"Putnam on Facts and Values"


In this paper I attempt to explicate Putnam's position on the fact-value distinction, while exposing one of his crucial presuppositions.  I attempt to reveal that contrary to his own assertions Putnam must accept some type of incommensurability thesis.  It is this thesis that does the heavy lifting for his denial of the fact value distinction.  I then present one argument against the incommensurability thesis.  The paper also shows Putnam's reliance on Carnap's principle of tolerance and the latter's internal-external distinction.  

Taking Hilary Putnam as a spokesman for a pragmatic conception of facts and values, I hope (a) to explicate his understanding of the issue and (b) to evaluate the presuppositions necessary for such an understanding.  Surely Putnam is correct when he states that the question of fact and value is one that even the "man on the street" must make a decision about;[1] the relevance of this topic is worn on its sleeve.  Swinburne is equally correct when he notes that there are three theses about the fact-value distinction.  He writes: "The first type of thesis is concerned with how we do use language, the second type with how we ought to use language, and the third type with what the world is like."[2]  With respect to the first type, Swinburne believes that two different ways of using languages can be unearthed.  Users of one way recognize a fact-value distinction, whereas users of the other way do not.  The second type of thesis enables us to make a decision about which way of using language is correct.  But as Swinburne points out the second type of thesis about fact and value depends on the third thesis.  The way we ought to use language can only be decided by the way the world is.  The second and third theses go hand in hand.  I think that Putnam agrees with the spirit of this claim, but not the letter.  What he disagrees over is that a clear distinction between how we ought to use language and how the world is can be drawn.  This disagreement is the heart of Putnam's argument for denying the fact-value distinction.

To say that science is concerned with what is and ethics is concerned with what ought to be is a platitude.  A more difficult issue is whether or not these two fields are wholly unrelated; Putnam suggests they are not.  By showing that science presupposes values and that these values cannot be reasonably jettisoned Putnam can both deny the fact-value distinction and maintain that values are objective (339).  That science presupposes some values is uncontroversial.  One such value is truth.  Science aims at truth.  But most do not regard truth as a robust ethical value akin to good.  Thus, the need to presuppose some values does not establish Putnam's strong conclusion.  The values have to be the right kind.  Putnam argues that truth is indeed a species of the right kind; his argument utilizes three premises, all of which are controversial.  The first is that truth is not correspondence to mind-independent reality (339).  The second is that truth is a function of what we hold to be rationally acceptable (340-41).  The third is that what we hold to be rationally acceptable is based on values akin to standard ethical ones (344, 347).  Thus, if it is the case that truth is not correspondence but is rather a derivative concept based on rational acceptability, which in turn is based on virtues like simplicity, Putnam's three stage argument works.

According to Putnam the correspondence theory of truth is dead.  But this claim does not mean that the quest for truth is dead as well.  On the contrary, the quest is clearer.  Rather than looking for some sort of belief-matching-reality criterion we look for criteria of rational acceptability.  Thus, to be true is to conform to criteria of rational acceptability.  Of course, we may and must ask if the criterion is itself rational.  This meta-criteria question however is confused.  Notice that the criteria for x's need not be x's themselves.  Self-referential absurdity does not always apply to criteria.  Putnam seems to suggest that in this case the criteria of rational acceptability need not themselves be rational.  Rather, they must yield results that all ordinary people consider rational to accept.  He writes, "...the reason we want this sort of representation [where the world is represented as instrumentally efficacious, coherent, comprehensive, and functionally simple] that having this sort of representation system is part of our idea of human cognitive flourishing, and hence part of our idea of total human flourishing, of Eudaemonia" (345).  Seen in this light "truth" depends on our criteria of rational acceptability.  To seek the truth is simply a matter of putting into practice such criteria.  For example, if simplicity and coherence are our criteria, then the truth of p will be established only if p meets these criteria.  Such a process of verification is not static.  The criteria may develop over time.  Hence, although p's truth does not vary, our criteria do. 

I do not wish to press this premise in Putman's argument too far.  For now I am content to say that the distance between truth and justification has been substantially diminished to the point of being negligible.  Of course, Putnam attempts to circumvent this problem by positing some ideal limit.  He writes, "If the notion of comparing our system of beliefs with unconceptualized reality to see if they match makes no sense, then the claim that science seeks to discover the truth can mean no more than that science seeks to construct a world picture which, in the ideal limit, satisfies certain criteria of rational acceptability" (341).  Unchanging truth would be found in the ideal world picture, whereas justification would could come and go as provisional world pictures come and go.  The key question is how one could ever know that one's present belief is contained in the ideal world picture.  Images of correspondence obviously will not do.[3] Furthermore, how could one ever know what the ideal world picture is?  It seems that the best strategy would be to eliminate talk of an ideal world picture and accept the identity of justification and truth.  Of course numerous (and arguably worse) problems arise for this view, but we can leave them aside as I do not think that these negative features of Putnam's overall system affect his present argument.  For instance, he could simply leave out the notion of truth entirely and simply opt for another uncontroversial value that science leans on.[4] 

The preceding paragraphs explicated the first two premises of Putnam's argument.  What remains is his third premise, the claim that the values presupposed in science are of the same kind as more common ethical values.  If Putnam is correct about this, the more common ethical values are as objective as those presupposed by science.  Putnam argues that if the values science presupposes are mere attitudes, then what is rationally acceptable must be as well.  "But...the view that rational acceptability itself is simply subjective is a self-refuting one.  So we are compelled to conclude that at least these value-terms have some sort of objective application, some sort of objective justification conditions" (347).  A natural response is to ask why the claim that "rational acceptability is subjective" is self-refuting.  By Putnam's own standards the criteria of rational acceptability could be different.  He writes, "The question: which is the rational conception of rationality itself is difficult.  There is no neutral conception of rationality to which to appeal" (347).  Of course Putnam's claim appears to be about the nature of rational acceptability and not about the criteria.  But it is difficult to see what the salient difference could be.  If truth is established by the criteria of rational acceptability, how is rational acceptability, his surrogate for truth, not established by the same criteria?  Yet even if the claim that "rational acceptability is subjective" is not self-refuting, this fact can do little damage to Putnam's point.  It makes little difference if the fact-value distinction collapses into the fact category -values are considered as objective as facts- or the value category -facts are considered as subjective as values.

This same worry appears from another angle.  Putnam notes that "If there is no conception of rationality one objectively ought to have, then the notion of a "fact" is empty" (348).  Putnam's reasoning is clear:  Science gets on with a certain criteria of rational acceptability.  These criteria have a hand in what counts as fact and what does not.  Putnam writes, "What I am saying is that we must have criteria of rational acceptability to even have an empirical world" (346).  Hence, if all criteria are equally good, and if at least some of these criteria are incommensurable (see below), then what the facts are, is indeed empty.

Before evaluating the merits of this third premise let me summarize the argument so far.  The main thrust of Putnam's argument is as follows:

If world pictures (criteria of rational acceptability) decide or delineate what the facts are and world pictures are themselves formed in accordance with certain values of what is and what is not rationally acceptable, then the facts are intimately related to values.  That is, values are necessary in constructing world pictures, which in turn are necessary for delineating the facts, thus values are necessary for delineating the facts.

Turning back to Swinburne's types of thesis about fact and value it is clear that Putnam disagrees with one of Swinburne's underlying assumptions.  Swinburne assumes that there is a great difference between how language is used and how the world is.  Putnam on the other hand argues that this difference cannot be maintained.  How the world is is largely a product of our world picture, a picture which is largely a product of our criteria of rational acceptability.  Swinburne appears to see this possibility when he writes, "I will give one popular example of a metaphysical system of each type [a metaphysical system that accepts the fact-value distinction and one that does not] in order to show that whether one is to consider that there is a fact/value dichotomy in the world depends upon one's weltanschauung" (306).  I think that Putnam would agree and add that one's weltanschauung itself depends on values.  If this is correct, Putnam's position amounts to claiming that adherence to a fact-value distinction is implicitly contradictory. 

Consider a theory that accounts for some phenomena.  Steve accepts the theory and maintains that the phenomena in question are a fact and that values play no role in either theory selection or the contents of such theory.  Granting that values do play a significant role in theory selection, Steve implicitly contradicts himself.  He wants to maintain that the facts that the theory is supposed to explain are unrelated to values.  But the theory itself is accepted because of values.  Perhaps Steve can maintain his fact-value distinction by accepting the claim that values play a role in theory selection, but they play no role in the contents of a theory.  Clearly Steve cannot get away with this, at least according to Putnam.  For Putnam, the theory determines the facts as much as the facts determine the theory (346).  Thus, if theory selection is based on values and the selected theory determines the facts, it follows that values play a significant role in determining the facts. Anyone who adopts a weltanschauung that attempts to maintain a fact-value distinction implicitly contradicts themselves by the same token. The final task is to evaluate Putnam's third and final premise, but before doing that I must state why I think Putnam's claim that facts are empty if there is no conception of rationality one ought to have, depends on the notion of incommensurability.[5] 

Assume that there is no conception of rationality that one ought to have.  Does it immediately follow that the notion of "fact" is empty?  I think not.  What if there is only one notion of rationality.  Does it follow that that is the one we ought to have?  It only follows if either (a) we have an ideal world picture or (b) our aims are the ones we ought to have.  Regarding (a) we have already seen that on Putnam's grounds knowing whether or not we have an ideal world picture is impossible.  His dismissing of the correspondence theory of truth guaranteed that.  Furthermore he chides the Brain-in-a-Vatists for not being able to give an account of how they know their statements to be true (343).  Thus, if we do have an ideal world picture, but we do not know it (and cannot per Putnam) we are in as deep incoherence as the Brain-in-a-Vatists.  Regarding (b) having right aims appears to presuppose having the right world picture.  Yet by dismissing the correspondence theory of truth one admits that a right world picture is not in the offing.  Hence, the only way for Putnam to maintain the idea that if there is no conception of rationality that one ought to have, then "facts" are empty, is to maintain the idea of incommensurable world pictures.  By doing so, he can conceivably argue that the conception of rationality we ought to have is the one that is simple, efficacious, coherent, etc.  But this strategy works only if the idea of incommensurability is itself coherent, but it is not.

Take two incommensurable world pictures each intra-consistent, but inter-inconsistent.  How can we account for the inconsistency?  If world picture 1 (WP1) is inconsistent with world picture 2 (WP2), the notion of consistency must be the same in each world.  If it is not, it could be the case that WP2's notion of consistency renders WP2 consistent with WP1.  This violates our assumption of inter-inconsistent worlds.      Thus, both world pictures must have the same notion of consistency.[6]  At this stage, incommensurability has been essentially deflated.  The only plausible incommensurability is local.  Will this help Putnam?  I think not.  Remember that one of his premises stated that criteria for rational acceptability are value based.  As has been shown consistency is both a criterion and is not value based; rather it is necessary for the very discernment of criteria of rational acceptability.  Perhaps Putnam could respond by saying that discourse and thought are themselves value-dependent, so that the very act of communication presupposes values, and the very act of thought presupposes values.  If this were the line Putnam took I would agree.  However, I do not think that this line is open to him.  He writes, "... [T]he notion of a transcendental match between our representation and the world in itself is nonsense. ... [T]he empirical world, depends upon our criteria of rational acceptability. ...In short, I am saying that the "real world" depends upon our values (and, again, vice versa)" (345-46).  And, according to Putnam our values can change for they are in large part dependent on "historical conditioning" (347).  My claim is that our notion of consistency cannot change.  It is known a priori, and is a requirement that all world pictures must meet.  By Putnam's own standards it is not a value of the kind he needs.  Thus, at least one of the criteria of rational acceptability is not a value of the relevant kind.  This conclusion is a significant blow to his attempt to deny the fact-value distinction. 

The point can be made another way.  Assume that there are some things knowable a priori.  These things are thus necessarily true;[7] they must be contained in every world picture, and every set of criteria of rational acceptability must meet them.  It is false then to claim that rational acceptability is based on values.  Rather, rational acceptability is in large part based on conformity to a priori (or necessary) truths.  The third premise stated that what we hold to be rationally acceptable is based on values similar to ethical ones.  What we have seen is that this is not the case.[8]

In this final section of the paper I hope to show the connection between Putnam's position and Carnap's.  In so doing the above objection to Putnam's position should become clearer.  In The Logical Syntax of Language[9] Carnap introduces his principle of tolerance (POT): Principle of Tolerance: It is not our business to set up prohibitions, but to arrive at conclusions (10).  According to POT there is no uniquely correct language.  Rather, there are many possible languages, each with its own merit.  Choosing a language is based on certain aims or goals.  Carnap writes, "The purposes for which the language is intended to be used, for instance, the purpose of communicating factual knowledge, will determine which factors are relevant for the decision [i.e. relevant for choosing a language].  The efficiency, fruitfulness, and simplicity of the use of the thing language may be among the decisive facts" (208).[10]  For Carnap languages are chosen because they fulfill some function.  For example, the language which contains middle-sized objects as existents is adopted because it is efficient, fruitful, and simple.  The language is not adopted because it is true.  In order to clarify this point Carnap distinguishes between internal questions and external questions.  An internal question is asked within the language.  The procedure for determining the answer is given by the language.  Thus, internal questions can be answered and therefore have truth-values.[11]  The same cannot be said with regard to external questions.  An external question is asked outside of the linguistic framework.  It is a question about the language.  As such, it cannot use the procedures of an internal question in order to come to an answer.  Such a question is not a theoretical question -one that has a determinate answer-, but is rather a practical question –one that is concerned with whether or not we should adopt a particular language for a particular end (206-07).

It should be clear that Putnam's position is quite close to Carnap's.  Carnap rejects the idea of a correct language.  By abandoning the correspondence theory of truth Putnam does the same.  Carnap relegates truth to a notion internal to a specified linguistic framework.  By making truth a derivative claim of rational acceptability Putnam does the same.  Lastly, Carnap makes the acceptance of a language a practical question; one decided by the aims of the languages user.  By making the criteria of rational acceptability dependent on the aims of the user Putnam does the same.  Thus, we must ask if Carnap's position succumbs to the same criticism leveled against Putnam's. 

The criticism against Putnam's position is that he presupposes that there are no a priori or necessary truths.  Thus, if one can demonstrate that there are a priori or necessary truths, Putnam's denial of the fact-value distinction cannot be maintained.  If Carnap's position also depends on a denial of the a priori or necessary truths[12], then his position will fail for the same reasons that Putnam's did.  It is not difficult to see that Carnap's position does assume a denial of the a priori.  Consider again his POT.  If there are a priori truths, POT makes little sense.  Each proposed language would have to include the a priori.  This is clearly a prohibition on languages.  Carnap's attempt to locate truth claims wholly within a language is also falsified if there are a priori truths.  Indeed, the distinction between internal and external questions is blurred if a priori truths exist.  According to Carnap (and Putnam) internal questions are answered via the procedures specified by a given language.  Thus, if some language specifies procedures P and another language specifies procedure P*, the internal question will be answered differently and could have a different answer.  According to Carnap (and Putnam) this is the case with all questions that have determinate answers.  But if there are a priori truths this is not the case.  Certain questions will have the same answer in each and every language.  Thus, truth is not solely determined by a language and the procedures it specifies (or per Putnam by the criteria of rational acceptability).  A Priori truths fall outside of any linguistic system of the kind envisaged by Carnap (and of the kind envisaged by Putnam).

Carnap's position fails for the same reason that Putnam's does.  The existence of a priori truths undermines both.  It is for this reason that my argument against Putnam's denial of the fact-value distinction is persuasive.  Despite the fact that my argument rest on one principle-consistency-one principle is all that is needed to show the error used to generate Putnam's denial of the fact-value distinction.

[1] Putnam, Hilary "Facts and Value" from Reason, Truth and History reprinted in Pragmatism: A Reader ed. Louis Menand, Random House: New York; 1997 338-362.

[2] Swinburne, R.G. "Three Types of Thesis About Fact and Value" The Philosophical Quarterly Vol. 11, No. 45 (Oct., 1961), 301-307.

[3] This criticism can be found in McDermid, Douglas. "Pragmatism and Truth: The Comparison Objection to Correspondence" The Review Of Metaphysics  98, 51(4); 775-811.

[4] If, however, truth cannot be freed from the correspondence theory, Putnam entire argument collapses.  In order to hold onto both the correspondence theory, a denial of the fact-value distinction, and cognitivism one would have to opt for some sort of supervenience theory or a priori theory.

[5] I too think that facts would be empty if there were no conception of rationality one ought to have but for entirely different reasons.  One is that I think that there is only one rational conception of rationality.  Two is that I think that the one conception of rationality is the one that corresponds to the facts.  I think that something like a transcendental argument can be given for this, but I will spare you the details. 

[6] This can be further clarified.  Consistency is symmetric.  Necessarily, if p is consistent with q, then q is consistent with p.  The only way for two world pictures to be intra-consistent and inter-inconsistent is if the notion of consistency in each is different.  But if this is the case the claim that they are inter-inconsistent is false.  For world picture one (WP1) may be inconsistent with world picture two (WP2) via WP1's notion of consistency and yet WP2 may be perfectly consistent with WP1 via WP2's notion of consistency.  On this rendering consistency is not symmetric, a truly absurd upshot.   

[7] It makes no difference to the argument if we deny their a priori status.  As long as my argument for the necessity of the same notion of consistency in all world pictures is sound, whether this necessity is a priori or a posteriori makes little difference.

[8] Unless Putnam decides to make ethical claims a priori (or necessary) as well. 

[9] Littlefield, Adams and Co., 1959. translated by Amethe Smeaton.

[10] Carnap, Rudolph "Empiricism, Semantics, and Ontology" reprinted in Meaning and Necessity: A Study in Semantics and Modal Logic.  Chicago; The University of Chicago Press: 1988.  205-221.

[11] Of course all of this depends on the language and the question. 

[12] For convenience I will henceforth speak of only the a priori, keeping in mind that this is meant to include necessary truths.