The worry about the ‘naturalizability’ of intentional properties is much harder to pin down. According to Fodor, the worry derives from a certain ontological intuition: That there is no place for intentional categories in a physicalistic view of the world, and thus, that the semantic and/or intentionality will prove permanently recalcitrant to integration in the natural order. If, however, intentional properties cannot be integrated into the natural order, then presumably they ought to be banished from serious scientific theorizing. Psychology should have no truck with them. Indeed, if intentional properties have no place in the natural order, then nothing in the natural world has intentional properties, and intentional states do not exist at all. So goes the worry. Unfortunately, neither Fodor nor anyone else has said anything very helpful about what is required to ‘integrate’ intentional properties into the natural order. There are, to be sure, various proposals to be found in the literature. But all of them seem to suffer from a fatal defect. On each account of what is required to naturalize a property or integrate it into the natural order, there are lots of perfectly respectable non - intentional scientific or common - sense properties that fail to meet the standards. Thus, all the proposals that have been made so far, end up being declined and thrown out.
 Now, or course, the fact that no one has been able to give a plausible account of what is required to ‘naturalize’ the intentional may indicate nothing more than that their project is a difficult one. Perhaps with further work a more plausible account will be forthcoming. But one might also offer a very different diagnosis of the failure of all accounts of ‘naturalizing’ that have so far been offered. Perhaps the ‘ontological intuition’ that underlies the worry about integrating the intentional into the natural order is simply muddled. Perhaps, there is no coherent criterion of naturalization or naturalizability that all properties invoked in respectable science must meet, as, perhaps, that this diagnosis is the right one. Until those who are worried about the naturalizability of the intentional provide us with some plausible account of what is required of intentional categories if they are to find a place in ‘a physicalistic view of the world’. Possibly we are justified in refusing to take their worry seriously.
 Recently, John Searle (1992) has offered a new set of philosophical arguments aimed at showing that certain theories in cognitive psychology are profoundly wrong - headed. The theories that are the target of computational explanations of various psychological capacities - like the capacity to recognize grammatical sentences, or the capacity to judge which of two objects in one ‘s visual field is further away. Typically, these theories are set out in the form of a computer program - a set of rules for manipulating symbols - and the explanations offered for the exercise of the capacity in question is that people’s brains are executing the program. The central claim in Searle’ s critique is that being a symbol or a computational stat e is not an ‘intrinsic’ physical feature of a computer state or a brain state. Rather, being a symbol is an ‘observer relative’ feature. However, Searle maintains, only intrinsic properties of a system can play a role in causal explanations of how they work. Thus, appeal to symbolic or computational states of the brain could not possibly play a role in a ‘casual account of cognition in knowledge’.
 All of which, the above aforementioned surveyed, does so that implicate some of the philosophical arguments aimed at showing that cognitive psychology is confusing and in need of reform. My reaction to those arguments was none too sympathetic. In each case, it was maintained to the philological argument that is problematic, not the psychology it is criticizing.
 It is fair to ask where we get the powerful inner code whose representational elements need only systematic construction to express, for example, the thought that cyclotrons are bigger and more than vast than black holes. Nonetheless, on this matter, the language of thought theorist has little to say. All that concept learning could be, assuming it is to be some kind of rational process and not due to mere physical maturation or a bump on the head. According to the language of thought theorist, is the trying out of combinations of existing representational elements to see if a given combination captures the sense (as evidenced in its use) of some new concept.  The consequence is that concept learning, conceived as the expansion of our representational resources, simply does not happen. What happens instead is that we work with a fixed, innate repertoire of elements whose combination and construction must express any content we an ever learn to understand. And note that it is not the trivial claim that in some sense the resources a system starts with must set limits on what knowledge it can acquire. For these are limits which flow not, for example, from sheer physical size, number of neurons, connectivity of neurons, and so forth. But from a base class of genuinely representational elements. They are more like the limits that being restricted to the propositional calculus would place on the expressive power of a system than, say, the limits that having a certain amount of available memory storage would place on one.
 But this picture of representational stasis in which all change consists in the redeployment of existing representational resources, is one that is fundamentally alien to much influential theorizing in developmental psychology. The prime example of a developmentalist who believed in a much stronger formsa much stronger form in genuine expansion of representational power at the very heart of a model of human development. In a similar vein, recent work in the field of connectivism seems to open up the possibility of putting well - specified models of strong representational change back into the centre of cognitive scientific endeavours.
 Nonetheless, the understanding of how the underlying combinatoric code ‘develops’ the deep understanding of cognitive processes, than understanding the structure and use of the code itself (though, doubtless the projects would need to be pursued hand - in - hand).
 The language of thought depicts thoughts as structures of concepts, for which in turn exist as elements (for any basic concept) or concatenations of elements (for the rest) in the inner code. The intentional states, as common - sense understands them, have both causal and semantic properties and that the combination appears to be unprecedented. However, a further problem about inferential role semantics is that it is, almost invariably, suicidally holistic. it seems, that, if externalism is right, then (some of) the intentional properties of thought are essentially ‘extrinsic’: They essentially involve mind - to - world relations. All and all, in assuming that the computational role of a mental representation is determined entirely by its intrinsic properties, such properties of its weigh t, shape, or electrical conductivity as it might be. , hard to see how the extrinsic properties: Which is to say, that it is hard to see how there could be computationally sufficient conditions for being in an intentional state, for which is to say that it is hard to see how the immediate implementation of intentional laws could be computational.
 However, there is little to be said about intrinsic relation s between basic representational items. Even bracketing the (difficult) question of which, if any words in our public language may express content s which have as their vehicles atomic items in the language of thought (an empirical question on which it is to assume that Fodor to be officially agnostic), the question of semantic relations between atomic items in the language of thought remains. Are there any such relations? And if so, in what do they consist? Two thought s are depicted as semantically related just in casse they share elements themselves (like the words of public language on which they are modelled) seem to stand in splendid isolation from one another. An advantage of some connectionist approaches lies precisely in their ability to address questions of the interrelation of basic representational elements (in act, activation vectors) by representing such items as location s in a kind of semantic space. In such a space related contents are always expressed by related representational elements. The connectionist’s conception of significant structure thus goes much deeper than the Fodorian’s. For the connectionist representations need never be arbitrary. Even the most basic representational items will bear non - accidental relations of similarity and difference to one another. The Fodorian, having reached representational bedrock, must explicitly construct any such further relations. They do not come for free as a consequence ee of using an integrated representational space. Whether this is a bad thing or a goo one will depend, of course, on what kind of facts we need to explain. But it is to suspect that representational atomism may turn out to be a conceptual economy that a science of the mind cannot afford.
 The approach for ascribing contents must deal with the point that it seems metaphysically possible for here to be something that in actual and counterfactual circumstances behaves as if it enjoys states with content, when in fact it does not. If the possibility is not denied, this approach must add at least that the states with content causally interact in various ways with one - another, and also causally produce intentional action. For most causal theories, however, the radical separation of the causal and rationalizing role of reason - giving explanations is unsatisfactory. For such theorists, where we can legitimately point to an agent’s reasons to explain a certain belief or action, then those features of the agent’s intentional states that render the belief or action reasonable must be causally relevant in explaining how the agent came to believe or act in a way which they rationalize.  One way of putting this requirement is that reason - giving states not only cause, but also causally explain their explananda.
 On most accounts of causation an acceptance of the causal explanatory role of reason - giving connections requires empirical causal laws employing intentional vocabulary. It is arguments against the possibility of such laws that have, however, been fundamental for those opposing a causal explanatorial view of reasons. What is centrally at issue in these debates is the status of the generalizations linking intentional states to each other, and to ensuing intentional acts. An example of such a generalization would be, ‘If a person desires ‘X’, believes ‘A’ would be a way of promoting  ‘X’, is able to ‘A’ and has no conflicting desires than she will do ‘A’. For many theorists such generalizations are between desire, belief and action. Grasping the truth of such a generalization is required to grasp the nature of the intentional states concerned. For some theorists the a priori elements within such generalization s as empirical laws. That, however, seems too quick, for it would similarly rule out any generalizations in the physical sciences that contain a priori elements, as a consequence of the implicit definition of their theoretical kinds in a causal explanation theory. Causal theorists, including functionalist in philosophy of mind, can claim that it is just such implicit definition that accounts for th a priori status of our intentional generalizations.
 The causal explanatory approach to reason - giving explanations also requires an account of the intentional content of our psychological states, which makes it possible for such content to be doing such work. It also provides a motivation for the reduction of intentional characteristics to extensional ones, on an attempt to fit intentional causality into a fundamentally materialist world picture. The very nature of the reason - giving relation, however, can be seen to render such reductive projects unrealizable. This, therefore leaves causal theorists with the task of linking intentional and non - intentional levels of description in such a way as to accommodate intentional causality, without either over - determination or a miraculous coincidence of prediction from within distinct causally explanatorial frameworks.
 The existence of such causal links could well be written into the minimal core of rational transitions required for the ascription of the contents in question. Yet, it is one thing to agree that the ascription of content involves a species of rational intelligibility. It is another to provide an explanation of this fact. There are competing explanations. One treatment regards rational intelligibility as ultimately dependent on or upon what we find intelligible, or on what we could come to find intelligible in suitable circumstances. This is an analogue of classical treatments of secondary qualities, and as such is a form of subjectivism about content. An alternative position regards the particular conditions for correct ascription of given contents as more fundamental. This alternative states that interpretation must respect these particular conditions. In the case of conceptual contents, this alternative could be developed in tandem with the view that concepts are individuated by the conditions for possessing them. These possession conditions would then function as constraints upon correct interpretation. If such a theorist also assigns references to concepts in such a way that the minimal rational transitions are also always truth - preserving, he will also have succeeded in explaining why such transitions are correct. Under an approach that treats conditions for attribution as fundamental, intelligibility need not be treated as a subjective property. There may be concepts we could never grasp because of our intellectual limitations, as there will be concepts that members of other species could not grasp. Such concepts have their possession conditions, but some thinkers could not satisfy those conditions.
 Ascribing states with content to an actual person has to proceed simultaneously with attribution of a wide range of non - rational states and capacities. In general, we cannot understand a person’s reasons for acting as he does without knowing the array of emotions and sensations to which he is subject: What he remembers and what he forgets, and how he reasons beyond the confines of minimal rationality. Even the content  - involving perceptual states, which play a fundamental role in individuating content, cannot be understood purely in terms relating to minimal rationality. A perception of the world as being a certain way is not (and could not be) under a subject’s rational control. Though it is true and important that perceptions give reasons for forming beliefs, the beliefs for which they fundamentally provide reasons - observational beliefs about the environment - have contents which can only be elucidated by referencing back to perceptual experience. In this respect (as in others) perceptual states differ from those beliefs and desires that are individuated by mentioning what they provide reasons for judging or doing: For frequently these latter judgements and actions can be individuated without reference back to the states that provide reasons for them.
 What is the significance for theories of content of the fact that it is almost certainly adaptive for members of a species to have a system of states with representational contents which are capable of influencing their actions appropriately? According to teleological theories of content, a constitutive account of content - one which says what it is for a state to have a given content - must make use of the notion of natural function and teleology. The intuitive idea is that for a belief state to have a given content ‘p’ is for the belief - forming mechanisms which produced it to have the function b(perhaps derivatively) of producing that state only when it is the case that ‘p’. One issue this approach must tackle is whether it is really capable of associating with states the classical, realistic, verification - transcendent contents which pre - theoretically, we attribute to hem. It is not clear that a content’s holding unknowably can influence the replication of belief - forming mechanics. Bu t even if content itself proves to resist elucidation in terms of natural function and selection. It is still a very attractive view that selection must be mentioned in an account of what associate ss something - such as sentence - with a particular content, even though that content itself may be individuated by other means.
 Contents are normally specified by ‘that . . . ‘ clauses, and it is natural to suppose that a content has the same kind of sequential and hierarchical structure as the sentence that specifies it. This supposition would be widely accepted for conceptual content. It is, however, a substantive thesis that all content is conceptual. One way of treating one sort of perceptual content is to regard the content as determined by a spatial type, the type under which the region of space around the perceiver’s must fall if the experience with that content is to represent the environment correctly. The type involves a specification of surfaces and features in the environment, and their distances are directed from the perceiver’s body as origin. Such contents lack any sentence - like structure at all. Supporters of the view that all content is conceptual will argue that the legitimacy of using these spatial type in giving the content of experience does not undermine the thesis that all content is conceptual. Such supporters will say that the spatial type is just a way of capturing what can equally be captured by conceptual components such as ‘that distance’, or ‘that direction’, where these demonstratives are made available by the perception in question. Friends of non - conceptual content will respond that these demonstratives themselves cannot be elucidated without mentioning the spatial types for which lack sentence - like structure.
 The actions made rational by content - involving states are actions individuated in part by reference to the agent’s relations to things and properties in his environment. Wanting to see a particular movie and believing that, that building over thee is a cinema showing it makes rational the action of walking in the direction of that building. Similarly, for the fundamental casse of a subject who has knowledge about his environment, a crucial factor in making rational the formations of particular attitude is the way the world is around him. One may expect, the n, that any theory that links the attribution of contents to states with rational intelligibility will be commit to the thesis that the content of a person’s states depends in part on his relations to the world outside him. We call this thesis the thesis of externalism about content.
 Externalism about content should steer a middle course. On the one had, it should not ignore the truism that the relations of rational intelligibility involve not things and properties in the world, but the way they are presented as being - an externalist should use some version of Frége’s notion of mode of presentation. On the other hand, the externalist for whom considerations of rational intelligibility are pertinent to the individuation of content is likely to insist that we cannot dispense with the notion of something in the world - being presented in a certain way. If we dispense with the notion of something external bing presented in a certain way, we are in danger of regarding attributions of content as having no consequence for how an individual relates to his environment, in a way that is quite contrary to our intuitive understanding of rational intelligibility.
 Externalism comes in more and fewer extreme versions. Consider a mind of a thinker who sees or perceives of a particular pear, and thinks a thought that the pear is ripe, where the demonstrative way of thinking of the pear expressed by ‘that pear’ is made available to him by his perceiving the pear. Some philosophers have held that the thinker would be employed of thinking were he perceiving a different perceptually based way of thinking were he perceiving a different pear. But externalism need not be committed to this. In the perceptual state that makes available the way on thinking pear is presented as being in a particular distance, and as having certain properties. A position will still be externalist if it holds that what is involved in the pear’s being so presented is the collective role of these components of content in making intelligible in various circumstances the subject’s relations to environmental directions distance and properties of object. This can be held without committed to the object - dependence of the way of thinking expressed by ‘that pear’. This less strenuous form of externalism must, though, address the epistemological arguments offered in favour of the more extreme versions, to the effect that only they are sufficiently world - involving.
 The apparent dependence of the content of belief on factors external to the subject can be formulated as a failure of supervenience of belief content upon facts about what is the case within the boundaries of the subject’s body. To claim that such supervenience fails is to make a model claim: That there can be two persons the same in respect of their internal physical states (and so in respect to those of their dispositions that are independent of content - involving states), who nevertheless differ in respect of which beliefs they have. Hilary Putnam (1926 - ), the American philosopher of science, who became more prominent in his writing about ‘Reason, Truth, and History’ (1981) marked of a subtle position that he call’s internal realism, initially related to a n ideal limit theory of truth, and apparently maintaining affinities with verificationism, but in subsequent work more closely aligned with minimalism. Putnam’s concern in the later period has largely been to deny any serious asymmetry between truth and knowledge as obtained in moral s, and even theology.
 Nonetheless, in the case of content - involving perceptual states. It is a much more delicate matter to argue for the failure of supervenience. The fundamental reason for this is answerable not only to factors ion the input side - what in certain fundamental cases causing the subject to be in the perceptual state - but also to factors on the perceptual state - but also to factors on the output side - what the perceptual state is capable of helping to explain amongst the subject’s actions. If differences in perceptual content always involve differences in bodily - described actions in suitable counter - factual circumstances, and if these different actions always will after all be supervenience of content - involving perceptual states on internal states. But if this should turn ut to be so, that is not a refutation of externalism for perceptual contents. A different reaction to this situation of dependence ads one of supervenience is in some cases too strong. A better is given by a constitutive claim: That what makes a state have the content it does are certain of its complex relations to external states of affairs. This can be held without commitment to the model separability of certain internal states from content - involving perceptual states.
 Attractive as externalism about content ma be, it has been vigorously contested notably by the American philosopher of mind Jerry Alan Fodor (1935 - ), who is known for a resolute realism about the nature of mental functioning. Taking the analogy between thought and computation seriously, Fodor believes that mental representations should be conceived as individual states with their own identities and structure, like formulae transformed by processes of computation or thought. His views are frequently contrasted with those of ‘Holist’ such as Herbert Donald Davidson (1917 - 2003), although Davidson is a defender of the doctrines of the ‘indeterminacy’ of radical translation and the ‘inscrutability’ of reference, his approach has seemed to many to offer some hope of identifying meaning as a respectable notion, even within a broadly ‘extensional’ approach to language. Davidson is also known for rejection of the idea of a ‘conceptual scheme’, thought of as something peculiar to one language or in one way of looking at the world, arguing that where the possibility of translation stops so does the coherence of the idea that there is anything to translate. Nevertheless, Fodor (1981) endorses the importance of explanation by content - involving states, but holds that content must be narrow, constituted by internal properties of an individual.
 One influential motivation for narrow content is a doctrine about explanation that molecule - for - molecule counter - parts must have the same causal powers. Externalists have replied that the attributions of content - involving states presuppose some normal background or context for the subject of the states, and that content - involving explanations commonly take the presupposed background for granted. Molecular counter - parts can have different presuppose d backgrounds, and their content - involving states may correspondingly differ. Presupposition of a background of external relations in which something stands is found in other sciences outside those that employ the notion of content, including astronomy and geology.
 A more specific concern of those sympathetic to narrow content is that when content is externally individuated, the explanatorial principles postulated in which content - involving states feature will be a priori in some way that is illegitimate. For instance, it appears to be a priori that behaviour is intentional under some description involving the concept ‘water’ will be explained by mental states that have the externally individuated concept about ‘water’ in their content.  The externalist about content will have a twofold response. First, explanations in which content - involving states are implicated will also include explanations of the subject’s standing in a particular relation to the stuff water itself, and for many such relations, it is in no way a priori that the thinker’s so standing has a psychological explanation at all. Some such cases will be fundamental to the ascription of externalist content on treatments that tie such content to the rational intelligibility of actions relationally characterized. Second, there are other cases in which the identification of a theoretically postulated state in terms of its relations generates a priori truths, quite consistently with that state playing a role in explanation. It arguably is phenotypical characteristic, then it plays a causal role in the production of that characteristic in members of the species in question. Far from being incompatible with a claim about explanation, the characterization of genes that would make this a priori also requires genes to have a certain casual explanatory role.
 Of anything, it is the friend of narrow content who has difficulty accommodating the nature content are fit to explain bodily movements in environment - involving terms. But we note, that the characteristic explananda of content - involving states, such as walking towards the cinema, are characterized in environment - involving terms. How is the theorist of narrow content to accommodate this fact? He may say, that we merely need to add a description of the context of the bodily movement, which ensures that the movement is in fact a movement toward the cinema. But mental property of an event to an explanation of that event does not give one an explanation of the event’s having that environmental property, let alone a content - involving explanation of the fact. The bodily movement may also be a walking in the direction of Moscow, but it does not follow that we have a rationally intelligible explanation of the event as a walking in the direction of Moscow. Perhaps the theorist of narrow content would at this point add further relational proprieties of the internal states of such a kind that when his explanation is fully supplemented, it sustains the same counter - factuals and predications as does the explanation that mentions externally individuated content. But such a fully supplemented explanation is not really in competition with the externalist’s account. It begins to appear that if such extensive supplementation is adequate to capture the relational explananda it is also sufficient to ensure that the subject is in states with externally individuated contents. This problem, however, affects not only treatments of content as narrow, but any attempt to reduce explanation by content - involving states to explanation by neurophysiological states.
 One of the tasks of a sub - personal computational psychology is to explain how individuals come to have beliefs, desires, perceptions and other personal - level content - involving properties. If the content of personal - level states is externally individuated, then the contents mentioned in the sub - personal psychology that is explanatory of those personal states must also be externally individuated. One cannot fully explain the presence of an externally individuated state by citing only states that are internally individuated. On an externalist conception of sub - personal psychology, a content - involving computation commonly consists in the explanation of some externally individuated states by other externally individuated states.
 This view of sub - personal content has, though, to be reconciled with the fact that the first states in an organism involved in the explanation - retinal states in the case of humans - are not externally individuated. The reconciliation is affected by the presupposed normal background, whose importance to the understanding of content we have already emphasized. An internally individuated state, when taken together with a presupposed external background, can explain the occurrence of an externally individuated state.
 An externalist approach to sub - personal content also has the virtue of providing a satisfying explanation of why certain personal - level states are reliably correct in normal circumstances. If the sub - personal computations that cause the subject to be in such states are reliably correct, and the final commutation is of the content of the personal - level state, then the personal - level state will be reliably correct. A similar point applies to reliable errors, too, of course. In either case, the attribution of correctness condition to the sub - personal state is essentially to the explanation.
 Externalism generates its own set of issues that need resolution, notably in the epistemology of attributions. A content - involving state may be externally individuated, but a thinker does not need to check on his relations to his environment to know the content of his beliefs, desires, and perceptions. How can this be? A thinker’s judgements about his beliefs are rationally responsive to his own conscious beliefs. It is a first step to note that a thinker’s beliefs about his own beliefs will then inherit certain sensitivities to his environment that are present in his original (first - order) beliefs. But this is only the first step, for many important questions remain. How can there be conscious externally individuated states at all? Is it legitimate to infer from the content of one’s states to certain general facts about one’s environment, and if so, how, and under what circumstances?
 Ascription of attitudes to others also needs further work on the externalist treatment. In order knowledgeably to ascribe a particular content - involving attitude to another person, we certainly do not need to have explicit knowledge e of the external relations required for correct attribution of the attitude. How then do we manage it? Do we have tacit knowledge of the relation on which content depends, or do we in some way take our own case as primary, and think of the relations as whatever underlies certain of our own content - involving states? In the latter, in what wider view of other - ascription should this point be embedded? Resolution of these issues, like so much else in the theory of content, should provide us with some understanding of the conception each one has of himself as one mind amongst many, interacting with a common world which provides the anchor for the ascription of content.
 There seems to have the quality of being an understandably comprehensive  characteristic as ‘thought’, attributes the features of ‘intentionality’ or ‘content’: In thinking, as one thinks about certain things, and one thinks certain things about those things - one entertains propositions that maintain a position as promptly categorized for the states of affairs. Nearly all the interesting properties of thoughts depend upon their ‘content’: Their being coherent or incoherent, disturbing or reassuring, revolutionary or banal, connected logically or illogically to other thoughts. It is thus, hard to see why we would bother to talk of thought at all unless we were also prepared to recognize  the intentionality of thought. So we are naturally curious about the nature of content: We want to understand what makes it possible, what constitutes it, what it stems from. To have a theory of thought is to have a theory of its content.
 Four issues have dominated recent thinking about the content of thought, each may be construed as a question about what thought depends on, and about the consequences of its so depending (or not depending). These potential dependencies concern: (1) The world outside of the thinker himself, (2) language, (3) logical truth (4) consciousness. In each casse the question is whether intentionality is essentially or accidentally related to the items mentioned: Does it exist, that is, only by courtesy of the dependence of thought on the aid items? And this question determining what the intrinsic nature of thought is.
 Thoughts are obviously about things in the world, but it is a further question whether they could exist and have the content they do whether or not their putative objects themselves exist. Is what I think intrinsically dependent on or upon the world in which I happen to think it? This question was given impetus and definition by a thought experiment due to Hilary Putnam, concerning a planet called ‘twin earth’. On twin earth there live thinkers who are duplicates of us in all internal respects but whose surrounding environment contain different kinds of natural objects. The suggestion then is that what these thinkers refer to and think about is individuality dependent upon their actual environment, so that where we think about cats when we say ‘cat’ they think about that word - the different species that actually sits on their mats and so on. The key point is that since it is impossible to individuate natural kinds like cats solely by reference to the way they strike the people who think about them cannot be a function simply of internal properties of the thinker. The content, here, is relational in nature, is fixed by external facts as they bear upon the thinker. Much the same point can be made by considering repeated demonstrative reference to distinct particular objects: What I refer to when I say ‘that bomb’, of different bombs, depends on or upon the particular bomb in front of me and cannot be deduced from what is going on inside me. Context contributes to content.
 Inspired by such examples, many philosophers have adopted an ‘externalist’ view of thought content: Thoughts are not antonymous states of the individual, capable of transcending the contingent facts of the surrounding world. One is therefore not free to think whatever one’s liking, as it was, whether or not the world beyond cooperates in containing suitable referents for those thoughts. And this conclusion has generated a number of consequential questions. Can we know our thoughts with special authority, given that they are thus hostage to external circumstances? How do thoughts cause other thoughts and behaviour, given that they are not identical with an internal states we are in? What kind of explanation are we giving when we cite thoughts? Can there be a science of thought if content does not generalize across environments? These questions have received many different answers, and, of course, not everyone agrees that thought has the kind of world - dependence claimed. Nonetheless, what has not been considered carefully enough, is the scope of the externalist thesis - whether it applies to all forms of thought, all concepts. For unless this questions be answered affirmatively we cannot rule out the possibility that though in general depends on there being some thought that is purely internally determined, so that the externally fixed thoughts are a secondary phenomenon. What about thoughts concerning one’s present sensory experience, or logical thoughts or ethical thought? Could there, indeed, be a thinker for whom internalism was generally correct? Is external individuation the rule or the exception? And might it take the rule or the exception? And might it take different forms in different cases?
 Since words are also about things, it is natural to ask how their intentionality is connected to that of thoughts. Two views have been advocated: One view takes thought content to be self - subsisting relative to linguistic content, with the latter dependent upon the former: the other view takes thought comment to be derivative upon linguistic content, so that there can be no thought without a bedrock of language. Thus, arise controversies about whether animals really think, being non - speakers, or computers really use language. , being non - thinkers. All such question depend critically upon what one is to mean by ‘language’. Some hold that spoken language is unnecessary for thought but that there must be an inner language in order for thought to be possible, while others reject the very idea of an inner language, preferring to suspend thought from outer speech. However, it is not entirely clear what it amounts to assert (or deny)that there is an inner language of thought. If it means merely that concepts (thought constituents) are structured in such a way as to be isomorphic with spoken language, then the claim is trivially true, given some natural assumptions. But if it means that concepts just are ‘syntactic’ items orchestrated into springs of the same, then the claim is acceptable only in so far as syntax is an adequate basis for meaning - which, on the face of it, it is not. Concepts no doubt have combinatorial powers compactable to those of words, but the question is whether anything else can plausible be meant by the hypothesis of an inner language.
 On the other hand, it appears undeniable that spoken language does have autonomous intentionality, but instead derives its meaning from the thought of speakers - though language may augment one’s conceptual capacities. So thought cannot postdate spoken language. The truth seems to be that in human psychology speech and thought are interdependent in many ways, but there is no conceptual necessity about this. The only ‘language’ on which thought essentially depends itself: Thought indeed, depends upon there being insoluble concepts that can join with others to produce complete propositional statements. But this is merely to draw attention to a property any system of concepts must have: It is not to say what concepts are or how they succeed in moving between thoughts as they so. Appeals to language at this point, are apt to flounder on circularity, since words take on the power of concepts only insofar as they express them. Thus, there seems little philosophical illumination to be got from making thought depend on or upon language.
 This third dependency question is prompted by the reflection that, while people are no doubt often irrational, woefully so, there seems to be sme kind of intrinsic limit to their unreason. Even the sloppiest thinker will not infer anything from anything. To do so is a sign of madness The question then is what grounds this apparent concession to logical prescription. Whereby, the hold of logic over thought? For the dependence there can seem puzzling: Why should the natural causal processes relations of logic, I am free to flout the moral law to any degree I desire, but my freedom to think unreasonably appears to encounter an obstacle in the requirement of logic? My thoughts are sensitive to logical truth in somewhat the way they  are sensitive to the world surrounding me: They have not the independence of what lies outside my will or self that I fondly imagined. I may try to reason contrary to modus ponens, but my efforts will be systematically frustrated. Pure logic takes possession of my reasoning processes and steers them according to its own indicates, variably, of course, but in a systematic way that seems perplexing.
 One view of tis is that ascriptions of thought are not attempts to map a realm of independent causal relations, which might then conceivably come apart from logical relations, but are rather just a useful method of summing up people’s behaviours. Another view insists that we must acknowledge that thought is not a natural phenomenon in the way merely, and physical facts are: Thoughts are inherently normative in their nature, so that logical relations constitute their inner essence. Thought incorporates logic in somewhat the way externalists say it incorporates the world. Accordingly, the study of thought cannot be a natural science in the way the study of (say) chemistry compounds is. Whether this view is acceptable, depends upon whether we can make sense of the idea that transitions in nature, such as reasoning appear to be, can also be transitions in logical space, i.e., be confined by the structure of that space. What must be thought, in such that this combination n of features is possible. Put differently, what is it for logical truth to be self - evident?
 This dependency question has been studied less intensively than the previous three. The question is whether intentionality ids dependent on or upon consciousness for its very existence, and if so why. Could our thoughts have the very content they now have if we were not to be consciousness beings at all? Unfortunately, it is difficult to see how to mount an argument in either direction. On one hand, it can hardly be an accident that our thoughts are conscious and that this content is reflected in the intrinsic condition of our state of consciousness: It is not as if consciousness leaves off where thought content begins - as it does with, say, the neural basis of thought. Yet, on the other hand, it is by no means clear what it is about consciousness that links it to intentionality in this way. Much of the trouble here stems from our exceedingly poor understanding of the nature of consciousness could arise from grain tissue (the mind - body problem), so that we fill to grasp the manner in which conscious states bear meaning. Perhaps content is fixed by extra - conscious properties and relations and only subsequently shows up in consciousness, as various naturalistic reductive accounts would suggest; Or perhaps, consciousness itself plays a more enabling role, allowing meaning to come into the word, hard as this may be to penetrate. In some ways the question is analogous to, say, the properties of ‘pain’: Is the aversive property of pain, causing avoidance behaviour and so forth, essentially independent of the conscious state of feeling, or is it that pain, could only have its aversion function in virtue of the conscious feedings? This is part of the more general question of the epiphenomenal character of consciousness: Is conscious awareness just a dispensable accompaniment of some mental feature - such as content or causal power - or is it that consciousness is structurally involved in the very determination of the feature? It is only too easy to feel pulled in both directions on this question, neither alterative being utterly felicitous. Some theorists, suspect that our uncertainty over such questions stems from a constitutional limitation to human understanding. We just cannot develop the necessary theoretical tools which to provide answers to these questions, so we may not in principle be able to make any progress with the issue of whether thought depends upon consciousness and why. Certainly our present understanding falls far short of providing us with any clear route into the question.
 It is extremely tempting to picture thought as some kind of inscription in a mental medium and of reasoning as a temporal sequence of such inscriptions. On this picture all that a particulars thought requires in order to exist is that the medium in question should be impressed with the right inscription. This makes thought independent of anything else. On some views the medium is conceived as consciousness itself, so that thought depends on consciousness as writing depends on paper and ink. But ever since Wittgenstein wrote, we have seen that this conception of thought has to be mistaken, in particular of intentionality. The definitive characteristics of thought cannot be captured within this model. Thus, it cannot make room for the idea of intrinsic world - dependence. Since any inner inscription would be individualatively independent of items outside the putative medium of thought. Nor can it be made to square with the dependence of thought on logical pattens, since the medium could be configured in any way permitted by its intrinsic nature, within regard for logical truth - as sentences can be written down in any old order one likes. And it misconstrues the relation between thought and consciousness, since content cannot consist in marks on the surface of consciousness, so to speak. States of consciousness do contain particular meanings but not as a page contains sentences: The medium conception of the relation between content and consciousness is thus deeply mistaken. The only way to make meaning enter internally into consciousness is to deny that it as a medium for meaning to be expressed. However, it is marked and noted as the difficulty to form an adequate conception of how consciousness does carry content - one puzzle being how the external determinants of content find their way into the fabric of consciousness.
 Only the alleged dependence of thought upon language fits the naïve tempting inscriptional picture, but as we have attested to, this idea tends to crumble under examination. The indicated conclusion seems to be that we simply do not posses a conception of thought that makes  its real nature theoretically comprehensible: Which is to say, that we have no adequate conception of mind? Once we form a conception of thought that makes it seem unmysterious as with the inscriptional picture. It turns out to have no room for content as it presents itself: While building in a content as it is leaves’ us with no clear picture of what could have such content. Thought is ‘real’, then, if and only if it is mysterious.
 In the philosophy of mind ‘epiphenomenalism’ means that while there exist mental events, states of consciousness, and experience, they have themselves no causal powers, and produce no effect on the physical world. The analogy sometimes used is that of the whistle on the engine that makes the sound (corresponding to experiences), but plays no part in making the machinery move. Epiphenomenalism is a drastic solution to the major difficulties the existence of mind with the fact that according to physics itself only a physical event can cause another physical event an epiphenomenalism may accept one - way causation, whereby physical events produce mental events, or may prefer some kind of parallelism, avoiding causation either between mind and body or between body and mind. And yet, occasionalism considers the view that reserves causal efficacy to the action of God. Events in the world merely form occasions on which God acts so as to bring about the events normally accompanying them, and thought of as their effects. Although, the position is associated especially with the French Cartesian philosopher Nicolas Malebranche (1638 - 1715), inheriting the Cartesian view that pure sensation has no representative power, and so adds the doctrine that knowledge of objects requires other representative ideas that are somehow surrogates for external objects. These are archetypes of ideas of objects as they exist in the mind of God, so that ‘we see all things in God’. In the philosophy of mind, the difficulty to seeing how mind and body can interact suggests that we ought instead to think of hem as two systems running in parallel. When I stub my toe, this does so cause pain, but there is a harmony between the mental and the physical (perhaps due yo God) that ensures that there will be a simultaneous pain, when I form an intention and then act, the same benevolence ensures that my action is appropriated to my intention. The theory has never been wildly popular, and many philosophers would say that it was the result of a misconceived ‘Cartesian dualism’. Nonetheless, a major problem for epiphenomenalism is that if mental events have no causal relationship it is not clear that they can be objects of memory, or even awareness.
 The metaphor used by the founder of revolutionary communism, Karl Marx (1805 - 1900) and the German social philosopher and collaborator of Marx, Friedrich Engels (1820 - 95), to characterize the relation between the economic organization of society, which is its base, an the political, legal, and cultural organizations and social consciousness of a society, which is the super - structure. The sum total of the relations of production of material life conditions the social political, and intellectual life process in general. The way in which the base determines of much debate with writers from Engels onwards concerned to distance themselves from that the metaphor might suggest. It has also in production are not merely economic, but involve political and ideological relations. The view that all causal power is centred in the base, with everything in the super - structure merely epiphenomenal. Is sometimes called economicism? The problems are strikingly similar to those that are arisen when the mental is regarded as supervenience upon the physical, and it is then disputed whether this takes all causal power away from mental properties.
 Just the same, for if, as the causal theory of action implies, intentional action requires that a desire for something and a belief about how to obtain what one desires play a causal role in producing behaviour, then, if epiphenomenalism is true, we cannot perform intentional actions. Nonetheless, in describing events that happen does not of itself permit us to talk of rationality and intention, which are the categories we may apply if we conceive of them as actions. Ewe think of ourselves not only passively, as creatures within which things happen, but actively, as creatures that make things happen. Understanding this distinction gives rise to major problems concerning the nature of agency, of the causation of bodily events by mental events, and of understanding the ‘will’ and ‘free will’. Other problems in the theory of action include drawing the distinction between the structures involved when we do one thing ‘by’ doing another thing. Even the placing and dating of action can give ruse to puzzles, as one day and in one place, and the victim then dies on another day and in another place. Where and when did the murder take place? The notion of applicability inherits all the problems of ‘intentionality’. The specific problems it raises include characterizing the difference between doing something accidentally and doing it intentionally. The suggestion that the difference lies in a preceding act of mind or volition is not very happy, since one may automatically do what is nevertheless intensional, for example, putting one’s foot forwards while walking. Conversely, unless the formation of a volition is intentional, and thus raises the same questions, the presence of a violation might be unintentional or beyond one’s control. Intentions are more finely grained than movements, one set of movements may both be answering the question and starting a war, yet the one may be intentional and the other not.
 However, according to the traditional doctrine of epiphenomenalism, things are not as they seem: In reality, mental phenomena can have no causal effects: They are casually inert, causally impotent. Only physical phenomena are casually efficacious. Mental phenomena are caused by physical phenomena, but they cannot cause anything. In short, mental phenomena are epiphenomenal.
 The epiphenomenalist claims that mental phenomena seem to be causes only because there are regularities that involve types (or kinds) of mental phenomena. For example, instances of a certain mental type ‘M’, e.g., trying to raise one’s arm might tend to be followed by instances of a physical type ‘P’, e.g., one’s arms rising. To infer that instances of ‘M’ tend to cause instances of ‘P’ would be, however, to commit the fallacy of post hoc, ergo propter hoc. Instances of ‘M’ cannot cause instances of ‘P’: Such causal transactions are casually impossible. P - typ e events tend to be followed by M - type events because instances of such events are dual - effects of common physical causes, not because such instances causally interact. Mental events and states can figure in the web of causal relations only as effects, never as causes.
 Epiphenomenalism is a truly stunning doctrine. If it is true, then no pain could ever be a cause of our wincing, nor could something’s looking red to us ever be a cause of our thinking that it is red. A nagging headache could never be a cause of a bad mood. Moreover, if the causal theory of memory is correct, then, given epiphenomenalism, we could never remember our prior thoughts, or an emotion we once felt, or a toothache we once had, or having heard someone say something, or having seen something: For such mental states and events could not be causes of memories. Furthermore, epiphenomenalism is arguably incompatible with the possibility of intentional action. For if, s the casual theory of action implies, intentional action requires that a desire for something and a belief about how to obtain what one desires lay a causal role in producing behaviour, then, if epiphenomenalism is true, we cannot perform intentional actions. As it strands, to accommodate this point - most obviously, specifying the circumstances in which belief - desire explanations are to be deployed. However, matter are not as simple as the seem. Ion the functionalist theory, beliefs are casual functions from desires to action. This creates a problem, because all of the different modes of psychological explanation appeal to states that fulfill a similar causal function from desire to action. Of course, it is open to a defender of the functionalist approach to say that it is strictly called for beliefs, and not, for example, innate releasing mechanisms, that interact with desires in a way that generates actions. Nonetheless, this sort of response is of limited effectiveness unless some sort of reason - giving for distinguishing between a state of hunger and a desire for food. It is no use, in that it is simply to describe desires as functions from belief to actions.
 Of course, to say the functionalist theory of belief needs to be expanded is not to say that it needs to be expanded along non - functionalist lines.  Nothing that has been said out the possibility that a correct and adequate account of what distinguishes beliefs from non - intentional psychological states can be given purely in terms of respective functional roles. The core of the functionalist theory of self - reference is the thought that agents can have subjective beliefs that do not involve any internal representation of the self, linguistic or non - linguistic. It is in virtue of this that the functionalist theory claim to be able to dissolve such the paradox. The problem that has emerged, however, is that it remains unclear whether those putative subjective beliefs really are beliefs. Its thesis, according to which all cases of action to be explained in terms of belief - desire psychology have to be explained through the attribution of beliefs. The thesis is clearly at work as causally given to the utility conditions, and hence truth conditions, of the belief that causes the hungry creature  facing food to eat what I in front of him - thus, determining the content of the belief to be. ‘There is food in front of me’, or ‘I am facing food’. The problem, however, is that it is not clear that this is warranted. Chances would explain by the animal would eat what is in front of it. Nonetheless, the animal of difference, does implicate different thoughts, only one of which is a purely directive genuine thought.
 Now, the content of the belief that the functionalist theory demands that we ascribe to an animal facing food is ‘I am facing food now’ or ‘There is food in front of me now’. These are, it seems clear, structured thoughts, so too, for that matter, is the indexical thought ‘There is food here now’. The crucial point, however, is that the casual function from desires to actions, which, in itself, is all that a subjective belief is, would be equally well served by the unstructured thought ‘Food’.
 At the heart of the reason - giving relation is a normative claim. An agent has a reason for believing, acting and so forth. If, given here to other psychological states this belief/action is justified or appropriate. Displaying someone’s reasons consist in making clear this justificatory link. Paradigmatically, the psychological states that prove an agent with logical states that provide an agent with treason are intentional states individuated in terms of their propositional content. There is a long tradition that emphasizes that the reason - giving relation is a logical or conceptual representation. In the case of reason for actions the premises of any reasoning are provided by intentional states other than belief.
 Notice that we cannot then, assert that epiphenomenalism is true, if it is, since an assertion is an intentional speech act. Still further, if epiphenomenalism is true, then our sense that we are enabled is true, then our sense that we are agents who can act on our intentions and carry out our purposes is illusory. We are actually passive bystanders, never the agent in no relevant sense is what happens up to us. Our sense of partial causal control over our exert no causal control over even the direction of our attention. Finally, suppose that reasoning is a causal process. Then, if epiphenomenalism is true, we never reason: For there are no mental causal processes. While one thought may follow anther, one thought never leads to another. Indeed, while thoughts may occur, we do not engage in the activity of thinking. How, the, could we make inferences that commit the fallacy of post hoc, ergo propter hoc, or make any inferences at all for that matter?
 As neurophysiological research began to develop in earnest during the latter half of the nineteenth century. It seemed to find no mental influence on what happens in the brain. While it was recognized that neurophysiological events do not by themselves casually determine other neurophysiological events, there seemed to be no ‘gaps’ in neurophysiological causal mechanisms that could be filled by mental occurrences. Neurophysiological appeared to have no need of the hypothesis that there are mental events. (Here and hereafter, unless indicated otherwise, ‘events’ in the broadest sense  will include states as well as changes.) This ‘no gap’ line of argument led some theorists to deny that mental events have any casual effects. They reasoned as follows: If mental events have any effects, among their effects would be neurophysiological ones: Mental events have no neurophysiological effects: Thus, mental events have no effect at all. The relationship between mental phenomena and neurophysiological mechanisms is likened to that between the steam - whistle which accompanies the working of a locomotive engine and the mechanisms of the engine, just as the steam - whistle which accompanies the working of a locomotive engine and the mechanisms of the engine: just as the steam - whistle is an effect of the operations of the mechanisms but has no casual influence on those operations, so too mental phenomena are effects of the workings of neurophysiological mechanisms, but have no causal influence on their operations. (The analogy quickly breaks down, as steam whistles have casual effects but the epiphenomenalist alleges that mental phenomenons have no causal effects at all.)
 An early response to this ‘no gap’ line of argument was that mental events (and states) are not changes in (and states of) an immaterial Cartesian substance e, they are, rather changes in (and states of) the brain. While mental properties or kinds are not neurophysiological properties or kinds, nevertheless, particular mental events are neurophysiological events. According to the view in question, a given events can be an instance of both a neurophysiological type and a mental type, and thus be both a mental event and a neurophysiological event. (Compare the fact that an object might be an instance of more than one kind of object: For example, an object might be both a stone and a paper - weight.) It was held, moreover, that mental events have causal effects because they are neurophysiological events with causal effects. This response presupposes that causation is an ‘extensional’ relation between particular events that if two events are causally related, they are so related however they are typed (or described). Given that assumption is today widely held. And given that the causal relation is extensional, if particular mental events are indeed, neurophysiological events are causes, and epiphenomenalism is thus false.
 This response to the ‘no gap’ argument, however, prompts a concern about the relevance of mental properties or kinds to causal relations. And in 1925 C.D. Broad tells us that the view that mental events are epiphenomenal is the view that mental events either (a) do not function at all as causal - factors, or hat (b) if they do, they do so in virtue of their physiological characteristics and not in virtue of their mental characteristics. If particular mental events are physiological events with causal effects, then mental events function as case - factors: They are causes, however, the question still remains whether mental events are causes in virtue of their mental characteristics. , yet, neurophysiological occurrences without postulating mental characteristics. This prompts the concern that even if mental events are causes, they may be causes in virtue of their physiological characteristics. But not in virtue of their mental characteristics.
 This concern presupposes, of course, that events are causes in virtue of certain of their characteristics or properties. But it is today fairly widely held that when two events are causally related, they are so related in virtue of something about each. Indeed, theories of causation assume that if two events ‘x’ and ‘y’ are causally related, and two other events ‘a’ and ‘b’ are not, then there must be some difference between ‘x’ and ‘y’ and ‘a’ and ‘b’ in virtue of which ‘x’ and ‘y’ are. But ‘a’ and ‘b’ are not, causally related. And they attempt to say what that difference is: That is, they attempt to say what it is about causally related events in virtue of which they are so related. For example, according to so - called ‘nomic subsumption views of causation’, causally related events will be so related in virtue of falling under types (or in virtue of having properties) that figure in a ‘causal law’. It should be noted that the assumption that casually related events are so related in virtue of something about each is compatible with the assumption that the causal relation is an ‘extensional’ relationship between particular events. The weighs - less - than relation is an extensional relation between particular objects: If O weighs less than O*, then O and O* are so related, have them of a typed (or characterized, or described, nevertheless, if O weighs less than O*, then that is so in virtue of something about each, namely their weights and the fact that the weight of one is less than the weight of the other. Examples are readily multiplied. Extensional relations between particulars typically hold in virtue of something about the particular. It is, nonetheless, that we will grant that when two events are causally related, they are so related in virtue of something about each.
 Invoking the distinction between types and tokens, and using the term ‘physical’, rather than the more specific term ‘physiological’. Of the following are two broad distinctions of epiphenomenalism:
  Token Epiphenomenalism: Mental events cannot cause anything.
  Type Epiphenomenalism: No event can cause anything in virtue of
  falling under a mental type.
So in saying. That property epiphenomenalism is the thesis that no event can cause anything in virtue of having a mental property. The conjunction of token epiphenomenalism and the claim those physical events cause mental events is, that, of course, the traditional doctrine of epiphenomenalism, as characterized earlier. Ton epiphenomenalism implies type epiphenomenalism, for if an event could cause something in virtue of falling under a mental type, then an event could be both epiphenomenalism would be false. Thus, if mental events cannot be causes, then events cannot be causes in virtue of falling under mental types. The denial of token epiphenomenalism does not, however, imply the denial of type epiphenomenalism, if a mental event can be a physical event that has causal effects. For, if so, then token epiphenomenalism may still be true. For it may be that events cannot be causes in virtue of falling under mental types. Mental events may be causes in virtue of falling under mental types. Thus, even if token epiphenomenalism is false, the question remains whether type epiphenomenalism is.
 Suppose, for the sake of argument, that type epiphenomenalism is true. Why would that be a concern if mental events are physical events with causal effects? In our assumption that the causal relation is extensional, it could be true, consist with type epiphenomenalism, that pains cause winces, that desires cause behaviour, that perceptual experience cause beliefs and mental states cause memories, and that reasoning processes are causal processes. Nevertheless, while perhaps not as disturbing a doctrine as token epiphenomenalism, type epiphenomenalism can, upon reflection, seen disturbing enough.
 Notice to begin with that ‘in virtue of’ expresses an explanatory relationship. In so doing, that ‘in virtue of’ is arguably a near synonym of the more common locution ‘because of’. But, in any case, the following seems true so as to be adequate: An event causes a G - event in virtue of being an F - event if and only if it causes a G - event because of being an F - event.’In virtue of’ implies ‘because of’, and in the case in question at least the implication seems to go in the other direction as well. Suffice it to note that were type epiphenomenalism consistent with its being the case that an event could have a certain effect because of falling under a certain mental type, then we would, indeed be owed an explanation of why it should be of any concern if type epiphenomenalism is true. We will, however, assume that type epiphenomenalism is inconsistent with that. We will assume that type epiphenomenalism could be reformulated as: No event can cause anything because of falling under a mental type. (And we will assume that property epiphenomenalism can be reformulated thus: No event can cause anything because of having a mental property.) To say that ‘a’ causes ‘b’ in virtue of being ‘F’ is too say that ‘a’ causes ‘b’ because of being ‘F’; that is, it is to say that it is because ‘a’ is ‘F’ that it causes ‘b’. So, understood, type epiphenomenalism is a disturbing doctrine indeed.
 If type epiphenomenalism is true, then it could never be the case that circumstances are such that it is because some event or states is a sharp pain, or a desire to flee, or a belief that danger is near, that it has a certain sort of effect. It could never be the case that it is because some state in a desire of ‘X’ (impress someone) and another is a belief that one can ‘X’ by doing ‘Y’ (standing on one’s head) that the states jointly result in one’s doing ‘Y’ (standing on one’s head). If type (property) epiphenomenalism is true, then nothing has any causal powers whatever in virtue of (because of) being an instance of a mental type, then, never be the case of a certain mental type that a state has the causal power in certain circumstances to provide some effect. For example, it could never the case that it is in virtue of being an urge to scratch (or a belief that danger is near) that a state has the causal power in certain circumstances to produce scratching behaviour (or fleeing behaviour) if type - epiphenomenalism is true, then the mental qua mental, so to speak, is casually impotent. That may very well seem disturbing enough.
 What reason is there, however, for holding type epiphenomenalism? Even if neurophysiology does not need to postulate types of mental events, perhaps the science of psychology does. Note that physics has no need to postulate types of neurophysiological events: But that may well not lead one tp doubt that an event can have effects in virtue of being (say) a neuron firing. Moreover, mental types figure in our every day, casual explanations of behaviour, intentional action, memory, and reasoning. What reason is there, then, for holding that events cannot have effects in virtue of being instances of mental types? This question naturally leads to the more general question of which event types are such that events have effects in virtue of falling under them. This more general question is best addressed after considering a ‘no gap’ line of argument that has emerged in recent years.
 Current physics includes quantum mechanics, a theory which appears able, in  principle, to explain how chemical processes unfold in terms of the mechanics of sub - atomic particles. Molecular biology seems able, in principle, to explain how the physiological operations of systems in living things in terms of biochemical pathways, long chains of chemical reactions. On the evidence, biological organisms are complex physical objects, made up of molecular particles (there are noo entelechies or élan vital). Since we are all biological organisms, the movements of our bodies and of their minute parts, including the chemicals in our brains, and so forth, are causally determined, too whatsoever subatomic particles and fields. Such considerations have inspired a lin e of argument that only events within the domain of physics are causes.
 Before presenting the argument, let us make some terminological stipulations: Let us henceforth use ‘physical events’ (states) and ’physical property’ in as strict and narrow sense to mean, respectfully, a type of event (state) physics (or, by some improved version of current physics). Event if they figure in laws of physics. Finally, by ‘a physical event (states) we will mean an even (state) that falls under a physical type. Only events within the domain of (current) physics (or, some improved eversion of current physics) count as physical in this strict and narrow sense.
Consider, then:
   The Token - Exclusion Thesis Only physical events can have
  causal effects (i.e., as a matter of causal necessity, only physical
  events have casual effects).
The premises of the basis argument for the token - exclusion thesis are:
   Physical Caudal Closure Only physical events can cause
  physical events.
   Causation by way of Physical Effects As a matter of at least
  casual necessity, an event is a cause of another event if and only if it
  is a cause of some physical event?
These principles jointly imply the exclusion thesis. The principle of causation through physical effects is supported on the empirical grounds that every event occurs within space - time, and by the principle that an event is a cause of an event that occurs within a given region of space - time if and only if it is a cause of some physical event that occurs within that region of space - time. The following claim is offered in support of physical closure:
   Physical causal Determination, For any (caused) physical
  event, ‘P’, there is a chain of entirely physical events leading to ‘P’,
  each link of which casually determines its successor.
(A qualification: If strict determinism is not true, then each link will determine the objective probability of its successor.) Physics is such that there is compelling empirical reason to believe that physical causal determination holds. Every physical event will have a sufficient physical cause. More precisely, there will be a deterministic casual chan of physical events leading to any physical event, ‘P’. Butt such links there will be, and such physical causal chains are entirely ‘gap - less’. Now, to be sure, physical casual determination does not imply physical causal closure, the former, but not the latter, is consistent with non - physical events causing physical events. However, a standard epiphenomenalist response to this is that such non - physical events would be, without exception, over - determining causes of physical events, and it is ad hoc are over - determining non - physical events. Nonetheless, a standard epiphenomenalist response of this is that such non - physical events would be, without exception, over - determining causes of physical events, and it is ad hoc to maintain that non - physical events are over - determining causes of physical events.
 Are mental events within the domain of physics? Perhaps, like objects, events can fall under many different types or kinds. We noted earlier that a given object might, for instance, be both a stone and a paper wight, however, we understand how a stone could be a paper - wight, but how, for instance could an event of subatomic particles and fields be a mental event? Suffice e it to note for a moment that if mental events are not within the domain of physics, then if the token - exclusion thesis is true, no mental event can ever cause anything: Token epiphenomenalism is true.
 One might reject the token - exclusion thesis, however, on the grounds that, typical events within the domains of the special sciences - chemistry, the life sciences, and so on - are not within the domain of physics, but nevertheless have causal effects. One might maintain that neuron firing, for instance, cause either neuron firing, even though neurophysiological events are not within the domain of physics. Rejecting the token - exclusion either, however, requires arguing either that physical causal closure is false or that the principle of causation by way of physical effects is.
 But one response to the ‘no - gap’ argument from physics is to reject physical casual closure. Recall that physical causal determination is consistent with non - physical events being over - determining causes of physical events. One might concede that it would be ad hoc to maintain that a non - physical event, ‘N’, is an over - determining cause of a physical event ‘P’, and that ‘N’ causes ‘P’ in a way that is independent of the causation of ‘P’ by other physical events. Nonetheless, ‘N’ can be a cause of another event, that ‘N’ can cause a physical event ‘P’ in a way that is dependent upon P’s being caused by physical events.  Again, one might argue that physical events ‘underlie’ non - physical events, and that a non - physical event ‘N’ can be a cause of anther event ‘X’ (physical or non - physical), in virtue of the physical event that ‘underlie’ ‘N’ being a cause of ‘X’.
 Another response is to deny the principle of causation through physical effects. Physical causal closure is consistent with non - physical events. One might concede physical causal closure but deny the principle of causation by way of physical effects, and argue that non - physical events cause other non - physical events without causing physical events. This would not require denying that (1) Physical events invariably ‘underlie’ non - physical events or that (2) Whenever a non - physical event causes another non - physical event, some physical event that underlies the first event causes a physical event that underlies the second. Clams of both tenets (1) and (2) do not imply the principle of causation through physical effects. Moreover, from the fac t that a physical event ‘P’, causes another physical event ‘P*’. It may not allow that ‘P’ causes every non - physical event that ‘P*’ underlies. That may not follow it the physical events that underlie non - physical events casually suffice for those non - physical events. It would follow from that, which for every non - physical event, there is a causally sufficient physical event. But it may be denied that causal sufficiency suffices for causation: It may be argued that there are further constraints on causation that can fail to be met by an event that causally suffices for another. Moreover, it ma be argued that given the further constraints, non - physical events are the causes of non - physical events.
 However, the most common response to the ‘no - gap’ argument from physics is to concede it, ad thus to embrace its conclusion, the token - exclusion these, but to maintain the doctrine of ‘token physicalism’, the doctrine that every event (state) is within the domain of physics. If special science events and mental events are within the domain of physics, then they can be causes consistent with the token - exclusion thesis.
 Now whether special science events and mental events are within the domain of physics depends, in part, on the nature of events, and that is a highly controversial topic about which there is nothing approaching a received view. The topic raises deep issues that are beyond the scope of this essay, yet the issues concerning the ‘essence’ of events and the relationship between causation and causal explanation, are in any case, . . .  suffice it to note here that it is believed that the sme fundamental issues concerning the causal efficacy of the mental arise for all the leading theories of the ‘relata’ of casual relation. The issues just ‘pop - up’ in different places. However, that cannot be argued at this time, and it will not be for us to be assumed.
 Since the token physicalism response to the no - gap argument from physics is the most popular response, is that special science events, and even mental events, are within the domain of physics. Of course, if mental events are within the domain of physics then, token epiphenomenalism can be false even if the token - exclusion is true: For mental events may be physical events which have causal effects.
 Nevertheless, concerns about the causal relevance of mental properties and event types would remain. Indeed, token physicalism together with a fairly uncontroversial assumption, naturally leads to the question of whether events can be causes only in virtue of falling under types postulated by physics. The assumption is that physics postulates a system of event types that has the following features:
   Physical Causal Comprehensiveness: When two physical
  events are causally related, they are so related in virtue of falling
  under physical types.
That thesis naturally invites the question of whether the following is true:
   The Type - Exclusion Thesis: An event can cause something
  only in virtue of falling under a physical type, i.e., a type
  postulated by physics.
The type - exclusion thesis offers one would - be answer to our earlier question of which effects types are such that events have effects in virtue of falling under them. If the answer is the correct one, it may, however, be in fact (if it is correct) that special science events and mental events are within the domain of physics will be cold comfort. For type physicalism, the thesis that every event type is a physical type, seems false. Mental types seem not to be physical types in our strict and narrow sense. No mental type, it seems, is necessarily coextensive (i.e., coextensive in every ‘possible world’) with any type postulated by physics. Given that, and given the type - exclusion thesis, type epiphenomenalism is true. However, typical special science types also fail to be necessarily coextensive with any physical types, and thus typical special science types fail to be physical types. Indeed, we individuate the sciences in part by the event (state) types they postulate. Given that typical special science types are not physical types (in our strict sense), then typical special science types are not such that even can have causal effects in virtue of falling under them.
 Besides, a neuron firing is not a type of event postulated by physics, given the type exclusion thesis, no event could ever have any causal effects in virtue of being a firing of a causal effect. The neurophysiological qua neurophysiological is causally impotent. Moreover, if things have casual powers only in virtue of their physical properties, then an HIV virus, qua HIV virus, does not have the causal power to contribute to depressing the immune system: For being an HIV virus is not a physical property (in our strict sense). Similarly, for the same reason the SALK vaccine, qua SALK vaccine, would not have the causal power to contribute to producing an immunity to polio. Furthermore, if, as it seems, phenotype properties are not physical properties, phenotypic properties do not endow organisms with casual powers conducive to survival. Having hands, for instance, could never endow nothing with casual powers conducive to survival since it could never endow anything with any causal powers whatsoever. But how, then, could phenotypic properties be units of natural selection? And if, as it seems, genotypes are not physical types, then, given the type exclusion thesis, genes do not have the causal power, qua genotypes, to transmit the genetic bases for phenotypes. How, then, could the role of genotypes as units of heredity be a causal role? There seem to be ample grounds for scepticism that any reason for holding the type - exclusion thesis could outweigh our reasons for rejecting it.
 We noted that the thesis of universal physical causal comprehensiveness or ‘upc - comprehensiveness’ for short, invites the question of whether the type - exclusion thesis is true. But does upc - comprehensiveness while rejecting the type - exclusion thesis?
 Notice that there is a crucial one - word difference between the two theses: The exclusion thesis contains the word ‘only’ in front of ‘in virtue of’, while thesis of upc - comprehensiveness does not. This difference is relevant because ‘in virtue of’ does not imply ‘only in virtue of’, I am a brother in virtue of being a male with a sister, but I am also a brother in virtue of being a male with a brother, and, of course, being a male with a brother, and conversely. Likewise, I live in the province of Ontario in virtue of living in the city of Toronto, but it is also true that I live in Canada in virtue of living in the County of York. Moreover, in the general case, if something ‘x’ bears a relation ‘R’, to something ‘y’ in virtue of x’s being ‘F’ and y’s being ‘G’. Suppose that ‘x’ weighs less than ‘y’ in virtue of x’s weighing lbs., and y’s weighing lbs. Then, it is also true that ‘x’ weighs less than ‘y’ in virtue of x’s weighing under lbs., and y’s weighing over lbs. And something can, of course, weigh under lbs., without weighing lbs. To repeat, ‘in virtue of’ does not imply ‘only in virtue of’.
 Why, then, think that upc - comprehensiveness implies the type - exclusion thesis? The fact that two events are causally related in virtue of falling under physical types does not seem to exclude the possibility that they are also causally related in virtue of falling under non - physical types, in virtue of the one being (say) a firing of a certain other neuron, or in virtue of one being a secretion of enzymes and the other being a breakdown of amino acids. Notice that the thesis of upc - comprehensiveness implies that whenever an event is an effect of another, it is so in virtue of falling under a physical type. But the thesis does not seem to imply that whenever an event vis an effect of another, it is so only in virtue of falling under a physical type. Upc - comprehensiveness seems consistent with events being effects in virtue of falling under non - physical types. Similarly, the thesis seems consistent with events being causes in virtue of falling under non - physical types.
 Nevertheless, an explanation is called for how events could be causes in virtue of falling under non - physical types if upc - comprehensiveness is true. The most common strategy for offering such an explanation involves maintaining there is a dependence - determination relationship between non - physical types and physical types. Upc - comprehensiveness, together with the claim that instances of non - physical event types are causes or effects, implies that, as a matter of causal necessity, whenever an event falls under a non - physical event type, if falls under some physical type or other. The instantiation of non - physical types by an event thus depends, as a matter of causal necessity, on the instantiation of some or other physical event type by the event. It is held that non - physical types in physical context: Although as given non - physical type might be ‘realizable’ by more than one physical type. The occurrence o a physical type in a physical context in some sense determines the occurrence of any non - physical type that it ‘realizes’.
 Recall the considerations that inspired the ‘no gap’ arguments from physics: Quantum mechanics seems able, in principle, to explain how chemical processes unfold in terms of the mechanics of subatomic particles: Molecular biology seems able, in principle, to explain how the physiological operations of systems in living things occur in terms of biochemical pathways, long chains of chemical reactions. Types of subatomic causal processes ‘implement’ types of chemical processes. Many in the cognitive science community hold that computational processes implement that mental processes, and that computational processes are implemented, in turn, by neurophysiological processes.
 The Oxford English Dictionary gives the everyday meaning of ‘cognition’ as ‘the action or faculty of knowing’. The philosophical meaning is the same, but with the qualification that it is to be ‘taken in its widest sense, including sensation, perception, conception, and volition’. Given the historical link between psychology and philosophy, it is not surprising that ‘cognitive’ in ‘cognitive psychology’ has something like this broader sense, than the everyday one. Nevertheless, the semantics of ‘cognitive psychology’, like that of many adjective - noun combinations, is not entirely transparent. Cognitive psychology is a branch of psychology, and its subject matter approximates to the psychological study that are largely historical, its scope is not exactly what one would predict.
 Many cognitive psychologists have little interest in philosophical issues, as cognitive scientists are, in general, more receptive. Fodor, because of his early involvement in sentence processing research, is taken seriously by many psycholinguistics. His modularity thesis is directly relevant to questions about the interplay of different types of knowledge in language understanding. His innateness hypothesis, however, is generally regarded as unhelpful, and his prescription that cognitive psychology is primarily ignored. Dennett’s recent work on consciousness treats a topic that is highly controversial, but his detailed discussion of psychological research findings has enhanced his credibility among psychologists. Overall, psychologists are happy to get on with their work without philosophers telling them about their ‘mistakes’.
 The hypotheses driving most of modern cognitive science is simple to state - the mind is a computer. What are the consequences for the philosophy of mind? This question acquires heightened interest and complexity from new forms of computation employed in recent cognitive theory.
 Cognitive science has traditionally been based on or upon symbolic computation systems: Systems of rules for manipulating structures built up of tokens of different symbol types. (This classical kind of computation is a direct outgrowth of mathematical logic.) Since the mid - 1980s, however, cognitive theory has increasingly employed connectionist computation: The spread of numerical activation across units - the view that one of the most impressive and plausible ways of modelling cognitive processes in by means of a connectionist, or parallel distributed processing computer architecture. In such a system data is input into a number of cells as one level, or hidden units, which in turn delivers an output.
 Such a system can be ‘trained’ by adjusting the weights a hidden unit accords to each signal from an earlier cell. The’ training’ is accomplished by ‘back propagation of error’, meaning that if the output is incorrect the network makers the minimum adjustment necessary to correct it. Such systems prove capable of producing differentiated responses of great subtly. For example, a system may be able to task as input written English, and deliver as output phonetically accurate speech. Proponents of the approach also, point pout that networks have a certain resemblance to the layers of cells that make up a human brain, and that like us. But unlike conventional computing programs, networks degrade gracefully, in the sense that with local damage they go blurry rather than crashed altogether. Controversy has concerned the extent to which the differentiated responses made by networks deserve to be called recognitions, and the extent to which non - recognizable cognitive function, including linguistic and computational ones, are well approached in these terms.
 Some terminology will prove useful: that is, for which we are to stipulate that an event type ‘T’ is a casual type if and only if there is, at least one type T*, such that something can case a T* in virtue of being a ‘T’. And by saying that an event type is realizable by physical event types or physical properties. For that of which is least causally possible for the event to be realized by a physical event type. Given that non - physical causal types must be realizable by physical types, and given that mental types are non - physical types, there are two ways that mental types might to be causal. First, mental types may fail to be realizable by physical types. Second, mental types might be realizable by physical types but fail to meet some further condition for being causal types. Reasons of both sorts can be found in the literature on mental causation for denting that any mental types are causal. However, there has been much attention paid to reasons for the first sort in this casse of phenomenal mental types (pain states, visual states, and so forth). And there has been much attention to reasons of the second sort in the case of intentional mental states (i.e., beliefs that P, desires that Q, intentions that R, and so on).
 Notice that intentional states figure in explanations of intentional actions not in virtue of their intentional mode (whether they are beliefs or desires, and so on) but also in virtue of their contents, i.e., what is believed, or desired, and so forth. For example, what causally explains someone’s doing ‘A’ (standing on his head) is that the person wants to ‘X’ (impress someone) and believes that by doing ‘A’ he will ‘X’. The contents of the belief and desire (what is believed and what is desired) sem essential to the causal explanation of the agent’s doing ‘A’. Similarly, we often causally explain why someone came to believe that ‘P’ by citing the fact that the individual came to believe that ‘Q’ and inferred ‘P’ from ‘Q’. In such cases, the contents of the states in question are essential to the explanation. This is not, of course, to say that contents themselves are causally efficacious, contents are not among the relata of causal relations. The point is, however, that we characterize states when giving such explanations not only as being as having intentional modes, but also as having certain contents: We type states for having certain contents, we type states for the purpose of such explanations in terms of their intentional modes and their contents. We might call intentional state types that might include content properties ‘conceptual intentional state types’, but to avoid prolixity, let us call them ‘intentional state types’ for short: Thus, for present purposes, b y ‘intentional state types’ we will mean types such as the belief that ‘P; the desire that ‘Q’, and so on, and not types such as belief, desire and the like, and not types such as belief, desire, and so forth.
 Although it was no part of American philosopher Hilary Putnam, who in 1981 marked a departure from scientific realism in favour of a subtle position that he called internal realism, initially related to an ideal limit theory of truth and apparently maintaining affinities with verification, but in subsequent work more closely aligned with ‘minimalism’, Putnam’s concepts in the later period has largely to deny any serious asymmetry between truth and knowledge as it is obtained in natural science, and as it is obtained in morals and even theology. Still, purposively of raising concerns about whether ideational states are causal, the well - known ‘twin earth’ thought experiment have prompted such concerns. These thought - experiments are fairly widely held to show alike in every intrinsic physical respect can have intentional states with different contents. If they show that, then intentional state type fail to supervene on intrinsic physical state types. The reason is that with contents an individual’s beliefs, desires, and the like, have, depends, in part, on extrinsic, contextual factors. Given that, the concern has been raised toast states cannot have effects in virtue of falling under intentional state types.
 One concern seems to be that state cannot have effects in virtue of falling under intentional state types because individuals who are in all and only the same intrinsic states must have all and only the same causal powers. In response to that concern, it might be pointed out that causal power ss often depend on context. Consider weight. The weight of objects do not supervene on their intrinsic properties: Two objects can be exactly alike in every intrinsic respect (and thus have the same mass) yet have different weights. Weight depends, in part on extrinsic, contextual factors. Nonetheless, it seems true that an object can make a scale read 10lbs in virtue of weighing 10lbs. Thus, objects which are in exactly the am e type of intrinsic states may have different causal powers due to differences in their circumstances.
 It should be noted, however, that on some leading ‘externalist’ theories of content, content, unlike weight, depends on a historical context, such as a certain set of content - involving states is for attribution of those states to make the subject as rationally intelligible as possible, in the circumstances. Call such as theory of content ‘historical - externalist theories’. On one leading historical - externalist theory, the content of a state depends on the learning history of the individual on another. It depends on the selection history of the species of which the individual is a member. Historical - externalist theories prompt a concern that states cannot have causal effects in virtue of falling under intentional state types. Causal state types, it might be claimed, are never such that their tokens must have a certain causal ancestry. But, if so, then, if the right account of content is a historical - externalist account, then intentional types are not casual types. Some historical - externalists appear to concede this line of argument, and thus to effects in virtue of falling under intentional state types. However, explain how intentional - externalists attempt to explain how intentional types can be casual, even though their tokens must have appropriated causal ancestries. This issue is hotly debated, and remains unresolved.
 Finally, by noting, why it is controversial, whether phenomenal state types can be realized by physical state types. Phenomenal state types are such that it is like something for a subject to be in them: It is, for instance, like something to have a throbbing pain. It has been argued that phenomenal state types are, for that reason, subjective to fully understand what it is to be in them. One must be able to take up is to be in them, one must be able to take up a certain experiential point of view. For, it is claimed, an essential aspect of what it is to be in a phenomenal state is what it is like to be in a phenomenal state is what it is like to be in the state, only by tasking up certain experiential point of view can one understand that aspect (in our strict and narrow sense) are paradigms’ objective state, i.e., non - subjective states. The issue arises, then, as to whether phenomenal state types can be realized by physicalate types. How could an objective state realize a subjective one? This issue too is hotly debated, and remains unresolved.  Suffice it to say, that only physical types and types realizable by physical types and types realizable by physical types are causal, and if phenomenal types are neither, then nothing can have any causal effects, so, then, in virtue of falling under a phenomenal type. Thus, it could never be the case, for example, that a state causally results in a bad mood in virtue of being a throbbing pain.
 Philosophical theories are unlike scientific ones, scientific theories ask questions in circumstances where there are agreed - upon methods for answering the question and where the answers themselves are generally agreed upon. Philosophical theory: They attempt to model the known data to be seen from a new perspectives, a perspective that promotes the development of genuine scientific theory. Philosophical theories are, thus, proto - theories, as such, they are useful precisely in areas where no large - scale scientific theory exist. At present, which is exactly the state psychology it is in. Philosophy of mind, is to be a kind of propaedeutics to a psychological science. What is clear is that at the moment no universally accepted paradigm for a scientific psychological science exists. It is exactly in this kind of circumstance for a scientific psychology exists. It is exactly in this kind of circumstance that the philosophers of mind in the present context is to consider the empirical data available and to ry to form a generalized, coherent way of looking at those data tat will guide further empirical research, i.e., philosophers can provide a highly schematized model that will structure that research. And the resulting research will, in turn, help bring about refinements of the schematized theory, with the ultimate hope being that a closer, viable, scientific theory, one wherein investigators agree on the question and on the methods to be used to answer them, and will emerge. In these respects, philosophical theories of mind, though concerned with current empirical data, are too general in respect of the data to be scientific theories. Moreover, philosophical theories aimed primarily at a body of accepted data. As such,  philosophical theories merely give as ‘picture’ of those data. Scientific theories not only have to deal with the given data but also have to make predictions, in that can be gleaned from the theory together with accepted data. This removal go unknown data is what forms the empirical basis of a scientific theory and allows it to be justified in a way quite distinct from the way in which philosophical theories are justified. Philosophical theories are only schemata, coherent pictus of the accepted data, only pointers toward empirical theory, and as the history of philosophy makers manifest, usually unsuccessful one - though I think this lack of success is any kind of a fault, these are different tasks.
 In the philosophy of science, a generalization or set of generalizations purportedly making reference to unobservable entities, e.g., atoms, genes, quarks, unconscious wishes, and so forth. The ideal gas law, for example, refers only to such observables as pressure, temperature and volume and their properties. Although an older usage suggests lack of adequate evidence in support thereof (‘merely a theory’), current philosophical usage does not carry that connotation. Einstein’s special theory of relativity, for example, is considered extremely well founded.
 There are two main views on the nature of theories. According to the ‘received view’ theories are partially interpreted axiomatic systems, according to the semantic view a theory is a collection of models.
 The axiomatization or axiomatics belongs of a theory that usually emerges as a body of (supposed) truths that are not neatly organized, making the theory difficult to survey or study as a whole. The axiomatic method is an idea for organizing a theory: One tries to select from among the supposed truths a small number from which all the others can be seen to be deductively inferrable. This make the theory rather more tractable since, in a sense, all the truths are contained in those few. In a theory so organised, the few truths from which all others are deductively inferred are called ‘axioms’. David Hilbert had argued that, just as algebraic and differential equations and physical precesses, could themselves be made mathematical objects, so axiomatic theories, like algebraic and differential equations, which are means of representing physical processes and mathematical structures, could be made objects of mathematical investigation.
 Wherein, a credibility programme of a speech given in 1900, the mathematician David Hilbert (1862 - 1943) identified 23 outstanding problems in mathematics. The first was the ‘continuum hypothesis’. The second was the problem of the consistency of mathematics. This evolved into a programme of formalizing mathematic - reasoning, with the aim of giving meta - mathematical proofs of its consistency. (Clearly there is no hope of providing a relative consistency proof of classical mathematics, by giving a ‘model’ in some other domain. Any domain large and complex enough to provide a model would be raising the same doubts.) The programme was effectively ended by Kurt Gödel (1906 - 78), whose theorem of 1931, which showed that any system of arithmetic would need to make logical and mathematical assumptions at least as strong as arithmetic itself, and hence be just as much prey to hidden inconsistencies.
 In the tradition (as in Leibniz, 1704), many philosophers had the conviction that all truths, or all truths about a particular domain, followed from a few principles. These principles were taken to be either metaphysically prior or epistemologically prior or both. In the first sense, they were taken to be entities of such a nature that what exist is ‘caused’ by them. When the principles were taken as epistemically prior, that is, as axioms, either they were taken to be epistemically privileged, e.g., self - evident, not needing to be demonstrated, or again, inclusive ‘or’, to be such that all truths do in need follow from them, in at least, by deductive inferences. Gödel (1984) showed - in the spirit of Hilbert, treating axiomatic theories as themselves mathematical objects - that mathematics, and even a small part of mathematics, elementary number theory, could not be axiomatized that more precisely, any class of axioms which is such that we could effectively decide, of that class, would be too small to capture all of the truths.
 ‘Philosophy is to be replaced by the logic of science - that is to say, by the logical analysis of the concepts and sentences of the sciences, for the logic of science is nothing other than the logical syntax of the language of science’, has a very specific meaning. The background was provided by Hilbert’s reduction of mathematics to purposes of philosophical analysis, any scientific theory could ideally be reconstructed as an axiomatic system formulated within the framework of Russell’ s logic. Further analysis of a particular theory could then proceed a the logical investigation of its ideal logical reconstruction. Claims about theories in general were couched as claims about such logical systems.
 In both Hilbert’s geometry and Russell’s logic had an attempt made to distinguish between logical and non - logical terms. Thus the symbol ‘&’ might be used to indicate the logical relationship of conjunction between two statements, while ‘P’ is supposed to stand for a non - logical predicate. As in the case of geometry, the idea was that underlying any scientific theory is a purely formal logical structure captured in a set of axioms formulated in the appropriated formal language. A theory of geometry, for example, might include an axiom stating that for ant two distinct P’s (points), ‘p’ and ‘q’, there exist a number ‘L’ (Line) such that O(p, I) and O(q, I), where ‘O’ is a two place relationship between P’s and L’s (p lies on I). Such axioms, taken all together, were said to provide an implicit definition of the meaning of the non - logical predicates. In whatever of all the  P’s and L’s might be, they must satisfy the formal relationships given by the axioms.
 The logical empiricists were not primarily logicians: They were empiricists first. From an empiricist point of view, it is not enough that the non - logical terms of a theory be implicitly defined: They also require an empirical interpretation. This was provided by the ‘correspondence rules’ which explicitly linked some of the non - logical terms of a theory with terms whose meaning was presumed to be given directly through ‘experience’ or ‘observation’. The simplest sort of correspondence rule would be one that takes the application of an observationally meaningful term, such as ‘dissolve’, as being both necessary and sufficient for the applicability of a theoretical term, such as ‘soluble’. Such a correspondence rule would provide a complete empirical interpretation of the theoretical term.
 A definitive formulation of the classical view was provided by the German logical positivist Rudolf Carnap (1891 - 1970), who divided the non - logical vocabulary of theories into theoretical and observational components. The observational terms were presumed to be given a complete empirical interpretation, which left the theoretical terms with only an indirect empirical interpretation provided by their implicit definition within an axiom system in which some of the terms possessed a complete empirical interpretation.
 Among the issues generated by Carnap’s formulation was the viability of ‘the theory - observation distinction’, of course, one could always arbitrarily designate some subset of non - logical terms as belonging to the observational vocabulary, but that would compromise the relevance of the philological analysis for an understanding of the original scientific theory. But what could be the philosophical basis for drawing the distinction? Take the predicate ‘spherical’, for example. Anyone can observe that a billiard ball is spherical. But what about the moon, on the one hand, or an invisible speck of sand, on the other. Is the application of the term? For which the ’spherical’ in these objects are ‘observational’?
 Another problem was more formal, as did, that Craig’s theorem seem to show that a theory reconstructed in the recommendations fashioned could be re - axiomatized in such a way as to dispense with all theoretical terms, while retaining all logical consequences involving only observational terms. Craig’s theorem continues as a theorem in mathematical logic, held to have implications in the philosophy of science. The logician William Craig at Berkeley showed how, if we partition the vocabulary of a formal system (say, into the ‘T’ or theoretical terms, and the ‘O’ or observational terms) then if there is a fully ‘formalized system’ ‘T’ with some set ‘S’ of consequences containing only ‘O’ terms, there is also a system ‘O’ containing only the ‘O’ vocabulary but strong enough to give the same set ‘S’ of consequences. The theorem is a purely formal one, in that ‘T’ and ‘O’ simply separate formulae into the preferred ones, containing as non - logical terms only one kind of vocabulary, and the objects. The theorem might encourage the thought that the theoretical terms of a scientific theory are in principle dispensable, since the same consequences can be derived without them.
 However, Craig’s actual procedure gives no effective way of dispensing with theoretical terms in advance, i.e., in the actual process of thinking about and designing the premises from which the set ‘S’ follows. In this sense ‘O’ remains parasitic upon its parent ‘T’. Thus, as far as the ‘empirical’ content of a theory is concerned, it seems that we can do without the theoretical terms. Carnap’s version of the classical vew seemed to imply a form of instrumentionalism, a problem which Carl Gustav Hempel (1905 - 97) christened ‘the theoretician’s dilemma’.
 In the late 1940s, the Dutch philosopher and logician Evert Beth published an alternative formalism for the philosophical analysis of scientific theories. He drew inspiration from the work of Alfred Tarski, who studied first biology and then mathematics. In logic he studied with Kotarinski, Lukasiewicz and Lesniewski publishing a succession of papers from 1923 onwards. Yet he worked on decidable and undecidable axiomatic systems, and in the course in his mathematical career he published over 300 papers and books, on topics ranging from set theory to geometry and algebra. And also, drew further inspiration from Rudolf Carnap, the  German logical positivist who  left Vienna to become a professor at Prague in 1931, and felt Nazism to become professor  in Chicago in 1935. He subsequently worked at Los Angeles from 1952 to 1951. All the same, Evert Beth drew inspirations from von Neumann’s work on the foundations of quantum mechanics. Twenty years later, Beth’s emigrant who left Holland around the time Beth’s and van Fraassen. Here we are consider the comprehensibility in following the explication for which as preconditions between the ‘syntactic’ approach of the classical view and the ‘semantic’ approach of Beth and van Fraassen, and further consider the following simple geometrical theory as van Fraassen in 1989, presented first in the form of:
  A1: For any two lines, at most one point lies on both.
  A2: For any two points, exactly one line lies on both.
  A3: On every line are at least two points.
Note first, that these axioms are stated in more or less everyday language. On the classical view one would have first to reconstruct these axioms in some appropriate formal language, thus introducing quantifiers and other logical symbols. And one would have to attach appropriate correspondence rules. Contrary to common connotations of the word ‘semantic’, the semantic approach down - plays concerns with language as such. Any language will do, so long as it is clear enough to make reliable discriminations between the objects which satisfy the axiom and those which do not. The concern is not so much with what can be deduced from their axioms, valid deduction being  matter of syntax alone. Rather, the focus is on ‘satisfaction’, what satisfies the axioms - a semantic notion. These objects are, in the technical, logical sense of the term, models of the axioms. So, on the semantic approach, the focus shifts from the axiom as linguistic entities, to the models, which are non - linguistic entities.
 It is not enough to be in possession of a general interpretation for the terms used to characterize the models, one must also be able to identify particular instances - for example, a particular nail in a particular board. In real science must effort and sophisticated equipment may be required to make the required identification, for example, of a star as a white dwarf or of a formation in the ocean floor as a transformed fault. On a semantic approach, these complex processes of interpretation and identification, while essential in being able t use a theory, have no place within the theory itself. This is inn sharp contrast to the classical view, which has the very awkward consequence that various innovations in instrumenting itself. The semantic approach better captures the scientist’s own understanding of the difference between theory and instrumentation.
 On the classical view the question ‘What is a scientific theory’‘? Receives a straightforward answer. A theory is (1) a set of uninterrupted axioms in a specific formal language plus (2) a set of correspondence rules that provide a partial empirical interpretation in terms of observable entities and processes. A theory is thus true if and only if the interpreted axioms are all true. To obtain a similarly straightforward answer a little differently. Return to the axiom for placements as considered as free - standing statements. The definition could be formulated as follows: Any set of points and lines constitute a seven - pointed geometry is not even a candidate for truth or falsity, one can hardly identify a theory with a definition. But claims to the effect that various things satisfy the definition may be true or false of the world. Call these claims theoretical hypotheses. So we may say that, on the semantic approach, a theory consists of (1) a theoretical definition plus (2) a number of theoretical hypotheses. The theory may be said to be true just in case all its associated theoretical hypotheses are true.
 Adopting a semantic approach to theories still leaves wide latitude in the choice of specific techniques for formulating particular scientific theories. Following Beth, van Fraassen adopts a ‘state space’ representation which closely mirrors techniques developed in  theoretical physics during the nineteenth century - techniques were carried over into the developments of quantum and relativistic mechanics. The technique can be illustrated most simply for classical mechanics.
 Consider a simple harmonic oscillator, which consists of a mass constrained to move in one dimension subject to a linear restoring force - a weight bouncing gently while from a spring provides a rough example of such a system. Let ‘x’ represent the single spatial dimension, ‘t’ the time., ‘p’ the momentum, ‘k’ the strength of the restoring force, ands ‘m’ the mass. Then a linear harmonic oscillator may be ‘defined’ as a system which satisfies the following differential equation of motion:
  dx/dt = DH/Dp. Dp/dt =  - DH/Dx, where H = (k/2)x2 + (1/2m)p2
The Hamiltonian, ‘H’, represents the sun of the kinetic and potential energy of the system. The state of the system at any instant of time is a point in a two - dimensional position - momentum space. The history of any such system is this state space is given by an ellipse, as in time, the system repeatedly traces by revealing the ellipse onto the ‘x’ axis  covering classical mechanics. It remains to be any real world system, such as a bouncing spring, satisfies this definition.
 Other advocates of a semantic approach defer from the Beth - van Fraassen point of view in the type of formalism they would employ in reconstructing actual scientific theories. One influential approach derives from the word of Pattrick Suppes during he 1950s and 1960s, some of which inspired Suppes was by the logician J.C.C. Mckinsey and Alfred Tarski. In its original form. Suppes’s view was that theoretical definition should be formulated in the language of set theory. Suppes’s approach, as developed by his student Joseph Sneed (1971), has been adopted widely in Europe, and particularly in Germany, by the late Wolfgang Stegmüller (1976) and his students. Frederick Suppe has shares features of both the state space and the set - theoretical approaches.
 Most of those who have developed ‘semantic’ alternatives to the classical ‘syntactic’ approach to the nature of scientific theories were inspired by the goal of reconstructing scientific theories - a goal shared by advocates of the classical view. Many philosophers of science now question whether there is any point in treating philosophical reconstructions as scientific theories. Rather, insofar as the philosophy of science focuses on theories at all, it is the scientific versions, in their own terms, that should be of primary concern. But many now argue that the major concern should be directed toward the whole practice of science, in which theories are but a part. In these latter pursuits what is needed is not a technical framework for reconstructing scientific theories, but merely a general imperative framework for talking about required theories and their various roles in the practice of science. This becomes especially important when considering science such as biology, in which mathematical models play less of a role than in physics.
 At this point, at which there are strong reasons for adopting a generalized model - based understanding of scientific theories which makes no commitments to any particular formalism - for example, state spaces or set - theoretical predicates. In fact, one can even drop the distinction between ‘syntactic’ and ‘semantic’ as a leftover from an old debate. The important distinction is between an account of theories that takes models as fundamental versus that takes statements, particularly laws, as fundamental. A major argument for a model - based approach is that just given. There seem in fact to be few, if any, universal statements that might even plausibly be true, let alone known to be true, and thus available to play the role which laws have been thought to play in the classical account of theories, rather, what have often been taken to be universal generalisations should be interpreted as parts of definitions. Again, it may be helpful to introduce explicitly the notion of an idealized, theoretical model, an abstract entity which answers s precisely to the correspondence theoretical definition. Theoretical models thus provide, though only by fiat, something of which theoretical definitions may be true. This makes it possible to interpret much of scientific’ theoretical discourse as being about theoretical models than directly about the world. What have traditionally been interpreted as laws of nature thus out to be merely statements describing the behaviour of theoretical models?
 If one adopts such a generalized model - based understanding of scientific theories, one must characterize the relationship between theoretical models and real systems. Van Fraassen (1980) suggests that it should be one of isomorphism. But the same considerations that count against there being true laws in the classical sense also count against there being anything in the real world strictly isomorphic in any theoretical model, or even isomorphic to an ‘empirical’ sub - model. What is needed is a weaker notion of similarity, for which it must be specified both in which respect the theoretical model and the real system are similar, and to what degree. These specifications, however, like the interpretation of terms used in characterizing the model and the identification of relevant aspects of real systems, are not part of the model itself. They are part of a complex practice in which models are constructed and tested against the world in an attempt to determine how well they ‘fit’.
 Divorced from its formal background, a model - based understanding of theories is easily incorporated into a general framework of naturalism in the philosophy of science. It is particularly well - suited to a cognitive approach to science. Today the idea of a cognitive approach to the study of science means something quite different - indeed, something antithetical to the earlier meaning. A ‘cognitive approach’ is now taken to be one that focuses on the cognitive structures and processes exhibited in the activities of individual scenists. The general nature of these structures and processes is the subject matter of the newly emerging cognitive science. A cognitive approach to the study of science appeals to specific features of such structures and processes to explain the model and choices of individual scientists. It is assumed that to explain the overall progress of science one must ultimately also appeal to social factors and social approaches, but not one in which the cognitive excludes the social. Both are required for an adequate understanding of science as the product of human activities.
 What is excluded by the newer cognitive approach to the study of science is any appeal to a special definition of rationality which would make rationality a categorical or transcendent feature of science. Of course, scientists have goals, both individual and collective, and they employ more or less effective means for achieving these goals. So one may invoke an ‘instrumental’ or ‘hypothetical’ notion of rationality in explaining the success or failure of various scientific enterprise. But what is it at issue is just the effectiveness of various goal - directed activities, not rationality in any more exalted sense which could provide a demarcation criterion distinguishing science from other activities, sch as business or warfare. What distinguishes science is its particular goals and methods, not any special form of rationality. A cognitive approach to the study of science, then, is a species of naturalism in the philosophy of science.
 Naturalism in the philosophy of science, and philosophy generally, is more an overall approach to the subject than a set of specific doctrines. In philosophy it may be characterized only by the most general ontological and epistemological principles, and then more by what it opposes than by what it proposes.
 Besides ontological naturalisms and epistemological type naturalism, it seems that its most probably the single most important contributor to naturalism in the past century was Charles Robert Darwin (1809 - 82), who, while not a philosopher, naturalist is both in the philosophical and the biological sense of the term. In ‘The Descent of Man’ (1871) Darwin made clear the implications of natural selection for humans, including both their biology and psychology, thus undercutting forms of anti - naturalism which appealed not only to extra - natural vital forces in biology, but to human freedom, values, morality, and so forth. These supposed indicators of the extra - natural are all, for Darwin, merely products of natural selection.
 All and all, among advocates of a cognitive approach there is near unanimity in rejecting the logical positivist leal of scientific knowledge as being represented in the form of an interpreted, axiomatic system. But there the unanimity ends. Many employ a ‘mental models’ approach derived from the work of Johnson - Laird (1983). Others favour ‘production rules’ if this, infer that, a long usage for which the continuance by researchers in computer science and artificial intelligence, while some appeal to neural network representations.
 The logical positivist are notorious for having restricted the philosophical study of science to the ‘context of justification’, thus relegating questions of discovery and conceptual change to empirical psychology. A cognitive approach to the study of science naturally embraces these issues as of central concern. Again, there are differences. The pioneering treatment, inspired by the work of Herbert Simon, who employed techniques from computer science and artificial intelligence to generate scientific laws from finite data. These methods have now been generalized in various directions, while appeals to study of analogical reasoning in cognitive psychology, while Gooding (1990) develops a cognitive model of experimental procedure. Both Nersessian and Gooding combine cognitive with historical methods, yielding what Neressian calls a ‘cognitive - historical’ approach. Most advocates of a cognitive approach to conceptual change are insistent  that a proper cognitive understanding of conceptual change avoids the problem of incommensurability between old and new theories.
 No one employing a cognitive approach to the study of science thinks that there could be an inductive logic which would pick out the uniquely rational choice among rival hypotheses. But some, such as Thagard (1991) think it possible to construct an algorithm that could be run on a computer which would show which of two theories is best. Others seek to model such judgements as decisions by individual scientists, whose various personal, professional, and social interests are necessarily reflected in the decision process. Here, it is important to see how experimental design and the result of experiments may influence individual decisions as to which theory best represents the real world.
 The major differences in approach among those who share a general cognitive approach to the study of science reflect differences in cognitive science itself. At present, ‘cognitive science’ is not a unified field of study, but an amalgam of parts of several previously existing fields, especially artificial intelligence, cognitive psychology, and cognitive neuroscience. Linguistic, anthropology, and philosophy also contribute. Which particular approach a person takes has typically been determined more by developing a cognitive approach may depend on looking past specific disciplinary differences and focussing on those cognitive aspects of science where the need for further understanding is greatest.
 Broadly, the problem of scientific change is to give an account of how scientific theories, proposition, concepts, and/or activities alter over the corpuses of times generations. Must such changes be accepted as brute products of guesses, blind conjectures, and genius? Or are there rules according to which at least some new ideas are introduced and ultimately accepted or rejected? Would such rules be codifiable into coherent systems, a theory of ‘the scientific method’? Are they more like rules of thumb, subject to exceptions whose character may not be specifiable, not necessarily leading to desired results? Do these supposed rules themselves change over time? If so, do they change in the light of the same factors as more substantive scientific beliefs, or independently of such factors? Does science ‘progress’? And if so, is its goal the attainment of truth, or a simple or coherent account (true or not) of experience, or something else?
 Controversy exists about what a theory of scientific change should be a theory of the change ‘of’. Philosophers long assumed that the fundamental objects of study of study are the acceptance or rejection of individual belief or propositions, change of concepts, positions, and theories being derivative from that. More recently, some have maintained that the fundamental units of change are theories or larger coherent bodies of scientific belief, or concepts or problems. Again, the kinds of causal factors which an adequate theory of scientific change should consider are far from evident. Among the various factors said to be relevant are observational data: The accepted background of theory, higher - level methodological constraints, psychological, sociological, religious, meta - physical, or aesthetic factors influencing decisions made by scientists about what to accept and what to do.
 These issues affect the very delineation of the field of philosophy of science, in  what ways, if any, does it, in its search for a theory of scientific change, differ from and rely on other areas, particularly the history and sociology of science? One traditional view was that those others are not relevant at all, at least in any fundamental way. Even if they are, exactly how do they relate to the interest peculiar to the philosophy of science? In defining their subject many philosophers have distinguished maltsters internal to scientific development - ones relevant to the discovery and/or justification  of scientific claims - from ones external thereto - psychological, sociological, religious, metaphysical, and so forth, not directly relevant but frequently having a causal influence. A line of demarcation is thus drawn between science and non - science, and simultaneously between philosophy of science, concerned with the internal factors which function as reasons (or count as reasoning), and other disciplines, to which the external, nonrational factors are relegated.
 This array of issues is closely related to that of whether a proper theory of scientific change is normative or descriptive. Is philosophy of science confined in description of what scientific cases be described with complete accuracy as it is descriptive, to what extent must scientific cases be described with compete accuracy? Can the theory of internal factors be a ‘rational reconstruction’ a retelling that partially distorts what actually happened in order to bring out the essential reasoning involved?
 Or should a theory of scientific change be normative, prescribing how science ought to proceed? Should it counsel scientists about how to improve their procedures? Or would it be presumptuous of philosophers to advise them about how to do what they would it be presumptuous of philosophers to advise them about how to do what they are far better prepared to do? Most advocates of normative philosophy of science agree that their theories are accountable somehow to the actual conduct of science. Perhaps philosophy should clarify what is done in the best science: But can what qualifies as ‘best science’ be specified without bias? Feyerabend objects to taking certain developments as paradigmatic of good science. With others, he accepts the  ‘Pessimistic induction’ according to which, since all past theories have proved incorrect, present ones can be expected to do so also, what we consider good science, eve n the methodological rules we rely on, may be rejected in the future.
 Much discussion of scientific change since Hanson centres on the distinction between context of discovery and justification. The distinction is usually ascribed to the philosopher of science and probability theorist Hans Reichenbach (1891 - 1953) and, as generally interpreted, reflective attitude of the logical empiricist movement and of the philosopher of science Raimund Karl Popper (1902 - 1994) who overturns the traditional attempts to found scientific method in the support that experience gives in suitably formed generalizations and theories. Stressing the difficulty, the problem of ‘induction’ put in front of any such method. Popper substitutes an epistemology that starts with the hold,, imaginative formation of hypotheses. These face the tribunal of experience, which has the power to falsify, but not to confirm them. A hypotheses that survives the ordeal of attempted refutation between science and metaphysics, that an unambiguously refuted law statement may enjoy a high degree of this kind of ‘confirmation’, where can be provisionally accepted as ‘corroborated’, but never assigned a probability.
 The promise of a ‘logic’ of discovery, in the sense of a set of algorithmic, content - neutral rules of reasoning distinct from justification, remains unfulfilled. Upholding the distinction between discovery and justification, but claiming nonetheless that discovery is philosophically relevant, many recent writers propose that discovery is a matter of a ‘methodology’, ‘rationale’, or ‘heuristic;’ rather than a ‘logic’. That is, only a loose body of strategies or rules of thumb - still formulable discoveries, there is content of scientific belief - which one has some reason to hope will lead to the discovery of a hypothesis.
 In the enthusiasm over the problem of scientific change in the 1960s nd 1970s, the most influential theories were based on holistic viewpoints within which scientific ‘traditions’ or ‘communities’ allegedly worked. The American philosopher of science Samuel Thomas Kuhn (1922 - 96) suggested that the defining characteristic of a scientific tradition is its ‘commitment’ to a shared ‘paradigm’. A paradigm is ‘the source of the methods, problem - field, and standards of solution accepted by any mature scientific community at any given time’. Normal science e, the working out of the paradigm, gives way to scientific revolution when ‘anomalies’ in it precipitate a crisis leading to adoptions of a new paradigms. Besides many studies contending that Kuhn’s model fails for some particular historical case, three major criticisms of Kuhn’s view are as follows. First, ambiguities exist in his notion of a paradigm. Thus a paradigm includes a cluster of components, including ‘conceptual, theoretical, instrumental, and methodological’ communities: It involves more than is capturable in a single theory, or even in words. Second, how can a paradigm fall, since it determine s what count as facts, problems, and anomalies? Third, since what counts as a ‘reason’ is paradigm - dependent, there remains no trans - paradigmatic reason for accepting a new paradigm upon the failure of an older one.
 Such radical relativism is exacerbated by the ‘incommensurability’ thesis shared by Kuhn (1962) and Feyerabend (1975), are, even so, that, Feyerabend’s differences with Kuhn can be reduced to two basic ones. The first is that Feyerabend’s variety of incommensurability is more global and cannot be localized in the vicinity of a single problematic term or even a cluster of terms. That is, Feyerabend holds that fundamental changes of theory lead to changes in the meaning of all the terms in a particular theory. The other significant difference concerns the reasons for incommensurability. Whereas Kuhn thinks that incommensurability stems from specific translational difficulties involving problematic terms. Feyerabend’s variety of incommensurability seems to result from a kin d of extreme holism about the nature of meaning itself. Feyerabend is more consistent than Kuhn in giving a linguistic characterization of incommensurability, and there seems to be more continuity in his usage over time. He generally frames the incommensurability claim in term’s of language, but the precis e reasons he cites for incommensurability are different from Kuhn’s. One of Feyerabend‘s most detailed attempts to illustrate the concept of incommensurability involves the medieval European impetus theory and Newtonian classical mechanics. He claims that ‘the concept of impetus, as fixed by the usage established in the impetus theory, cannot be defined in a reasonable way within Newton’s theory’.
 Yet, on several occasions Feyerabend explains the reasons for incommensurability by saying that there are certain ‘universal rules’ or ‘principles of construction’ which govern the terms of one theory and which are violated by the other theory. Since the second theory violates such rules, any attempt to state the claims of that theory in terms of the first will be rendered futile. ‘We have a point of view (theory, framework, cosmos, modes of representation) whose elements (concepts, facts, picture) are built up in accordance e with certain principles of construction. The principle s involve e something ;like a ‘closure’, there are things that cannot be said, or ‘discovered’, without violating the principles (which does not mean contradicting them). Stating such terms as ‘universal’ he states: ‘Let us call a discovery, or a statement, or an attitude incommensurable with the cosmos (the theory, the framework) if it suspends some of its universal principles’.  As an example, of this phenomena, consider two theories, ‘T’ and T*, where ‘T’ is classical celestial mechanics, including the space - time framework, and ‘T’ is general relativity theory. Such principles as the absence of an upper  limit for velocity, governing all the terms in celestial mechanics, and these terms cannot be expressed at once such principles are violated, as they will be by general relativity theory. Even so, the meaning of terms is paradigm - dependent, so that a paradigm tradition is ‘not only incompatible but often actually incommensurable with that which has gone before’. Different paradigms cannot even be compared, for both standards of comparison and meaning are paradigm - dependent.
 Response to incommensurability have been profuse in the philosophy of science, and only a small fractions can be sampled at this point, however, two main trends may be distinguished. The first denies some aspects of the claim, and suggests a method of forging a linguistic comparison among theories, while the second, though not necessarily accepting the claim of linguistic incommensurability proceeds to develop other ways of comparing scientific theories.
 Inn the first camp are those who have argued that at least one component of meaning is unaffected by untranslatability: Namely, reference, Israel Scheller (1982) enunciates this influential idea in responses to incommensurability, but he does not supply a theory of reference to demonstrate how the reference of terms from different theories can be compared. Later writers seem to be aware of the need for a full - blown theory of reference to make this response successful. Hilary Putnam (1975) argues that the causal theory of reference can be used to give an account of the meaning of natural kind terms, and suggests that the same can be done for scientific terms in general, but the causal theory was first proposed as a theory of reference for proper names, and there are serious problems with the attempt to apply it to science. An entirely different language response to the incommensurability claim is found in the American philosopher Herbert Donald Davidson (1917 - 2003), where the construction takes place within a generally ‘holistic’ theory of knowledge and meaning. A radial interpreter can tell when a subject holds a sentence term and using the principle of ‘charity’ ends up making an assignment of truth conditions to individual sentences, although Davidson is a defender of the doctrine of the ‘indeterminacy’ of radical translation and the in reusability ‘ of reference, his approach has seemed to many to offer some hope of identifying meaning as an extensional approach to language. Davidson is also known for rejection of the idea of a conceptual scheme, thought of s something peculiar to one language or one way of looking at the world.
 The second kind of response to incommensurability proceeds to look or non - linguistic ways of making a comparison between scientific theories. Among these responses one can distinguish two main approaches. One approach advocates expressing theories in model - theoretic terms, thus espousing a mathematical mode of comparisons. This position has been advocated by writers such as Joseph Sneed and Wolfgang Stegmüller, who have shown how to discern certain structural similarities among theories in mathematical physics. But the methods of this ‘structural approach‘ do not seem applicable t any but the most highly mathematized scientific theories. Moreover, some advocate of this approach have claimed that it lends support to a model - theoretic analogue of Kuhn’s incommensurability claim. Another trend which has scientific theories to be entities in the minds or brains of scientists, and regard them as amendable to the techniques of recent cognitive science, proponents include Paul Churchlands, Ronald Gierre, and Paul Thagard. Thagard’s (1992) s perhaps the most sustained cognitive attempt to rely to incommensurability. He uses techniques derived from the connectionist research program in artificial intelligence, but relies crucially from a linguistic mode of representing scientific theories without articulating the theory of meaning presupposed. Interestingly, neither cognitivist who urges acing connectionist methods to represent scientific theories. Churchlands (1992), argues that connectionist models vindicate Feyerabend’s version of incommensurability.
 The issue of incommensurability remains a live one. It does not arise just for a logical empiricist account of scientific theories, but for any account that allows for the linguistic representation of theories. Discussions of linguistic meaning cannot be banished from the philosophical analysis of science, simply because language figures prominently in the daily work of science itself, and its place is not about to be taken over by any other representational medium. Therefore, the challenge facing anyone who holds that the scientific enterprise sometimes requires us to mk e a point - by - point linguistic comparison of rival theories is to respond to the specific semantic problem raised by Kuhn and Feyerabend. However, if one does not think that such a piecemeal comparison of theories is necessary, then the challenge is tp articulate another way of putting scientific theories in the balance and weighing them against one - another.
 The state of science at any given time is characterized, in part at least, by the theories that are ‘accepted’ at that time. Presently, accepted theories include quantum theory, the general theory of relativity, and the modern synthesis of Darwin and Mendel, as well as lower level (but still clearly theoretical) assertions such as that DNA has a double helical structure, that the hydrogen atom contains a single electron and so firth. What precisely involves the accepting of a theory?
 The commonsense answer might appear to be that given by the scientific realist, to accept a theory means, at root, to believe it to be true for at any rate, ‘approximately’ or ‘essentially’ true. Not surprising, the state of theoretical science at any time is in fact too complex to be captured fully by any such single notion.
 For one thing, theories are often firmly accepted while being explicitly recognized to be idealizations. The use of idealizations raises as number of problems for the philosopher of science. One such problem is that of confirmation. On the deductive nomological model of scientific theories, which command virtually universal assent in the eighteenth and nineteenth centuries, is that confirming evidence for a hypothesis of evidence which increases its probability. Nonetheless, presumably, if it could be shown that any such hypothesis is sufficiently well confirmed by the evidence, then that would be grounds for accepting it. If, then, it could be shown that observational evidence could confirm such transcendent hypotheses at all, then that would go some way to solving the problem of induction. Nevertheless, thinkers as diverse in their outlook as Edmund Husserl and Albert Einstein have pointed to idealizations as the hall - mark of modern science.
 Once, again, theories may be accepted, not be regarded as idealizations, and yet be known not to be strictly true  -  for scientific, rather than abstruse philosophical, reasons. For example, quantum theory and relativity theory were uncontroversially listed as among those presently accepted in science. Yet, it is known that the two theories, yet relativity requires all theories are not quantized, yet quantum theory say that fundamentally everything is. It is acknowledged that what is needed is a synthesis of the two theories, a synthesis which cannot of course (in view of their logical incommutability) leave both theories, as presently understood, fully intact, (This synthesis is supposed to be supplied by quantum field theory, but it is not yet known how to articulate that theory fully) none of this means, that the present quantum and relativistic theories regarded as having an authentically conjectural character. Instead, the attitude seems to be that they are bound to survive in modified form as limited cases in the unifying theory of the future   -  this is why a synthesis is consciously sought.
 In addition, there are theories that are regarded as actively conjectured while nonetheless being accepted in some sense: It is implicitly allowed that these theories might not live on as approximations or limiting cases in further sciences, though they are certainly the best accounts we presently have of their related range of phenomena. This used to be (perhaps still is) the general view of the theory of quarks, few would put these on a par with electrons, say,, but all regard them as more than simply interesting possibilities.
 Finally, the phenomenon of change in accepted theory during the development of science must be taken into account: But from the beginning, the distance between idealization and the actual practice of science was evident. Karl Raimund Popper (1902 - 1994), the philosopher of science, was to note, that an element of decision is required in determining what constitute a ‘good’ observation, a question of this sort, which leads to an examination of the relationship between observation and theory, has prompted philosophers of science to raise a series of more specific questions. What reasoning was in fact used to make inferences about light waves, which cannot be observed from diffraction patterns that can be? Was such reasoning legitimate? Are they to be construed as postulating entities just as real as water waves only much smaller? Or should the wave theory be understood non realistically as an instrumental device for organizing the predicting observable optical phenomena such ass the reflection, refraction, and diffraction of light? Such questions presuppose that here is a clear distinction between what can and cannot be observed. Is such a distinction clear? If so, how is it to be drawn? As, these issues are among the central ones raised by philosophers of science about theory that postulates unobservable entities
 Reasoning begins in the ‘context of justification’, as this is accomplished by deriving conclusions deductively from the assumptions of the theory. Among these   conclusions at least some will describe states of affairs capable of being establish ed as true or false by observations. If these observational conclusions turns out to be true, the theory is shown to be empirically supported or probable. On a weaker version due to Karl Popper (1959), the theory is said to be ‘corroborated’, meaning simply that it has been subjected to test and has not been falsified. Should any of the observational conclusions turn out to be false, the theory is refuted, and must be modified or replaced. So a hypothetico - deductivist can postulate any unobservable entities or events he or she wishes in the theory, so long as all the observational conclusions of the theory are true.
 However, against the then generally accepted view that the empirical science are distinguished by their use of an inductive method. Popper’s 1934 book had tackled  two main problems: That of demarcating science from non - science (including pseudo - science and metaphysics), and the problem of induction. Again, Popper proposed a falsifications criterion of demarcation: Science advances unverifiable theories and tries to falsify them by deducing predictive consequences and by putting the more improbable of these to searching experimental tests. Surviving such testing provided no inductive support for the theory, which remain a conjecture, and may be overthrown subsequently. Popper’s answer to the Scottish philosopher, historian and essayist David Hume (1711 - 76), was that he was quite right about the invalidity of inductive inference, but that this does not matter, because these play no role in science, in that the problem of induction drops out.
 Then, is a scientific hypothesis to be tested against protocol statements, such that the basic statements in the logical positivist analysis of knowledge, thought as reporting the unvanishing and pre - theoretical deliverance of experience: What it is like here, now, for me. The central controversy concerned whether it was legitimate to couch them in terms of public objects and their qualities or whether a less theoretical committing, purely phenomenal content could be found. The former option makes it hards to regard then as truly basic, whereas the latter option ,makes it difficult to see how they can be incorporated into objectives science. The controversy is often thought to have been closed in favour of a public version by the ‘private language’ argument. Difficulties at this point led the logical positivist to abandon the notion of an epistemological foundation altogether, and to flirt with the ‘coherence theory’ of truth’, it is widely accepted that trying to make the connection between thought and experience through basic sentences depends on an untenable ‘myth of the given’.
 Popper advocated a strictly non - psychological reading of the empirical basis of science. He required ‘basic’ statements to report events that are ‘observable’ only in that they involve relative position and movement of macroscopic physical bodies in certain space - time regions, and which are relatively easy to tests. Perceptual experience was denied an epistemological role (though allowed a causal one),: Basic statements are accepted as a result of a convention or agreement between scientific observers. Should such an agreement break down, the disputed basic statements would need to be tested against further statements that are still more ‘basic’ and even easier to test.
 But there is an easy general result as well: Assuming that a theory is any deductively closed set of sentences as assuming, with the empiricist, that the language in which these sentences are expressed has two sorts of predates (observational and theoretical) and, finally, assuming that the entailment of the evidence is the only constraint on empirical adequacy, then there are always indefinitely many different theories which are equally empirically adequate as any given theory. Take a theory as the deductive closure of some set of sentences in a language in which the two sets of predicates are differentiated: Consider the restriction of ‘T’ to quantifier - free sentences expressed purely in the observational vocabulary, then any conservative extension of that restricted set of T’s consequences back into the full vocabulary is a ‘theory’ co - empirically adequate with  -  entailing the same singular observational statement as  -  ‘T’. Unless very special conditions apply (conditions which do not apply to any real scientific theory), then some of these empirically equivalent theories will formally contradict ‘T’. (A similarly straightforward demonstration works for the currently a  fashionable account of theories as set of models.)
 Many of the problems concerning scientific change have been clarified, and many new answers suggested. Nevertheless,, concepts central to it (like ‘paradigm’, ‘core’, ‘problem’, constraint’, ‘verisimilitude’) still remain formulated in highly general, even programmatic ways. Many devastating criticisms of the doctrine based of them have not been answered satisfactory.
 Problems centrally important for the analysis of scientific change have been neglected, there are, for instance, lingering echoes of logical empiricism in claims that the methods and goals of science are unchanging, and thus are independent of scientific change itself, or that if they do change, they do so for reasons independent of those involved in substantive scientific change itself. By their very nature, such approaches fail to address the change that actually occur in science. For example, even supposing that science ultimately seeks the general and unalterable goal of ‘truth’ or ‘verisimilitude’, that injunction itself gives guidance ass to what scenists should seek or others should go about seeking it. More specific goals do provide guidance, and, as the transition from technological mechanistic to gauge - theoretic goals illustrate, those goals are often altered in light of discoveries about what is achieved, or about what kinds of theories are promising. A theory of scientific change should account for these kinds of goal changes, and for how, once accepted, they alter the rest of the patterns of scientific reasoning and change, including ways in which mor general goals and methods may be reconceived.
 Traditionally, philosophy has concerned itself with relations between propositions which are specifically relevant to one another in form or content. So viewed, philosophical explanation of scientific change should appeal to factors which are clearly more scientifically relevant in their content to the specific direction of new scientific research and conclusions than are social factors whose overt relevance lies elsewhere. However, in recent years many writers, especially in the ‘strong programme’ in the sociology of science have maintained that all purported ‘rational’ practices must be assimilated to social influences.
 Such claims are excessive. Despite allegations that even what is counted as evidence is a matter of mere negotiated agreement, many consider that the last word has not been said on the idea tat there is in some deeply important sense a ‘given’, inn experience in terms of which we can, at least partially, judge theories. Again, studies continue to document the role of reasonably accepted prior beliefs (‘background information’) which can help guide those and other judgements. Even if we can no longer naively affirm the sufficiency of ‘internal’ givens and background scientific information to account for what science should and can be, and certainly for what it is often in human practice, neither should we take the criticisms of it or granted, accepting that scientific change is explainable only by appeal to external factors.
 Equally, we cannot accept too readily the assumption (another logical empiricist legacy) that our task is to explain science and its evolution by appeal to meta - scientific rules or goals, or metaphysical principles, arrived at in the light of purely philosophical analysis, and altered (if at all) by factors independent of substantive science. For such trans - scientific analysis, even while claiming to explain ‘what science is’, do so in terms ‘external’ to the processes bty which science actually changes.
 Externalist claims are premature: Not enough is yet understood about the roles of indisputable scientific consecrations in shaping scientific change, including changes of method and goals. Even if we ultimately cannot accept the traditional ‘internalist’ approach in philosophy of science, as philosophers concerned with the form and content of reasoning we must determine accurately how far it can be carried. For that task, historical and contemporary case studies are necessary but insufficient: Too often the positive implications of such studies are left unclear, and their too hasty assumption is often that whatever lessons are generated therefrom apply equally to later science. Larger lessons need to be a systematic account integrating the revealed patterns of scientific reasoning and the ways they are altered into a coherent interpretation of the knowledge  - seeking enterprise  -  a theory of scientific change. Whether such efforts are successful or not, it only nr=e through attempting to give sch a coherent account in scientific terms , or through understanding our failure ton do so, that it will be possible to assess precisely the extent to which trans - scientific factors (meta - scientific, social, or otherwise) must be included in accounts of scientific change.
 That for being on one side, it is noticeable that the modifications for which of changes have conversely been revealed as a quality specific or identifying to those of something that makes or sets apart the unstretching obligation for ones approaching the problem. That it has echoed over times generations in making different or become different, to transforming substitution for or among its own time of change. Finding in the resulting grains of residue that history has amazed a gradual change of attitudinal values for which times changes in 1925, where the old quantum mechanics of Planck, Einstein, and Bohr was replaced by the new (matrix) quantum mechanics of Born, Heisenberg, Jordan, and Dirac. In 1926 Schrödinger developed wave mechanics, which proved to be equivalent to matrix mechanics in the sense that they ked to the same energy levels. Dirac and Jordan joined the two theories into pone transformation quantum theory. In 1932 von Neumann presented his Hilbert space formations of quantum mechanics and proved a representation theorem showing that sequences in transformation theory were isomorphic notions of theory identity are involved, as theory individuation of theoretical equivalence and empirical equivalences.
 What determines whether theories T1 and T2, are instances of the same theory or distinct theories? By construing scientific theories as partially interpreted syntactical axiom system TC, positivism made specific of the axiomatization individuating factures of the theory. Thus, different choices of axioms T or alternations in the correspondence rules - say, to accommodate a new measurement procedure - resulting in a new scientific meaning of the theorized descriptive terms τ. Thus, significant alternations in the axiomatization would result not only in a new theory T’C’ but one with changed meaning τ’. Kuhn and Feyerabend maintained that the resulting change could make TC and T’C’ non-comparable, or ‘incommensurable’. Attempts to explore individuation issues for theories through the medium of meanings change or incommensurability proved unsuccessful and have been largely abandoned.
 Individuation of theories in actual scientific practice is at odds with the positivistic analyses. For example, difference equation, differential equations, and Hamiltonian versions of classical mechanics, are all formulations of one theory, though they differ in how fully they characterize classical mechanics. It follows that syntactical specifics of theory formulation cannot be undeviating features, which is to say that scientific theories are not linguistic entities. Rather, theories must be some sort of extra-linguistic structure which can be referred to through th medium of alterative and even in equivalent formulations (as with classical mechanics). Also, the various experimental designs, and so forth, incorporated into positivistic correspondence rules cannot be individuating features of theories. For improved instrumentation or experimental technique  does not automatically produce a new theory. Accommodating these individuation features was a main motivation for the semantic conception of theories where theories are state spaces or other extra-linguistic structures standing in mapping relations to phenomena.
 Scientific theories undergo developments, are refined, and change. Both syntactic and semantic analysis of theories concentrate on theories at mature stages of development, and it is an open question either approach adequately individuates theories undergoing active development.
 Under what circumstances are two theories equivalent? On syntactical approaches, axiomatizations T1 and T2 having a common definitional extension would be sufficient Robinson’s theorem which says that T1 and T2 must have a model in common t be compatible. They will be equivalent if theory have precisely the same (or equivalent) sets of models. On the semantic conception the theories will be two distinct sets of structures (models) M1 and M2. The theories will be equivalent just in case we can prove a representation theorem showing that M1 and M2 are isomorphic (structurally equivalent). In this way von Neumann showed that transformation quantum theory and the Hilbert Space formulation were equivalent.