Back to HomeInternalism/Externalism Entry

(Last revised 09/10/99

Putnam, H. (1975), "The Meaning of ‘Meaning’ ", in Mind, Language and Reality. Philosophical Papers, vol. 2, Cambridge University Press, Cambridge), pp. 215-271. 

By relying on a thought experiment according to which in a remote part of our universe there lies a planet which is a physical duplicate of the Earth but for the fact that H20 is replaced by a different substance however phenomenologically indistinguishable from it, Putnam discards the Fregean assumption to the effect that sense determines reference. A new picture of meaning is provided, which fits the main claim of the so-called new theory of reference, i.e. that the meaning of directly referential terms such as names and indexicals collapses on their reference. According to this picture, meanings are no longer in the head, but are primarily determined by the external physical environment an individual who uses the relevant term is in relation with. A residual mentalistic component of meaning is however individuated in terms of the stereotype. This is a generic descriptive characterization which may be satisfied not exclusively by the extension of the relevant term and which moreover it may turn out to be even false of that extension. It must however be possessed by anyone to whom a linguistic competence regarding that term is ascribed. According to whether one is interested in focussing upon an individual’s competence or his/her relation to his/her physical environment, a certain mental state of his/her can be taxonomized either as narrow or as broad. 

Burge, T. (1979), "Individualism and the Mental", Midwest Studies in Philosophy 4, pp. 73-121. 

A famous thought-experiment is presented in order to show that individualism, i.e. the view according to which nothing which obtains outside an individual’s body is relevant in order to individuate his mental states, is false. An individual whose internal (i.e., neurological) make-up remains the same may have different thoughts insofar as he is respectively placed in a certain actual an in a certain counterfactual social environments. For depending on these different locations his thoughts will have different contents, i.e. will concern different concepts, and therefore they will be different mental state types. 

Fodor, J.A. (1980) "Methodological Solipsism Considered as a Research Strategy in Cognitive Psychology", Behavioral and Brain Sciences 3, pp. 63-109. 

Internalism's cornerstone, which conjugates the representational theory of mind, according to which propositional attitudes are relations between organisms and their mental representations, and the computational theory of mind, according to which mental processes have access only to formal-non-semantic, syntactic, properties of the mental representations over which they are defined. Propositional attitudes are type-distinct insofar as the representations they are related with are formally distinct. To identify intentional states only in terms of the formal properties of the representations they are related with amounts to adopting methodological solipsism, according to which the semantic properties of these representations such as being about something and being true - hence, what determines the representations' truth-conditions - are psychologically inert. Psychological explanation entails both a) an explanation of behavior in terms of representations' content and b) an opaque taxonomy of mental states, according to which the truth-conditionally identical beliefs that a is F and that b is F are type-distinct. Both points are accounted for by the representational-computational theory of mind. For according to it the representations the afore-mentioned type-distinct beliefs are related with are have different behavior-explaining content only insofar as they formally distinct; and they may be such even though they share their truth-conditions. Finally, Putnam's (1975) classification of a narrow mental state coincides with its opaque individuation; as a consequence, twins' truth-conditionally different states are type-identical. 

Pettit, P. (1980), "Wittgenstein, Individualism and the Mental", in E. Morscher, R. Stranzinger (eds.), Ethics - Foundations, Problems, and Applications, Hölder-Pichler-Tempsky, Vienna, pp. 446-455. 

Unlike Burge’s (1979) anti-individualism, Wittgenstein’s anti-individualism holds that the content of a belief is guideline-, but not grasp-independent; that is, it will be such that we can ascribe it to a belief only where the believer is responsive to appropriate guidelines for the use of words which express that content, but not such as not to presuppose that the believer has a proper understanding of it. This implies that, unlike Burge, if we imagine to positionate a subject endowed with certain true beliefs involving a certain concept in a counterfactual community where the corresponding term is differently used, we are not able to attribute him beliefs with contents involving the concept of the counterfactual community; for that subject would have to be said to lack proper grasp of this latter concept. A further consequence is that, unlike the Burgean one, this latter thought-experiment applies to any beliefs whose contents require linguistic behavior for their manifestation, not to those beliefs only whose content involve concepts which are complex enough for us to be to able to ascribe possession of them in the absence of proper understanding. 

Peacocke, C. (1981), "Demonstrative Thought and Psychological Explanation", Synthese 49, pp. 187-217. 

A neo-Fregean theory of demonstrative thoughts. Like Frege, one such thought has both an informative value (is the bearer of truth-values) and a cognitive significance (is the content of propositional attitudes). Unlike Frege, however, it is object-dependent insofar as the mode of presentation type it contains (not a Kaplan’s (1989) character, for it is not essentially linguistic) is appropriately indexed with an object in order to get a mode of presentation token. Hence, no object, no mode of presentation token. This mode of presentation is employed by the thought’s subject but merely referred to by that thought’s reporter. Putnam (1975) cases do not force a bifurcated account of demonstrative thought, for neo-Fregean object-dependent demonstrative thoughts figure in psychological explanation insofar as they cause Twins’ different object-dependent actions. An indispensability thesis is finally defended, according to which a satisfactory explanation of a subject’s acting on a given object must include, within that subject’s attitudes, a demonstrative thought containing a demonstrative mode of presentation of that object. 

Fodor, J.A. (1982), "Cognitive Science and the Twin-Earth Problem", Notre Dame Journal of Formal Logic 23, pp. 98-118. 

A first attempt to bifurcate narrow and broad content is provided, by distinguishing between the content of a belief, given in both universal and phenomenological (or stereotypical) terms, and its truth-conditions. This make Putnamian twins share a 'phenomenological' de dicto belief, i.e. hold a universally quantified belief, although the truth-conditions of their beliefs' tokenings differ, for different things in their respective environment satisfy the description the universally quantified belief contains. Meaning  is both in and outside the head; the internal content determines the extension of the relevant belief-expressing terms) only provided a context. 

Harman, G. (1982), "Conceptual Role Semantics", Notre Dame Journal of Formal Logic 23, pp. 242-256. 

Two main theses are presented: a) the meaning of linguistic expressions are determined by the concepts they are used to express b) the contents of these concepts is determined by the functional role in a person’s psychology. Mainly, this is the role the concept plays in inference; however, it also includes the functional relation with the external world in connection with perception on the one hand and with action on the other. Insofar as it is such, functional role is not solipsistic. Moreover, its identity is given by the normal context of its production; as a consequence, the content of a thought is not determined by the present context of its production (for present context =/= normal context). 

Loar, B. (1982), "Conceptual Role and Truth-Conditions", Notre Dame Journal of Formal Logic 23, pp. 272-283. 

The bases for a bifurcational theory of content are settled, by arguing that conceptual role semantics cannot explain both why two thoughts with the same conceptual role have different truth-conditions in different contexts and why we assign (certain) truth-conditions to a thought’s conceptual role. As to the former point, it is further claimed that to assign conceptual role also the ability of determining a meaning function is an ad hoc move. As to the latter, Loar remarks that it is true that knowing a thought’s conceptual role plus knowing the conceptual role of the disquotational predicate "true" amounts to knowing that thought’s truth-conditions; still this does not explain why that thought has those truth-conditions, for these cannot be determined by such conceptual roles. It is a mere coincidence that these roles put together ascribe that thought the same truth-conditions with respect to which that thought is a reliable indicator. Insofar as such a reliability is what fixes a thought’s truth-conditions, it is independently accounted to by causal theories of reference at a semantic level different from that which conceptual role semantics deals with. 

McGinn, C. (1982), "The Structure of Content", in A. Woodfield (ed.), Thought and Object, Oxford University Press, Oxford, pp. 207-258. 

A paradigmatic defense of bifactorial theories of mental as well as linguistic content. Truth-conditions of a mental state yield its broad content; the explanatory role of such a state requires its narrow content. Truth-conditions go hand in hand with the state's aboutness, which is extra-individually determined by means of a causal relation between an external object and a subject; explanatory role goes hand in hand with the (feasible) way that object is represented in that state. Broad content has to do with the information role of either a mental state or its linguistic expression (communication is elicited by an intention to convey information about the external world; in order for this intention to be fulfilled the linguistic signs must already be in a representational relation with the world). Narrow content has to do with the thinker's propension to assign a certain probability value to his/her thoughts/statements conditionally upon certain other thoughts/statements. On the linguistic side, a term's occurring transparently gives us only knowledge about the truth-conditions of the sentences where it figures; if it occurs opaquely, it gives us informati on also on the relevant subject's behavior-explaining representations. A theory of truth is only partially a theory of meaning, for truth-conditionally identical sentences may not be synonymous (they may differ in their narrow content). Transparency of meaning only regards narrow content. 
McDowell, J. (1984), "De Re Senses", The Philosophical Quarterly 34, pp. 283-294. 

A defense of object-dependent thoughts is given by claiming that de re thoughts include in their content object-dependent de re modes of presentation. Insofar as a switch of context brings to one's attention a different object, such a switch forces a switch in the corresponding object-dependent thought. This does not make the object be a constituent of the thought, however. For this would violate the Fregean requisite that (the content of) a thought is a sense whose constituents are only senses, not objects. De re modes of presentation can be grouped into de re types; a de re type mode of presentation may determine a de re token mode, or else it may determine nothing (this will happen whenever no object is suitably involved in the thought's context). 

Fauconnier, G. (1985), Mental Spaces, The MIT Press, Cambridge MA. 

Focussing on a sentence’s truth-conditions is irrelevant if we want to grasp how a subject represents the world to himself in uttering that sentence. As far as the referential terms of that sentence are concerned, we must on the contrary rely on a pragmatic function which, as with metonymical or metaphorical reference, enables one to shift from the trigger, i.e. the semantic referent of one such term, to its target, namely the mental image which the involved subject associates with that term in his mental space. 

Block, N. (1986), "Advertisement for A Semantics for Psychology", Midwest Studies in Philosophy 10, pp. 615-678. 

A bifactorial version of a conceptual role semantics is put forward. Conceptual role is what a Mentalese term originally possesses and the corresponding natural language term inherits. To know the meaning of a term is therefore to instantiate in one’s mind a certain conceptual role. Two elements must be traced in it: i) what determines the function from context to broad content, hence what contextually the referential factor of meaning. This element includes causal relations with the contextual referent of the term; ii) what accounts for a subject’s behavior, namely its (holistically based) inferential role. This element is the narrow content of the term properly speaking. It is not ‘long-armed’ as Harman’s (1982) conceptual role, i.e. it does not include also sensorial inputs and behavioral outputs. 

Chomsky, N. (1986), Knowledge of Language, Praeger, New York. 

If one wants to scientifically study what being competent of a language is, one has to focus on I-languages, i.e. internally-based representational structures whose grammar is given by the modification, induced by the exposure of a subject to a certain linguistic environment, of the innate Universal Grammar. An I-language requires nothing like a broad content. For such a content is infected by scientifically untractable notions such as reference and truth, as these are called upon E-languages, i.e. languages seen as external artifacts determined by certain environmental relations. At most, I-language representations may be ascribed a syntactic counterpart of reference. 

Davies, M. (1986), "Externality, Psychological Explanation, and Narrow Content", Proceedings of the Aristotelian Society, Suppl. vol. 60, pp. 263-283. 

Fodor's individualist theory of narrow content  (paradigmatically exposed in the later Fodor (1987)) is mainly considered, in order to show first that if one à la Fodor taxonomizes mental states in terms of causal powers then externalism is not ruled out yet; hence, it is no longer the case that these powers remain the same across contexts. In this respect, the externalist's idea that environmental differences cause differences in the causal powers of a person's psychological states without changing the person physiologically is claimed to be perfectly maintainable. This does not make narrow content an unviable notion; a coarser taxonomy of mental states which generalize across environmental differences is still possible; the problem of narrow content's supposed inexpressibility may be overcome by appealing to the distinction between referring to and expressing a mode of presentation. To be sure, however, by comparing narrow content with the neo-Fregean notion of an (unanchored) mode of presentation it turns out that the former is not identical with the latter, for this does not supervene upon the neurophysiological. 

Jackendoff, R. (1986), "Conceptual Semantics", Versus 44/45, pp. 81-97. 

By virtue of their form alone, i.e. of their computational features, mental representations provide meanings in compositional terms for linguistic utterances, by thereby constituting semantic or conceptual structure. This structure is articulated in conceptual constituents whose basic elements are ontological category features which allows the mind to carve up the world as a constructed entity. They indeed provide the sense of linguistic terms which allows the latter to refer to constructed slices of reality. Moreover, their number is indefinite for the mind can generate new conceptual items at will. Finally, they are either decomposable in primitives or analyzable in terms of preference rule systems. 

Lakoff, G., "Cognitive Semantics", Versus 44/45 (1986), pp. 119-154. 

Beyond internalism and externalism. Against the picture of mental content dictated by the metaphysical realism advanced e.g. by Putnam (1975), a theory of experientialist cognition is put forward, according to which both innate capacities and the experience given to a subject by its bodily interaction with its natural and social environment determine one’s conceptual structure. Basic-level concepts, which turn out of the interaction with one’s environment, and image-schemas are what our innate capacity operates on in order to get abstract cognitive models. This is obtained by means of various imaginative processes of projections: schematization, metaphor, metonymy and categorization. 

Lepore, E., Loewer, B. (1986), "Solipsistic Semantics", Midwest Studies in Philosophy 10, pp. 595-614. 

Although Kripke's and Kaplan's reflections on proper names and indexicals are hardly reconcileable with a solipsistic semantics, i.e. a theory of meaning according to which the representation of one thinker have meaning independently of the existence of any other mental or physical individual than the thinker itself or its thought, the Cartesian intuition according to which one may have all the thoughts one has independently of whether they are veridical or not seems to support it. This leaves the way open to a solipsistic conception of narrow content, for this is the only content which supervenes on the thinker's body. Three are the possible ways of conceiving such a content: conceptual role theory, the "phenomenological" and the "indexical" strategy. The former however fails to really attribute a content to one's thought, while the latter two actually presuppose a non-solipsistic semantics. Moreover, if one endorses an informational theory of meaning the only way to make it fit solipsistic semantics is to implausibly claim that one's thoughts are about one's nervous system. Transparency of meaning may however be retained even without sticking to solipsistic semantics if it can be shown that one knows the content of one's thought by having the corresponding disquotational thought, i.e. by knowing the former thought's truth-conditions. This latter knowledge is made possible by the facts that that thought has a certain functional role and it is related in an appropriate way to the object it is about. 

McDowell, J. (1986), "Singular Thought and the Extent of Inner Space", in P. Pettit and J. McDowell (eds.), Subject, Thought and Context, Clarendon Press, Oxford, pp. 137-68. 

Object-dependent, or singular, thoughts conceived à la Frege (i.e. those thoughts whose objects figure in them without constituting them) remove the idea of a gap between mind and reality. This idea is a Cartesian legacy which arises when one moves from the epistemological doubt that one's thought may only seem to be a singular one to the claim that there are inner facts completely separated from outer ones, as witnessed by the fact that they are privilegedly accessed by their subject. Once a thought is conceived as a separated inner entity, however, it can no longer be maintained that it has a content by means of which it directs itself upon an external object. Hence, both internalist and dualist theories of intentional states are utterly unable to account for their intentionality. A so-called inner (or an allegedly narrow component of) content has no intentionality; whatever is subjectively perceived as being directed upon an external object cannot be something radically inner. Intentionality may instead be accounted for in terms of acquaintance for a subject having a certain object-dependent thought with that very object. 

McCulloch, G. (1986), "Scientism, Mind, and Meaning", in Pettit, P. & McDowell, J. (eds.), Subject, Thought, and Context, Clarendon Press, Oxford, pp. 59-94. 

The traditional individuation of propositional attitudes via their broad content, taken as the meaning expressed by the sentence which figures as nested in a propositional attitude ascription, is held by means of a defense of folk psychology against a methodologically-solipsistically oriented superpsychology. One cannot dispense with folk psychology insofar as it irreducibly accounts for the phenomenology of mental states, i.e. the way mental states appear to their subjects who take themselves as persons communicatively interacting with other persons. 

Fodor, J.A. (1987), Psychosemantics, The MIT Press, Cambridge MA. 

The most famous dualist theory of content, shaped in a contextualist version. In order for folk-psychological explanation of behavior to be acceptable, content must supervene on internal (ultimately neurophysiological) factors. Putnam (1975) and Burge (1979) Twins' cases show that this is not the case as far as broad content is concerned. Hence, a different notion of content is needed, narrow content. This is what allows an individuation of intentional states in terms of their causal powers. It is moreover treated à la Kaplan (1989) as a function from contexts of thought to broad contents. Unlike traditional dualist theories such as Block (1986), Loar (1988), and McGinn (1982), narrow content is a content only potentially, for it really is what broad content is provided a content. On its turn, broad content is given denotationally, along the lines of a naturalistic theory of intentionality which makes it what adequately covaries with the mental token which denotes it. To explain what such an adequate covariance is, Fodor puts forward an asymmetrical dependence theory of meaning. According to this theory, occasional causes of mental tokens are ruled out as such tokens' meanings insofar as the causal connection they enact is asymmetrically dependent on another causal connection between the above tokens and another type of entity, which is what therefore figures as the meaning of those tokens. 
LePore, E., Loewer, B. (1987), "Dual Aspect Semantics", in E. LePore (ed.), New Directions in Semantics, Academic Press, London, pp. 83-112. 

A detailed criticism of bifurcationalist theories of content is given, by arguing again the 'conceptual role conception' of narrow content. Conceptual role is a) hardly intersubjective b) useless for a theory of interpretation (one's assigning a certain conceptual role to a sentence does not count as evidence for one's believing that p); c) unable to account for psychological explanation, for psychological explanation involves content (S did r for she believed that p and desired that q) whereas to assign a sentence a certain conceptual role fails to individuate a content for the corresponding belief (in the inference "S bears a relation R* to a certain mental presentation/sentence M which has conceptual role C, then she believes that __", blanks cannot be filled). A suitable version of a Davidsonian theory of truth may instead be able to provide a unitary notion of content by both providing an account of language understanding and an account of the relations between meaning and reality. This is done by supplyng sentences with not reified truth-conditions, so that an individual may accept a certain biconditional without accepting another, even though the respective right-hand side members possess co-designative terms. 

Burge, T. (1988), "Individualism and Self-Knowledge", The Journal of Philosophy 85, pp. 649-665. 

One of the most famous attempts to discard the typical critique raised against externalism, according to which externalism violates the principle of the first-person authority about one’s mental states insofar as content externistically determined may be unknown to a subject who entertains a state with that content. Burge’s strategy is that of separating one’s knowledge of the content of one’s thought from one’s knowledge of the individuation conditions of such a thought. Insofar as one may fail to have the latter while keeping the former knowledge, first-person authority is saved. A doubtful defense, if Glock-Preston (1995) are right: content belongs to the individuation conditions of a thought. 
Loar, B.  (1988) “Social and Psychological Content” [with “Comments” by A.Bilgrami and a “Reply”], in Grimm, A.H. & Merrill, D. (eds.), Contents of Thoughts, The University of Arizona Press, Tucson. 

One of the most powerful defenses of a bifurcational theory of mental content. According to Loar, sameness of a de dicto ascription (expressed via the relevant oblique ‘that’-clause) does not imply sameness of conceptual role; nor do differences in de dicto ascriptions imply differences in conceptual role. This shows that there is a content - psychological content - which oblique ‘that’-clauses fail to express. As it is the mere conceptual role of a thought, this content must be conceived internalistically; insofar as it is such, it moreover is what is responsible for the explanation of action. What the oblique ‘that’-clauses rather express is social content, i.e. a content whose elements are determined social-externistically in the sense of Burge (1979). Insofar as different de dicto ascriptions match different truth-conditions of the thoughts ascribed, social content can be also considered as those thoughts’ respective truth-conditions. Although psychological content is not truth-conditional, it may be seen as the realization conditions of a certain thought, i.e. as the set of possible worlds in which that thought would be true if it were not a misconception. In a first-person perspective, this content seem representational (as social content actually is); in other terms, the thought it is the psychological content of appears to be in this perspective internally intentional; but this internal intentionality of a thought disappears in a third-person perspective, where that content precisely turns out to be mere conceptual role. 

Castañeda, H-N. (1989), Thinking, Language and Experience, University of Minnesota Press, Minneapolis. 

Mind is diaphanous to objecthood insofar as a thought content consists of guises, i.e. intensional objects which are made out of set of properties. Guises are however also what determines outer reality, for they may be connected in such a way - consubstantiation - as to form a system of guises, which roughly corresponds to what it is commonsensically held as a massive external object. 

Jackendoff, R. (1989), "What is a Concept, that a Person May Grasp It?", Mind and Language 4, pp. 68-102. 

Echoing Chomsky's (1986) distinction between I- and E-languages, Jackendoff's Conceptual Semantics only deals with I-concepts, i.e. both lexical concepts, i.e. mental schemas decomposable in terms of primitives and principle of combination, and sentential concepts, analogously generated out of lexical concepts and principles of combination. Conceptual Semantics differs from Fodor's Language of Thought insofar as it fails to attribute I-language sentences propositional content. For representations' combinatorial character does not have to mimic a supposed propositional content's combinatoriality. Although there are aspects of conceptual structure wich are not regularly expressed in syntax, each major syntactic constituent of a sentence corresponds to a conceptual constituent in the meaning of the sentence. Each conceptual constituent belongs to one of a small set of major ontological categories. 

Kaplan, D. (1989), "Demonstratives", in Almog, J., et al. (eds.), Themes from Kaplan, Oxford University Press, Oxford, pp. 481-563. 

By showing that Frege’s unitary notion of semantic meaning cannot work in the case of indexicals and that it must be replaced by content (the truth-conditional import of a tokening containing indexicals, which amounts to a Russellean singular proposition made out of the denotata of the tokenings’ subsentential terms) and character (a function from contexts of utterance to contents, which is intended to express both the linguistic and the cognitive meaning of an utterance containing indexicals), this paper influenced bifurcational theories of mental content. For, whereas content is what differs as far as tokenings of one and the same indexical sentences are concerned insofar as these tokenings are about different things (cf. broad content) character is what all these tokenings share (cf. narrow content). 

McGinn, C. (1989), Mental Content, Blackwell, Oxford. 

One of the best defenses of an externalist position. The external environment determines what an intentional state is insofar as what that state is about figures within the individuation conditions of such a thought. Weak externalism, according to which a state only depends on the existence of the external objects it is about, is preferable to strong externalism, which holds that a state depends on factors lying in the subject’s external environment. For it allows to keep misperceptions and local hallucinations the relational content one would intuitively ascribe them. A teleological account of intentional states, moreover, enables one to find a theoretical rationale for externally individuated content. For the relational proper function of an intentional state, namely what that state is supposed to do, coincides with such a content.. 

Segal, G. (1989), "The Return of the Individual", Mind 98, pp. 39-57. 

An internalist picture of singular thoughts is provided, first by criticizing McDowell's (1984, 1986) conception of object-dependent singular thoughts and second by arguing for a contextualist perspective on such thoughts. On the negative side, Segal relies on empty singular thoughts in order to show that the way McDowell accounts for their role in psychological explanation, i.e. by appealing to second-order beliefs (in which one believes himself to have a certain object-dependent singular belief), makes object-dependent singular thoughts superfluous as far as psychological explanation is concerned. For even when one has one such thought one also has a similar second-order belief. On the positive side, Segal says that a singular thought is such that its subject position is occupied not by an object-dependent sense, but rather by something like a free variable to which the context of the thought possibly provides a referent. As a consequence, different contexts provide different objects for one and the same singular thought. Whether it is filled or not with a contextual object, this thought is what figures in psychological explanation. 

Macdonald, C. (1990), "Weak Externalism and Mind-Body Identity", Mind 99, pp. 387-404. 

Both internalism and externalism can divide into two versions, depending on whether the qualifying claim of these doctrines is taken to hold with respect to either mental types or their tokenings. By being externalist with respect to mental types and internalist with respect to mental tokens, Macdonald labels her own position a weak externalism. Mental type externalism is argued for on explanatory bases. First, in order also for acts narrowly construed to be, one must already have acts which are broadly construed insofar as they are individuated by reference to appropriate objects located in the subject's environment. Second, these latter acts are explained only insofar as their mental reasons are analogously individuated. Third, when different mental states are, unbeknownst to their subject, about the same environmental object, it is this fact which explains the failure of that subject to act as expected. Moreover, mental token internalism is defended in order to account for the token identity between a mental state and its physical (neurological) correlate. The fact that a mental state type has its relational component as one of its essential features, which is what obtrudes the type identity between that state and its physical correlate (for the latter is definitely non-relational), does not force a token of its to have that component essentially as well. 

Bilgrami, A.(1992), Belief and Meaning, Blackwell, Oxford. 

Against bifactorialist theories, a unitary conception of linguistic and mental content is proposed. Content is constituted out of concepts. These are both holistically and relationally determined. Holistically, for a concept consists of a global set of beliefs which a subject has on a certain theme. As such, concepts are highly idiosincratic. Relationally, for at least some of those beliefs has to be determined in relation with the outer world. This makes Bilgrami hold himself for an externalist; no contact with the outer world, no content. To be sure, inasmuch as it exhausts to such a thesis, it is definitely a weak form of externalism. Frege-cases are accounted for by saying that distinct co-referential tokens are linked for a subject to different sets of (weakly relationally determined) beliefs. Explanation of behavior however requires a functional distinction between a global and a local content. Global, or aggregative, content is conceived along the afore-mentioned lines; local content amounts to selecting in global content a subset of beliefs responsible for a certain behavior. Commonality of behavior is thus explained by saying that distinct subjects may share a local content although they differ in the global content they associate with a certain term. 

Davies, M. (1992), "Perceptual Content and Local Supervenience", Proceedings of the Aristotelian Society 92, pp. 21-45. 

The intuition according to which perceptual content supervenes upon phenomenological character (which makes the former be an existentially quantified, not object-involving, content) is compatible with the externalist idea that twins belonging to different possible worlds have different perceptual contents insofar as they are exposed to different environments. For the supervenience called upon by that intuition is a weak, intra-worldly and intra-subjective, one. Whereas that idea implies the rejection of a strong, inter-worldly and intersubjective, supervenience thesis of perceptual content upon phenomenological content. Phenomenological character, however, may still be strongly supervenient upon inter-worldly subjects' internal states. The conjunction of the latter thesis plus the above rejection allows one to characterize phenomenological character as having phenomenal but not representational properties. Moreover, the ascription to inter-worldly twins of different perceptual contents does not violate behavioral constraints. For it quite possible to imagine those twins as placed in not only environmentally, but also nomically, different worlds, so that although they are internally identical they are disposed to behave in different ways. 

Fodor, J.A., LePore, E. (1992) Holism. A Shopper Guide, Blackwell, Oxford. 

Holistic conceptions of content presuppose the following argument: a) a semantic property is anatomic, i.e. is such that if anything has it, then at least one other thing does; b) there is no principled analytic/synthetic distinction; hence, c) a semantic property is holistic. A suitable way of blocking this argument is to reject premise a). This allows Fodor-Lepore to discredit a notion of narrow content as inferential role, which definitely has a holistic nature. Such a notion is independently problematic, for it makes possible for one and the same thought to have a narrow and a broad content which are utterly irrelated. As it has been questioned, however, if we reformulate the master argument for holism as follows: a’) some of an expression’s inferential liaisons are relevant to fixing its meaning; b’) there is no principled distinction between the meaning-constitutive and the non-meaning-constitutive inferential liaisons; hence, c’) all of an expression’s inferential liaisons are relevant to fixing its meaning, does c’) really follows from b’)? 

Rey, G. (1992), "Semantic Externalism and Conceptual Competence", Proceedings of the Aristotelian Society 92, pp. 315-331. 

Any externalist semantics which disregards to account for conceptual competence is subject to the problem of how to rule out fortuitous lockings between Mentalese terms and external phenomena. In order for one's Mentalese term to have a certain concept or property as its broad content one needs 'conceptions' which one is disposed to holistically confirm by attempting to discover what in fact satisfies them. Conceptions lock a Mentalese term to a certain property by means of its containing a Kaplanian 'dthat' operator. Due to differences in one's whole cognitive system, one and the same conception may lock a term to different properties. On the other hand, different conceptions may lock a term to the same property. Insofar as the having of the afore-mentioned confirmation disposition locks a Mentalese term to a certain property, one may be said to be conceptually competent with respect to that term. 

Stoneham, T. (1992), "Comments on Davies: A General Dilemma?", Proceedings of the Aristotelian Society 92, pp. 225-231

By commenting Davies (1992), it is argued that his externalism is faced with a dilemma: either it must allow a certain kind of individualism about perceptual content, or deny that experiences with distinct perceptual contents necessarily have any phenomenal character which is preserved across duplicates. 

Dretske, F. (1993), "The Nature of Thought", Philosophical Studies 70, pp. 185-199.  

A defense of natural against social externalism is provided, by appealing to a info-teleological view of intentional content. Once this view is meshed with a distinction between triggering and structural causes of behavior, moreover, it is able to account for what would otherwise seem to require an internalist conception of content, namely the causal explanation of a thinker’s behavior. 

Recanati, F. (1993), Direct Reference, Blackwell, Oxford. 

The content of a thought is not exhausted by that thought’s truth-conditions, if this coincides with a Kaplanean-Russellean singular proposition. A narrow component is also required, made out of the de re modes of presentation of what the thought is about. Such modes are to be conceived as truth-conditionally irrelevant mental files. Any "x"-file, where "x" is the label for what the thought is about, consists of a list of "x"-beliefs. As such, the narrow component fits a non-solipsistic internalism, for the list’s beliefs are to be individuated not only holistically, but also relationally. 

Boghossian, P.A. (1994), "The Transparency of Mental Content", Philosophical Perspectives 8, pp. 33-50. 

Content conceived externistically violates epistemic transparency both of sameness and of difference, namely the theses according to which if two of a thinker's token thoughts possess the same/distinct contents, then the thinker must be able to know a priori that they do. This occurs not only with respect to de re, but also to de dicto thoughts. Insofar as the latter is the case, de dicto contents externistically individuated fail to be suitable for the purpose of psychological explanation and assessment of rationality. As a consequence, we are faced with a dilemma: either we give up the externalist conception of mental content or we have to revise our conception of rationality and the practice of psychological explanation it underwrites. 

Fodor, J.A. (1994), The Elm and the Expert, Cambridge MA. 

To be sure, typed in terms of broad content thoughts do not supervene on computational processes. However, the two can be reliably co-istantiated; cases in which no such co-instantiation occur are taken to be unreliable. These are most cases which prima facie force one to defend a notion of narrow content in order to defend a content-computation supervenience, i.e. Twin-cases and Frege-cases. As to the former cases, they are either nomologically impossible or accidental, i.e. irrelevant for the purposes of scientific generalizations. As to the latter cases, they are either marginal or such that the computational difference they prompt in one’s thoughts is compatible with the broad content identity of these thoughts. A further narrow-content inducing case, the Putnam (1975) one of the Experts, is also ruled out by Fodor: by deferring to the Experts we really distinguish between phenomenologically indistinct substances and therefore assign the corresponding Mentalese terms for these substances different broad content. Broad content is still treated in purely informational terms along the lines inaugurated by Fodor (1987); such a treatment is applied to Mentalese proper names. 

Glock, H., Preston, J. (1995), "Externalism and First-Person Authority",The Monist 78, pp. 515-533. 

Wittgenstein’s conception of intentional states is neither internalist nor externalist. It is not internalist, for Wittgenstein urges that what goes on inside the mind/brain is neither a necessary nor a sufficient condition for an intentional states. It is not externalist either, for although it recognizes a certain role to context in the individuation of intentional states, this is not the externalist’s role of determining a state’s content, but rather that of ruling out certain awowals or third-person ascriptions of an intentional state as unintelligible. Moreover, unlike externalism, such contextual circumstances are what a state’s subject is aware of. In actual fact, the main vice of externalism is that of making content undiaphanous to a state’s subject. This vice is unamendable in the way Burge (1988) attempts to do. For it is unacceptable to say that one can be certain that one believes that p although one may be uncertain that it is p that one believes. The non transparency of content is rejected by Wittgenstein insofar as he i) links meaning with understanding ii) sees awowals and explanations e.g. of what one believes as the main criteria for third-person ascription of a belief state. These are only defeasible criteria; for Wittgenstein discards any full-fledged dispositional account of intentional states. 
Aydede, M. (1997), "Has Fodor Really Changed His Mind on Narrow Content?", Mind and Language 12, pp. 422-458.  avalaible online

A (too?) extremely detailed critique of Fodor (1994). Against Fodor's treatment of the cases which prima facie induce one to embrace narrow content: Frege-, Experts-, and Twin-cases. As to Frege-cases, it is acutely shown that Fodor's 'mapping function'- notion of narrow content (cf. e.g. Fodor (1987)) never contemplated them; hence, they cannot be used to dismiss such a notion. As to the Expert-cases, Mentalese terms whose distinct referents are indistinguishable for a non-expert subject still have different narrow content in terms of mapping functions. As to the Twin-cases, it is preliminarily said that Fodor's pure-informational semantics must properly attribute Twin terms one and the same disjunctive broad content. Insofar as this is the case, narrow content would become a constant function from contexts to broad content; hence, narrow and broad content collapse. Finally, insofar as pure-informational semantics assigns meaning to Mentalese expressions in dispositional terms, it gives an internalist rather than an externalist account of broad/narrow content. 

Brook, A., Stanton, R.J. (1997), "Fodor's New Theory of Content and Computation" avalaible online, Mind and Language 12, pp. 459-474. 

Fodor's new (1994) conviction that infointentionalism and computationalism are compatible insofar as computational and intentional states reliably co-istantiated is critically evaluated. In spite of Frege-, Putnam- and Quine-cases, where no such a co-instantation occurs, Fodor does not provide us with the mechanisms he promises in order to guarantee the reliability of this covariation. Moreover, both Frege- and Putnam-cases do not square with the dispositional account of broad content his pure informational theory of meaning sustains. In the former cases, it's hard to assign a single disposition for the co-referential terms involved; in the latter cases, a single disposition is what one finds instead of the different two the theory would predict, given the non-coreferentiality of the terms involved. Fodor's underlying motivation for holding the above conviction seems to be that either we are able to show that content is natural (as the claim of compatibility between infointentionalism and computationalism maintains) or we must give it up. But we are not forced to such an alternative. 

Jacob, P., What Minds Can Do, Cambridge University Press, Cambridge 1997.

A mixed teleological-informational view of mental content is argued for, according to which the content of a representational state is the state of affairs that state nomically covaries with, provided that the state precisely has the proper function of indicating such a state of affairs (that is, such a function is what enabled the ancestors of that state’s possessor to survive and proliferate). Jacob’s ambition is that this intermediate view better accounts than its competitors in the naturalistic camp for the problem of misrepresentation. As far as the internalism/externalism debate is concerned, insofar as it acknowledges that a state’s content does not supervene on the state’s physical basic, it has many similarities with an externalistic account of content, with however some caveats. First, it allows for a many-layered view of content. Second, as far as propositional attitudes having the same truth-conditions are concerned, they don’t have an identical content (notably, a broad one), for the logical potentials of their respective informationally equivalent carrying-information structures differ.  

Marconi, D. (1997), Lexical Competence, MIT Press, Cambridge MA. 

Marconi presents a dual theory of lexical competence, according to which knowing the meaning of a word is being able both to draw inferences from sentences containing that word to other sentences and to make both the world fit the word and vice versa: in other terms, to be both inferentially and referentially competent. Inferential and referential competence are separate skills, for one may have one without the other. This distinction is confirmed by neurological data. Although both inferential and referential competence are to be conceived internistically, abnormality makes us withdraw an attribution of competence. This means that competence may also be conceived communitarianly, i.e. as regarding the correct use of a term, where correctness is taken as convergence with the linguistic use of speakers inasmuch as they are regarded as competent. The individualistic and the communitarian picture of competence are compatible (we might take Marconi’s position as a whole as outlining a weak form of social-externalism regarding competence), for the latter enters the stage only insofar as the issue of correctness enters too. This picture of competence is however incompatible with some theories of meaning and reference, notably the causal-externalistic ones. Putnam cases (at most) show that not only competence, but also reference is social-externalistically oriented. 

Segal, G. (1997), "Content and Computation: Chasing the Arrows", Mind and Language 12 (1997), pp. 490-501. 

Critical notice of Fodor (1994). Against Fodor's treatment of Putnam-cases, it is shown that there are cases in which Twins' concepts have different contents even though the failure to distinguish between the substances the Twins refer to is not accidental. Frege-cases, which are such that different syntactic realizations of one and the same content respectively cause further different synctatic implementations which can no longer be the realizations of one and the same content, are moreover shown to be only a small aspect of a more general problem, i.e. how can there be a coordination between the information-carrying properties of a Mentalese expression, i.e. those which involve backward-looking causal sensitivities, and the properties of this expression that determine its forward looking causal role? 

Voltolini, A. (1997), “Is Narrow Content The Same As Content of Mental State Types Opaquely Taxonomized?", in Meggle, G., Analuomen 2. Proceedings of the 2nd Conference “Perspectives in Analytical Philosophy”, volume III: Philosophy of Mind. Practical Philosophy. Miscellanea, de Gruyter, Berlin, pp. 179-185. 

Fodor’s conviction that de dicto content (i.e. the content of mental state types opaquely taxonomized) and narrow content (the function from contexts of thought onto broad contents modeled upon Kaplan’s notion of character) are identical is criticized, by showing that de dicto content cannot have a character-like nature. One needs therefore to appeal to two distinct notions of non-broad content: de dicto content, which helps to explain intentional behavior, and narrow content, which accounts for linguistic competence. 

Top of page