I examine the claim, made by some authors, that we sometimes acquire knowledge from falsehood. I focus on two representative cases in which a subject S infers a proposition q from a false proposition p. If S knows that q, I argue, S's false belief that p is not essential to S's cognition. S's knowledge is instead due to S's belief that p′, a proposition in the neighbourhood of p that S believes . S thus knows despite her false belief. (...) The widely accepted and plausible principle that inferential knowledge requires known premises is unscathed. (shrink)

According to the worm theory, persons are (maximal) aggregates of person-stages existing at different times. Personites, on the other hand, are non-maximal aggregates of stages that are very much like persons. Their existence appears to make instances of prudential self-sacrifice morally problematic: the personites that exist at the time of the sacrifice but not at the time of the reward seem not to receive future compensation for their sacrifice. Instances of punishment appear to give rise to a similar problem. We (...) argue that these impressions arise from a mistaken assumption about the primary bearers of properties such as suffering, receiving compensation (in the future) and having (previously) committed a crime. According to the worm theory, stages, rather than persons or personites, possess these properties. Persons and personites have these properties only derivatively. As we show, once this clarification and related ones are made, the apparent moral problems raised by the existence of personites dissolve. (shrink)

We argue that four-dimensionalists should adopt a contextualist semantics, according to which ordinary speakers’ judgments may concern person-stages, person-segments or person-worms, depending on the context. We explain how context helps select the boundaries of the temporal parts we refer to or quantify over and show that contextualism offers the best treatment of ordinary predications and ordinary counting judgments. Contextualism implies an error theory; however, we explain why this error theory is less problematic than those entailed by the worm and stage (...) theories. (shrink)

Epistemic contextualism, many critics argue, entails that ordinary speakers are blind to the fact that knowledge claims have context-sensitive truth conditions. This attribution of blindness, critics add, seriously undermines contextualism. I show that this criticism and, in general, discussions about the error theory entailed by contextualism, greatly underestimates the complexity and diversity of the data about ordinary speakers? inter-contextual judgments, as well as the range of explanatory moves that are open to both invariantists and contextualists concerning such judgments. Contextualism does (...) entail that some speakers suffer from semantic blindness; however, at its roots, this blindness concerns not the context-sensitivity of knowledge claims, but the question whether knowledge sentences possess context-independent truth conditions. I argue that this blindness should not be deemed problematic, but that invariantism entails an error theory that is, by comparison, much more troubling. (shrink)

I argue that assertion and practical reasoning must be governed by the same epistemic norm. This is because the epistemic rule governing assertion derives from the epistemic rule governing practical reasoning, together with a plausible rule regarding assertion, according to which assertion must manifest belief.

I defend a contextualist account of bare epistemic modal claims against recent objections. I argue that in uttering a sentence of the form ‘It might be that p,’ a speaker is performing two speech acts. First, she is (directly) asserting that in view of the knowledge possessed by some relevant group, it might be that p. The content of this first speech act is accounted for by the contextualist view. But the speaker's utterance also generates an indirect speech act that (...) consists in a weak suggestive that p. Since this second speech act is typically the main point of a bare epistemic modal utterance, our (negative or positive) responses to the utterance actually target this second speech act. I show how this two-speech-act account can explain the data recently adduced against contextualism. (shrink)

Some authors have recently claimed that relativism about knowledge sentences accommodates the context sensitivity of our use of such sentences as well as contextualism, while avoiding the counterintuitive consequences of contextualism regarding our inter-contextual judgments, that is, our judgments about knowledge claims made in other contexts. I argue that relativism, like contextualism, involves an error theory regarding a certain class of inter-contextual judgments.

Causal contextualism holds that sentences of the form ‘c causes e’ have context-sensitive truth-conditions. We consider four arguments invoked by Jonathan Schaffer in favor of this view. First, he argues that his brand of contextualism helps solve puzzles about transitivity. Second, he contends that how one describes the relata of the causal relation sometimes affects the truth of one’s claim. Third, Schaffer invokes the phenomenon of contrastive focus to conclude that causal statements implicitly designate salient alternatives to the cause and (...) effect. Fourth, he claims that the appropriateness of a causal statement depends on what is contextually taken for granted or made salient. We show that causal invariantism can explain these linguistic data at least as well as contextualism. We then argue that pace Schaffer, some causal sentences are always correct and can never be plausibly denied, regardless of the context. (shrink)

We propose an original response to Derk Pereboom’s four-case manipulation argument. This response combines a hard-line and a soft-line. Like hard-liners, we insist that the manipulated agent is blameworthy for his wrongdoing. However, like soft-liners, we maintain that there is a difference in blameworthiness between the manipulated agent and the non-manipulated one. The former is less blameworthy than the latter. This difference is due to the fact that it is more difficult for the manipulated agent to do the right thing. (...) We explain how we can make sense of this notion of difficulty in terms of Fischer and Ravizza’s notion of reasons-responsiveness. (shrink)

I defend the principle that a person is blameworthy for her action only if that action was morally wrong. But what should we say about an agent who does the right thing based on bad motives? I present three types of cases that have these features. In each, I argue, the agent is not culpable for her action; however, she violates the norm of moral responsibility, and thus acts in a morally irresponsible way. This analysis, I show, has several virtues. (...) It also has important theoretical ramifications: it allows for a more nuanced approach to moral luck and paves the way for an unexplored line of defense in favor of a particular norm of moral responsibility. (shrink)

W. V. Quine is commonly read as holding that there are no analytic truths and no a priori truths. I argue that this is a misreading. Quine’s view is that no sentence is determinately analytic or determinately a priori. I show that my reading is better supported by Quine’s arguments and general remarks about meaning and analyticity. I then briefly reexamine the debate between Quine and Carnap about analyticity, and show that the nature of their disagreement is different than what (...) it is usually thought to be. (shrink)

I consider three challenges to the traditional view according to which moral responsibility involves an epistemic condition in addition to a freedom condition. The first challenge holds that if a person performs an action A freely, then she thereby knows that she is doing A. The epistemic condition is thus built into the freedom condition. The second challenge contends that no epistemic condition is required for moral responsibility, since a person may be blameworthy for an action that she did not (...) know was wrong. The third challenge invokes the quality of will view. On this view, a person is blameworthy for a wrong action just in case the action manifests a bad quality of will. The blameworthy person need not satisfy an additional epistemic condition. I will argue that contrary to appearances, none of these challenges succeeds. Hence, moral responsibility does require a non-superfluous epistemic condition. (shrink)

Contextualism, in its standard form, is the view that the truth conditions of sentences of the form ‘S knows that P’ vary according to the context in which they are uttered. One possible objection to contextualism appeals to what Keith DeRose calls a warranted assertability maneuver (or WAM), according to which it is not our knowledge sentences themselves that have context-sensitive truth conditions, but what is pragmatically conveyed by the use of such sentences. Thus, proponents of WAMs argue, the context (...) sensitivity of knowledge attributions is not a semantic phenomenon but a pragmatic one. I examine a number of WAMs and show that each (i) is seriously flawed, or (ii) undercuts standard contextualism if we hold a minimalist conception of semantic content. I propose an alternative form of contextualism that accommodates minimalism and is immune to the second type of WAM, and show that this new form of contextualism shares the virtues of standard contextualism. (shrink)

According to a popular objection against epistemic contextualism, contextualists who endorse the factivity of knowledge, the principle of epistemic closure and the knowledge norm of assertion cannot coherently defend their theory without abandoning their response to skepticism. After examining and criticizing three responses to this objection, we offer our own solution. First, we question the assumption that contextualists ought to be interpreted asassertingthe content of their theory. Second, we argue that contextualists need not hold that high epistemic standards govern contexts (...) in which they defend their theory. (shrink)

I explore the question of when an agent is derivatively, rather than directly, culpable for an undesirable outcome. The undesirable outcome might be a harmful incompetent or unwitting act, or it might be a harmful event. By examining various cases, I develop a sophisticated account of indirect culpability that is neutral about controversies regarding normative ethical issues and the condition on direct culpability.

Analytic Philosophy, Volume 63, Issue 1, Page 28-43, March 2022.

Would a tiny contribution such as the addition of one drop of water to the canteen of a thirsty person relieve her suffering? According to Barnett’s 2017 paper ‘No free lunch’, the answer is ‘yes’: even tiny contributions can make a morally relevant difference. To defend this answer, Barnett raises an objection against the rival view that tiny contributions never make any difference. I argue that we should reject both Barnett’s and the rival view. I propose an alternative account that (...) reflects the vagueness at play in the outcome of tiny contributions and entails that, in a range of cases, it is indeterminate whether these contributions make any difference. (shrink)

Abstract: Several critics of contextualism claim that this view cannot consistently maintain its advertised neutrality between skepticism and anti-skepticism. Some critics contend that contextualists are forced to side with the skeptic, since any defense of contextualism unavoidably puts in place the skeptic's high requirements for knowledge; others hold that the contextualists' claim to have knowledge of what their own view entails forces them to reject the skeptic's knowledge denial. I show that these arguments misconstrue the role of context in contextualism, (...) and explain how we are to understand the contextualist's proposed agreement with both the skeptic's knowledge denial and the anti-skeptic's knowledge attribution. (shrink)

According to a view widely held by epistemic contextualists, the truth conditions of a knowledge claim depend on features of the context such as the presuppositions, interests and purposes of the conversational participants. Against this view, I defend an intentionalist account, according to which the truth conditions of a knowledge attribution are determined by the speaker’s intention. I show that an intentionalist version of contextualism has several advantages over its more widely accepted rival account.

Davidson claims that a creature that has spent its entire life in isolation cannot have thoughts. His two reasons for this claim are that interaction with another creature is required to locate the cause of the creature's responses, and that linguistic communication is necessary to acquire the concept of objective truth, which is itself required in order to have thoughts. I argue that, at best, these two reasons imply that in order to have thoughts a creature must be capable of (...) participating in triangulation, not that it must have already participated in triangulation. I then argue that triangulation doesn't solve the ambiguity problem; that is, it doesn't entail that a being's thoughts and utterances are about distal objects rather than proximal patterns of stimulation. Fortunately, ambiguity, like other forms of indeterminacy, doesn't entail that we cannot have thoughts. /// Davidson afirma que una criatura que ha pasado su vida entera aislada no puede tener pensamientos. Aduce dos razones para esto: que se requiere la interacción con otra criatura para localizar la causa de las respuestas de la criatura, y que la comunicación lingüística es necesaria para adquirir el concepto de verdad objetiva, que a su vez es necesario para tener pensamientos. Argumento que, en el mejor de los casos, estas dos razones implican que para tener pensamientos una criatura debe ser capaz de participar en una triangulacion, no que ya debe haber participado en ella. Arguyo, además, que la triangulación no resuelve el problema de la ambigüedad, esto es, no implica que lo que un ser piensa y dice sea acerca de objetos distantes, más que de patrones de estimulacion próximos. Por fortuna, la ambigiiedad, como otras formas de indeterminación, no implica que no podamos tener pensamientos. (shrink)

Shakespeare's Coriolanus is one of the most brilliant political plays ever written. Despite its ancient Roman setting, it remains a perennially relevant study of the relationship between personality and politics. The Introduction to this new edition illuminates its relevance to Shakespeare'sown time and to later ages while also emphasizing the wide range of interpretations that are possible in performance.

According to what I call theidentity view, omissions are actual events. For example, the nominal ‘Ali's non-jogging’ denotes whatever Ali is doing at the time she is said not to be jogging. Some have objected that omissions (and more generally absences) cannot be events, since the two do not have the same causal relations. I show how advocates of the identity view can offer a pragmatic account of the data the objection relies on.

David Sosa argues that the knowledge account of assertion is unsatisfactory, because it cannot explain the oddness of what he calls dubious assertions. One such dubious assertion is of the form ‘P but I do not know whether I know that p.’ Matthew Benton has attempted to show how proponents of the knowledge account can explain what’s wrong this assertion. I show that Benton’s explanation is inadequate, and propose my own explanation of the oddness of this dubious assertion. I also (...) explain what’s wrong with other dubious assertions mentioned by Sosa. (shrink)

Morgan's canon can be construed as claiming that an intentional explanation of a behavior should be ruled out if there exists an explanation of this behavior in terms of 'lower' mechanisms. Unfortunately, Morgan's conception of higher and lower faculties is based on dubious evolutionary considerations. I examine alternative interpretations of the terms 'higher' and 'lower', and show that none can turn the canon into a principle that is both correct and useful in drawing the line between thinkers and non-thinkers. In (...) the process, I identify a number of problems that an adequate formulation of the canon should avoid. I then consider two more recent versions of the canon, proposed by Elliott Sober and Jonathan Bennett. Both are found unsatisfactory, but I argue that a version of Bennett's unity condition that is restricted to the attribution of recognitional concepts is on the right track. (shrink)

Contextualism about vagueness holds that the content of vague predicates is context sensitive. I contrast this view with a similar view called nonindexical contextualism, and explain why my brand of contextualism should be preferred to it. I then defend contextualism against three objections that have been recently raised against it. I show that these objections are actually more damaging to rival views than to contextualism itself.Quanto ao fenômeno da vagueza, o contextualism defende a tese de que o conteúdo dos predicados (...) vagos é sensível ao contexto. Contrasto essa visão com uma visão similar chamada contextualismo não-indexical, e explico por que minha versão dever ser preferida à outra. Depois, eu defendo o contextualismo contra três objeções que foram apresentadas recentemente contra ele. Argumento que essas objeções valem muito mais, na verdade, não contra o contextualismo, e sim contra as visões rivais. (shrink)

I consider the rule of assertion according to which knowledge is sufficient for epistemically proper assertion. I examine a counterexample to this rule recently proposed by Jennifer Lackey. I present three responses to this counterexample. The first two, I argue, highlight some flaws in the counterexample. But the third response fails. The lessons I draw from examining these three responses allow me to propose two counterexamples to the sufficiency rule that are similar to Lackey’s but avoid its problems.

I argue that a capacity‐based account of blameworthiness and praiseworthiness is superior to an account based on quality of will. I focus on four types of cases about which the two accounts disagree and show that the capacity‐based view offers a better treatment. As part of my argument, I motivate the distinction between an assessment of a person's moral character, as reflected by her action, and an assessment of her blameworthiness or praiseworthiness for that action.

Contrastivism is the idea that knowledge is question-relative: to know is to be able to answer a contextually salient question. Constrastivism's main selling point is that it promises to respect ordinary speakers' judgments about knowledge claims made in various contexts. I show that contrastivism fails to fulfill this promise, and argue that the view I call epistemic pluralism does much better in this respect.

I argue that pace Quine, indeterminacy of translation affects observation sentences. I illustrate this indeterminacy with examples and show how it is tied to the indeterminacy affecting the analytical status of observation categoricals. I propose my own construal of the thesis of indeterminacy of translation, according to which indeterminacy is based on the inextricability of meaning and belief. I explain why this construal should be favored over Quine’s.

Donald Davidson holds that intentional concepts are not reducible to physical or dispositional ones. This is due, he claims, to the constitutive role of normativity in the principles that govern the application of intentional concepts. According to Davidson, the specific way in which norms of rationality and coherence are mobilised by our interpretative principles sets mental concepts off from those of the natural sciences. I agree with Davidson on the irreducibility of the mental. However, I show that irreducibility is due (...) not to the normative character of intentional concepts, but to holism and the flexibility of interpretative principles. I then consider three arguments that Davidson has put forth to support the irreducibility of intentional properties on the basis of their normative character, and show that none of them goes through. (shrink)

An excuse is an event or condition that exonerates an agent for a wrongdoing. An excuse may be an event or condition that interferes with the exercises of a person’s rational capacities, thereby preventing them from doing the right thing. I argue that a person who fails to do the right thing always has an excuse for their failure. This puzzle has troubling consequences, for it means that we are never to blame for our wrongdoings.

I raise three new objections against Robert Kane’s account of ultimate responsibility based on what he calls self-forming actions (sfa s). First, the ultimate responsibility that we have for our character is very limited, since, according to Kane’s model of character development, our character is shaped by sfa s for which we are only minimally responsible. Second, it is not desirable to rely on sfa s to shape our character. There are much better alternatives. Third, given what typically motivates our (...) sfa s, we only rarely satisfy the conditions on responsibility for the changes in character that our sfa s give rise to. (shrink)

I consider cases in which a person’s action causes a foreseeable harm, but does so through an unforeseeable causal path. According to a common view, the person is blameless for the harm in such cases. I argue that any defense of this common view incurs serious costs. I then show how a popular view about resultant luck can make the rejection of the common view palatable.

Résumé J’examine la thèse défendue par Donald Davidson selon laquelle un être ne peut avoir des pensées que s’il a été en communication linguistique avec quelqu’un d’autre par le passé. Cette thèse, que j’appelle « l’interprétationnisme radical », dérive de la thèse A selon laquelle il est nécessaire d’avoir les concepts de croyance et de vérité objective pour avoir des croyances, et de la thèse B voulant que la communication linguistique soit requise pour l’acquisition du concept de vérité objective. En (...) réponse à A, je préconise un point de vue contextualiste, selon lequel les normes d’attribution de croyances dépendent du contexte conversationnel. Le contextualisme entraîne non pas que A est fausse, mais qu’elle doit être relative à un contexte. Je montre par ailleurs que contrairement à ce qu’affirme B, l’interaction avec autrui n’est pas nécessaire pour acquérir le concept de vérité objective. Je conclus que les arguments de Davidson soutiennent au mieux l’interprétationnisme modéré, c’est-à-dire l’idée selon laquelle un être a des pensées si et seulement si ces pensées sont interprétables par un interprète pleinement informé.I examine Donald Davidson’s thesis that a creature cannot have thoughts unless it has had a linguistic communication with another creature. This thesis, which I call “Radical Interpretationism”, derives from the claim that it is necessary to have the concepts of belief and objective truth in order to have beliefs and the claim that linguistic communication is required to acquire the concept of objective truth. In response to, I put forward a contextualist viewpoint, according to which the norms of belief attribution are context-dependent. Contextualism entails not that is false, but that it must be relativised to a context. I then show that contrary to what says, interaction with others is not necessary to acquire the concept of objective truth. I conclude that Davidson’s arguments at best support moderate interpretationism, that is, the view that a creature has thoughts if and only if these thoughts are interpretable by a fully-informed interpreter. (shrink)

Quine’s negative theses about meaning and analyticity are well known, but he also defends a positive account of these notions. I explain what his negative and positive views are, and argue that Quine’s positive account of meaning entails that two of his most famous doctrines, namely the claim that there are no analytic statements and the indeterminacy of translation thesis, are false. But I show that the falsity of these doctrines doesn’t affect his criticisms of traditional conceptions of meaning. This (...) is because the class of analytic statements that his account of meaning enables us to isolate is of no philosophical interest, and because we can hold that translation is determined without admitting that meaning is. (shrink)

According to contextualism about vagueness, the content of a vague predicate is context sensitive. On this view, when item a is in the penumbra of the vague predicate ‘F’, speakers may utter ‘Fa’, or they may utter ‘not-Fa’, without contravening the literal meaning of ‘F’. Unlike its more popular variants, the version of contextualism I defend rejects the principle of tolerance, a principle according to which small differences should not affect the applicability of a vague predicate. My goal is to (...) show how such a rejection allows for a plausible treatment of higher-order vagueness, and for a dissolution of paradoxes of higher-order vagueness. (shrink)

I propose a version of inferential role theory which says that having a concept is having the disposition to draw most of the inferences based on the stereotypical features associated with this concept. I defend this view against Fodor and Lepore.

Abstract: Despite all the critical scrutiny they have received recently, contextualist views in philosophy are still not well understood. Neither contextualists nor their opponents have been entirely clear about what contextualist theses amount to and what they are based on. In this article I show that there are actually two kinds of contextualist view that rest on two very different semantic phenomena, namely, semantic incompleteness and semantic indeterminacy . I explain how contextualist approaches can be used to dissolve certain debates (...) in philosophy. According to such approaches, the same philosophical thesis can be correctly endorsed in some contexts and correctly denied in others: it is thus pointless to seek a context-independent solution to debates about this thesis. My purpose is not to defend particular contextualist views but to lay out the general framework on which they rest: this allows us to see more clearly the similarities and differences among contextualist views defended in various areas of philosophy. (shrink)

Pooh rubbed his nose with his paw, and said that the Heffalump might be walking along, humming a little song, and looking up at the sky, wondering if it would rain, and so he wouldn't see the Very Deep Pit until he was half-way down, when it would be too late. Jerry Fodor is loath to have content be constituted, even in part, by inferential relations. This loathing, I will argue, gets him into trouble. In his latest book, Concepts, Fodor (...) contrasts informational atomism, his view of concepts and their content, with inferential role theory. The latter, he argues, is almost certainly false. This strikes cognitive science at its core, he adds, since the major theories of concepts currently in vogue in cognitive science are all variants on inferential role theory. (shrink)

I examine in detail the argument vaguely suggested by Davidson to the effect that holism entails the irreducibility of the mental. I defend this argument against two objections often made against arguments that attempt to derive metaphysical theses from premises that concern our ordinary criteria for applying terms. I appeal to two-dimensional semantics to explain the links between these criteria and issues about reference and reduction. I show how the irreducibility of the mental follows from the holistic and flexible character (...) of our criteria for attributing beliefs and intentional states. (shrink)

La thèse de l'indétermination de la traduction de W. V. O. Quine est certainement une des thèses les plus controversées de la philosophie du langage. Le présent article explique en quoi consiste cette thèse et examine les liens qu'elle entretient avec la thèse de la sous-détermination des théories scientifiques. La première section montre comment la thèse de l'indétermination de la traduction découle de la conception behavioriste du langage de Quine. Les sections suivantes exposent deux façons de dériver la thèse de (...) l'indétermination: par le holisme et le vérificationnisme dans la section 2 et par la thèse de la sous-détermination dans la section 3. La section 4 précise la formulation de la thèse de la sous-détermination proposée dans «On Empirically Equivalent Systems of the World» et la section 5 expose les critiques exprimées par W. Bechtel à propos de cette formulation. Ces critiques témoignent d'une mauvaise compréhension de la position de Quine et j'essaierai de montrer en quoi elles sont fautives. La section 6 traite des deux façons de dériver la thèse de l'indétermination, à la lumière des modifications apportées à la formulation de la thèse de la sous-détermination. Enfin, la section 7 s'intéresse aux questions de fait relatives à la sous-détermination des théories et à l'indétermination de la traduction. (shrink)

This paper explores the relationships between Davidson's indeterminacy of interpretation thesis and two semantic properties of sentences that have come to be recognized recently, namely semantic incompleteness and semantic indecision.1 More specifically, I will examine what the indeterminacy thesis entails for sentences of the form 'By sentence S (or word w), agent A means that m' and 'Agent A believes that p.' My primary goal is to shed light on the indeterminacy thesis and its consequences. I will distinguish two kinds (...) of indeterminacy that have very different sources and very different consequences. But this does not purport to be an exhaustive study: there may well be other forms of indeterminacy that this .. (shrink)

Donald Davidson considère qu'une théorie de l'interprétation doit êtreradicale, c'est-à-dire qu'elle ne doit présupposer aucune connaissance de la langue à interpréter. Cette exigence repose sur l'idée suivante: si une théorie de l'interprétation pour une langue L présuppose une certaine compréhension de L, alors elle perd son pouvoir explicatif et échoue à rendre compte de ce qui permet la compréhension de L. L'interpr'tation radicale a l'avantage de nous forcer à rendre explicite ce qui est à l'œuvre dans le processus d'interprétation du (...) comportement verbal d'autrui. Elle nous permet aussi de nous concentrer sur la nature des données empiriques qui rendent possible la communication. Il importe de noter que Davidson ne veut aucunement suggérer qu'un interprète compétent de Lconnaϊtoupossèdeune théorie de l'interprétation pour L: la théorie set plutôt à donner une description de sa compétence. (shrink)

L'objectif de cette these est d'etudier les fondements empiriques de notions semantiques telles que la signification, l'analyticite et la reference. J'essaie entre autres choses de defendre l'idee qu'il n'existe pas de critere permettant de definir strictement ces notions. Dans un premier temps, j'examine les critiques formulees par W. V. Quine contre la distinction analytique-synthetique. Il n'existe pas d'argument general pour rejeter cette distinction. Pour critiquer celle-ci, on doit examiner individuellement les diverses tentatives visant a l'etablir. ;Tout d'abord, j'examine les objections (...) de Quine contre le point de vue selon lequel la distinction analytique-synthetique aurait un caractere fondationnel. Ces objections sont rapprochees du paradoxe sceptique de Ludwig Wittgenstein, qui concerne le concept normatif de signification, selon lequel la signification d'un enonce justifierait son usage. ;J'aborde ensuite la question de savoir s'il est possible de tracer la distinction analytique-synthetique dans le cadre d'une theorie verificationniste de la signification, qui est fondee sur la maniere dont les locuteurs s'y prennent pour etablir la verite de leurs enonces. Une telle approche ne permet pas d'assigner aux enonces theoriques une signification determinee a cause du holisme, qui affirme que les enonces theoriques ne sont pas liees individuellement aux experiences qui les confirment , mais seulement en tant que fragments plus ou moins importants de theorie. Je tente de clarifier la these du holisme, et en etudie la source. ;Je presente ensuite le compte rendu de la signification propose par Quine, qui est fonde sur l'experience de pensee de la traduction radicale. Un manuel de traduction radicale doit organiser les donnees empiriques a partir desquelles il est construit, i.e. les dispositions verbales des locuteurs, conformement a certaines contraintes comme la contrainte de la communaute et le principe de charite. J'explique comment ces contraintes doivent etre comprises et appliquees. Le compte rendu de Quine est compare et contraste a celui de Donald Davidson, qui presente plusieurs points en commun. J'examine les principales differences qui les separent et essaie d'en decouvrir les motivations. Je montre par exemple pourquoi la theorie distale de Davidson est preferable a la theorie proximale de Quine. ;Enfin, la these de l'indetermination de la traduction est presentee. Je propose une formulation de cette these, selon laquelle l'indetermination de la traduction est directement liee a l'absence de critere empirique permettant de tracer la distinction analytique-synthetique. J'illustre ensuite cette these au moyen d'exemples et examine certaines de ses consequences, qui n'ont pas toutes ete envisagees par Quine, notamment celles qui concernent la notion de contenu empirique. (shrink)