SORITES, ISSN 1135-1349

Issue #08. June 1997. Pp. 33-60.

Framework of an Intersubjetivist Theory of Meaning

Copyright © by SORITES and Cristina Corredor

Framework of an Intersubjetivist Theory of Meaning

Cristina Corredor


Very recently, some internal criticisms have been directed towards core theses of the Theory of communicative action and some posterior modifications of them.<23>Foot note 3_1 This theory aims at reconstructing the unavoidable and most general presuppositions, of a formal-pragmatic character, which underlie and are constitutive for our processes of human understanding (Einverständnissprozessen). One of the core theses in the theory affirms that the particular interaction form represented by the communicative use of language is the original modus of use of language; furthermore, this notion becomes the most basic one for a general theory of human rationality -- understood as communicative rationality. This latter is considered to be a universal competence allowing to take part in processes of human understanding. These are claimed to share, from the point of view of their formal-pragmatic structure, a common triple reference to three validity claims -- truth, rightness and sincerity, resp. correlated to the assertive, communicative and expresive uses of language -- to the effect that these validity claims establish an internal connection between meaning and validity redeemable in rational argumentation, and entail a correlative triple reference to three «worlds» or dimensions of reality: namely, an objective world (of facts or states of affairs), an intersubjective world (of valid norms and values), and a subjective world.

In fact the TkH entails a wide theory of rationality in which -- following its most recent formulation -- three types (or roots) of rationality are to be distinguished, the three of them being nevertheless connected to each other through the discoursive rationality embodied by the proccess of argumentation in the sense stated above. These three types, internally related to knowledge, action and speech, are: epistemic, strategic-teleological, and communicative rationality. The present discussion deals in particular wih epistemic rationality and its internal connection to knowledge. Within the framework of the TkH, an epistemic conception of truth as rational acceptability is assumed, according to which «Wissen (...) unterstellt die Möglichkeit einer diskursiven Einlösung entsprechender Wahrheitsansprüche».<24>Foot note 3_2 This means that we can be said to know facts, provided we know why the corresponding sentential judgements are true. Thus the following claim can be seen as a definition of the notion of epistemic rationality: «Um eine Meinung as [epistemically, C.C.] rational zu qualifizieren, genügt es, daß sie im gegebenen Rechtfertigungskontext aus guten Gründen für wahr gehalten, d.h. rational akzeptiert werden kann».<25>Foot note 3_3 Yet rationality in a judgement does not presuppose its truth, but only its justified acceptability in a given context. In order to keep track of the fundamental distinction between been true and been accepted-as-true, a reflexive ascent is required. This is assumed to be possible, due to the double propositional structure (Austin, Searle) of epistemic claims within scientific theories -- or in the epistemic use of language in general-. Nevertheless, the necessity and lack of a further detailed logical-semantic analyse of this structure is explicitly recognised.<26>Foot note 3_4

The criticism referred to above<27>Foot note 3_5 aims at showing that the internal logic in the TkH, together with an insufficient account of the epistemic use of language, have finished in a form of meaning holism and a subsequent relativism that threat the universalism intended by the theory. This relativism is claimed to depend on the (maybe non-explicit, but acritical) acceptance of two thesis: firstly, the preeminence of meaning upon reference, and secondly a form of meaning holism. Under these premisses, neither a reflexive revision of the Welterschliessung (world disclosure) we belong with, nor a critical distinction between our meaning-knowledge and our world-knowledge are to be possible. This conclusion becomes unavoidable, since the claimed internal connection between meaning and intersubjectively shared validity does not allow one to isolate the acceptability conditions of speech acts from the background knowledge determining them and making them possible. So long as any intelligible use of language, jointly with the corresponding validity claims and standards of rationality, are stated to come from the pre-reflexive, linguistically structured lifeworld of intersubjectively shared meanings and practices, it can only be claimed for the former the same contingent and historical character possessed by the latter -- the very instance constitutive for them. Even a discoursive use of language is condemmed to such relativism, given the fact that the standards of rationality and validity criteria to be revised only become reflexively accessible by means of their own application.<28>Foot note 3_6 The conclusion to be drawn seems to be that of questioning the universality claimed by the TkH and which constitutes an unrenounceable premisse for the Diskursethik.

In order to overcome these difficulties, however, the criticism stated above is accompanied by an explicit proposal. The suggested solution finds theoretical support in semantical theories of direct reference (Kripke, McDowell, previous works by Putnam) and, in Putnam's internal realism. The theories firstly mentioned emphasize the function of rigid designation fulfilled by certain terms and complex expressions, which are emploied with a referential value in inductive epistemic contexts and learning processes, and such that it is not possible to associate to them, as their meaning, a complete linguistic description of all the properties, relations, etc that will be virtually predicated of the entities they name. Although in the TkH this epistemic use of singular terms is not ignored, the communicative use of language (especially the discoursive one) does always suppose a previous understanding of sense that makes impossible, according to the comented criticism, to reflexively question and put under control the received Weltanschauung structures that constitute their own conditions of possibility. To this extent, a meaning theory of direct reference, re-elaborated within a form of internal (pagmatical) realism (Putnam) and the presupposition of an objective world entailed by it, appears to be -- so the referred proposal assumes -- the unique theoretical solution in order to avoid meaning holism and the subsequent relativism.<29>Foot note 3_7

An important theoretical contribution implicit in this proposal seems to be the idea, already present in the TkH,<30>Foot note 3_8 that it is the breakage points or «breaches» between language (our theories or constituted meanings), on the one hand, and the objective world (supposed to be independent of the theory), on the other, what makes of these problematic situations the very constitutive instances for a new Welterschliessung. To the extent that these breakage points, or problematic contexts, as they become manifest in the course of epistemic practices (paradigmatically, in learning processes), are compelling for a critical revision of our background knowledge, they cannot be referred back to the schon immer of previously established senses arising from the lifeworld. For the problematic situation promptes the counterfactual, thus normative presupposition of a unique objective world. Moreover, the revision it motivates is underlied by a fallibilistic intuition concerning our knowledge and by a competence allowing us to distinguish between our knowledge of already constituted meanings and our knowledge of the world, the latter only counterfactually anticipated.

Yet there are in this proposal some other elements which I find more difficult to agree with. In particular, the stated criticism seems to adscribe to the TkH an acritical embodiment of a theory of indirect reference. This would be brought about by an explicit appeal to formal semantics, considered as the theoretical device adequate to analyse the structure of the propositional content element in speech acts and the referential function of langage in epistemic contexts. A meaning theory of indirect reference and an attributive interpretation of designating expressions are seen as necessarily correlative to this Fregean formal semantics.

Nevertheless, to this argument it can be objected that the pregnancy of formal semantics is undeniable, given the fact that the most recent linguistic theories since Chomsky<31>Foot note 3_9 have made use of these formal-semantical developments in order to construct models intended to be adequate counterparts for their sintactical and grammatical accounts of natural language. These structural descriptions of linguistic expressions should be seen as strictly instrumental and just a useful device to make it explicit structural-semantical relations among expressions. A standard formal semantics is interpreted as related to a referential theory of meaning; the definition of a semantical model is carried out by introducing a set of «entities» (or semantical domain) and a set of properties and relations defined onto it; finally, the «meaning» of linguistic expressions is seen as a relation between the linguistic symbols and the set of entities, this latter being considered independent of the language. (No matter whether the logical proper names or constant symbols are interpreted as naming Wittgenstein's objects, Carnap's Erlebnisse or Putnam's stereotypes).

The formal semantics developed from the seminal works of Frege, Tarski and Montague (among others) does not convey as the theory of direct reference does, any «ontological commitment» in itself. Though it is undeniable that Frege, Russell or Carnap argued for a theory of indirect reference, they did it in a very different context. They were in search for a vollkommene Sprache, in which all names played the role of genuine proper names -- what amounted to saying that the minimal linguistic categories should name the corresponding minimal categories of reality, the minimal expressions for predicates should name the most basic properties and relations, and so on. But the formal language in itself, the semantics, was just to be seen as an instrumental device, able to consistently and adequately model some theory. The semantical domain (or universe of discourse) and the set of relations and fuctions defined on it were not in themselves «ontological»; only when this instrumental device was applied in the pragmatic level as the formal correlate of a theory, could the meta-linguistic expressions be seen as intented to convey information on the «ontology» or «structure» of reality.

The fact that this semantical structure was read by its own authors as a «real» one has been denounced as an «ontological contamination of semantics».<32>Foot note 3_10 What is of importance here is that formal semantics is just a useful device to describe structural relations among linguistic expressions, and its embodiment within the framework of empirical and theorical linguistics validates them. Only a «realist» reading (remember that a semantical model is always defined «a posteriori» and fixes just the available knowledge on the matter) can assign to formal languages a different character, not intended by its creators. To a certain extent, the direct reference theorists have tried to exaggerate the consequences of such a realist reading, in order to emphasize what their own reconstruction is able to provide: i.e. the fact that, in practical contexts and for certain referring expressions, we cannot assume the interpretation function to be substantively given -- this precisely because it is in course of elaboration. From this point of view, it seems essential to keep in mind once more that we are dealing with actual contexts of praxis, where some conflict opposes our previously established knowledge. If it is an undeniable fact that formal semantics is usually related to referential theories of meaning (whether direct or indirect), neither a realist reading of the model nor meaning holism should be seen as logical consequences of it; on the contrary, these count as previous theoretical assumptions guiding the introduction of a particular analysis and the preference for a definite basic logic.

These considerations send us back to the communicative (discoursive) domain in which our own criteria of validity and meta-theoretical principles are reflexively accesible and critically revised. A refusal of such a possibility, the adscription of logical or theorical priority to a particular use of language (the epistemic one) or to a particular semantical theory (the direct reference theory) is subject to an almost unavoidable risk: that of carrying out an objectivist or functionalist explanation, this fact in spite of all efforts to keep track of the normative dimension.


The assert that the TkH acritically embodies an indirect reference theory of meaning (because of its explicit acceptance of formal-semantical analyses, its implicit substaining of the preeminence-of-meaning-upon-reference thesis and a subsequent meaning holism) finds legitime support in some of Prof. J. Habermas' writings -- and in particular in the 1st volume of TkH-, as C. Lafont's critical reconstruction has clearly shown. Nevertheless, there are also in the TkH asserts and analyses that seem to make it explicit the same kind of intuition motivating the criticism. Thus in the 2nd volume of this work, and in the course of an interpretive study on G.H. Mead, an explicit treatment of the identity-of-meaning problem is carried out -- i.e. the problem of the intersubjective constitution of shared meanings. Here a re-elaboration of Wittgenstein's last meaning theory allows Prof. Habermas to formulate what is intended to be an intersubjectivist theory of meaning as use, as opposed to any referentialist meaning theory in general -- no matter whether «direct» or «indirect». For, as it is explained, «Die bedeutungskonstante Verwendung desselben Symbols muß nicht nur an sich gegeben, sondern für die Symbolbenutzer selbst erkennbar sein. Und diese Identität der Bedeutung kann nur durch die intersubjektive Geltung einer Regel, die die Bedeutung eines Zeichens «konventionell» festlegt, gesichert werden.»<33>Foot note 3_11

There are two ideas to notice here. The first one is that meaning identity or the intersubjective sharing of meanings do not possess the character of that schon immer given. It depends on the validity of rules which in a practical context fix or warrant the use of the term. In the case of a referential usage in learning processes, for instance, neither should a preeminence of meaning upon reference be assumed, nor the inverse. For what counts as reference and what counts as meaning are both dependent on the rules that are constitutive of the epistemic praxis. Here a conceptual distinction is assumed between what is constitutive of meaning (or sense) and to what justifies its validity -- to put it in Prof. Apel's terms.

The second noticeable idea in the quotation is that meaning identity, as given through the validity of a rule, is not credited by the empirical fact of a continuity in its use, it is not something merely «given»; meaning identity requieres the possibility of a reflexive access to it. The reason for that lies on the fact that these rules are counterfactual, that is to say, they have a normative character. This allows for mutual instruction and reciprocal criticism: «Indem sie die kritische Stellungsnahme des anderen zur fehlschlagenden Deutung eines kommunikatives Aktes sich selbst gegenüber einnehmen, bilden sich Regeln der Symbolverwendung aus (...) Auf diese Weise bilden sich Bedeutungskonventionen, und bedeutungsidentisch verwendbare Symbole aus.»<34>Foot note 3_12 Here the choice is made for a theory of meaning as use, as different from referential (direct or indirect) theories of meaning. Notwithstanding this, from the point of view of the discussed criticism this moving is not enough to avoid relativism. For it does not grant that a reached accord on the validity of a rule does not arise from standards schon immer given within the lifeworld background.

It is in fact difficult to find in the TkH an explicit discussion of this problem. The arguments against relativism seem to focuse on another question, undoubtedly a core one: namely, the way in which «die formalpragmatische Analyse, die an hoch idealisierten, vereinzelten und elementaren Sprechhandlungen ansetzt», can be applied to factual, empirically given «kommunikativ strukturierter Lebensformen».<35>Foot note 3_13 I would say that, in an implicit way, the idea stated here is as follows. The risk of falling back to relativism lies on the step leading the way from intersubjectively identical meanings (as counterfactual presuppositions) to their effective performance (Einlösung) in a particular lifeworld context. Here Wittgenstein's notion of eine Regel folgen is appplied to show how this is possible. For constant or «literal» meanings, and invariant semantical traits can be reconstructed as bearing a normative character: they constitute inevitable presuppositions for the participants in any linguistically mediated interaction.

Yet to «take as granted» -- as it seems to be the case in the TkH -- the intersubjective validity of shared meanings in this formal-pragmatic level of analysis, in which all conditions of possibility (the inevitable and maximally general presuppositions) are integrated, is not enough to grant universalism in the level of ordinaire communicative practices. For in this case «handelt es sich erstens um das grundsatzliche Problem, wie sie die situative Bedeutung eines Sprechakts zur wörtlichen Bedeutung seiner Satzkomponenten verhält.»<36>Foot note 3_14 Therefore the idealization of a literal meaning seems to overcome contextuality in Einverständnis processes, so long as this very presupposition counts as such for all participants and is reflexively accessible whenever a problem interrupts the process, so that the performances redeeming literal, warrantly shared meanings do not «open the door» to insuperable forms of contextualism or relativism. It is evident that the claim for universalism rests on the possibility to reach an agreement on this ideally presupposed meaning, when this is necessary for the subsequent interaction.<37>Foot note 3_15 Thus the problem was seen as a de iure problem, namely that of identifiying what makes a justification of the validity of factual, constituted meanings possible -- and that previously to, and with independence of, the particular contents to which this validity should be accorded.

It may be in this sense that in the TkH it was argued: «Diese Relativierung der Bedeutung sprachlich standardisierter Ausdrücke fuhrt freilich nicht zur kontextualistischen Auflösung semantischer Invarianzen, also zu einem konsequenten Bedeutungsrelativismus; denn die particularen Lebensformen weisen nicht nur Familienähnlichkeiten auf: in ihnen kehren vielmehr die allgemeinen Infrastrukturen von Lebenswelten überhaupt wieder.»<38>Foot note 3_16

The assert that all lifeworlds share a common lifeworld structure (the formal-pragmatic structure of communicative rationality) represents, as I see it, a «strong» thesis central to the TkH. Here lies as well the plausibility of its claim for universalism. Yet what the discussed criticism aims at showing from a philosophical-linguistic perspective is that meaning relativism does not only emerge in the transition from intersubjectively valid, identical meanings, to their effective realization in situative lifeworld contexts, in which breakages of human understanding take place. From this critical prespective, it is the very moment of constitution of intersubjectively valid meanings what is in need of further elaboration. The intuition and subsequent commitment to internal realism could be seen as an effort to answer this insufficient development in the restricted domain of the epistemic use of language. Nevertheless, in some respects this proposal seems to be in conflict with basic TkH premisses. Here, the fundamental remark is made that a «wahrgenommene Konstanz der Bedeutung» (or of the reference in epistemic contexts) is not enough; this constancy (or rigidity) can be seen as purely functional and identifiable just from an observer's perspective. In order to legitimately speak of shared meanings, a further step is needed: that leading the way to the constitution of valid rules for reference-fixing and herewith to the interactive sphere of meaning constitution.

In this point, however, it can be anticipated that an answer cannot come so much from the sphere of meaning constitution as from the normative sphere of validity justitication. Here it is where these formal-pragmatic, unavoidable and general rules enter into play, including that of referring to (or anticipating) a common objective world. But for that it is necessary that some problematic situation make it compelling a questioning of the previous accepted validity and a searching for a new justification of it.


The burden of the proof is now for the TkH to show that shared meaning, as based in the intersubjetive validity of a rule, can escape the relativism of being immersed in a unassailable linguistic Welterschliessung. The commented critical proposal subsumes a double betting. Firstly, it posits a direct reference theory (as theoretical reconstruction of semantical structures of language) as a possibility to overcome the «totality» represented by a particular Welterschliessung and its hipostatization. It order for that to be possible, it is necessary to identify some instance, of a normative or regulative character, not pre-determined by this linguistic world disclosure. What is required is that this instance, counterfactually operating as an inevitable presupposition in epistemic contexts, should encourage, make possible and regulate such a linguistic disclosure. This instance is identified as the counterfactual presupposition of a common objective world -- which presents itself in epistemic contexts as previous and inevitable with respect to any symbolic mediation at disposal.

Henceforth the second betting subsumes an epistemic commitment with internal realism. If this position is assumed as valid, then -- although this conclusion is not drawn as such by the discused criticism -- the epistemic use of language, which allows for an explicit treatment of truth claims implicit in the propositional component of speech acts and which is seen as based in a direct reference theory, inevitably becomes the anchoring point for a formal-pragmatic reconstruction of (maximally general and unavoidable) structures of speech aiming at overcoming the risk of meaning holism -- i.e. the risk of not being able to distinguish our world-knowledge from our language-knowledge. This seems to highlight the insufficiency of a Wittgensteinian account of meaning publicity.

Yet I thik -- as advanded before -- that something essential could turn out to be lost in this double betting. For in epistemic contexts it is not only the presupposition of a common objective world -- not yet linguistically open -- what enters into play. It is required as well the implicit assumption that the realization of such a linguistic disclosure -- towards which the epistemic practice is directed -- can take place. This amounts to saying that the objetive world is accessible to liguistic disclosure, thus that the structure of reality is, toghether with the structure of language, rational. Or, equivalently, that the objective world can be counterfactually interpreted as articulated in a way similar to the semantical structure of language, i.e. the referential, predicative, attributive language structures. But now, even if the intuition differentiating world and language is preserved, it is so only by means of this structure -- at once empirical and quasi-transcendental, for it prefigures the formal structure of a not-yet-disclosed world.

What the previous remarks attempt to show is the risk subsumed by the commented proposal and which it shares -- paradoxically enough -- with all referentialist theories of meaning. Namely, the risk of performing a realist reading that eventually absolutizes what is only a semantical fuction of language. If a direct reference theory is seen just as a descriptive reconstruction of the epistemic use of language in learning processes, then it is undoubtedly accurate and pregnant. So restricted, the theory can legitimately dispense with two necessary complementations: firstly, a global reconstruction of all sintactical and semantical structures of language to which reference, attribution and predication belong, and secondly, a theory of knowledge accounting for the relationship between these structural elements and what counts as empirical evidence and/or fenomenological experience.

I think that it was this second, methodological perspective, and not the first, «absolutizing» one, what H. Putnam adopted in his initial formulation. But if he was far away from metaphisically absolutizing the referential function of language, his proposal could not help but turning his reconstruction of learning processes into a functionalist explanation.<39>Foot note 3_17 Among the core facts that convey the use of a word and help to fix reference are «pieces of empirical information», «purely linguistic information» and sometimes the extension of the word. What seems clear is that this assignment of semantical value, even for rigid designators, does not escape a linguistic mediation and a dependence of background knowledge and of a set of constituted practices. For the use of a referring expression is explained in its turn in terms of the function it fulfils -- possibly from an observer's perspective -- within a more comprehensive epistemic framework and thus in dependence of other elements within it.

Certainly, the critical proposal here discussed has a different character. A reconstruction of the (direct) referential function in purely semantic terms is here intended to count as a reconstruction of the inevitable rational presuppositions in epistemic contexts. However, in such a case it should not be enough to show that the result of such a reconstruction «does the work». The fact that this fixation of a direct reference fulfils or satisfies a function in our learning processes or inductive practices (or better: to be able to explain the latter in terms of the former) does not confer a rational character to these epistemic practices -- if rational is to preserve its philosophical relevance here.

In order to clarify concepts, it could be of use here to make it explicit what is to be understood under «functionalist explanation»:<40>Foot note 3_18

«An institution or behavioral pattern X is explained by its function Y for a group Z if, and only if,

1) Y is an effect of X

2) Y is beneficial for Z

3) Y is unintended by the actors producing X

4) Y or at least the causal relation between X and Y, is unrecognized by the actors in Z

5) Y maintains X by a causal feedback loop passing through Z.»

What makes a functional explanation different from the type of quasi-transcendental deduction brought into play is, firstly, that the kind of presupposition identified by the latter cannot be said to «fulfil» a function Y, but to constitute a necessary (inevitable) condition for this epistemic practice -- there are, so to say, no «alternative models». Furthermore, they should be previous to, and independent from, the particular realizations to which epistemic rationality is accorded. Secondly, the relation between this practice and the presupposition making it possible should be accessible to the participants -- if this linkage is not to be condemmed to the pre-rational status of what is given in inherited practices (points 3. and 4. in the definition).

If this is right, then the discussed proposal is to be seen as involving two «strong» assumptions. On one side, a notion of rationality which should be different from mere operative or teleological «blind» rationality and able to acquire reflexive knowledge of its own conditions. On the other side, a universality claim which is not to be restricted to the contextual character pregnant in the quoted definition.


I think this latter idea is what the commented proposal attempts to show and therefore what puts it to the proof. In a recent paper,<41>Foot note 3_19 Axel Müller accomplishes a reconstruction, from a formal-pragmatic viewpoint, of the structural properties present in those epistemic contexts in which we talk of «learning by experience». After the linguistic turn in philosophy, this investigation can only be philosophical-linguistic. The author shows that there is a connection between certain predicates applied in inductive practices and singular terms for natural kinds when applied as rigid designators. This connection is claimed to lie on the rational presuppositions underlying the epistemic output of such expressions. These rational presuppositions would have in themselves an aprioric character,<42>Foot note 3_20 since they regulate normatively the correct usage of these singular and predicative terms in their empirical application. It is by means of presupposing a fixed reference in the case of singular terms, or the projectability of inductive predicates onto the domain of discourse, that this rationality underlying the epistemic use of language becomes effective. Thus both practices acquire a quasi-constitutive status when learning by experience, to the extent that they are underlied by two inevitable, epistemically rational presuppositions. A first, quasi-ontological one, is to the effect that the «entities» correlated to the concepts (a natural kind, or the class of all entities satisfaying an inductive predicate) are legitimately supposed to be «real», no matter what method of identification has been emploied. A second, meta-epistemic presupposition allows for a permanence of the objectual domain, notwithstanding the fact that a continuous change in the knowledge assocciated with the concept is to take place.<43>Foot note 3_21 Finally, it is important to notice that these presuppositions of epistemic rationality are to be reconstructed as a universal competence for following rules -- precisely, the kind of rules that make the epistemic output of these concepts possible. Moreover, this competence permits to distinguish «zwischen Zeichen und Bezeichnetem, zwischen Wirklichkeit und Konstruktion und zwischen Bezugnahme und Mitteilung».<44>Foot note 3_22

I think that this reconstruction of the pragmatic structures basic for the epistemic rationality represents precisely the kind of necessary complementation which a general Wittgensteinian account lacks, as stressed above; yet it is indispensable for the commented proposal to overcome the two mentioned risks. But the rigour of the analysis makes some difficulties in it manifest. The intuition I would like to make good here is that a reconstruction of learning processes in terms of a direct reference theory is not enough by itself to avoid meaning holism, i.e. to identify the linguistic game allowing for a distinction between meaning knowledge and world knowledge.

This in its turn amounts to questioning the sufficiency of the two presuppositions identified in A. Müller's analysis as constitutive for epistemic rationality. The expression sufficiency is not intended to put into question their mostly general and necessary character, but to suggest that, in order for the two presuppositions to effectively acquire such character, it is not enough to place them within the domain of epistemic rationality in its application to learning by experience in general. This move suffers from a certain ambiguity, since these processes, so far as they take place in a non-problematic way and following the usual patterns of scientific practices are not necessarily guided by counterfactual presuppositions but rather by the factual criteria at work in «normal science» periods (to speak along with Kuhn). At this stage, and as long as no reflexive or critical revision of these criteria is needed, a reconstruction in terms of underlying presuppositions of rationality is not different from a «reconstruction» (explanation) in functionalist terms. Only when a «breach» or a conflict arises, is there also a need for the rational, reflexive competence which bestows to these presuppositions their character. Yet this «screw turn» appeals for a discursive use of language and no longer for a mere «blind» epistemic rationality.

I would say that, to some extent, it is this moment of conflict which Follesdall's comment hints at, although not intendedly -- in fact, with an opposite intention-, when he says: «All our talk (...) presupposes that we can keep our singular terms referring to the same objects. To the extent that we fail, these notions [change, causation, knowledge] become incoherent.»<45>Foot note 3_23 Only in such moments does it become possible to reflexively identify the presuppositions of rationality «brought into play». Consequently, only then is it possible to adscribe to epistemic rationality the import of a universal competence, able to separate the referential function of language from the presupposition -- now of a non-ambiguous normative character -- of a common objective world. Insofar as no conflict interrupts the continuum of epistemic practices in normal periods, nothing transforms reference-fixing practices and inductive reasoning into an «anchoring point» for facing meaning holism. Contrary to it, and so long as these practices take place non-problematically within the framework of previously established practices, they do not escape either a falling back into the pre-instituted Welterschliessung.

The fact that this is so, i.e. that those presuppositions of rationality only acquire their character in the context of a reflexive revision forced by a «breach» or conflict, becomes manifest in the folowing sense. If these epistemic breakages are considered to be constitutive for the immediate («blind») epistemic rationality, their output cannot be separated from particular contents and from a contextual dependence on the epistemic asserts they contribute to establish -- in the same sense as the corresponding reconstruction of the process cannot be kept appart from a functionalist explanation. I think that this becomes apparent in A. Müller's paper, and I will try to hint at some relevant points in his analysis in order to support this claim.

The first difficulty is almost terminological. For a post-metaphysical notion of experience is only to be understood as a result of linguistic mediations and as an abstraction -- since to talk about «our» experience already presupposes intersubjectivity and linguisticity in its transmition and interpretation.<46>Foot note 3_24 This problem is only an aspect of a more basic difficulty, as explained by the author himself: «Damit ergibt sich (...) die Notwendigkeit, den Sprachverwendern die Einsicht in die Möglichkeit der «direkten», aber nicht unmittelbaren Bezeichnung von etwas von Bestimmungswissen unabhängigen mittels eines Zeichens zuzumuten.»<47>Foot note 3_25

It seems to me that the logical grammar of singular terms and inductive predicates is here accurately analysed, but it does not answer the question of how the application and correct use of these terms are communicated and acquire intersubjetive validity, and in what way a particular (provisoire) fixation of reference turns out to be publically accesible and the correlated concept reaches intersubjectivity. If it is true that this reference-fixing is said to be accomplished through a variety of procedures (e.g. operational, contextual, ostensible, theoretical ones), nevertheless -- as it happened in Putnam's first formulation -- all these procedures are instances of established practices, linguistically mediated and context relative. The possibility of a reference-fixing is underlied by a presupposition not mentioned, namely the possibility of its public transmision and of some form of consensus with respect to it -- herewith, the initial question on the constitution of intersubjective validity reappears once more. To the extent that introducing a rigid designator -- and its associated concept -- is guided by internal criteria, nothing in it guarantees an access to the difference between world knowledge and meaning knowledge. What to a «naive» intuition appears to be a new entity in the world, not reducible to pre-existing meanings, may turn out to be a result of our background knowledge and practices, methodological criteria, etc., to the effect that all these things together produce «new» objects of experience.<48>Foot note 3_26

The fact that the crucial distinction between natural kind terms and general terms, as correlative to the opposition between a referential and an attributive use of language, do not seem to rest on a wholly rational criterion -- as the author himself notes-, makes this determination of reference to depend again on internal criteria coming from the concerned empirical theories and henceforth language-relative. (From this perspective, to distinguish between the assignment of content to a symbol and the structural function it fulfils within the epistemic practice is not enough, as previously remarked, since a functionalist explanation does not reconstruct the formal-pragmatic structures of the epistemic rationality).

This difficulty -- and the subsequent fact that distinguishing between a referential and an attributive use of language does not depend on formal traits of linguistic expressions, but on their usage in the particular context in which they play such a role -- is present again in the distinction between inductive (or projectable) and non-inductive predicates. Here, a risk of circularity -- projectable predicates are identified precisely as those which «do the work» in learning processes -- is only to be overcome if this identification is not formulated in formal terms -- since in this way, no acces to them exist -- but by taking into account the particular context within which a predicate turns out to be inductive. In this sense, it seems correct the decision in favor of a precise notion of induction<49>Foot note 3_27. For this notion supposes an empirical or experimental context where a hypothesis has been proposed, before the inductive process oriented to its cheking and the assignment of a projectable character to the involved predicate can take place.

That a hypothesis has already been proposed means that some regular and (assumed as) complete event has been observed whose repetition is expected, and that some set of entities has been picked out as possible «universe of discourse» (therefore, as objectual set virtually satisfying the predicate), as well as some property or relation identified as that named by the predicate. In this respect, A. Müller critically discusses Goodman's paradoxe -- a counterexample proposed by this author in order to show that we tend to use as projectable those predicates whose output in previous inductive practices has conferred them this character. Müller refutes Goodman's thesis by showing in his turn that Goodman's paradoxe is brought about by an incorrect application of induction in the case of a non-inductive predicate.

Yet it seems to me that the paradoxe can be solved even adopting Goodman's perspective. For his sophisticated counterexample can be said to lack the contextual determination required, if it is seen as a case of «too quick» induction -- thus of «blind» epistemic rationality-, formulated before the complete sequence of events had taken place. In order to define his non-inductive predicate «grot», Goodman needs to adopt a «God's eye» perspective -- or an observer's a posteriori viewpoint. The previous remarks are intended to suggest that the attribution of an inductive (projectable) character to a predicate is epistemically reliable (therefore epistemically rational) only when one has successfully isolated the «correct» sequence of events. The burden of the proof for the validity of the induction -- thus for the new knowledge it introduces -- lies on the correctness in the synthetic moment of abduction, or formulation of the hypothesis. But this in its turn implies entering into the sphere of validity justification, hence of discursive-reflexive rationality.

What from the side of language counts as intersubjectively valid meaning includes, from the practical side, some procedure or criterion allowing to introduce such terms by a fixation of reference (in a communicable way) and/or by establishing (though provisionally) their application. So long as the implicit expectations do not become problematic -- that is to say, as long as «we can keep our singular terms referring to the same objects», or we can keep our inductive predicates projectable-, the kind of epistemic rationality «set to work» follows the pattern of that given in our background practices. Furthermore, and unless we resort to a notion of experience previous to the linguistic turn, what counts as such is also a result of mediations that bring nearer the experiential and the experimental. Here, it seems difficult to distinguish between the inevitable presuppositions of rationality, in a normative sense, and what counts as functional, operative devices from a descriptive point of view.

The situation changes as soon as the concerned presuppositions (fixity of reference, projectability of predicates) become problematic. Only then is there a need for the kind of competence that makes a revision of the implicit presuppositions possible, hence an access -- no longer unintentional or unconcious -- to the rules of use which normatively guide the application of empirically interpreted terms. This problematic moment -- that presumably will lead to a new hypothesis -- acquires a constitutive character for the new synthetic terms to be introduced. Moreover, it makes legitime attach to them the counterfactual, formal-pragmatic presupposition at stake: as A. Müller puts it, «die Fortsetzbarkeit and Festgelegtheit vorauszusetzen bedeutet nicht, etwas über die Welt gelernt zu haben, sondern die Differenzierung von Sprache und Welt reflexiv durchführen zu können.»<50>Foot note 3_28 I think that in this sense it becomes possible to speak, in relation to learning processes in general, of Prozessen des Verlernens.<51>Foot note 3_29

For these presuppositions have a double character. As far as they provide rules of use in the course of learning processes, within which the application of synthetic categories is non-problematic (since their reference/their unambiguous implementation is paradigmatically determined), these presuppositions belong to the competence that is to be attributed to the agents as epistemic rationality, so long as they are able to make a correct use of the correlative terms, empirically interpretable. These rules of use belong then to the background knowledge of the involved practices.<52>Foot note 3_30 Nevertheless, at this stage a conceptual distance remains open between the participants' pre-understanding of the rules, on the one hand, and the normative and rational character to be attributed to them, on the other. This becomes apparent whenever their paradigmatical application turns out to become problematic, because the same questioning seems to affect (or be extensible to) their rationality -- in the sense, e.g., in which the attribution of projectability to Goodman's «grot» happens to lose its «rational» character. Yet this problematic situation motivates a critical revision of the implicit presuppositions and a re-elaboration of the epistemic devices in question. And this is only possible through a form of reflexion that -- within the framework of a practice enjoying intersubjective validity -- is to take place uniquely as communicative rationality. Only in the context of a problematic situation -- i.e., whenever the kind of «blind» epistemic rationality accounted for by a functionalist explanation happens to fail -- do the rules of use corresponding to singular terms and projectable predicates acquire the validity of a rational presupposition. For only then are they to be reflexively recovered and a critical revision of the epistemic devices presently into play is made possibleand necessary. Only then too does the un-learning process correspond to a competent (i.e. rational) decision.

I think also that these reflexive language games, in which the problems arosen in learning processes are revised, can be seen as constitutive for our learning of the difference between language knowledge and world knowledge. Yet this assert leads the way to a more complex question; namely, that concerning the status to be assigned to this shared objective world that we counterfactually presuppose -- and to which we have access through reflexion in the referred problematic contexts. Finally, another difficulty arises in relation to the three basic concepts of knowledge, reality and truth, and the relationship among them. This question is dealt with in two very recent papers by C. Lafont.<53>Foot note 3_31


C. Lafont argues for a non-epistemic concept of truth and shows that the inconditional validity we attribute to it comes from its internal connection with the concept of reality, a non-epistemic one as well. So long as true/false are conceived of as exclusively dependent of the (absolute) opposition between it is the case/it is not the case, true can preserve its inconditional validity with respect to any epistemic criteria whatsoever of rational acceptability.<54>Foot note 3_32 The inconditional validity of truth rests, to this extent, on the logical condition expressed by the tertium non datur. By reconstructing this basic intuition, it is noticed that we are confronted with a formal aspect inherent to the concept of reality that is not exhausted by its epistemic counterpart, namely the absolute, totalizing character we assign to it. This character of the notion of reality becomes manifest in the inevitable presupposition related to the practices concerned with the revision of our believes: namely, the counterfactual presupposition of a unique, shared objective world.<55>Foot note 3_33

This perspective necessarily entails a very critical position with respect to epistemic views on truth, which characterize this notion as «rational acceptability under ideal conditions». Among those are Putnam's, Dummett's and the Diskurstheorie der Wahrheit defended by Prof. J. Habermas and critically revised by Prof. A. Wellmer. C. Lafont's criticism seems to be based on the observation that, in order for an epistemic notion of truth to preserve the inconditionality we intuitively associate to the concept, we are forced to suppose some kind of emphatic knowledge that enters into conflict with the fallibilism we assign to ours.<56>Foot note 3_34 This last fact implies, quoting Prof. A. Wellmer, that «es gegen das, was wir jetzt als wahr einsehen, auch in Zukunft keine triftigen Gegenargumente geben wird.»<57>Foot note 3_35 According to the discussed criticism, this view unavoidably entails the presupposition of a consensus on that which is acceptable on rational grounds, a consensus which has to be seen as definitive or irrevisable.<58>Foot note 3_36 This inconditionality, even if we understand it as counterfactual presupposition or strong idealization, cannot be brought into agreement with the fallibilistic reservation we maintain with respect to all factual knowledge and all factual rational consensus. If truth is to preserve its normative value in relation to our epistemic criteria, then it cannot be itself -- so the discussed critical proposal runs -- an epistemic concept, and this amounts to saying that it does not need embody the counterintuitive anticipation of the incorregibility of such criteria.<59>Foot note 3_37

Prof. J. Habermas has explicitly refused to assign a realist and non-epistemic import to the concept of truth. This refusal forces him, according to the commented criticism, to the presupposition -- inherent to any epistemic view on truth -- of a final true theory or a ultimate opinion that would be so metaphysical as incompatible with fallibilism. Certainly, if C. Lafont is right here, then this has also relativist consequences incompatible with the claimed universalism.

The criticism and its conclusions appear to be indisputable. Yet there are elements in the epistemic conception of truth that seem not to receive a wholly fair treatment. For, on the one hand, the speaker's intuitive pre-understanding of the concept seems to be underlied by something more than the formal-pragmatic presupposition of a unique objective world -- with no need in this «something more» for a substantive rationality. This is to be observed in the different ways in which a statement considered as true can turn out to be problematic, henceforth in the kind of commitment assumed by asserting it. As C. Lafont shows, the statement held as true anticipates the obligation to correct it, whenever some pregnant counterargument arises. Yet what counts as relevant counterargument cannot always be formulated in terms of «it is the case»/«it is not the case» (as the discussed criticism claims) and prevents us from considering that the statement truth or falsity just depend on this opposition. For neither the search for a rational justification of it seems to adopt always the form of an exclusion of one of the two cases (as the strictly logical bivalence of the tertium non datur would require),<60>Foot note 3_38 nor should two different commitments be identified, namely that of correcting the statement if relevant counterarguments arise and that of revising the affected belief when confronted to the opposite one. A counterargument does not need to adopt the form of a logical negation of the statement in question, as a strictly formal identification of truth with bivalence, without regard to its epistemic import, would require, and as the commented criticism seems to assume.<61>Foot note 3_39

On the other hand, and even if it is right that the epistemic view on truth introduces strong idealizations that it attributes to the participants in the epistemic game, these idealizations concern the inconditionality associated with the predicate true and this trait is also nuclear to the non-epistemic account.<62>Foot note 3_40 Obviously enough, the difference lies on the justification for it; and the presupposition of a shared objective world -- seen as contitutive for the epistemic use of language by this non-epistemic account -- is not free from some idealizations either -- as it is argued below.

In relation to the idealizations embodied by the epistemic view, in the explanation of truth as «rational acceptability under ideal conditions», these inevitable idealizations and the inconditionality associated with true are present in the two adjectives: ideal and rational. In the case of the expression «under ideal conditions», it becomes apparent that a determination of such conditions is not independent of the particular context and the theory («language») for which they are formulated. The referred idealization can be seen as purely internal and immanent -- as Putnam's analogy with ideal conditions in physics or chemistry suggests. At the same time, however, the expression is intended to trascend this contextual dependence. For it expresses, from a formal-pragmatic perspective, our intuition that our knowledge of the context suffices to grant that the attribution of truth to the statement is not misled, that it has not «got astray». Perhaps this supposition represents another way to introduce a form of idealization. Yet this latter does not refer to our (substantive) knowledge as such, but to the competence allowing us to know the preparatory and general contextual conditions necessary and sufficient to assert the statement. And this knowledge presupposes in its turn the ability to distinguish between language-knowledge (that allowing for the enunciation of the statement) and world-knowledge (i.e. of the contextual conditions making the assert possible). So understood, the resort to «ideal conditions» is not intended to account for a presupposition of infallibility that we would attach to our knowledge, but for the independence that speakers attribute to their own competence in which concerns virtual situative and contextual conditions that could have an effect upon them, affecting their capability of judgement.

To realize this does not mean refusing the discussed criticism, but rather displacing its focus. For the competence referred to -- if it is accepted as playing a role whenever a statement is held to be true -- forms part of our epistemic rationality so much as the knowledge of the statement truth-conditions. Furthermore, this competence makes the burden of the reconstruction of the inconditional validity attributed to true rest on the adjective rational. And here again, if rational is assumed to belong to the domain of the epistemic rationality, then its universalism seems in danger, since we are sent back to a criterial rationality with restricted validity. Yet there is also a possibility that it belongs to communicative rationality. For -- as it is evident in the TkH as well as in C. Lafont's papers -- true enters explicitly into play in problematic contexts; and in such cases only some form of consensus or final accord, communicatively attained, can (provisionally) settle the question. Nevertheless, even if this communicative rationality is claimed to be restricted to the domain of intersubjective relationships and to a reference to the intersubjective world<63>Foot note 3_41 -- henceforth, in the present context, to be oriented to a conjoint definition of the situation-, its application (of communicative rationality) depends on assuming the competence and knowledge of the context conditions alluded to by the expression «under ideal conditions» -- as analysed above.

This fact suggests that rational should be understood here in a wider sense, i.e. as referring to a competence capable of integrating both types of rationality, the epistemic and the communicative ones, and to articulate them. A «conjoint definition of the situation» should include not only an evaluation of the problematic statement and the concurring relevant counterarguments, but also the virtual revision of the criteria of rationality playing a role in these formulations, together with the competence that permits to jointly find a final accord on the matter. Yet this seems to mean «dividing» the notion of communicative rationality -- to some extent, in analogy to Kant's two uses of Vernunft. In the present context, the strong commitment underlying the adjective rational, within the framework of the Diskurstheorie der Wahrheit, is concerned with formal-pragmatic traits (e.g. a knowledge of the rules for participating in a discussion) that a quasi-transcendental reflexion would find to be maximally general and inevitable conditions present not only in the epistemic use of language, but also in any forum of discussion of a reflexive character. From this perspective, the anticipation of a shared objective world cannot be distinguished from the presupposition of a possible consensus on the definition of the situation and a conjoint determination of the preparatory and general context conditions related to the problematic statement.

Although this picture undoubtedly subsumes a strong idealization, it does not entail the idea of a «great final theory» or an infallible knowledge «in the long run». For both things would have a substantive character, whereas the above idealization concerns the formal-pragmatic conditions in which the discussion is to take place.<64>Foot note 3_42 This presupposition, which points at a virtual final «substantivizing» -- whereof the imputation of making appeal to a «great final theory», or an «infallible knowledge», or to a «definite consensus in the long run» seems justified -- could only be seen as feasible if another strong, somewhat gratuitous assumption is introduced, namely that no ulterior problematic situation or questioning of our established knowledge will take place. Without this assumption, the fallibility in our knowledge and the necessity to renew the consensus w.r.t. its conditions of possibility cannot be brought to any final point, without a unjustified categorial «jump». (In such a case, even the «objective world» would lose its counterfactual character).

It is undeniable that, in the explanation of the discursive theoty of truth itself, some elements are rightly susceptible of C. Lafont's criticism. Thus, as Prof. Habermas notices, with the TkH «wird der scheinbar klare Unterschied zwischen der Explikation der Bedeutung von Wahrheit und der Angabe von Kriterien für die Feststellung der Wahrheit empfindlicht relativiert.»<65>Foot note 3_43 What is more, «jeder, der ernsthaft einen theoretischen Satz aufstellt, findet sich unvermeidlich in der Rolle des `letzten' Theoretikers.»<66>Foot note 3_44 In spite of this risk of «substantivizing», thus of relativizing the notion of truth, as objected by the commented criticism, the final intention underlying the epistemic reconstruction of our pre-understanding of the notion, i.e. of the logical grammar of true, should not be accounted for -- so do I think -- in realist terms. For a realist interpretation and its alluded objection supposes the kind of categorical «jump» that only the disappearance of all epistemic breaks would permit. As I see it, the Diskurstheorie der Wahrheit makes the inconditionality of truth depend not on the presupposition of a non-fallible knowledge, but on the competence that allows for a revision of our own criteria of validity and for searching an agreement on them, whenever the arising of a problem demands it. This latter competence would make possible, in principle, to gain access to our own standards of rationality by a process of rational argumentation. And it is not different from the unique type of knowledge to which the Diskurstheorie der Wahrheit bestows a non-fallible character, namely «jenes vorgängige, von allen kompetenten Sprechen geteilte, freilich bloss intuitive, d.h. der Nachkonstruktion bedürftige Wissen, auf das wir rekurrieren, wenn wir sagen sollen, was es bedeutet, in einer Argumentation einzutreten.»<67>Foot note 3_45

As I tend to see it, the intuition underlying this epistemic concept of truth is not only that true, with the inconditionality we attribute to it, allows us to reflexively distinguish between our beliefs and an objective world counterfactually presupposed; this predicate true, understood as «rationally acceptable under ideal conditions», makes it possible for us to assign a fallible character to the very criteria deciding its application.<68>Foot note 3_46 But the idea underlying C. Lafont's proposal is -- if I am not mistaken -- that the counterfactual presupposition of a common objective world is in effect constitutive as well as normative of the epistemic use of language -- and not merely the obligatoire reference embodied by any epistemic truth claim, as asserted within the framework of the TkH. Yet within the framework of the TkH it is only possible to attribute a constitutive character to language games, themselves part of the lifeworld. And, «weil sich alle Lebenswelten über das Medium verständigungsorientierten Handelns reproduzieren müssen, kommt in der Mannigfaltigkeit konkreter Lebensformen zugleich die Allgemeinheit kommunikativer Rationalität zur Geltung.»<69>Foot note 3_47

My impression is that, by assigning a constitutive role to the common objective world, C. Lafont tries to avoid an anchoring of the epistemic use of language in a reference to or a theoretical dependence on intersubjectivity -- thus avoiding the relativism that she sees as correlative to it. But attributing a constitutive value to the realist presupposition of a common objective world is not exempt of other consequences as well. If it is to fulfil a role at the same time constitutive and regulative of the epistemic praxis, then a conceptual precision seems to be required here on the meaning of the word constitutive. For according to Kant, and in relation to the theoretische Vernunft, constitutive is opposed to regulative (normative); but in the opposition between Verstand and Vernunft, constitutive principles become regulative and are those fundamental principles that, as rules of the objective use of synthetic a priori categories and meaning postulates (either in the form of axioms or anticipations that can be constructed, or in the form of analogies and postulates that only in the discourse attain validity), such constitutive principles make the knowledge of the objects of experience possible.

Similarly, the presupposition of a common objective world appears to have to be considered more a postulate than an anticipation -- if it is to remain as constitutively present and counterfactual; for, as anticipation to be constructed, it would send us back to the «big true theory» previously denounced. But this seems to entail, in its turn, that only through a discursive-reflexive use of language is it possible to gain access to this presupposition, together with the corresponding rules for its empirical interpretation. Hence, only in a discourse can this presupposition, as postulate, reach argumentative validity. It is this reflexive, communicative use of rationality, thus the discoursive use of language -- the «discursive game» -- what is to be seen as constitutive: by Kant, of the thinking subject; and within the present framework, of the set of presuppositions playing a role for the epistemic use of language.

If the above conclusion is right, the reference-to-an-objective-world subsumed in the raising of a validity claim in general becomes a postulate of a kind that only acquires validity through discoursive (reflexive) argumentation. And the corresponding process takes place precisely in contexts in which a previously established knowledge happens to be questioned. But this reading seems to imply for the commented proposal a risk of self-contradiction. Namely, that of assuming intersubjectivity to be constitutive not only for the standards of rationality, but even for the formal-pragmatic presuppositions that are present in the epistemic language game -- and all language games in general. Nevertheless, dispensing with this perspective seems to imply the opposite risk, and one which -- in my opinion -- the commented proposal cannot avoid. Namely, that of absolutizing,<70>Foot note 3_48 by attributing a normative and (at the same time) a constitutive character to the presupposition of a common objective world, a postulate that is only to be redeemed in certain contexts -- in those contexts in which the questioning of a previous knowledge demands a reflexive, communicative use of language.

Here the commented proposal could object that even the discussion prompted by the problematic situation is guided by the realist thesis at stake -- even more, that this latter is condition of possibility for the former to arise. And now the argumentation seems to bring up into a circularity difficult to overcome. For on the one hand, if the commented proposal is right -- and the presupposition of a common objective world is anchored in the epistemic use of language and reconstructed by means of a theory of direct reference jointly with a realist view on truth-, then some form of (non-linguistic) causalty is required to account for our relationship to this world,<71>Foot note 3_49 together with some experience of the commensurability of different interpretations on it. Yet on the other hand, what is assumed to be a shared objective world needs to be anchored in the possibility (still counterfactual presupposed) of a shared experience, and in the communicability and possible public testing of this experience. This possibility can only be consistently supposed, in its turn, against a background of common and shared practices, within an intersubjectively validated linguistic context.

Henceforth the `uniqueness' and objectivity of the world -- understood as independence w.r.t. aprioric, «already given» meanings -- is only accessible, as counterfactual presupposition, in the public context of the lifeworld we share -- and which includes epistemic practices of fixation of reference in learning processes, whereof the causalty relation is reversed: it is us ourselves who «act» upon the world. The presupposition of an objective world, as it is present in the intuition of speakers, can be seen as arising from these shared public practices linguistically mediated, thus from the «breach» between our practices (actions, activities, practical behaviour, etc.) and what constitutes their object. It is in the context of the problems brought about in the course of these practices where the possibility to distinguish between our meaning-knowledge and our world-knnowledge is to be based upon. Likewise, the fact that practical problems cannot be produced «on purpose» does not allow us to see these epistemic discontinuities -- in the continuum constituted meanings / counterfactually presupposed world -- as arising from a pre-existing intersubjectivity. Furthermore, not any arbitrary form of intersubjectivity can be seen as constitutive for validity claims. Yet conversely, any reference to the counterfactual assumption of a shared world subsumes already the very idea of a form of legitimately (rational-communicatively) constituted intersubjectivity.


A theory of direct reference could be, as C. Lafont and A. Müller's analyses have shown, the best device in the formal-semantical level to account for our use of language in learning processes and epistemic contexts in general. This reconstruction would allow us to avoid the kind of meaning holism that prevents distinguishing between language knowledge and meaning knowledge. But this proposal embodies a risk as well, that of absolutizing the referencial function of language and giving priority to the corresponding designative use of it, thus falling back into an instrumentalist (functionalist) view -- something that Frege was trying to avoid when he introduced the distinction Sinn/Bedeutung.

Thus the presupposition of a common objective world that underlies these epistemic practices -- and the epistemic use of language in general -- can only be consistently introduced and considered as revisable (reflexively accessible) if it belongs to the sphere of communicative rationality and under the assumption that we are reflexively competent to gain access to our own «rules of the game». Yet this point of view, together with the idea that this reflexive competence can only be redeemed in the context of a legitimately constituted intersubjectivity, is close to claiming that the internal realism connected with the epistemic use of language becomes a sort of `idealism of the intersubjectivity' on the level of the communicative rationality, on which it depends.

And yet a problem remains open, which must play the role of (or replace) any final conclusion. The central question underlying the present discussion concerns the universality claim essential for the TkH and threatened by meaning holism. On the one hand, any emphatical notion of rationality -- thus substantive -- or any approximation to it in terms of a notion of non-fallible knowledge appears to be not universalizable. On the other hand, however, the intuition related to the non-epistemic concept of truth, no matter how much it claims for its universality, does not grant what for the Diskurstheorie der Wahrheit seems a main concern: the possibility to `substantivize' a consensus with respect to our interpretation on the objective world. For the universality claimed by the realist view, purely formal, can only be substantivized through linguistic mediations whose potential universality is dependent upon the universality of a virtual consensus, rational-communicatively attained.

A non-epistemic view on truth is semantically modelled. Its addoption from the part of the commented critical proposal, jointly with the correlated realist assumption, represents an attempt to reconstruct the domain of (factual) meaning constitution, granting from this domain on that universalism is preserved. Nevertheless, and as the present discussion has tried to show, this proposal carries with it other difficulties as well. The TkH instead focused on a de iure question, namely what makes possible to justify validity for those already constituted meanings in problematic contexts.

Cristina Corredor
University of Valladolid
E--47071 Valladolid, Spain