What, if anything, is the difference What, if anything, is the difference between asserting a proposition and asserting that it is true? What does the answer to this question reveal about the concept of truth?

The claim that assertion and truth are equivalent is the basis of the redundancy theory of truth, first proposed by F.P. Ramsey. It claims that the truth predicate is logically superfluous to the language, and that that is why we experience difficulty when trying to explain the notion of truth. Ramsey claims that truth is nothing more than linguistic shorthand, being used in place of other sentences. Many people have objected to the notion that just because it can be said to be logically equivalent it therefore follows that it has no linguistic meaning. There are also certain cases where it is unclear what propositions the shorthand would be referring to, which seems to lead to the idea that the theory is unable to explain all cases of truth, a severe flaw if it is effectively trying to remove the whole notion of truth. In addition to this, it has been noted that if we wish to remove the allegedly empty notion of truth from our language, then how are we to assign notions of 'correctness' to propositions. It seems counterintuitive to completely remove the idea of truth, even if we can say that it is an idle wheel in language. I personally think that we could remove the truth predicate from language with only trivial semantic loss, but that we would still need a concept of truth, and that that would find other ways to express itself through language. Thus, we would effectively be doing nothing more than shifting the truth predicate from one term to another, a seemingly pointless exercise.

The most obvious difference between asserting a proposition, p, and asserting that p is true is that the latter contains the truth predicate. This may seem like a trivial statement, but it seems to me that it is the most fundamental one. I think that by saying that p is true, one is in fact saying two things. Firstly, one is indeed asserting that p, but one is also asserting that you believe p to be true. Thus, the issue here is whether or not the notion of believing a proposition to be true has any effect upon its semantic meaning. The claim of the redundancy theorist would be that saying you believe that p is true is saying nothing in addition to asserting p, and that whilst it can be semantically derived, there is no need for it to be there. Thus, the elimination of the truth predicate would not alter the meaning of the statement. It is indeed true that the two statements are in fact logically equivalent, that is they are true and false under exactly the same truth conditions, but that assumes that there is no meaning in language outside logic, and I think that most people would dispute that. Ramsey himself acknowledges that saying something is true can sometimes be a way of saying that you agree with someone's statement. That seems to me to involve more meaning than is expressed through the logical translation of the redundancy theory. It is generally recognised that logic is vastly inferior to natural languages when it comes to conveying semantic meaning, despite the fact that it allows more precision and clarity. That is in fact its function, to remove the ambiguities of natural language, and that comes at the cost of certain implicit parts of language. Ramsey also acknowledges that to say that it is true that p implies that there is a person who has a belief that p. It seems that in removing the truth predicate, we would be affecting more than simple logical syntax.

However, the problems with this theory are not purely confined to its somewhat vague effects upon natural language, there are certain logical problems with the concept. For example, when someone uses the truth predicate to refer to propositions unknown to them at that time, as in "What x said yesterday is true", where the speaker does not in fact know what x said yesterday, but rather believes in the general truth of x's statements. Here it seems difficult to replace 'is true' with the statement that a redundancy theorist would claim it is asserting, since the speaker would not be able to use any of the actual propositions that x asserted on the previous day. It is worth noting here that Ramsey did in fact shy away from the notion of propositions, instead preferring to use only the term "propositional reference". He did this mostly to avoid complicating his ontology with the notion of propositions as real entities, but it does mean that we need to be careful when discussing his attitude to propositions.

Another logical problem that is somewhat related to the first is how to eliminate the truth predicate in cases where innumerable propositions would have to be used to replace it. If, for example one were to assert that everything said by one race of people was true, then it seems that one could not reasonably replace such a sentence with an assertion of every single statement made by these people. One attempted solution to this is to reformulate the statement logically as follows: For any proposition p, if a person of race x asserts p, then p is true. This would seem to have the same function as the natural language sentence, but it still contains the truth predicate. Thus, the redundancy theorist would say that we could remove the truth predicate from the argument, and still have a complete and meaningful statement, indeed, one of the same meaning. However, many have argued that the elimination of the truth predicate renders the sentence grammatically incorrect, and thus definitely alters the meaning. Their argument runs that by saying "For any proposition p, if a person of race x asserts p, then p", one leaves the final segment, "then p", without a verb, and thus grammatically incomplete. The response that Ramsey gives to this is that whilst that may be the case when one considers the sentence with unfilled variables in it, the variable p does in fact stand for a complete proposition, one that would necessarily contain a verb, and thus would fulfil the necessary grammatical rules. However, this brings us back to the point that the speaker does not in fact know all the propositions that he is referring to, and so cannot properly replace the variable p.

Even assuming that the redundancy theorist can answer these objections, there is still the intuitive notion that there is a concept of truth. All that the redundancy theorist would be achieving is the elimination of the truth predicate from the language, not the removal of the whole notion from our system of logic. It would seem that if we were to eliminate the notion of truth, then we would rule out the possibility of error existing anywhere, an absurd idea. Whilst it would perhaps be helpful be able to do that, it's ludicrous to suppose that everyone could be right the whole time, since that would effectively require everyone to know everything about any topic which they asserted something on, or for everyone to keep their speech within the realms of opinion, and thus for no-one to ever be sure of anything. Here we are straying into reasons for valuing the concept of truth, and that is a separate issue. Suffice it to say that I cannot see the elimination of the notion of truth as reasonable, nor can I see the elimination of the truth predicate from the language as significant.

We must now ask what this has revealed about the concept of truth, and I think that it has in fact revealed several points. Firstly, we both value and need the concept of truth, since it seems absurd to try and say anything meaningful without it. Were the predicate to be eliminated from the language, we would simply find an alternative method of conveying our indication of truth. What Ramsey was seeking was "the reduction of truth to reference", but as we have seen, there are some cases where saying that something is true cannot be explained in terms of reference. Where the referent is unknown, or unfeasibly complex, it is not sufficient to substitute the assertion of the propositions for the claim that they are true. In addition, the idea that 'is true' is used purely for semantic emphasis is a double-edged sword, since one can then argue that it does in fact fulfil a meaning other than purely reference. What is not pointed out is that it is still a reasonable definition of truth, but that it is simply not one that we would intuitively accept. We want truth to be more than an exclamation or similar, and that applies to Ramsey's redundancy theory as well.

It seems then that there is a definite difference between the assertion of a proposition and the assertion of its truth. Exactly what that difference is is subject to some debate. It is essentially believed that something additional is being asserted when we make a claim about the truth of a proposition, perhaps something as minor as the assertion that there is someone who believes that p, but that is still a difference in meaning. It seems that the problem is that Ramsey is mistaking logical equivalence for complete equivalence, whereas most would recognise that logic is a more limited language than natural languages. However, the objections to his theory do help us to determine certain aspects of the concept of truth that we seem to require be present. Thus, in formulating a theory that gave rise to a barrage of criticism, he has elicited a more complete understanding of the concept of truth from us, even if it was not the one he wished us to have.


P. Horwich (Ed), Theories of Truth (Cambridge: Dartmouth Publishing Company, 1994) A. White, Truth (Norwich: Macmillan, 1971)

File translated from TEX by TTH, version 2.92.
On 24 Mar 2002, 23:49.