Hesperus is Bosphorus

A group blog by philosophers in and from Turkey

Is truth beneficial and/or socially constructed?

with 5 comments

There are several varieties of the pragmatist theory of truth. Since, according to the pragmatist theories, what we take to be truth is dependent on our pragmatic interests, rather than being “representations of reality,” the pragmatist theories can be regarded as anti-realist theories of truth. According to C.S. Peirce, who is one of the important figures in the pragmatic tradition, truth is, briefly, beliefs socially agreed upon in the long run. Hence “reality” is something socially constructed and is based on consensus. Another important figure, William James, thought that truth is something that has instrumental value. For James “facts” are our mental constructs which prove beneficial in the long run. The popular versions of especially the instrumentalist variety of pragmatism can be found in slogans like, “The truth is what works,” “Truth is what is convenient to believe,” “A proposition is true if believing it has advantageous results.” For the purposes of what follows, I will take the pragmatist theory to be claiming the following:

(PT) Proposition S is true IFF believing that S yields beneficial results in the long run.

I assume that (PT) is shared fully or partly by all pragmatist theories. (One could substitute “has pragmatic value” for “yields beneficial results” in (PT) to stay closer to the letter of the title “pragmatic theory of truth.”)

First off, a counterexample. Suppose I had a terrible nightmare. Every time I remember it, I get the creeps.* I don’t tell anyone about it, because my nightmare is also kind of embarrassing and I am afraid people will make fun of me or will insist that I go see a shrink to get it analyzed—which I’d hate to do. So I keep silent about it for the rest of my life. Thus my belief that I had that nightmare produces no ostensible benefits whatsoever in my life—if anything, every time my belief is enlivened by my recollection of the nightmare, this does nothing but disturb me. As I tell no one about it, my belief yields no useful results for anyone else either. It might even make me edgy in my dealings with some other people at least for a while, and this is not going to be beneficial for any of the parties. So, no useful outcome ensues from my belief either for myself or for any portion of humanity. I have no idea what caused my nightmare, so I don’t have a clue how I can prevent my or someone else’s having a similar nightmare in the future. And since I refuse to consult a shrink about it, she is not getting any monetary or academic benefits out of it either.

Yet, it is true that I had the nightmare, even though no one reaps any benefits out of my belief that it happened.

I think a second problem with the pragmatist theory concerns probability assertions. How does the pragmatist analyze the concept of being “probably true”? ‘Pr(S) = a’ is standardly taken to mean “probability of the truth of proposition S is a.” In accordance with (PT), the pragmatist would presumably translate this into (omitting from now on the “in the long run” part and simplifiying further for ease of comprehension):

(PR) Pr(believing that S is beneficial) = a.

But the expression ‘Pr(S) = a’ is itself a proposition, asserting the probability of something being a. How would the pragmatist translate such an assertion into his language? Perhaps the notion of “probability of the truth of proposition S” becomes for the pragmatist “degree of beneficiality of believing that proposition S is true.” (This is somewhat analogous to the Bayesian construal of probabilities as degrees of belief of perfectly rational agents.) If so, asserting (PR) would amount to the following for the pragmatist:

(PR’) The degree of beneficiality of believing that (believing that S is beneficial) is a.

(PR’) seems to verge on incomprehensibility: it involves second-order beliefs and second-order beneficialities with degrees.

A more important worry is whether the notion of “degree of beneficiality of belief” satisfies the Kolmogorov axioms of probability, which is a requirement if “degree of beneficiality of belief” is to count as an adequate interpretation of the notion of probability. By comparison, the Bayesian interpretation of probability as degree of belief of a rational agent finds its grounding in Dutch-Book arguments. Can the pragmatist similarly find a way to demonstrate the adequacy of the notion of probability interpreted as degree of beneficiality of belief? Or, as an alternative, can he adapt the Dutch-Book approach for the purpose, somehow incorporating the notion of degree of beneficiality of belief in Dutch-Book style arguments? Unless either one of these is done, the pragmatist framework will be unfit to employ the notion of probability—and as a result, will be deprived of such a beneficial conceptual tool as probability.

It might be suggested that the pragmatist can translate a probability statement like ‘Pr(S)= a’ in a simpler way, by converting it into:

(PR”) The degree of beneficiality of believing that S is a.

We could quibble about the point that “believing that S” here is only a condensed rendering of “believing that S is true,” and that therefore (PR”) takes us back to (PR’). But even granting (PR”) to be a simpler pragmatist construal of probability that is different from and not reducible to (PR’), the challenge of showing that “degree of beneficiality of belief” satisfies the Kolmogorov axioms remains to be met.**

A third problem with the pragmatist theory is the following. Notice that the right-hand side of the ‘IFF’ in (PT) is a proposition itself, i.e. what we would regard as a statement of a supposed fact.*** But what is a “fact” for the pragmatist? It is not what ordinarily we or the truth-realist means by ‘fact.’ For the pragmatist, a fact is something belief in which is beneficial, and not something that obtains in “reality.” Now, if the right-hand side of (PT) referred to what the truth-realist would call an “objective” fact that had nothing to do with the beneficiality of believing, then the pragmatist would be guilty of admitting some facts in the truth-realist’s sense, viz. those mentioned in the right-hand side of (PT). More explicitly put, the pragmatist would not want to think of (PT) as implicitly stating the following:

(PT1) Proposition S is true IFF it is an objective fact that believing that S is beneficial.

Instead, the pragmatist would have to construe (PT) as follows:

(PT2) Proposition S is true IFF it is a fact in pragmatism’s sense that believing that S is beneficial.

The pragmatist’s way of explicating (PT2) would then be:

(PT3) Proposition S is true IFF believing that (believing that S is beneficial) is beneficial.

Question: Is (PT3) logically equivalent to (PT), which says the following (in its truncated form)?

(PT) Proposition S is true IFF believing that S is beneficial.

No, I think. For the right-hand side of the ‘IFF’ in (PT) talks about the beneficiality of believing a proposition, whereas the right-hand side of the ‘IFF’ in (PT3) talks about the beneficiality of believing the beneficiality of an act of believing a proposition. So, (PT3) is not reducible to (PT).

Now, the entire right-hand side of the ‘IFF’ in (PT3) purports to be an expression of a fact as well—a fact in pragmatism’s sense, of course. Then, by a further step of explication we get:

(PT4) Proposition S is true IFF believing that [believing that (believing that S is beneficial) is beneficial] is beneficial.

The full expansion of (PT) thus becomes an infinitely long analysis:

(PT5) Proposition S is true IFF … believing that {believing that [believing that (believing that S is beneficial) is beneficial] is beneficial} is beneficial … .

This is the content of the pragmatist theory of truth made fully explicit. It involves infinite regress. For the reasons I gave a moment ago, the infinite regress cannot be dispelled by arguing that (PT5) is reducible to the simple (PT): neither (PT3) nor (PT5) can be so reduced.

A similar result follows for the social-consensus or social-construct version of the pragmatist theory of truth. That version claims:

Proposition S is true IFF it is socially agreed or socially constructed that S.

The last thing this version would want to admit, of course, is that this analysis is implicitly making the following claim:

Proposition S is true IFF it is an objective fact that it is socially agreed or socially constructed that S.

Instead, the advocates of the socialized version would want to say (after some truncation for simplicity) the following:

Proposition S is true IFF it is socially agreed that it is socially agreed that S.

Going through a parallel reasoning to the one above, the socialized version can be seen to be really giving us this infinitely long account:

Proposition S is true IFF … it is socially agreed that {it is socially agreed that [it is socially agreed that (it is socially agreed that S)]} … .

Given this picture, the pragmatist theory couldn’t yield any beneficial or socially acceptable results, philosophically speaking…

—————————–

* We are told by psychiatrists and brain scientists that dreams serve a beneficial function for the brain. But I don’t know if every nightmare we had had to happen for our benefit. Maybe some of our nightmares are just mishaps. After all, not everything our body does to us is beneficial to us—think of kidney stones or inflamed appendices, for example. If so, maybe the nightmare I had was not all that necessary for my brain. Or whatever benefits it brought to my brain might have been brought by more docile dreams, without the rest of my waking life being troubled by that nightmare.

** As a possible resort, the pragmatist might wish to argue that Kolmogorov’s axioms are not really beneficial to believe. But it is hard to see how such an argument could be supported in the face of the fact that the probability theory has yielded highly beneficial results in sciences and other spheres of life.

*** Indeed, (PT) as a whole also needs to be taken by the pragmatist to be a true proposition. But I am not going to dwell on that point.

Advertisements

Written by Erdinç Sayan

March 19, 2012 at 4:11 am

Posted in Epistemology

5 Responses

Subscribe to comments with RSS.

  1. Pragmatism has no place (and indeed is not pragmatic 🙂 in a purely analytical epistemology, where all is certain as to truth value or numeric probability.

    Pragmatism is an empirical tool for steering between the two extremes of excessive skepticism and excessive gullibility, in choosing what to trust and how to investigate or to act, when things are _not_ certain.

    Skip

    March 20, 2012 at 11:49 am

    • Thanks for your comment. Yes, the pragmatic theory doesn’t have much of a pragmatic value, given that how it defines truth is with an infinitely long sentence. Plus, what it says simply cannot be true, if my nightmare counterexample is correct.

      I think “purely analytical epistemology,” if I understand what you mean by that term, has got some tools to deal with imprecision and uncertainty, e.g. probability, fuzzy logic, fuzzy probability, truthlikeness, error calculus, rational decision under uncertainty in Decision Theory, etc. Analytic philosophers are not gullible enough to mistake genuine imprecision for precision. I don’t know what tools pragmatism has got to deal with the phenomena of uncertainty and imprecision.

      Erdinç Sayan

      March 20, 2012 at 3:43 pm

  2. You make an extensive use of T-schema: (T’p’ iff p) in your argument.
    So, if a pragmatist denies s T-schema ie. if one holds that some instances of T-scheme is wrong(in the case of an answer to your counter-example (if p, then T’p’) part is enough) then one can consistently admit that you had a nightmare and it is not true that you had a nightmare.

    For your objection of regress, denying T-schema stops the regress.
    Analysis would be
    (PT) Proposition S is true IFF believing that S yields beneficial results in the long run. [full stop]

    Since; ‘believing that S yields beneficial results in the long run’ will not imply (in the context of rejecting T-schema) and will not implied by, ‘it is true that ‘believing that S yields beneficial results in the long run”.

    You can say that denial of T-schema seems rather ad hoc, but a pragmatist can cite different motivations for her denial, one example would be, semantic paradoxes.

    Moreover, admitting T-schema and asking for an analysis of truth seems to me an impossible task,
    Consider,(1) S is true iff ‘S’ has P ( substitute ‘…has P’ with your favorite analysis of truth, for example …corresponds to the facts, …coherent with the total amount of beliefs,… is entertaining to believe.)
    Since we admit (T’p’ iff p), we get
    ‘S’ has P. IFF ‘ ‘S’ has P’ is true. IFF ‘ ‘ S has P’ ‘ has P.
    Then by simply reiterating your argument, we get (2) S is true iff ‘ ‘S’ has P’ has P.
    So, once the argument goes off there is no way to stop, and T-scheme is so powerful that can fuel any analysis whatsoever.

    adil

    March 20, 2012 at 9:43 pm

    • Thanks for your comment. Sorry it took me long to reply. In my reply below, I made slight modifications on your punctuation to make it easier for me to understand your assertions. I hope I didn’t distort your meaning.

      You make an extensive use of T-schema: (T‘p’ iff p) in your argument.
      So, if a pragmatist denies T-schema ie. if one holds that some instances of T-scheme is wrong

      Actually in my post I don’t presuppose that the pragmatist accepts the T-schema. I don’t quite make an extensive use of the T-schema either. (PT) is simply the pragmatist definition of “true proposition” that is easily understandable to someone conversant with biconditional sentences. If for some reason the pragmatist isn’t fond of biconditional statements (perhaps because they don’t produce beneficial results in the long run?), we can state the pragmatist theory in more plain (though vaguer) English:

      A belief that is beneficial in the long run IS true.

      (in the case of an answer to your counter-example (if p, then T‘p’) part is enough) then one can consistently admit that you had a nightmare and it is not true that you had a nightmare.

      I take your suggestion to be that a pragmatist might intend his theory to be what we would call a sufficient-condition analysis of truth, rather than a necessary-condition or necessary-and-sufficient-condition analysis of truth. I.e. the pragmatist’s theory might just be:

      IF a belief is beneficial in the long run, THEN it is true.

      Although I seriously doubt that this is the intent of the pragmatist, if we are to construe his theory in that way, then we have other counterexamples to it: for example, Russell’s famous example in which a powerful tyrant brings it about that believing a certain stupid religion proves beneficial to all of the tyrant’s subjects. I think we can even improve Russell’s example. Instead of a tryrant trying to ensure it (which may not be very effective because the subjects may only pretend to believe in the stupid religion without actually believing in it), we can use more fool-proof methods to instill beneficiality to people. For example, we can induce drug-altered mental states in the members of the society or tamper with the genetic material of every individual in such a way that they will happily and beneficially believe some outrageously false statement—say, the belief that everybody gets reincarnated after death as an insect in some distant planet. What is beneficial to us is to a large degree a function of our biological constitution. But changing the biology shouldn’t change the truth. I think our notion of truth does not allow such arbitrariness.

      For your objection of regress, denying T-schema stops the regress.

      As I said, I didn’t presuppose that the pragmatist is committed to the T-schema. Infinite regress results even without (PT), as rehashed below.

      Since ‘believing that S yields beneficial results in the long run’ will not imply (in the context of rejecting T-schema) and will not implied by, ‘it is true that believing that S yields beneficial results in the long run’.

      I am not sure but you may be thinking that I should give up my formulation of the pragmatist theory as (PT). OK, but let me have at least the following version (which is open to any one of the sufficient-condition, necessary-condition, and necessary-and-sufficient-condition readings), otherwise I don’t know how to even formulate the pragmatist theory:

      (PTV) To say that proposition P is true IS to say that belief that P is beneficial.

      I want to ask the pragmatist now: When you say of a belief that P that it is beneficial, do you mean it is beneficial in an “objective” sense, i.e. independently of beneficiality considerations? Or is it beneficial in the pragmatist’s sense? In other words, could you clarify which one of the following are you claiming?

      (PTV.1) To say that proposition P is true IS to say that belief that P is beneficial as a matter of “objective fact”

      (PTV.2) To say that proposition P is true IS to say that belief that P is beneficial as a matter of fact in the sense of pragmatism.

      A consistent pragmatist would choose the claim (PTV.2) of course. For (PTV.1) would be an admission of existence of facts that are independent of beneficiality to society. (PTV.2) can be rephrased as (using parantheses for visual clarity instead of quotes):

      (PTV.2’) To say that proposition P is true IS to say that belief that (belief that P is beneficial) is beneficial.

      I will next run a similar questioning about (PTV.2’). How are we supposed to understand (PTV.2’)? Which one of the following two interpretations of (PTV.2’) is acceptable to the pragmatist?

      (PTV.2’.1) To say that proposition P is true IS to say that belief that (belief that P is beneficial) is beneficial as a matter of “objective fact”

      (PTV.2’.2) To say that proposition P is true IS to say that belief that (belief that P is beneficial) is beneficial as a matter of fact in the sense of pragmatism.

      The consistent pragmatist would once again pick (PTV.2’.2), which can be rephrased as:

      (PTV.2’.2’) To say that proposition P is true IS to say that belief that [belief that (belief that P is beneficial) is beneficial] is beneficial.

      And so on ad infinitum. Hence infinite regress…

      You can say that denial of T-schema seems rather ad hoc, but a pragmatist can cite different motivations for her denial, one example would be, semantic paradoxes.

      It would be much costlier to accept pragmatism than keeping up the hope of solving those paradoxes, I’d say.

      In the remainder of your comment you may be raising this interesting point: Couldn’t other truth theories be shown to also fall into infinite regress by a similar argument? I had thought about this question myself and I think I have a reply to it. Consider, for example, the correspondence theory:

      (CT) ‘P’ is true IFF ‘P’ corresponds to facts.

      Suppose, just as we did with the pragmatist, we questioned the correspondence theorist as to which one of the following would be the correct interpretation of (CT):

      (CT.1) ‘P’ is true IFF ‘P’ corresponds to facts as a matter of “objective fact”

      (CT.2) ‘P’ is true IFF ‘P’ corresponds to facts as a matter of “subjective fact”

      (for some sense of the phrase ‘subjective fact’). The correspondence theorist would not hesitate to pick (CT.1), which can be rephrased as:

      (CT.1’) ‘P’ is true IFF (‘P’ corresponds to facts) corresponds to facts.

      Is this the first step to the kind of infinite regress that infects the pragmatist theory? I think the correspondence theorist could argue that (CT.1’) is not different from (CT) in the sense that the two are logically equivalent. In other words:

      ‘P’ corresponds to facts IFF (‘P’ corresponds to facts) corresponds to facts

      is a conceptual truth. Hence no infinite regress ensues for the correspondence theorist. Or, at least, any infinite regress to ensue will be harmless.

      On the other hand, the pragmatist is not that lucky. For

      Belief that P is beneficial IFF belief that (belief that P is beneficial) is beneficial

      does not seem to be a conceptual truth. Suppose that believing in reincarnation is beneficial to me (as it allays my fear of death, etc.). But my second-order belief that my belief in reincarnation is (or would have been) beneficial to me may not be beneficial to me: as a rational and nonsuperstitious person I find the beneficiality of my belief in reincarnation alarming and something that I must watch out not to be lured by. This would be a case of conflict between one belief’s beneficiality and a second-order belief’s nonbeneficiality, and not a case of logical impossibility. A social example can also be generated: imagine a society which is composed of individuals who are all just like me with regard to their psychological bent towards belief in reincarnation and their intellectual vigilance against it.

      Erdinç Sayan

      March 26, 2012 at 10:35 am

  3. […] Turkish philosophers have strange dreams. Erdinc’s are so weird and wonderful that he cannot even tell us about them. And after our two day workshop […]


Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out / Change )

Twitter picture

You are commenting using your Twitter account. Log Out / Change )

Facebook photo

You are commenting using your Facebook account. Log Out / Change )

Google+ photo

You are commenting using your Google+ account. Log Out / Change )

Connecting to %s

%d bloggers like this: