after the moment of the act of uttering. The least likely to display lability and diversity is spatial location. The most likely to change from encounter to encounter and from time to time is one's sense of having a unique location in relation to other persons in the local social order or orders.
What do the four indexicalities we have identified in reflexive talk mean in terms of psychological functions? We can summarize the main points briefly, though each implies much in the way of developmental psychology of the Vygotsky/Wittgenstein sort.
Vygotsky's thesis stands out clearly: the native powers and tendencies to sensation, action, dependency and memory become the powers of an ordered human mind by the acquisition of first person skills of all sorts, prominent amongst which are linguistic and practical competencies (Shotter, 1983). Sensation becomes perception, action becomes agency, dependency becomes respect and memory becomes autobiography. Complementing these linguistic skills we can expect to find motor skills of all sorts, contributing in their way to a person's sense personal being.
We can also see which aspects of the sense of personal uniqueness are likely to vary culturally and historically and which are very unlikely to do so (Yoon & Choi, 1994). We can also see where pathologies of the self could take root, and even conjecture what might be their expressive forms (such as the oft-reported favouring of 'me' over 'I' in the talk of schizophrenics). Anthropology generally gives us the index of spatial location in the embodied person as a transcultural and robust feature of all indexical grammars. Even if multiple personality syndrome were to be a real disordering of mind in some cases, each of set of such multiple 'Is' is embodied in just one body (Thigpen & Cleckley, 1957). The indexing of social 'standing' seems to be the most labile and the most multiple in human individuals.
Though the use of grammatical analysis as a psychological
method depends on the Wittgensteinian distinction between
descriptive and expressive uses of language not all contexts
of application of the technique are alike. While there is an
ethology of feeling there is no ethology of the expression of
selfhood, no natural expression of being just one and the same
person. In accord with the Vygotskian view of the development
of the higher cognitive faculties, the self-organization that
is expressed in the grammars of indexical pronouns and other
functionally equivalent grammatical devices comes about in the
very learning of the grammar and other intentional,
normatively constrained skills with which it is expressed.
However, as many have argued, there are natural powers, some
of which are listed in this discussion too, which, so to say,
play the part of an ethology of self. One such power would be
that of making self-ordered movements in a physical
environment, a power which is partly perceptual and partly
What is the relationship between discursive psychology and
post-modernism? This question has often been obscured by the
use of the phrase 'social constructionism' to characterise
both discursive psychology and some versions of post-modernism. I will try to make the relation clearer by setting
out those principles and insights that discursive psychology
shares with post-modernism and those which it does not.
As a general rule I think it would right to say that whereas a
post-modernist author such as K. J. Gergen (1992) claims that
there are no viable dichotomies, no determinate meanings and
no facts of the matter, and no fixed human characteristics
that are more than locally valid, discursive psychologists
would take a more moderate position. Each method of enquiry
reveals some aspect of a reality that is independent of
individual persons, and some methods are more usefully
revealing than others. While no meanings are given as
determinate they can be made so in such a way that these
determinate meanings are shared by the participants in an
episode. While it would be a mistake to apply traditional
dichotomies in all contexts of enquiry there are none-the-less
some contexts of enquiry in which they may have a point.
Finally while discursive psychologists would be disinclined to
make any a priori assumptions about the existence of species-wide psychological processes and phenomena, independent of
history, most would regard the question as an empirical issue.
There may be ways of acting and being that are first locally
discernible, but are found step by step to be present in all
human forms of life. For example it seems plausible to claim
that some form of social hierarchy is found among all groups
of people, species wide, just as a sense of spatial location
is universal among human beings in their sense of self.
It is a striking, and, at first glance, surprising fact about mainstream academic psychology, that not only does it employ causal metaphysics, but also uses statistical, nomothetic methods to arrive at causal hypotheses. The historical origins of this methodology have been thoroughly explored by Danziger (1990). His studies have shown how the use of statistical analyses and sampling developed, not from any scientific considerations, but from the needs of the educational and military services which were the main clients for the work of psycholoogists. 'Confidence levels' are used to ensure that correlations do not occur by chance. But how should the non-random correlations revealed by statistical methods be interpreted? There are two main issues here.
The first is general and is not especially to do with the causal versus agentive issue in psychology. It is this: can anything be said about the propensities of individual human beings, on the basis of a statistical distribution of relations between type of response and type of situation found by using a statistical analysis on the results of studying a population? The answer, of course, is no. This is not a matter for debate, but simply a mathematical feature of statistics, DuMas's Theorem. It is astonishing how often this mistake is made. If 90% of a population, all treated in the same way, have a certain characteristic or reaction, it does not follow than any one them has a propensity, strength 0.9, to react that way. Those in the major segment have a propensity 1.0 and those in the minor segment a propensity 0.0. The chances of an unexamined person falling in the major segment is 0.9 indeed, but that is a quite different conclusion. This and other issues concerning the relation between statistical results and knowledge of individually located processes have been superbly analysed by Lamiell (1987). Statistical methods can yield only demographic hypotheses.
The second issue is very much a matter of rival
interpretations. Bearing in mind that a great deal of
empirical work in psychology is carried out using
questionnaires, vignettes and so on, to which 'subjects' are
asked to respond verbally, what must be the interpretation of
correlations, between question-types and answer-types and
between different aspects of answer-types? The use of
expressions like 'instrument' and 'measure', borrowed from the
natural sciences (and ultimately from everyday activities
liking cooking) has misled psychologists into thinking that by
these means causal relations could be discovered. But, given
the nature of the methods, the analysis of patterns of
questions and answers, the only possible explanation of such
correlations must be discursive, linguistic or narratological.
A correlation between a question and an answer, or between two
components of an answer, either displays a semantic rule, a
narrative convention or, rarely, some local experience of
'what goes with what'. It would be absurd to suppose that
asking a question of a certain sort causes someone to give a
certain kind of answer. According to the principles of
discursive psychology contemporary methods of enquiry reveal
discursive rules and conventions. They do not, since they
could not, reveal causal relations. If people tend to say that
liking someone is a factor in being friendly with them, we
have not found one of the causes of friendship, but a semantic
rule for the use of the word 'friend' or some synonym: namely,
that we apply the word to those we like, ceteris paribus. Far
from this being a defect in contemporary methodology it is to
be welcomed. A full and systematic development of the semantic
interpretation of this kind of 'experimenting' can be found in
Smedslund's many studies, summed up in his (1988) work,
Abelson, R. P. (1976) 'Script processing in attitude formation
and decision making'. In J. S. Carroll & J. W. Payne,
Cognition and Social Behavior New York: Earlbaum.
Billig, M. (1987) Arguing and Thinking Cambridge; Cambridge
Brown R. W. & Gillman, A. (1964)  'Pronouns of power
and solidarity'. In R. Brown, (Ed.) Psycholinguistics New
York, Free Press, Ch. 1.
Bruner, J. S. (1980) Beyond the Information Given London:
Allen & Unwin.
Bruner, J. S. (1983) In search of mind New York: Harper and
Chomsky, N. (1997) 'Language and cognition'. In Johnson and
Erneling, op. cit. pp. 1 - 31.
Cole, M. (1996) Cultural Psychology Cambridge, Mass.: Harvard
Danziger, K. (1990) Constructing the Subject Cambridge:
Cambridge University Press.
Danziger, K. (1997) Naming the Mind London: Sage.
Dixon, R. A. (1995) Compensating for Psychological Deficit and
Decline Mahwah, N. J.: Erlbaum.
Dilthey, W. (1985) Introduction to the Human Sciences: Wilhelm
Dilthey. Selected Works. Volume 1. Trans. R. A. Makkreel & F.
Rodi, Princeton, N. J.: Princeton University Press.
Edwards, D. & Potter, J. (1992) Discursive Psychology London:
Fodor, J. G. (1976) The language of thought Hassocks:
Freud, S. (1895 - 1938) Project for a Scientific Psychology.
See Standard Edition of the Complete Psychological Works of
Sigmund Freud London: Hogarth, Vol. 1.
Freud, S. (1933) New Introductory Lectures on Psychoanalysis
Garfinkel, H. (1957) Studies in Ethnomethodology Englewood
Cliffs: Prentice Hall.
Gergen, K. J. (1992) The saturated self New York: Basic Books.
Harré, R. and Secord, P. F. (1972) The explanation of social
behaviour Oxford: Blackwell.
Harré, R. & van Langenhove, L. (eds.) (1998) Positioning
Theory Oxford: Blackwell.
Johnson, D. M. & Erneling, C. E. (1997) The Future of the
Cognitive Revolution New York: Oxford University Press.
Lamiell, J. T. (1987) The Psychology of Personality: An
Epistemological Inquiry New York: Columbia University Press.
Luria, A. R. (1973) The Working Brain Harmondsworth: Penguin.
MacCorquodale, K. & Meehl, P. (1948) 'On a distinction between
hypothetical constructs and intervening variables'.
Psychological Review 55, 95 - 107.
Miller, G., Gallanter, and Pribram, K. (1960) Rules and the
structure of behavior New York: Holt
Muhlhausler P. and Harré, R. (1993) Pronouns and people
Shotter, J. (1984) Selfhood and social accountability Oxford:
Shweder, R. A. (1991) Thinking through cultures Cambridge,
Mass.: Harvard University Press.
Silver, M., Sabini, J. & Parrott, G. J. (1987) 'Embarrassment:
A Dramaturgic Account' Journal for the Theory of Social
Behaviour 17: 47 - 61.
Smedslund, J. (1988) Psychologic Heidelberg and New York:
Stearns, P. N. & Stearns, C. Z. (1986) Anger: The Struggle for
Emotional Control in America's History Chicago: Chicago
Stern, W. (1925) Psychology from a Personalist Point of View
Strawson, P. F. (1959) Individuals London: Methuen
Thigpen C. H. & Cleckley H. M. (1957) The Three faces of Eve
New York: McGraw Hill.
Vygotsky, L. S. (196 ) Thought and Language Cambridge, Mass.:
Harvard University Press.
Vygotsky, L. S. (1978) Mind in society Cambridge, Mass.:
Harvard University Press.
Wierzbicka, A. (1992) Semantics, Culture and Cognition Oxford:
Oxford University Press.
Wittgenstein, L. (1953) Philosophical investigations. Trans.
G. E. M. Anscombe & G. H. von Wright, Oxford: Blackwell.
Wundt, W. (1917 - 1926) Volkerpsychologie (10 volumes),
Yoon, G. & Choi, S-C. (1994) Psychology of the Korean People
Seoul: Dong-A Publishing.
I confess to my chagrin that I taught a course on information theory as a psychology, going on about the entropy of channels and so on, without proper attention to the ontology such theorising assumed.