The book proposes a new analysis of mental causation and applies it to the problem of mental disorder and its etiology. The proposed analysis grounds mental causation in information- processing, which characterizes not only mental but also biological functions. It involves driving Brentano's thesis downwards, with the argument that biological processes have (genuine) intentionality and that there is a seamless transition through developmental biology and psychology to mature human cognition. This is primarily a philosophical and conceptual case, drawing on examples from biological and psychological research. It aims to provide a justification for, and explicit statement of the implications of, assumptions that are already commonplace in the behavioural and neurosciences. In contrast with current philosophical analyses of the role of intentionality in behaviour, the analysis shows how the conditions for intentionality in the normal case contain the conditions for disordered functioning. These proposals are worked out in relation to major areas of psychopathology, including depression, anxiety disorders, schizophrenia, and personality disorders.
PSYCOLOQUY CALL FOR BOOK REVIEWERS: Below is the Precis of "Mind,
Meaning & Mental Disorder: The Nature of Causal Explanation in
Psychology & Psychiatry" by D.Bolton & J.Hill (590 lines). This
book has been selected for multiple review in Psycoloquy. If you
wish to submit a formal book review please write to
psyc@pucc.princeton.edu indicating what expertise you would bring
to bear on reviewing the book if you were selected to review it.
(If you have never reviewed for PSYCOLOQUY or Behavioral & Brain
Sciences before, it would be helpful if you could also append a
copy of your CV to your inquiry.) If you are selected as one of the
reviewers and do not have a copy of the book, you will be sent a
copy of the book directly by the publisher (please let us know if
you have a copy already). Reviews may also be submitted without
invitation, but all reviews will be refereed. The author will reply
to all accepted reviews.
FULL PSYCOLOQUY BOOK REVIEW INSTRUCTIONS AT:
http://www.princeton.edu/~harnad/psyc.html http://www.cogsci.soton.ac.uk/psycoloquy/
Psycoloquy reviews are of the book not the Precis. Length should be
about 200 lines [c. 1800 words], with a short abstract (about 50
words), an indexable title, and reviewer's full name and
institutional address, email and Home Page URL. All references that
are electronically accessible should also have URLs.
CHAPTER 1. MIND, MEANING, AND THE EXPLANATION OF ACTION
1. The main theme of this chapter is that meanings are causes because they predict. Explanations in terms of meaningful, mental states are effective in prediction, and theory-driven predictive efficacy is prima facie strong evidence that explanations are causal (Fodor, 1987). Various aspects and illustrations of this idea are laid out in the first chapter, including theory of mind in application to others and to oneself. The next few chapters pick up the problems involved in the claim that meanings are causes: How do they relate then to the brain? In what sense 'causal'? And so on. The resulting analysis provides the basis of a model not only of mind but also of mental disorder.
CHAPTER 2. MIND, MEANING AND NEURAL CAUSATION
2. The problem of 'double causality', by meaningful states on the one hand and by the brain on the other, can be resolved on the assumption that the meaningful states in question are causal because they are encoded in the brain. The broad area of cognitive neuroscience is happy enough to proceed on the basis of this proposal, but nevertheless it runs into a great deal of philosophical flack. Historical considerations would predict this. In effect the proposal deconstructs two great dichotomies: the idea that the brain encodes meaning -- is a semantic, representational system -- collapses the distinction between 'res extensa' and 'res cogitans', and the inference that meaningful states are causal collapses the turn of the century distinction between meaning and causality (von Wright 1971).
3. The philosophical objections and rejoinders to them are run through in Chapter 2. The encoding thesis itself, in its strong form in Fodor's Language of Thought hypothesis, has been rejected because of its incompatibility with connectionist models of cognitive function (Ramsey et al, 1990). Another objection is that brain processes themselves, insofar as they involve only syntax, and perhaps specifically in the absence of consciousness, cannot yet be semantic (Cf. Searle, 1980). Or again, even if neural syntax does in some sense encode semantic properties, it must be the syntactic properties which really do the causing of behaviour, since these are alone proximate to the behaviour (McGinn, 1989; Fodor, 1987; Jacob, 1992). A familiar expression of this line of thought, which has generated a great deal of controversy and commentary, uses Putnam's "twin-earth" argument to show that meaning does not supervene on the brain, and therefore cannot be causal (Putnam, 1975; Fodor, 1982; McGinn, 1989; Bilgrami, 1992).
4. As against all this the proposal that meaningful states are causal because they are encoded in the brain has the advantages that it resolves the problem of 'double causality', and is innocently used to this effect throughout cognitive neuroscience. The arguments used to defend the proposal against all the above criticisms all turn essentially on the assumption, already prepared for in Chapter 1, that there is a logical, conceptual connection between meaning and its expression in intentional activity. This broadly Wittgensteinian approach to meaning, though currently somewhat unfashionable in the philosophy of cognitive science, in fact can do a great deal of constructive work in this area. Its effect is to ground meaning in the (higher-level, intentional) interactions between the living being and its environment, and this prevents problematic definition in other terms that are too subjective or too objective.
5. Consider for example the question of the definition of syntax and semantics, which is more or less directly involved in several of the above lines of thought. It is commonly assumed that syntax in general, and neural syntax in particular, should be defined in physicalist terms. But this relies simplistically on the model from classical computation theory. Outside this, the definition of 'sign' is better made in terms of it being what has meaning, and then the syntactic properties of a sign should be defined as whatever contributes to its having the particular meaning that it does. In this context it is speculation that syntax can be specified by physicalist descriptions. In particular, this option is ruled out by the theory of meaning which grounds it in intentional activity. As is well-known, intentional activity generally does not map onto physically defined processes. The implication is, then, that both neural syntax and neural semantics involve the regulation of intentional activity, and are to be described in terms of intentional, information-carrying states.
6. This conclusion undermines the argument that neural syntax as opposed to neural semantics must be what is causally relevant to behaviour. The distinction between syntax and semantics is indeed not really of any great significance here; what matters is rather the difference between intentional and non-intentional descriptions of brain processes. It is of course quite true that the brain processes can be described in non-intentional, physicalist terms, but what gets explained in these terms are neither more nor less than physical happenings in or by the body. On the other hand, if the task is to explain interactions between the brain and environment, then the appropriate form of description of what is happening in the brain is the intentional, i.e. that it is engaged in the pick-up, processing, and regulation of behaviour by, information. Both forms of explanation are possible and legitimate, and can be seen, e.g. in studies on functional neural scanning by P.E.T. or M.R.I.
7. The problem of 'double causality' -- meaningful on the one hand and neural on the other -- is thus resolved by a form of the 'dual explanandum' strategy envisaged by Kim (1991). In the form proposed here however it does not imply adherence to dualism. On the contrary, meaningful causality as envisaged here is a matter of neural processing, and the brain is of course a material substance. But this neural processing is of information (and meaning), and this generally cannot be specified in physicalist terms.
CHAPTER 3. RELATIVITY IN KNOWLEDGE OF MIND AND MEANING; AND IN KNOWLEDGE GENERALLY.
8. The main theme of the third chapter is the relativity of meaning, the fact that it involves the representing agent. A particular case is of course the relativity involved in knowing other minds, recently clarified by the so-called Simulation Theory (Davies & Stone, 1995). Relativity has been one reason why meaning has in the past been excluded from the scientific domain, in general (von Wright 1971), and in the case of the new cognitive science in particular (Stich 1983). A background assumption here is that relativity, because it involves subjectivity, excludes objectivity. Dennett on the other hand has argued that it is possible to have both that meaning is relative to us and that it has objective reality (Dennett, 1988), and we agree with the controversial principle of having and eating this cake. Relativity does not exclude objectivity, but rather requires a relativistic rather than an absolutist definition of it. Such a relativistic definition runs in terms of constancies and comparisons between observations made from different points of view, and is familiar in contemporary physical science. The general point is that relativity, conceived as interaction between the measure and what is measured, is familiar in the C20th paradigms of science, whereas (pure) subjectivity had no place at all in the absolutist, C17th paradigms. This is another aspect of the deconstruction of dualism the last century.
CHAPTER 4. THE DEFINITION OF MEANING: CAUSAL AND FUNCTIONAL SEMANTICS
9. If meaning is a causal relation then an obvious and parsimonious option is to assimilate it to causal relations generally, for the definition of which we can look to current orthodoxy in the philosophy of science and metaphysics. There we find roughly and briefly this: A causes B iff A is always followed by B and the fact that it is, is covered by a natural, specifically a physical, law (Davidson, 1967; Fodor, 1980). The proposal that this can serve to define "B means (or carries the information that) A" goes under the name of Causal Semantics (Dretske, 1981; Fodor, 1990; also e.g. Baker, 1989). Causal Semantics runs up against a lot of well-known problems, chief of which are the following three. First, most meanings cannot be specified in physical terms nor then are they covered by physical laws. Second, there is no obvious account here of the meaning of empty terms, such as 'phlogiston'. Third, correlations between events can be more or less reliable, but there is no apparent sense to a MIS-correlation, and hence none so far to the notion of false or incorrect information.
10. According to the account so far proposed, the diagnosis of these difficulties is in each case the same, namely, that Causal Semantics neglects, in fact goes out of its way to deny, the subjective involvement to meaning, the fact that the representing system contributes to it. This deficit is made good by so-called Functional Semantics, which specifies the causal relations of content by reference to the design of the representing system, this being understood in evolutionary theoretic terms (Millikan, 1984; Papineau, 1987). While this grounding in evolutionary theory is critical and important, it is possible to construct definitions of content that refer primarily to the (intentional) activity of the system, and this behaviour-based approach is consistent with the line of argument so far. It supplies definitions of correctness/error and of content along the following lines:
11. A system S emits a certain response R, regulated by an information- carrying state Si with a particular content. The response (we assume) is meant to achieve a particular result, typically some change in the environment to a condition CR. The response is appropriate to achieving CR if the initial condition of the environment is CI. In this sense the response is appropriate to it being the case that CI. The informational state Si is then true if CI is in fact the case, and is otherwise false. Whether the state is true or false will then (tend to) show up in the success or otherwise of the behaviour to which it gives rise.
12. As would be expected, the above account of truth and error in informational states defines also informational content. In brief, an information-carrying state Si has a particular content, i.e. carries the information that the environment is in condition C, in case it tends to cause, other things being equal, behaviour appropriate to it being the case that C. Content is defined in this way in terms of its use in regulating systemic output. Content so defined cannot be circumscribed, as in Empiricism: it is not a matter of pre-given sensory qualities, but is as diverse and elaborated as is systemic function through phylogeny and ontogeny. To put the point another way, content is specific to particular kinds and instances of systems. It includes what is common to all or many living beings, such as edge-detection, but also, eventually, what is specific to individual human beings, so-called 'personal meanings'. This is the great span of the content that regulats human behaviour. Hence also the specificity of content corresponds to specificity of causal principles. The causal principles required to predict the behaviour of functional systems become increasingly specific. The end-point are principles which refer only to the individual and these define the 'self-caused' actions of the agent.
CHAPTER 5. TWO FORMS OF CAUSALITY IN BIOLOGICAL AND PSYCHOLOGICAL PROCESSES
13. It was implicit in the discussion of causal and functional semantics that causal principles that invoke content come into play already in biological systems generally, as opposed to being applicable to mental functions only. In effect we are driving Brentano's thesis downwards, proposing that intentionality is the mark of bios not just of psyche. Generally speaking this is the result of rejecting dualism. Dualism split the mechanical body from the intelligent mind, and post-dualism inevitably has to make bodily processes more mind-like and mental processes more embodied.
14. According to Searle (1992), 'intentionality is that property of many mental states and events by which they are directed at or about or of, objects and states of affairs in the world.' We argue that intentionality defined in this way is a causal property, not just of mental states, but of biological systems in general, so that the problem of mind and body is no more a problem in principle than that of intentional causal processes in protein synthesis or the regulation of blood pressure. We specify fifteen distinctive features of intentional causality, and show how they apply in biological and psychological systems. Reference is made to some of them in the following summarized account.
15. We take the regulation of blood pressure as a starting point. The rise and fall in frequency of impulses in the nerves involved in the regulation of blood pressure can be predicted only in terms of their intentionality with respect to that pressure. By this we mean not that the changes in frequencies are a consequence of the pressure exerted by the blood, rather it is determined by the specification within the system of the way in which frequencies provide information about blood pressure. This is no less causal but the underlying principles of causality are different. In relation to the regulation of blood pressure these include that there is reference to the possibility that the regulation can occur correctly or incorrectly, and that the nerve impulses are intentional only with respect to those key aspects of the blood pressure that are important to its regulation through alteration of heart rate and blood vessel volumes. More generally we argue that wherever we identify intentional causal process we will also find that there can be normal and abnormal function, and that particular features of stimuli are selectively represented in the service of action.
16. A further feature of intentional processes helps to consolidate this argument, namely that they entail rules, conventions and agreement. This is relatively non-controversial in relation to human psychological functioning. Attributions refer to the rules used to interpret events, language entails the rule bound use of words, and successful interpersonal functioning requires an understanding of the rules of social interactions. The rules stipulate how events are to be interpreted and what signs stand for. Whilst these rules are sometimes explicit, relatively fixed and followed consciously, they are usually unconscious and fluid. Children similarly follow social rules before they are old enough to know what they are. It is characteristic of rules that they are conventionalised and require agreement among the elements of the system in which they operate. For instance different languages work as conventions, and many additional languages employing further conventions can be envisaged. The rules and the conventions work only if they are held in common among those who use them.
17. It is perhaps less immediately evident that this applies also to biological systems in general. Take again the transmission of nerve impulses in the regulation of blood pressure. The frequency shifts that accompany changes in blood pressure carry information, so that a change from A to B in blood pressure produces an alteration in frequency of X to Y. However that alteration is not given by the physical characteristics of the blood pressure change but by the rule linking the change to the alteration in frequencies. In other words, unlike causal equations in the physical sciences, the force or energy entailed in a change of blood pressure does not enter into the equation specifying its representation. This is defined by the rule, and where there is one rule different rules can be envisaged. The same change in blood pressure could lead to a different change of frequency from X1 to Y1, and provided it bore the same systematic relationship to blood pressure, it would carry the same information about it and lead to the same regulatory actions. In other words what guarantees the information is the following of the rule that is conventionalised, in that it is not given by the physical facts of the causal agent (the blood pressure). This convention linking blood pressure changes to frequency alterations must be followed at other points in the system if the intentionality is to be preserved.
18. This analysis leads to the identification of two different kinds of breakdown in biological systems. Physical non-intentional interruption of intentional causal processes are familiar in medicine and often identifiable as pathology. Examples such as physical damage to the cardiovascular system, apparently entail reduction of the intentional to the physical, but it is the behaviour that lacks intentionality that is explained, and hence it is not a reduction. The second form of disruption occurs where the system is deceived, or applies inappropriate rules. Examples are infrequent in biological systems where rules are hard wired, but become critical to an account of function and dysfunction in human minds where rules and representations are significantly acquired and multiple, and where links to action are not guaranteed.
CHAPTER 6. INTENTIONALITY CAUSALITY, NEUROBIOLOGY, AND DEVELOPMENT
19. We take the key aspects of intentionality forward from non-psychological biological systems to more or less complex organisms in their environments, by showing that rule bound, conventionalised processes requiring agreement in the system are seen throughout. Intentionality is present in the simplest biological systems but its logical possibilities are realized only in the most complex.
20. We discuss physiological systems and then turn to the social systems of relatively simple organisms. Bees returning to the hive perform a dance in which the straight run through a figure of eight gives the direction, in its angle from the vertical, and distance, through its duration, of nectar bearing flowers (Von Frisch 1967). However there is no reason why a species of bees should not have evolved that used, for instance, the angle with the horizontal to give direction, or a different ratio of time of run to distance of flowers; or that used the same dance to indicate a different set of information. There must therefore exist a set of rules, which are conventionalised within the species and taken to mean certain things. Not only for instance which direction, but that it is direction which is being referred to. Not only what distance, but also that distance is what is referred to. Not only where and how far, but also that what is referred to is nectar. Thus before the evolution of human minds there evolved the operation of rules, requiring conventions, and these have been utilised generally within social organisations so that both what is referred to and the content of that reference is shared.
21. The logic of rule bound representational systems is that if one set of rules leads to a particular sequence of actions, then another set of rules could lead to different actions. Therefore an organism that could move among several, or many, different rule bound representational systems could elaborate a repertoire of responses to environmental challenges. This, we argue is the crucial evolutionary step that underpins the flexible creative intelligence of humans. This creativity is further enhanced because, as social organisms, the principal environmental challenges are other people, who also are capable of multiple representations of social events. We argue therefore that where there is a capability to acquire multiple rules and representations in a setting where other organisms also have this capacity there is a need for skills that include: firstly, capacity to generate and test rules and representations; secondly, capacity to monitor one's own representations; thirdly, capacity to monitor the rules and representations of others; and fourthly, the ability to establish joint rules of representation and action with others, and monitor their progress. Human minds have evolved the ability to develop these capabilities. However by definition they must be developed through the experience of others in social interaction. There is considerable evidence that infants and young children establish these skills in early face to face interactions with caregivers, through joint attention to the external world, in the establishment of selective attachments, and in imaginative play with other children (Stern 1985). As a corollary they cannot be 'wired in' at birth and hence human infants are characterised by their neuronal, physical, cognitive and emotional immaturity extending over many years. It follows from this that the immature human young require protection over longer periods than the young of other species. Thus, we argue, there is a link between the realization, in the human brain, of the potential already inherent in intentional processes throughout biological systems, the prolonged immaturity and openness to social experience of children, and their need for protection.
22. Given the extent of acquisition of diverse representations and actions it is crucial to maintain an integrative function over place and time and relationship, that is capable of reviewing not only separate representations and actions, but their similarities, differences and compatibilities. Subjectively the requirement is that the individual experiences his/her self, and is experienced by others, as a whole. The monitoring metarepresentational capacity typically involves consciousness and this contributes to the experience of the self. The experienced metarepresentational system, the self, is probably built up both through innate integrative capacities, and repeated experiences of actions performed under different circumstances. The capacity to represent the self and world in language and hence to construct theories about them are critical in the development of reason and reach maturation in human beings around adolescence (Inhelder & Piaget, 1958). The fundamental links between intentionality and action are preserved through development, existing not only in the biological beginnings, but also in human knowledge (Inhelder & Piaget, 1958; Wittgenstein, 1953, 1969). The capacity to represent one's own intentional states, in language, may be called second-order intentionality, and is discussed under this heading in the first chapter. Second-order intentionality may be distinctive to human beings (and mature ones at that), but (first-order) intentionality is a broader and deeper biological phenomenon.
CHAPTER 7. PSYCHIATRIC DISORDER AND ITS EXPLANATION
23. The identification of disorder of the body or the mind entails an analysis of intentional processes. Abnormality is not given by the physical facts but by the extent to which the integrity of the intentional causal system has been disrupted. It is useful to talk of levels of function in order to demarcate, not neuroscience or physiology from psychology, but different levels of intentionality. For instance, typically the symptoms of anxiety arise where the individual faces stressful or uncontrollable circumstances, and then the intentionality is clearly psychological. However anxiety symptoms occurring in the absence of such circumstances may arise from low blood sugar and then they have intentionality with respect to the regulation of blood glucose but not sources of stress, and so this is an intrusion from one level into another. Intrusion can also occur at the same level of intentional analysis. For instance anxiety might be provoked unexpectedly when an event triggers a memory in a person's mind hence introducing a new set of representations to the situation. These are all examples of disruption arising through the operation of intentional causal processes. Disruption can ALSO arise through non- intentional causal routes, for instance where anxiety is caused by thyroid dysfunction.
24. 'Biological psychiatry' has focussed on possible neurochemical and genetic causes of psychiatric disorder, hence creating an apparent distinction from psychological and social causes. Our analysis of intentional and non-intentional causality renders this a meaningless and misleading distinction. Concepts and research arising from ethological and 'social brain' hypotheses, and embedded within evolutionary biology, are much closer to the framework outlined here.
CHAPTER 8. INTENTIONALITY IN DISORDER
25. Where representations and actions are wired in the response to threat is relatively straightforward to predict. Either the threat falls within the repertoire of the organism and appropriate action ensues, or it falls outside and there is failure of, or ineffective, action. In humans there is the possibility of more than one representation and of distortion. This analysis of intentional-causal pathways in disorder is elaborated with reference to contemporary psychological models of disorder. In order to illustrate some points we may consider the case of child sexual abuse. A child faced with sexual abuse may represent accurately the events and the emotional responses, but this representation cannot underpin action where there is need for secrecy, and if the perpetrator is a parent, because the source of fear and the needed source of protection are the same person. Actions can however be preserved where there is another representation that omits the distress and distorts the perception of the facts. There is scope then for the development of a range of representations that are painful and in many respects accurate, but maintained as separate from action and possibly unconscious, and others that are distorted but, at least in childhood, compatible with action. Whilst the inactivated, 'forgotten' representations are in some respects accurate about external facts, they are open to distortions regarding the individual and action. For instance the child may associate the fear with self blame and helplessness. These aspects of the representation are not open to testing in part because they are maintained as inactive, and in part because of the hostile environment. Representations that preserve action may distort the facts along the lines 'this is not happening', or through incorrect attributions such as those of responsibility. If the child sees the abuse as his/her fault then it is possible to show affection to, and look for protection from, the perpetrator.
26. Whilst isolated traumas may be dealt with effectively using such strategies, it is likely that if they are repeated then the organisation of the individual's cognitive-emotional life and the regulation of action in relationships may be determined by such distortions. Vulnerability to psychological disorder may then arise from a range of intentional- causal pathways. Firstly, representations maintained out of consciousness and unconnected to action may disrupt current representations and actions giving rise to distress and undermining of actions. Secondly, conflicting and unintegrated representations may lead to an experience of self that is fragmented, unpredictable and inconsistent and hence disturbed. Thirdly, information processing may be impaired leading to inappropriate interpretations of events and of the individual's capabilities, with implications for mood and actions. Fourth, distorted perceptions about the self and others may lead to difficulties in relationships and consequently social isolation, or conflicted, unsupportive or violent relationships.
CHAPTER 9. PSYCHIATRIC CONDITIONS
27. The biological, developmental account of intentionality places representation and action at the heart of explanations of disorders irrespective of their presumed origins. This approach is developed in relation to a variety of psychiatric conditions including depression, schizophrenia, anxiety disorders and personality disorders. Contrast, for example, depression and schizophrenia. Routes to depression may include the intrusion of painful representations associated with helplessness, distorted attributions of the meaning of events and their implications for action, and misrepresentations of the self and others leading to relationship problems and hence lack of support. Current research supports the proposition that childhood traumas often make a substantial contribution, and the causal mechanisms are the subject of current research. In schizophrenia there is little evidence that childhood trauma plays a part. It is possible however that deficits in information processing leading to difficulties in understanding social contexts, and deficits in the monitoring of one's own actions may contribute to the development of symptoms. Symptoms such as delusions and hallucinations may be seen as sacrificing accuracy of representation in the preservation of action, and negative symptoms of lack of motivation and inactivity as failure of action in the face of disrupted representations. If such deficits are present from early in life then the metarepresentational and self functions may be limited, and then confusing but not traumatic environments may play a part in the genesis of symptoms.
Bolton, D. & Hill, J. (1996). Mind, Meaning & Mental Disorder: The Nature of Causal Explanation in Psychology & Psychiatry. Oxford University Press.
Bilgrami, A. (1992). Belief and meaning: the unity and locality of mental content. Oxford: Blackwell
Davies & Stone, 1995) Davies, M. & Stone, A. (1994) (ed). Mental simulation: Philosophical and psychological essays. Oxford: Blackwell.
Dennett, D. (1988). Precis of The intentional stance with peer commentary. The Behavioral and Brain Sciences, 11, 495-546.
Dretske, F. (1981). Knowledge and the flow of information. Cambridge Mass: MIT Press.
Fodor, J. (1980). Methodological Solipsism considered as a research strategy in cognitive psychology. With peer commentary. The Behavioral and Brain Sciences, 3, 63-109.
Fodor, J. (1982). Cognitive Science and the twin earth problem. Notre Dame Journal of Formal Logic, 23, 116-17.
Fodor, J. (1987). Psychosemantics. Cambridge, Mass.: M.I.T.
Inhelder, B. and Piaget, J. (1958). The growth of logical thinking from childhood to adolesence. New York: Basic Books.
Jacob, P. (1992). Externalism and mental causation. Proceedings of the Aristotelian Society, 92, 203-219.
Kim, J. (1991). Dretske on how reasons explain behavior. In B. McLaughlin (ed.) Dretske and his critics, Oxford: Blackell, 1991, pp.52- 72.
McGinn, C. (1989). Mental content. Oxford: Blackwell.
Millikan, R. (1984). Language, thought, and other biological categories. Cambridge, Mass.: M.I.T. Press.
Papineau, D. (1987). Reality and representation. Oxford: Blackwell.
Putnam, H. (1975). The meaning of 'meaning'. In Mind, language and reality. Cambridge University Press, 215-271.
Ramsey, W., Stich, S., and Garan, J. (1990). Connectionism, eliminativism and the future of folk psychology. In Cole D., Fetzer, J. and Rankin, T. (eds) Philosophy, mind, and cognitive enquiry. Dordrecht: Kluwer, 1990, 117-144.
Searle, J. R. (1992). The rediscovery of the mind. Cambridge, Mass.: M.I.T. Press.
Searle, J. (1980). Minds, brains and programs. With peer commentary. Behavioral and brain sciences 3, 417-57.
Stern, D. (1985). The interpersonal world of the infant. New York: Basic Books.
Stich, S. (1983). From folk psychology to cognitive science. Cambridge, Mass. MIT Press.
von Frisch, K. (1967). The dance, language and orientation of bees. Cambridge, Mass.: Harvard University Press
Von Wright, G. H. (1971). Explanation and understanding. London: Routledge and Kegan Paul
Wittgenstein, L. (1953). Philosophical Investigations. Ed. G.E.M. Anscombe and R. Rhees. Trans. G.E.M. Anscombe. Oxford: Blackwell.
Wittgenstein, L. (1969). On Certainty. Ed. G.E.M. Anscombe and G.H. von Wright. Trans. D. Paul and G.E.M. Anscombe. Oxford: Blackwell.