"The more the ordinary mind takes the opposition between true and false to be fixed, the more it is accustomed to expect either agreement or contradiction with a given philosophical system, and only to see reason for the one or the other in any explanatory statement concerning such a system. It does not conceive the diversity of philosophical systems as the progressive evolution of truth; rather, it sees only contradiction in that variety. The bud disappears when the blossom breaks through, and we might say that the former is refuted by the latter" G.W.F Hegel1
Truth, language, and reality: an essay on Donald Davidson’s philosophy of truth.
Dissertation for BA Philosophy, Candidate Number: B01050, Heythrop College, University of London, 2010.
Table of Contents:
The Nature of Truth 3
Realism and the Correspondence Theory 4
Objections: sceptic and explanatory 5
Epistemic Theories 8
Objections: absurd and explanatory 10
Davidson on Truth 12
Concluding Remarks 15
Introduction: Early on the preface to his Phenomenology Hegel brushes on a metaphilosophical reflection which characterizes the line of thought I wish to pursue in this essay. Fruitful investigation of a philosophical system demands tracing historical veins of thought, intuitions, and acquisitions. A theory which pays no heed to its own genealogy or which manifests itself without articulating its own genesis is hard pressed to find dialectical leverage.
It is a platitude to admit that the concept of truth plays a large role in our epistemological or metaphysical outlook. Despite its surface conspicuousness it is by no means evident how to explicate the intricacies of this relationship. The nuanced interplay between truth, ontology, and knowledge surfaces quite demonstrably in the Modern period and in 20th Century analytic philosophy, where debates between Realism and idealism are largely motivated by sceptical worries or worries of explanatory vacuity, debates which Rorty2 has characterized as between systems of ideal representation and systems of ideal coherence. Donald Davidson appears to me to be able to offer us a way of characterizing truth which promises to avoid many of the interminable debates which surface from these historical periods.
In what follows I shall attempt to re-trace these sceptical and explanatory concerns which appear in the dialectic between two of the most dominant theories of truth, Realist and epistemic theories, distil from them their motivational impetus’, and demonstrate how Davidson’s work provides a way of accommodating the intuitive force of both whist evading their disquieting epistemological or explanatory repercussions. Realist theories hold onto the idea that truth is objective and mind-independent but, as we have learned from Berkeley, Hume, and Kant, this provokes scepticism about our ability of ascertain anything about what is true. Epistemic theories salvage the idea we can have knowledge of what is true but, as we have learned from Plato to Russell, provoke scepticism that the world is merely a reification of our means of justification. As Wright has put it “we want the mountain to be climbable, but we also want it to be real”3. Moreover, as Ramsey and Gödel have shown, most definitions of truth end up falling into explanatory vacuity.
Although, as it will become apparent, Davidson’s theory is not altogether without objections concerning its own ability to address scepticism, if I am correct it is already in a better position to address these worries purely in virtue of being able to accommodate the intuitions driving both theories.
The nature of truth:
Aristotle in the Metaphysics has been cited by Davidson and Tarski as giving the most foolproof characterization of truth in saying:
“To say of what is that it is not, or of what is not that it is, is false, while to say of what is that it is, or of that is not that it is not, is true” Metaphysics Γ, 7, 27.
This reiterates what Socrates and Hermogenes agree to in Plato’s Cratylus:
“And those [statements] that say of the things that they are that they are, are true, while those that say of the things that they are that they are not, are false” 387b2
The general theme in both these characterizations seems to be a relation between entities with truth-conditions or judgeable contents (e.g. propositions, beliefs, utterances, etc.) and something that provides the satisfaction of conditions for judgement. It is from this minimal characterization that the Realist, epistemic, Davidson’s theory will find common ground.
Realism and The Correspondence Theory: Philosophers often find in the intuitive notion of truth the idea of correspondence to facts or states of affairs; which has come to constitute the correspondence theory of truth. Despite the numerous variations of the correspondence theory over the centuries most can be summed up as holding that something is true if and only if it corresponds to some state of affairs of fact in reality.
Plato’s Forms can be seen as ushering an early implicit utilization of the correspondence theory. In the Phaedo, Symposium, Theaetetus, and Republic4 we are reminded that sensible things are wrought with the compresence of opposites and consequently can’t be said to have being. But as we see from the argument in (476-480) of Republic V knowledge must be of “what is” or has being. Plato’s conclusion is that we can only have knowledge if it is of that which can be said to have being: Forms. What is arguably working in the background here is the assumption that truth (as a condition for knowledge) must consist of a correspondence between statements and Reality. This is important because the worry of scepticism which results from the Heraclitean flux of sensible particulars is largely what is motivating a robust Realist theory; a theme which will be reappear in the course of this essay. Aristotle in his earlier Categories5 makes more explicit the correspondence theory by isolating the relations that truth needs as between propositions and facts (although it’s interesting that “facts” or “states of affairs” play little role in his later inquiries in the Metaphysics).
Medieval philosophers such as Aquinas and Ockham characterized the nature of correspondence as metaphysical or semantic respectively but more or less assumed a version of the correspondence theory of truth. A significant development in the correspondence theory is in Empiricism’s utilization of phenomenal experience/sense data; which sought to keep the notion of correspondence to reality but buttressed our epistemic entitlement by placing experience in as an epistemic intermediary. Locke says of truth:
“[truth] consists in knowing what ideas the words stand for and perceiving the agreement or disagreement of those ideas according as it is marked by those words”6
This provides a sophisticated conception of truth as correspondence between signifier and signed with the epistemically foundational role being played by experience/perception.
The correspondence theory began to explicitly involve the notion of facts (construed as structured composite entities) after responding to Frege’s point that direct word-world relations alone don’t constitute an expression of a thought (or anything truth-evaluable). Only sentences can express something which is truth-evaluable. This manifests itself in the works of Russell, Moore, and early Wittgenstein. Moore encapsulates correspondence as:
“To say that this belief is true is to say that there is in the Universe a fact to which it corresponds; and that to say that it is false is to say that there is not in the Universe any fact to which it corresponds”7.
Russell, in the Problems of Philosophy:
“A belief is true when there is a corresponding fact, and is false when there is no corresponding fact...and this fact does not (except in exceptional circumstances) in any way involve the mind of the person who has the belief”8.
And Wittgenstein asserts the world be “the totality of facts not things” and continues to mention states of affairs and truth as being a relation to such entities9.
Objections: sceptical and explanatory One historically pervasive challenge faced by the correspondence theory lies in the epistemic consequences of holding such a theory. Epistemological concerns are arguably what motivated Berkeley’s subjective idealism. He argues in the Principles XVIII:
“But though it were possible that solid, figured, moveable Substances may exist without the Mind, correspondingto the Ideas we have of Bodies, yet how is it possible for us to know this? Either we must know it by Sense, or by Reason. As for our Senses, by them we have the Knowledge only of our Sensations, Ideas, or those things that are immediately perceived by Sense” (my italics)10 Berkeley’s argument is clearly motivated by concerns about for the role material facts play in what it is possible for us to know. He concludes that if there were mind independent facts we couldn’t know them. Moreover in the “master-argument”11 Berkeley gives us reason to suppose mind-independent facts are inconceivable.
Berkeley’s response to this worry was to advocate immaterialism, which retains correspondence but re-interprets that to which ideas correspond into the mental. Historically Berkeley’s solution (and phenomenalism’s for that matter), although a solution to this epistemological worry, replaces one form of scepticism for another, which turns reality into nothing but representations, and the desire for objective truth drives correspondence back to beyond the veil of perception.
Kant’s Critique expressed the culmination of this worry about the role of truth bearers to truth-aiming assertions:
“the possibility of such noumenal [judgement] is quite incomprehensible, and beyond the sphere of phenomena, all is for us a mere void; that is to say, we posses an understanding whose province does problematically extend beyond this sphere, but we do not possess an intuition, indeed, not even the conception of a possible intuition, by means of which objects beyond the region of sensibility could be given to us, and in reference to which the understanding might be employed assertorically”12.
Kant’s point is that the very idea of judgements between objects conceptualized by the categories of experience and things-in-themselves beyond the transcendental limits of experience is impossible. However once we grant this point it begins to seem harder to characterize the kind of robust transcendent correspondence sought by Realists. Kant’s solution of transcendental idealism present a way of meeting these sceptical worries without losing track of the world, but by the turn of the 20th century we find worries that it errs too much on the side of making the mountain climbable rather than real.
Logical atomism, endorsed by Russell, Moore, and early Wittgenstein, may meet the sceptical worries for those whom Hume left behind but Berkeley goes too far. Russell at one point13 proposes of a method of analysis in which assertions are broken down into their constitutive logical components and then these components (the truth conditions of which are provided by his logical apparatus of the Principia) are ultimately satisfied by logical atoms or facts. We can have faith that these facts are graspable (meeting Berkeley) because they are logical (rather than material.
But Kant in his Logic gives us reason to suppose even logical atomism might be fundamentally flawed. He points out:
“Truth is said to consist in the agreement of knowledge with the object. According to this mere verbal definition, then, my knowledge, in order to be true, must agree with the object. Now, I can only compare the object with my knowledge by this means, namely, by taking knowledge of it. My knowledge, then, is to be verified by itself, which is far from being sufficient for truth. For as the object is external to me, and the knowledge is in me, I can only judge whether my knowledge of the object agrees with my knowledge of the object.”14 Kant’s argument here goes beyond his metaphysical scruples with the idea of things-in-themselves. He is taking issue with the very idea of truth being apprehended by atomic correspondence. Blanshard argues against Russell by pointing this out:
“Suppose that such [logically atomic] facts exist, and ask how we can know them. We plainly do not sense them; to say we sense that A is B is mere confusion. We apprehend the Bness of A through the judgement, A is B. But judgements, as we have seen, are in principle capable of error; this judgement therefore stands in need of verification”15 The Pyrronian aspect of this objection becomes more and more evident; Agrippa’s five modes demonstrate how judgements can only be supported by more judgements, if not we either recourse to dogmatism, circularity, or unjustified claims. This point has been argued at even more length by Sellars16, Davidson17, Rorty18 and recently Fogelin, Stroud, and Williams19. Although the sophisticated correspondence theories of Russell, Moore, and Wittgenstein may elude the sceptical grasp felt so powerfully by their empiricist ancestors it does not easily evade the Kant style objection that the notion of corresponding to atomic facts is, in principle, epistemically incoherent.
This critique however is irrevocably tied to the idea that truth is an epistemic concept; that a theory of truth must ultimately tie into a theory of justification (an idea seemingly shared by most correspondence theories and their critics). Kant’s point can be taken to dissuade us from atomic epistemic foundations, but why necessarily extend it to the nature of truth? The correspondence theorist could reply that they aren’t proposing a theory of truth which is epistemic; they are merely offering a definition of the nature of truth. But now epistemic theories could argue that truth must be connected to justification, and so both theories are simply rejecting the first dialectical step of their opponent.
A second worry for the correspondence theory is that the notion of a fact or state of affairs lacks any clear explanatory purpose. This objection hinges purely on the idea that explanatorily empty theories aren’t philosophically enlightening theories. Davidson argues “the real objection to correspondence theories is simpler: it is that there is nothing interesting or instructive to which true sentences correspond”20.
One way of demonstrating the explanatory vacuity of the correspondence theory is found in a formal argument that has its roots in Frege, and has been reiterated and refined by Church, Gödel, Quine, Davidson, and Neale: the slingshot argument. This argument aims to show how (assuming facts refer to anything at all) either all true sentences/propositions correspond to the same thing, in which case we have failed to locate anything beyond truth for true statements to correspond, or there is a unique fact to which every true sentence corresponds which again means we have failed to locate anything for a sentence to correspond to other than a disquoted sentence.
Neale21 constructs the argument as follows22.
Suppose we grant the following assumptions:
A1: “Fa” and “a= (x)(x=a & Fx)” stand for the same fact: the unique x such that x is identical to a and x is an F (where a is a singular referring term and F is some arbitrary predicate)
A2: Compositionality: that the signification of an expression with constituents which themselves have signification depends only on the signification of those constituents
We can construct the argument. Suppose that the following three premises are all true:
Fa corresponds to fact F1
a ≠ b corresponds to fact F2
Gb corresponds to fact F3
It follows that:
a = (x)(x=a & Fx) corresponds to fact F1 (by A1)
a = (x)(x=a & x ≠ b) corresponds to fact F2 (by A1)
(x)(x=a & Fx) & (x)(x=a & x ≠ b) stand for the same fact (by 4,5, and A2)
F1 = F2 (by 1, 2, and 6)
b = (x)(x=b & Gx) corresponds to the fact F3 (by A1)
b = (x)(x=b & x ≠ a) corresponds to fact F2 (by A1)
(x)(x=b & Gx) & (x)(x=b & x ≠ a) stand for the same fact (by 8,9, and A2)
F2 = F3 (by 2,3, and 10)
F1 = F2 = F3 (by 7 and 11)
It follows that for any statement which is true, if it corresponds to a fact, it corresponds to the same fact as every other true expression. C. I. Lewis points out, if we accept denotation of propositions “the important extensional property of any proposition is its truth value”23 and “the intension of a proposition comprises whatever the proposition entails; and it includes nothing else”24. If this is the case the role of facts becomes negligible, if only for it means that in the end “x is a fact” indicates nothing more than “x is true”.
If we reject any assumption we would most likely want to reject A1. The principle of compositionality (A2) is something we are hard pressed to reject without undermining the whole of formal languages along with it. Rejecting A1 implies that we have to get rid of the idea that logically equivalent terms refer to the same fact, and thus must accept that facts are as fine grained as expressions. As Gödel pointed25 out Russell’s logical atomism (wielded by the technical apparatus of his theory of descriptions) can avoid the conclusion of the slingshot argument by simply rejecting that unique descriptions refer to anything at all (assumption A1). Russell can then provide an account of the ontological entities that must be present for a belief (defined as n-ary relational entities) to be true26.
A more simplified version of this argument, attributed to Ramsey27, proves problematic for Russell. If asked to specify which fact it is that makes a sentence true, and in virtue of what that fact makes a sentence true, it’s hard to avoid the conclusion that what makes it the case that something is a fact that “x” is simply that “x” is true. If the reason “Snow is white” is true is because there is a composite relation; belief (someone, snow, whiteness); and the reason that relation holds is ultimately that it is true that there is snow, whiteness, and a person’s believing snow is white, we have yet to explain what it is that makes the sentence “snow is white” beyond noting that it is true.
The historical debate over the correspondence theory leads to two general points. Firstly that after responding to the critiques of Berkeley and Kant the correspondence theorist ends up at a dialectical impasse with the epistemic theorist. Secondly that the correspondence theorist finds itself hard pressed to offer us an explanatory account for what truth is; which is exactly what the theory purports to accomplish.
Kant’s Critique of Pure Reason leaves us with very strong reasons to suppose judgements, characterized in terms of being verified by a correspondence between reality-in-itself beyond the senses are impossible. Putnam28 makes the point that “internalist” (epistemic) theories of truth generally accept Kant’s basic point about the limits of reason and the transcendental character of experience, and consequently the limits of reality (as we can conceivably describe, predict, analyse, and understand it).
Early coherence theories of truth take on board this Kantian insight and conceive of truth systematically in terms of the coherence of judgement within the structure of experience as a whole. H.H. Joachim conceived ideal knowledge “a system, not of truths, but of Truth”29. Bradley asserts “Truth is a mode of self-realization of myself and of the Universe in one”30 echoing Hegel’s conception of Absolute knowledge as the relationship between spirit and self-conscious selves. Blanshard, perhaps most digestible to the analytic mind, characterizes coherence as a system “in which every judgement is entailed, and was entailed by, the rest of the system”31.
Coherence theorists argue we must jettison idea we can take a judgement and then somehow, without reference to all other judgements, discern whether or not it is true. We can have reason to believe that some judgements are more likely to be true than others, but this is again in virtue of their coherence. As Blanshard conceives it coherence is not only the nature of truth, but it is the only test we can give to judging something’s likelihood of being true.
Pragmatic theories of truth adopt this notion, but characterize truth far more explicitly in terms of more humane concepts such as value, adaptation, usefulness, etc. James encapsulates the pragmatist’s central insight as “True ideas are those that we can assimilate, corroborate, and verify. False ideas are those that we can not”32. This is no great conceptual break from the coherence theory; at the very least because coherence would seem to imply relations between our judgements and the world such that usefulness, verifiability, etc., and ideally pragmatic judgements point in the direction of coherence (putting aside worries of incommensurability). The major point of difference between the pragmatic and the coherence theory is whether coherence or pragmatic concern is more conceptually basic.
In the writings of later Vienna Circle philosophers such as Neurath, Carnap, and Hempel we find a far more straightforwardly epistemic theory of truth. These theories begin with the assumption that philosophy is ultimately the logic of scientific language. However science gives us no unambiguous mark of truth. Rather science refines hypotheses on the basis of how they cohere with the rest of our operating hypotheses as well as theoretical virtues which mediate between them. As Neurath points out in his debate against Schlick “we can formulate several groups of content statements that are free from contradictions; among these groups, verified by protocol statements accepted by us, we make a selection on the basis of extra-logical factors”33 . The idea that we can have metaphysical truth which serves to justify all logical inquiry is seen as remnants from an outdated form of philosophy34. We ought to, if we accept this metaphilosophical characterization, redefine our concept of truth such that it more or less translates into mutatis mutandis35 “a statement is adopted as true if it is sufficiently supported by protocol statements”36.
Dummett37 has argued that many of the debates outlined earlier can be seen as debates over the nature of the truth conditions for given classes of statements. He has defended an “anti-realist” position with regards to truth which defines truth in terms of a statement’s verification conditions; where something is true when it can be proven within an accepted theoretical framework (the standard example for this is intuitionism in the philosophy of mathematics). In a similar vein Putnam has also endorsed a form of epistemic theory (internalist theory) which equates truth with what is coherent within an idealized conception of rationality38 and so places emphasis on idealized verification conditions.