Putnam and Davidson on coherence, truth, and justification

The Science of Mind (精神科学) 54 (2016), pp. 51-70. Putnam and Davidson on Coherence, Truth, and Justification Lajos L. Brons Nihon University In the early 1980s both Donald Davidson and Hilary Putnam defended theories that superficially appear to be coherence theories of truth, but that upon closer inspection turn out to be better understood as coherence theories of justification (or perhaps, not as coherence theories at all). In Reason, Truth and History (1981a), Putnam argued that “truth is an idealization of rational acceptability” (55), and that rational acceptability of statements is their “coherence and fit” (id.). And in the same year Davidson gave a talk that would be published two years later as “A Coherence Theory of Truth and Knowledge” (1983) in which he defended “what may as well be called a coherence theory of truth and knowledge” (137). There are more interesting similarities between Putnam’s and Davidson’s theories than their closeness in dates of birth, however, and in later work Putnam’s theory further converged with Davidson’s. Furthermore, the most conspicuous difference – the role and nature of “conceptual schemes” – turns out to be smaller and more subtle than it may seem at a first glance. This paper introduces Putnam’s and Davidson’s coherence theories, and discusses the most important differences and similarities. The following two sections summarize “middle” Putnam’s and Davidson’s coherence theories respectively. The two sections after that look into the apparent disagreement on the possibility of different conceptual schemes, and discuss further developments in Putnam’s view and the consequent convergence of Putnam’s and Davidson’s coherentisms. The final section briefly compares the resulting – largely shared – 51 theory with more recent conceptions of coherentism and concludes that by current standards, Putnam’s and Davidson’s theories are not coherence theories but hybrids between coherentism and foundationalism. As will become apparent in the following, both Putnam’s and Davidson’s writings resist easy interpretation, albeit for very different reasons. Interpreting Putnam is complicated by his changes of mind; interpreting Davidson by his obscurity. Putnam belonged to the rare breed of thinkers that aren’t afraid to reject their own ideas upon realization of their incoherence. For some time in the 1960s and early 1970s, he was at the forefront of analytic philosophy, but when he became increasingly aware of problems and contradictions in the mainstream view he had helped to build, he left it behind. And when he found problems in the views he defended next (in his “middle” period), he left those behind as well. In this process he became more and more critical of mainstream analytic philosophy.1 More importantly, because of this continuing development, one cannot look for clarification of Putnam’s ideas in his own (much) earlier or later writings. In contrast, in case of Davidson, one has to look for clarification in earlier and later writings. Contrary to Putnam, changes in Davidson’s philosophical views are few and subtle. Davidson was a systematic thinker, and as he suggested in (1990), much of his system was already in place in the early 1960s. That system was built on Quine’s (with some modifications), and consequently, like Quine, Davidson rejected many of the core beliefs of mainstream analytic philosophy. 2 By implication, his philosophical ideas and terminology (like Quine’s) cannot be taken at face value (let alone lifted out of their context), and can only be properly understood within the broader program they are part of (e.g. Malpas 2011a; Brons 2014). But there are further complications in interpreting Davidson’s 1 For example, in “After Empiricism” (1985), Putnam writes that analytic philosophy is “total shambles”, that its achievements are entirely negative and that it has only “succeeded in destroying the very problem with which it started”, that it has “failed”, and that it is “at a dead end” (51). 2 But unlike Quine (and Putnam) it appears that Davidson was not always fully aware of this. In fact, he was even surprised when he was characterized as “post-analytic” (1993a). 52 Putnam and Davidson on Coherence, Truth, and Justification writings. “His style was too cryptic and elliptical, so that it was often unclear what his claims or arguments were,” said Timothy Williamson about Davidson in an interview (Williamson & Bo 2009: 60).3 Williamson is not a Davidsonian, of course, but similar remarks, differing more in tone than in content, can be found in the writings of philosophers closer to Davidson.4 Davidson himself once described his style as “breathless”,5 and this seems a fitting adjective. Most of his essays were talks originally, and even those that were not read like lecture transcripts: style and tone are informal, and usually the same applies to the reasoning. Because of this, upon a first and cursory reading many of Davidson’s essays seem to make sense (perhaps to be convincing even), but second, third, and further readings to figure out what exactly his argument is often result in puzzlement and more questions than answers. Sometimes the answers to such questions can be found elsewhere in Davidson’s writings, but this only points at a related problem already hinted at above: understanding a single text by Davidson usually requires (at least some) familiarity with all of his writings. Putnam on rational acceptability Hilary Putnam developed a kind of coherentism as part of the “internal realism” that he defended from 1976 until he gave it up in favor of “natural realism” in the early 1990s (see especially 1999). It is not clear, however, whether this 3 Williamson continues: “in response to questions or objections, he was defensive and guarded, doing little to articulate his ideas more explicitly,” but this is not entirely accurate in my opinion. Rather, I would say that Davidson’s “responses” were not responses at all: he rarely addressed the issues raised by those he was responding to. Nevertheless, in some “responses”, he did (somewhat) clarify his ideas, and consequently, these are valuable sources. Examples of this are many of the responses in Hahn (1999) and (Davidson 2001a). 4 For example, Ernest Lepore and Kirk Ludwig (2005) write about some apparently critical aspect of Davidson’s theory of triangulation that “the suggestion is not spelled out enough for it to be clear what is intended” (408n), and Peter Pagin’s (2001) influential analysis of that same theory is littered with statements like “if I understand him” and other expressions of uncertainty and puzzlement. 5 “… let me continue in my breathless way through one more chapter” (1988a: 43). 53 change also implies a rejection of his coherence theory, or merely an amendment. I will interpret it as the latter below (after introducing Davidson’s theory), and will briefly introduce the original, internal realist version of Putnam’s coherentism in this section. Before proceeding, it must be mentioned, however, that Putnam did not use the term “coherence theory” himself. In fact, he rejected the label “coherence theory of truth” because of its (to him) unacceptable connotations. In Reason, Truth and History (1981a) Putnam argues that there are two opposing philosophical perspectives. One is that of metaphysical realism or externalism, which assumes that “the world consists of some fixed totality of mindindependent facts” and that “there is exactly one true and complete description of ‘the way the word is’” (49). The externalist theory of truth is correspondentist: truth is some kind of correspondence relation of words (etc.) with external things. Externalism requires and assumes the possibility of an external point of view, which Putnam aptly characterizes as a “God’s Eye point of view” (id.). The other perspective – the one Putnam preferred – is that of internalism or internal realism. According to internalism, a question like “What objects does the world consist of?” can only be sensibly asked (and answered) “within a theory or description” (id.), and an internalist theory may hold “that there is more than one ‘true’ theory or description of the world” (id.). About the notion of truth in the internalist perspective, he wrote that: ‘Truth’ … is some sort of (idealized) rational acceptability – some sort of ideal coherence of our beliefs with each other and with our experiences as those experiences are themselves represented in our belief system – and not correspondence with mind-independent or discourse-independent ‘states of affairs’. There is no God’s Eye point of view that we can know or usefully imagine; there are only the various points of view of actual persons reflecting various interests and purposes that their descriptions and theories subserve. (1981a: 49-50; italics in original)6 6 54 On “truth” in internal realism, see also (Putnam 1983a; 1983b). Putnam and Davidson on Coherence, Truth, and Justification While Putnam noted that internalism does not necessarily hold that there are multiple true descriptions of the world, his internal realism does hold this claim. The externalist (or Aristotelian/medieval/Kripkean essentialist) idea of selfidentifying objects makes no sense.7 Rather, “‘objects’ do not exist independently of conceptual schemes. We cut up the world into objects when we introduce one or another scheme of description” (52). And there are multiple ways to do that “cutting up”: there are (at least sometimes) “equally coherent but incompatible conceptual schemes which fit our experience equally well” (73). That doesn’t mean that “anything goes”, however, or that any conceptual scheme is as good as any other. Internal coherence is not the only constraint on knowledge – there are “experiential inputs” to knowledge as well. These inputs are partially shaped by our concepts (or conceptual schemes), but only partially, and rational acceptability of some set of beliefs also requires coherence with these experiential inputs. What makes a statement, or a whole system of statements – a theory or conceptual scheme – rationally acceptable is, in large part its coherence and fit; coherence of “theoretical” of less experiential beliefs with one another and with more experiential beliefs, and also coherence of experiential beliefs with theoretical beliefs. (1981a: 54-5) Despite the above quoted claim that, for the internalist, truth is some sort of rational acceptability, Putnam rejects the identification of truth with rational acceptability. “Truth cannot simply be rational acceptability for one fundamental reason; truth is supposed to be a property of a statement that cannot be lost, whereas justification can be lost” (55). In other words, it is justification that is identified with rational acceptability (and thus coherence) rather than truth, which means that Putnam’s coherentism is a coherence theory of justification. 7 See also (1981b) and (1983a): “The idea … – i.e., that nature itself determines what our words stand for – is totally unintelligible. At bottom, to think that a sign-relation is built into nature is to revert to medieval essentialism, to the idea that there are ‘self-identifying objects’ and ‘species’ out there” (xii). 55 Nevertheless, Putnam does associate truth with “some sort of” rational acceptability: “truth is an idealization of rational acceptability. We speak as if there were such things as epistemically ideal conditions, and we call a statement ‘true’ if it would be justified under such conditions” (55; underlining added). 8 “Truth” is not defined as idealized rational acceptability (or “idealized justification”; p. 122) either, however. The logical nature of underlined “is” and “if” in the last quote is that of coincidence and material equivalence respectively, not that of conceptual identity and logical equivalence. The relation between truth and idealized justification must be one of material equivalence because it makes no sense to suppose that a statement could be justified under ideal epistemic conditions and false nevertheless,9 but there is no reason to suppose it is stronger than material equivalence. This raises the question how – according to Putnam – “truth” should be defined, but unfortunately, in Putnam’s writings of the last two or three decades it tends to be (much) clearer what he rejects than what (exactly) he accepts, and this is the case for “truth” as well. Davidson’s “coherence theory” Donald Davidson defended his brand of coherentism most explicitly in “A Coherence Theory of Truth and Knowledge” (1983), but it is connected to many other of his writings and ideas. Despite the title of the ’83 paper, Davidsonian coherentism is not a coherence theory of truth and/or knowledge. As Davidson 8 In (1983b) he writes: “A statement is true, in my view, if it would be justified under epistemically ideal conditions” (84). 9 Putnam first made the claim that “the supposition that even an ‘ideal’ theory (…) might really be false appears to collapse into unintelligibility” in (1976). That it is, is easy to see if one considers the implications of this supposition (but it should be noted that the following is not Putnam’s argument). If there would be something that is undetectable even under ideal epistemic conditions, then that (kind of) thing would be absolutely isolated from anything in the ideally detectable universe, meaning that it would have no effect or influence of any other kind on anything (in the same ideally detectable universe). That would require it to be non-physical and without physical causes or effects. Hence, it would not just be unknowable, but also utterly mysterious, and given its absolute lack of causal powers also completely irrelevant. The idea of supposing such a (kind of) thing indeed makes no sense. 56 Putnam and Davidson on Coherence, Truth, and Justification later admitted (1987b), the title is misleading and chosen badly. The sloganesque phrase “coherence yields correspondence” (1983: 137) in the first paragraph is equally misleading, but at the same time oddly accurate as a summary of Davidson’s argument. What makes it misleading – like the paper’s title – is Davidson’s idiosyncratic use of terms like “coherence” or “coherence theory” and “correspondence”. Firstly, Davidsonian coherentism is not a coherence theory of truth because there can be no theory of truth, or at least no theory that is not “empty or wrong” (1987b: 155). Truth is a primitive, it is “as clear and basic a concept as we have” (id.). A theory of truth as a definition of truth and/or a specification of what makes a belief true adds nothing to our understanding of what truth is. “Truth is correspondence with the way things are” (1983: 139), but correspondence does not make a belief true: correspondence is what “truth” means, not what truth requires. According to (the most common version of) the correspondence theory, correspondence to facts makes beliefs or sentences true. Davidson rejects the truth-making element thereof, but also the specific idea of facts as truth-makers. According to Davidson, there are no distinct facts; there is at most one fact (1969), and consequently, “there is nothing for sentences to correspond to” (1987b: 183). Hence, the intentionally vague phrasing as “correspondence with the way things are”. Secondly, Davidson rejects the empiricist notion of justification of beliefs by sense data (or something similar). This argument is related to his rejection of conceptual schemes and the “third dogma of empiricism” in (1974), and further elaborated in “The Myth of the Subjective” (1988a). The core idea is that the relation between sensations and beliefs is causal: sensations cause beliefs, but a cause is not a justification. Furthermore, beliefs are propositional and the only thing that can justify a proposition is another (set of) proposition(s). Sensation, however, is not propositional and there is no special class of intermediate beliefs such as “observation sentences”. If there is a distinction between observation sentences and other kinds of beliefs, it is a distinction of their causes, not of their 57 justifications (1983: 145-6). Regardless of the kind or nature of a belief, “nothing can count as a reason for holding a belief except another belief” (141). Davidson suggests a few times in the first pages of his “A Coherence Theory” that coherence is a test of truth, but these suggestions should not be confused with the idea that coherence makes a belief true, however – that would be getting things the wrong way around. Coherence can be a test of truth because according to Davidson, each of our beliefs may be false, but not all of them can be wrong (140), and consequently, coherence with our necessarily mostly true beliefs, justifies the belief that a belief is true. More precisely: coherence does not make a first-order belief true, but justifies a second-order belief x that a first-order belief y is true. Hence, Davidson’s theory is a coherence theory of justification. The question, of course, is why coherence would justify the second-order belief that a first-order belief is true. “Mere coherence, no matter how strongly coherence is plausibly defined, cannot guarantee that what is believed is so” (1983: 138). However, it can and does if it is the case that “most of the beliefs in a coherent set of beliefs are true” (id.). According to Davidson this is the case because “belief is in its nature veridical” (146). In “a Coherence Theory” (and elsewhere) he offered two closely related arguments for this claim. What relates the two arguments is that both depend on the triangle of speaker, interpreter, and a shared world. In Davidson’s early writings this triangle plays a central role in his theory of radical interpretation; in his later writings the triangular figure is used to explain intersubjectivity, among others. The “omniscient interpreter” argument for the veridicality of belief (which first appeared in 1977) is based on the former, and most commentators focus on this argument. However, Davidson himself later rejected it as a “sortie into science fiction” (1999b: 192). The second – and more important – argument is closely related to the aforementioned causal theory of mental content and belief. Davidson summarizes this argument in a key passage near the end of “A Coherence Theory” as follows: 58 Putnam and Davidson on Coherence, Truth, and Justification What stands in the way of global skepticism of the senses is, in my view, the fact that we must, in the plainest and methodologically most basic cases, take the objects of a belief to be the causes of that belief. And what we, as interpreters, must take them to be is what they in fact are. Communication begins where causes converge: your utterance means what mine does if belief in its truth is systematically caused by the same events and objects. (1983: 151) Communication and language learning require the same triangle of speaker (or teacher), interpreter (or learner), and a shared world as that required by (radical) interpretation. The interpreter/learner can not make sense of a word if she does not share an experience of the object or event referred to by the speaker/teacher in using that word. At the most basic level, our words are (and must be) grounded through a social process in a shared, real, external world. Although this is the same triangle as that of radical interpretation, Davidson’s theory of “triangulation” or “ostensive learning” turns that triangle on its head. The theory has weaker and stronger versions or applications. Weaker versions are intended to explain the source and possibility of intersubjective truth (e.g. 1982a). Stronger versions elaborate Davidson’s claim that “successful communication proves the existence of a shared, and largely true, view of the world” (1977: 201). In its strongest form, the theory of triangulation can be interpreted as a transcendental argument: from the fact that we do communicate it follows that there is such a shared world, that there are other minds, and that our beliefs about the shared world are mostly true (e.g. Davidson 1987a; 1991; see also Sosa 2003).10 It’s a variant of the stronger form that provides the second argument for the veridicality of belief: the facts that we do have language and are able to communicate prove that most of our most basic beliefs must be true. 10 Davidson tended to retreat from stronger versions of triangulation to weaker ones in response to criticism. See, for example, various replies in Hahn (1999) and (Davidson 2001a). 59 conceptual schemes There is much that Putnam and Davidson would have been able to agree upon (provided that differences in background, style, and terminology wouldn’t obstruct mutual understanding). Most fundamentally, both argue that – as Putnam (1981a) phrases it – “the notion of comparing our systems of beliefs with unconceptualized reality to see if they match makes no sense” (130), and that we can only justify our beliefs (etc.) by appealing to other beliefs and coherence therewith. But there are also important differences, and there are problems especially in Putnam’s theory, which lead him to give up “internal realism” in the 1990s (see next section). The most conspicuous difference between Putnam’s internal realism and Davidson’s philosophy concerns conceptual schemes. Putnam claims that description is relative to conceptual schemes and that there are “incompatible conceptual schemes”, while Davidson famously rejects the “very idea” of conceptual schemes in (1974). The extent of difference on this issue should not be exaggerated, however. The gap is made to look wider than it is by terminological and stylistic differences, and by established preconceptions about Putnam and Davidson; it can be narrowed by looking below the surface. Davidson did not reject all notions of conceptual schemes (and thus not “the very idea”, despite the title of 1974), but only a particular notion of conceptual schemes that he ascribed to Whorf, Kuhn, and Quine, among others. 11 That particular notion is one of untranslatable schemes. Additionally, he rejected an idea that he assumed to be inherently related to the notion of conceptual schemes and that he called “the third dogma of empiricism”: the idea of epistemic intermediaries between the world and our minds. The latter rejection was further developed in “The Myth of the Subjective” (1988a). (We’ll turn to this topic in the next section.) 11 In (Brons 2011) I show that Davidson misinterpreted Whorf and Kuhn and claimed that he only rejected a Quinean notion of conceptual schemes. I have come to believe that the latter claim is wrong, and that Davidson misunderstood Quine as well. 60 Putnam and Davidson on Coherence, Truth, and Justification That Davidson did not completely reject the notion of conceptual schemes is evidenced by his own occasional use of the term. For example, in “Seeing Through Language” (1997) he recognized that there are (or can be) “differences or provincialisms in our conceptual schemes. But these are variants or features we can explain to one another, or could, given enough time, adequate attention, and sufficient intelligence on both sides” (128; see also 1999c). In (1974) and elsewhere Davidson rejected the idea of untranslatable or incommensurable conceptual schemes, but even though the above quoted term “incompatible” may seem to suggest otherwise, Putnam (1981a) rejected incommensurability between schemes as well (see 114ff). Hence, the difference between Putnam and Davidson with regards to schemes is not one of rejection versus acceptance, but one of the degree of possible difference between schemes. According to Quine (1960) different conceptual schemes commit us to the existence of different things; that is, different schemes posit different things. Putnam and Davidson appear to disagree about our freedom in positing things. More specifically, they appear to give different answers to the question whether our posits (or positings) follow inherent “joints” in nature or external reality. “Appear to” because Putnam’s answer to this question is clear – it is a negative answer12 – but Davidson’s is not. One of Davidson’s arguments against conceptual schemes in (1974) is that those “organize” what is already organized. This argument makes sense only if it is interpreted as meaning that nature or reality provides that prior organization. His arguments in (1992) and (1993b) similarly seem to depend on the presupposition of an external reality consisting of (or pre-organized into) discrete objects and events, and in (1999d) Davidson argued for the existence of “divisions in nature” explicitly: 12 According to Putnam there are no joints or self-identifying objects: “we cut up the world into objects” (1981a: 52). See also above. 61 Nature is pretty much how we think it is. There really are people and atoms and stars, given what we mean by the words. The infertility of hybrids defines real species, though this matters only to those interested in the relevant concepts. This explains why it is foolish to deny that these divisions exist in nature, whether or not anyone entertains the thought. Even if no one had ever had a concept, there would be species, though of course this is our concept and our word, born of our interests. (38) Furthermore, Davidson’s notion of causality as a law-like relation between kinds of events presupposes that events come in discrete natural kinds (and perhaps even kinds with causal essences). And the fact that he called the irreducibility of kinds of mental events to kinds of brain events “anomalous monism” strongly suggests that he not just took natural kinds for granted, but their reducibility to more basic, physical kinds as well (otherwise there wouldn’t be anything “anomalous” about the mind), and as John Dupré (1993) has shown, such reducibility requiries natural kinds to have structural essences. Although this may seem to settle the question – Davidson’s answer to the question whether our posits follow inherent joints in nature was positive – that may be too hasty. None of the “textual evidence” is unambiguous, and Samuel Wheeler (2014), for example, has argued that Davidson actually held the opposite position (i.e. that he denied joints in reality). Furthermore, Frank Farell (1994), Jeff Malpas (2001b), and me (Brons 2012; 2013) have defended interpretations or extensions of Davidson’s philosophy that give (moderately) negative (i.e. joint-denying) answers as well. If these interpretations are right (or these extensions don’t deviate all too radically from their source), then the difference between Putnam and Davidson on the matter of conceptual schemes turns out to be very small; but even if they are wrong, Putnam and Davidson are considerably closer than they seemed at first hand. 62 Putnam and Davidson on Coherence, Truth, and Justification convergence In the early 1990s Putnam became increasingly aware of a number of problems with internal realism, which resulted in his rejection of that theory in favor of “natural realism” (in or before the Dewey Lectures given in 1994 and published in 1999). Some of these problems relate directly to his coherence theory, but as mentioned before, it is not clear whether the switch to natural realism implies a rejection or amendment of his coherentism. Perhaps, the lack of a clear positive theory in Putnam’s writings of this period should be taken to mean that he intended natural realism to be an amendment rather than a substitute. Because in the present context it is more productive to treat it as an amendment indeed, that is what I will assume it is. The most common objection to coherence theories is that they seem to imply that any coherent system of beliefs is as good as any other coherent system of beliefs. Putnam (1981a) and Davidson had different answers to that objection. Putnam’s answer was that the choice between conceptual schemes is constrained by experiential inputs which are partially conceptual, and that we judge whole schemes by – among others – their coherence with experiential inputs. But this suggestion results in a dilemma. If these experiential inputs are (even partially) conceptual indeed, then they are already (largely) determined by a scheme and thus cannot function as scheme-independent constraints. If they are not conceptual (to a significant extent) then this (testing for) coherence with experiential inputs is nothing but another variety of “comparing our systems of beliefs with unconceptualized reality to see if they match” (1981a: 130), which “makes no sense” (id.). Putnam (1999) opted for the first horn of this dilemma, and thus gave up on the idea of external coherence of a scheme with experience (or experiential beliefs or inputs).13 13 It is a peculiar twist that what lead to the rejection of internal realism is that it wasn’t (consistently) internal enough. 63 A few pages back I mentioned that Davidson’s (1974) rejection of the notion of conceptual schemes was related to his rejection of “the third dogma of empiricism”: the idea of epistemic intermediaries between the world and our minds. In The Threefold Cord (1999) Putnam focuses his attention on perception to develop a similar view. He characterizes the belief in an interface between the mind and the (external) world as “Cartesian” and “disastrous” (43). In perception we have direct and unmediated contact with the world, but it is a conceptual contact. Davidson similarly rejects the idea of non-conceptual (or even nonpropositional) perception. “What the senses ‘deliver’ (i.e., cause) in perception is perceptual beliefs”, he argues (1999a: 106). And both authors argue that with giving up the “disastrous” idea of an intermediary between our mind and the world, there is no room left for skepticism or massive error. According to Davidson, “it is impossible for most of our perceptual beliefs to be false” (1998b: 189). Hence, for both Davidson and Putnam (real)14 perception (or perceptual belief) is inherently veridical as well as conceptual.15 It is from this idea that Davidson’s answer to the aforementioned most common objection to coherence theories follows: internal coherence is sufficient. 16 Because most of our (most basic) beliefs are necessarily true (because they are caused by the world), any completely internally coherent set of beliefs must be true. And therefore, internal coherence (with other beliefs) is sufficient to justify (the belief in the truth of) some particular belief. Putnam (1999) gives 14 As opposed to illusions and hallucinations. Putnam (1999) rejects the common suggestion that these are indistinguishable from perception. 15 There are differences, however. For example, Davidson describes the relation between the world and our perceptual beliefs as causal, while Putnam appears to reject this suggestion. The disagreement may be largely terminological, however, although this is difficult to judge as neither author is sufficiently clear on this issue. 16 Davidson attributes this “most common objection” to Moritz Schlick in (1982b). He quotes Schlick arguing that “the coherence theory is shown to be logically impossible … for by means of it I can arrive at any number of consistent systems of statements which are incompatible with one another”, to which he replies that “it’s not clear what it means to say I could ‘arrive’ at various systems, since I do not invent my beliefs; most of them are not voluntary” (173). 64 Putnam and Davidson on Coherence, Truth, and Justification approximately the same answer for approximately the same reasons: the veridicality of perception implies that internal coherence is sufficient. Putnam’s philosophical development can be characterized as a step-wise process in which more constructive phases of theory development and elaboration are interspersed with shorter, more destructive phases resulting from the realization that the theory developed and elaborated in the preceding constructive phase can no longer be maintained (see introduction). Hence, a comparison between “middle” and “late” Putnam is a comparison between two constructive phases, which means an analysis of the destructive phase in between. Putnam gave up his (1981a) “middle” coherence theory because it was incoherent in requiring what it rejected, namely to compare (systems of) beliefs to reality. In (1999) he suggests that he “went astray” because he “was still assuming something like the sense datum picture” (18), the “disastrous idea that has haunted Western philosophy since the seventeenth century” (43). The key point of Davidson’s coherence theory – simultaneously accurately and misleadingly summarized as “coherence yields correspondence” – is that coherence justifies the belief that a belief is true, and truth means correspondence. The former is the case because most of our (most basic) beliefs are necessarily true; the latter is primitive.17 Between (1981a) and (1999) Putnam gradually moved closer to Davidson. 18 If (1999) is read as an amendment of his (1981a) coherentism, then the differences between his later coherentism and Davidson’s are small and subtle. 17 Or more accurately, the concept of truth is a primitive, and that primitive concept means a kind of correspondence with the way things are. 18 Unfortunately not just in substance, but in style as well, considering that some of Putnam’s later writings surpass even Davidson in inscrutability. 65 But is it coherentism? Putnam’s and Davidson’s coherence theories have been largely ignored in the debate on coherentism about justification of the last decades. 19 Since the end of the 1990s much of that debate is phrased in probabilistic terms. One of the most influential arguments against a coherence theory of justification is presented by Erik Olsson in (2002). It’s most recent version can be found in Olsson (2012). The following summary is based on that version. If E1 is the proposition that witness 1 reports that A and E2 is the proposition that witness 2 reports that A, then coherentism claims that: [Coherence Justification] P(A | E1, E2) > P(A) . However, if the two testimonies are independent from each other, then it is the case that: [Conditional Independence] P(E2 | E1,A) = P(E2 | A) & P(E2 | E1,¬A) = P(E2 | ¬A) . And if coherentist justification and foundationalist justification are mutually exclusive, then the following applies: [Non-foundationalism] P(A | E1) = P(A) & P(A | E2) = P(A) . And from [Conditional Independence] and [Non-foundationalism] it follows that: P(A | E1, E2) = P(A) , which contradicts [Coherence Justification]. Therefore, coherentism is wrong. Of course, neither Putnam nor Davidson ever responded to this objection, but what their response would be can be easily inferred from the foregoing. Putnam’s (1999) thesis of the veridicality of perception and Davidson’s thesis of 19 For example, in Olsson (2005) Putnam is mentioned only once in a footnote and Davidson is quickly discarded after a brief recapitulation of the usual misunderstandings about his view. 66 Putnam and Davidson on Coherence, Truth, and Justification the veridicality of (perceptual) belief imply a rejection of [Non-foundationalism] in favor of something like: [Weak Foundationalism] P(A | E1) > P(A) & P(A | E2) > P(A) , provided that the probabilistic framework and conceptualization in terms of witness reports would make sense to them, of course (which is far from obvious). And from [Conditional Independence] and [Weak Foundationalism] the conclusion [Coherence Justification] does follow. However, if Putnam’s and Davidson’s theories adhere to something like Foundationalism], [Weak then they are not coherence theories, because coherentism supposedly adheres to [Non-foundationalism]. Rather, if the theses of the veridicality of perception and belief are interpreted as a kind of (weak) foundationalism, then Putnam’s and Davidson’s theories of justification are hybrids between coherentism and foundationalism,20 and are, therefore, more closely related to other hybrids such as those defended by Susan Haack (2009) 21 and Paul Thagard (2000) than to the “pure” coherentism that Olsson and others argue against. This shouldn’t come as a surprise, however, considering that Davidson wrote about “a pure coherence theory” that “perhaps no one has ever held such a theory, for it is mad” (2005a: 43). A potentially more serious problem for Davidson’s and Putnam’s coherence theories (I’ll continue to call them that, even if they really are hybrids) than that posed by Olsson is the ambiguity of the notion of coherence. Neither Putnam, nor Davidson ever specified what exactly coherence means and/or requires in this context. What does it mean for a set of beliefs to be coherent? 20 Not just Putnam’s and Davidson’s epistemological theories are hybrids; so is their metaphysics. Both opt for a hybrid of – or intermediate position between – realism and anti-realism. Davidson (2005a) rejects both realism and anti-realism and argues that these are not the only options, and Putnam (1999) aims for what he calls “a middle way between reactionary metaphysics and irresponsible relativism” (5). 21 Haack spends many pages arguing against Davidson’s coherentism, but in those pages she only rejects the “omniscient interpreter” argument, which Davidson later rejected himself. 67 Mere non-contradiction (in a logical sense) is usually (if not universally) considered to be insufficient. If there is a set of beliefs B1 about a set of “things” T1, and another set of beliefs B2 about T2, such that there is no contradictory belief in B1 and no contradictory belief in B2, and the intersections of B1 and B2 and of T1 and T2 are empty, then the union of B1 and B2 will contain no contradictory beliefs. It will, however, consist of two subsets that are completely isolated from each other. Any belief that is isolated in this sense will be perfectly non-contradictory with any set of beliefs it is isolated from, but this means that any completely unconnected belief is coherent – by this definition of “coherence” – and thus justified. For Davidson, however, this objection to identifying coherence with noncontradiction is irrelevant for two reasons. Firstly, the objection assumes the existence of discrete, individual beliefs, but according to Davidson there is no meaningful way to individuate beliefs (1983). And secondly, even if you could individuate beliefs, there could be no isolated beliefs (in the above sense): Davidson’s holism entails that every one of our beliefs is necessarily related to many others (e.g. 1999d; 1995). The brief considerations in this last section show that Putnam’s and Davidson’s coherence theories aren’t just of historical interest, but are viable alternatives to the positions commonly recognized by mainstream analytic epistemology (although their dependence on unfashionable views such as – but not limited to – anti-essentialism in case of Putnam and social externalism (i.e. triangulation) and holism in case of Davidson make it doubtful that they will be welcomed or even understood by mainstream analytic philosophers). There may very well be other problems for their theories, of course, but that is a matter for further research. 68 Putnam and Davidson on Coherence, Truth, and Justification references Brons, L.L. 2011. “Applied Relativism and Davidson’s Arguments Against Conceptual Schemes. The Science of Mind 49: 221-40. —— 2012. “Dharmakīrti, Davidson, and Knowing Reality”. Comparative Philosophy 3.1: 30-57. —— 2013. “Meaning and Reality: a Cross-Traditional Encounter”. In Bo Mou & R. Tieszen Eds. Constructive Engagement of Analytic and Continental Approaches in Philosophy. Leiden: Brill. pp. 199-220. —— 2014. Book review of: Samuel Wheeler III (2014), Neo-Davidsonian metaphysics: from the true to the good. In International Journal of Philosophical Studies 22.2: 286-91. Davidson, D. 1969. “True to the Facts”. In (1984). pp. 43-54. —— 1974. “On the Very Idea of a Conceptual Scheme”. In (1984). pp. 183-98. —— 1977. “The Method of Truth in Metaphysics”. In (1984). pp. 199-214. —— 1982a. “Rational animals”. In (2001b). pp. 95-105. —— 1982b. “Empirical Content”. In (2001b). pp. 159-76. —— 1983. “A Coherence Theory of Truth and Knowledge”. In (2001b). pp. 137-53. —— 1984. Inquiries into Truth and Interpretation. Oxford: Clarendon. —— 1987a. “Knowing One’s Own Mind”. In (2001b). pp. 15-38. —— 1987b. “Afterthoughts”. In (2001b). pp. 154-8. —— 1988a. “The Myth of the Subjective”. In (2001b). pp. 39-52. —— 1988b. “Epistemology and Truth”. In (2001b). pp. 177-92. —— 1990. ‘Epistemology externalized’. In: (2001b). pp. 193-204. —— 1991. “Three Varieties of Knowledge”. In (2001b). pp. 205-20. —— 1992. “The Second Person”. In (2001b). pp. 107-22. —— 1993a. “The Third Man”. In (2005b). pp. 159-66. —— 1993b. “Locating Literary Language”. In (2005b). pp. 167-81. —— 1995. “The Problem of Objectivity”. In (2004). pp. 3-18. —— 1997. “Seeing through Language”. In (2005b). pp. 127-42. —— 1999a. “Reply to John McDowell”. In Hahn (1999). pp. 105-8. —— 1999b. “Reply to A.C. Genova”. In Hahn (1999). pp. 192-4. —— 1999c. “Reply to Stephen Neale”. In Hahn (1999). pp. 667-9. —— 1999d. “Interpretation: Hard in Theory, Easy in Practice”. In M. De Caro Ed. Interpretations and Causes. Dordrecht: Kluwer. pp. 31-44. —— 2001a. “Comments on Karlovy Vary Papers”. In Kotatko et al. (2001). pp. 285-307. —— 2001b. Subjective, Intersubjective, Objective. Oxford: Oxford University Press. —— 2004. Problems of Rationality. Oxford: Oxford University Press. —— 2005a. Truth and Predication. Cambridge MA: Belknap. —— 2005b. Truth, language, and history. Oxford: Oxford University Press. Dupré, J. 1993, The Disorder of Things: Metaphysical Foundations of the Disunity of Science. Cambridge MA: Harvard University Press. 69 Farell, F.B. 1994. Subjectivity, Realism and Postmodernism: the Recovery of the World in Recent Philosophy. Cambridge: Cambridge University Press. Haack, S. 2009. Evidence and Inquiry: a Pragmatist Reconstruction of Epistemology. New York: Prometheus. Hahn, L.E. Ed. 1999. The philosophy of Donald Davidson. Chicago: Open Court. Kotatko, P., P. Pagin, & G. Segal Eds. 2001. Interpreting Davidson. Stanford: CSLI. Lepore, E. & K. Ludwig 2005. Donald Davidson: Meaning, Truth, Language, and Reality. Oxford: Oxford University Press. Malpas, J. 2011a. “Introduction” in Malpas, J. ed.. Dialogues with Davidson. Cambridge MA: MIT Press. pp. xvii-xxv. —— 2011b. “Triangulation and philosophy: a Davidsonian landscape”. In Amoretti, M.C. & G. Preyer eds.. Triangulation from an Epistemological Point of View. Frankfurt: Ontos. 257-279. Olsson, E.J. 2002. “What is the Problem of Coherence and Truth?”, The Journal of Philosophy 99: 246-72. —— 2005. Against Coherence: Truth, Probability, and Justification. Oxford: Clarendon. —— 2012. “Coherentist Theories of Epistemic Justification”. In E.N. Zalta Ed. The Stanford Encyclopedia of Philosophy, Spring 2014 Edition. <http://plato.stanford.edu/archives/spr2014/entries/justep-coherence/> . Pagin, P. 2001. “Semantic Triangulation”. In: Kotako, Pagin, & Segal (2001). pp. 199-212. Putnam, H. 1976. “Realism and Reason”. In 1978. Meaning and the Moral Sciences. London: Routledge. pp. 123-40. —— 1981a. Reason, Truth and History. Cambridge: Cambridge University Press. —— 1981b. “Why There Isn’t a Ready-Made World”. In (1983c). pp. 205-28. —— 1983a. “Introduction”. In (1983c). pp. vii-xviii. —— 1983b. “Reference and Truth”. In (1983c). pp. 69-86. —— 1983c. Realism and Reason: Philosophical Papers Volume 3. Cambridge: Cambridge University Press. —— 1985. “After Empiricism”. In 1990. Realism with a Human Face. Cambridge MA: Harvard UP. pp. 43-53. —— 1999. The Threefold Cord: Mind, Body, and World. New York: Columbia University Press. Quine, W.V.O. 1960. Word & Object. Cambridge MA: MIT Press. Sosa, E. 2003. “Knowledge of self, others, and world”. In K. Ludwig Ed. Donald Davidson Cambridge: Cambridge University Press. pp. 163-82. Thagard, P. 2000. Coherence in Thought and Action. Cambridge MA: MIT Press. Wheeler, S.C. 2014. Neo-Davidsonian Metaphysics: from the True to the Good. New York: Routledge. Williamson, T. & Chen Bo. 2009. “Thinking Deeply, Contributing Originally: an Interview with Timothy Williamson”. Annals of the Japan Association for Philosophy of Science 18: 57-87. 70