Williams (1973) casually asserts that

to believe that P is to believe that P is true.

He explains what he means by that:

To believe that so and so is one and the same as to believe that that thing is true.

Since the dual component view of belief is said to be the standard account of belief and it includes the view that

To believe that p is to believe that p is true. To believe ‘she is late’ is to believe that it is true that ‘she is late’. So if you recognise that p is false (you realise she is not late), you abandon your belief that p.

it would seem thus that Williams' position is the standard view.

On the other side, I found a passage in Kvanvig (2003):

[1] First, to believe p is not the same as to believe that p is true, even if we grant the logical equivalence of 'p' and 'p is true'. To hold otherwise is to hold that having the concept of truth is a precondition of thought, that no one can think or believe anything without having the concept of truth. […]

[2] The best thing that can be defended about the relationship between believing and believing the truth is that if a person believes p, has the concept of truth, and considers whether p is true, that person cannot believe p and fail to believe that p is true.

I interpret the difference in the following way: Williams holds that "to believe that p" equals "to believe that p is true", while Kvanvig holds concedes the weaker view that "to believe that P" implies "to believe that P is true" (under certain premises).

I have the following questions:

  • Is there currently a consensus in the epistemology literature that "to believe that p" implies "to believe that "p" is true", i.e. the weaker position expressed by Kvanvig?

  • Has anybody argued against this weaker position?

Update: I enlarged my question to clarify it in response to the answers given so far.

  • Now that you have added the context of your question, quoting Wiliams and Kvanvig, one should first clarify: What do they mean by respectively, p and P? I agree with Cort Ammon that one has to discriminate between a proposition and a state of affairs. Do the two authors believe propositions or do they believe state of affairs? Can you derive the answer from their texts?
    – Jo Wehler
    Jan 24, 2016 at 4:49
  • Assuming that "having the concept of truth" minimally requires accepting the T-sentences (or accepting the capture/release rules for truth), and assuming some form of doxastic closure (i.e. the agent believes what she has inferred from her beliefs), it's hard to see how you could argue against the weaker thesis. I have seen the P->T("P") direction of the T-equivalence questioned (I can't remember where), so that might be a way to argue against it. Also, if the belief is some sort of partial belief then maybe you could have a case against it.
    – Johannes
    Jan 24, 2016 at 12:51
  • I don't have an answer to this since it's well outside of my own specialization, but it seems like the rub is going to be about the sort of entailment meant by "implies" and whether one considers believes that P and believes that P is true to be distinct beliefs (which again mirrors whether one considers the implication intuitively automatic or true upon reflection).
    – virmaior
    Jan 24, 2016 at 14:49
  • 1
    @Johannes: Good reasoning! The recent paper Rationally held ‘P, but I fully believe ~P and I am not equivocating' by Frances in Philosophical Studies got me thinking and I'm trying to formulate what rubs me the wrong way about it. I've got a feeling that granting Kvanvig's view might render Frances' point moot. But it's just an intution for now :)
    – DBK
    Jan 24, 2016 at 15:42
  • @virmaior: I agree. I think I interpreted Kvanvigs claim to stringently. It may be as simple as requiring doxastic closure, as Johannes mentioned. I don't know.
    – DBK
    Jan 24, 2016 at 15:46

7 Answers 7


One way of approaching this question is by noting the variety of mental states that can be included in 'belief':

Evidenced belief

  1. I believe that Queen Victoria died in 1901. (I am convinced)

  2. I believe that Trump will win in 2020. (I believe 60: 40 that Trump will win: I ascribe this order of probability to it.)

  3. I believe that the burglar crossed the lawn diagonally and broke in though the basement window. (This is my inference to the best explanation of the evidence.)

I don't agree that in any of these cases - not specially chosen - to believe that P is to believe that P is true. In every one, to believe that P is to believe that P is probabilistically true. The reason for this is that in each case I know perfectly well that the evidence on which my belief rests does not entail its truth. Put the point this way : epistemologically I realise that because of the inherent vulnerability of evidence to error, there is an unclosable gap between belief and truth where belief is based on evidence.

And there's another point. I have good inductive grounds that not all my beliefs are true. But I don't know which are false. Therefore I don't know if my current - or the OP's current - belief is true. This could be one of those that are false. This reasonably qualifies my belief - shifts a gear down from the simple belief that it is true.

Unevidenced belief

What might such belief look like ? I cite William Alston:

On considering the proposition that two quantities equal to the same quantity are equal to each other, this seems obviously true to me; and I shall suppose, though this is hardly uncontroversial, that in those circumstances I am justified in believing it. But where are the adequate grounds on which my belief is based? It is not that there are grounds here about whose adequacy we might well have doubts; it is rather that there seems to be nothing identifiable as grounds. (W. Alston, Epistemic Justification. Ithaca: Cornell University Press, 1989: p. 106)

There seem to be no evidential grounds here. Alston is appealing to intuition. I'm inclined to say that to have this kind of intuitionist belief does entail belief in its truth. And yet, haven't we learned to distrust our intuitions? The 'hardly controversial' is not necessarily the true. It used to be widely believed, clear by the 'natural light of reason', that space is Euclidean. A dodgy character, intuition ...


Alfred Tarski worked on phrasings like that. However, he added some quotes to the phrasing to make it more grammatically correct. He defined his material adequacy condition:

"P" is true if, and only if, P.

The quotes, of course, are needed because you'd never say "to believe that the sky is blue is true," but you would say "to believe that 'the sky is blue' is true."

Tarski was only interested in formal languages at the time, because their truthhood can be explored using nothing but mathematics. Donald Davidson later extended it to natural languages.

Tarski's material adequacy condition works rather well up until pronouns get involved. once P has a pronoun in it, you can get curious and sometimes paradoxical constructions.

  • 1
    Thanks, but this doesn't address my question about current epistemology literature. (I have expanded and, hopefully, made my questions clearer.)
    – DBK
    Jan 24, 2016 at 2:13

One has to discriminate between a state of affairs, which belongs to the real world, and a proposition, which belongs to the realm of mind.

A state of affairs happens or does not happen. On the opposite, a proposition is true or false.

In "To believe that P", P denotes a state of affairs. In "to believe that P is true", P denotes a proposition. Hence you use P in two different meanings.

In order to discriminate both, I recommend: If P denotes a state of affairs, e.g. it rains, then "P" (enclosed in quotation marks) is the proposition "It rains".

Hence: "To believe that P (state of affairs)" equalizes "To believe that "P" (proposition) is true".

  • I'd say propositions are linguistic entities rather than mental entities (more neutral with regards to the philosophy of language) Jan 23, 2016 at 15:31
  • @quen_tin I am not quite happy myself with the term "mental" domain. In German I say "geistiger" Bereich. How to translate this term? "Spiritual" would come with the improper association of spirituality. What about "realm of mind"?
    – Jo Wehler
    Jan 23, 2016 at 15:37
  • I am not sure that "to believe that p" denotes a state of affairs. You believe a proposition, not some state of affairs. After all, "to believe" is called a "propositional attitude" for a reason: it is an attitude wrt a proposition. But I specifically did not ask if "to believe that P (state of affairs)" equals "To believe that "P" (proposition) is true", but whether "to believe that P (state of affairs)" implies "to believe that "P" (proposition) is true". I will try to edit my question to make this clearer.
    – DBK
    Jan 24, 2016 at 2:21
  • @JoWehler sorry I can't help you with this but my point was that prepositions are not necessarily mental or spiritual or "in the head", in particular because meaning is public. Frege for example insisted that meaning has nothing to do with psychology Jan 24, 2016 at 12:54
  • @JoWehler now it depends on the author (cf. Internalism vs externalism about meaning) so "linguistic" is more neutral. Jan 24, 2016 at 12:55

There are at least three different ways to believe something, (loosely following the Lacanian version of three worlds) only one of which is to believe that it is true.

(In the 'real' world.) The most practical definition of belief is ethical. A belief is a statement with consequences that one is willing to act upon. This requires no notion of truth, only an attachment to consequences.

(In the imaginary world:) The most strict definition of belief is logical. Logic generally treats belief as the acceptance of the truth of a fact. And obviously this requires a standard of truth. Of course, the logic itself has to back up the notion of truth in order to make this meaningful.

(In the symbolic world.) Between these is the customary, everyday definition of belief as holding a position in a 'language game'. One believes something if one is willing to reason from it, at risk to one's standing, self-image and power positions, whether or not the reasoning ultimately leads to action. This also does not require a notion of truth. What matters is various players' positions within the game and the contribution to the evolution of the game's ultimate purpose. Religious and political beliefs are the form of belief that most obviously have this flavor. Religious and political ideologies have their own language games in which adherents negotiate orthodoxy and challenge or support different freedoms of faith. The game allows individuals to trust one another and seek effects that they communally want.

This last form of belief, although the most complex, and hardest to understand, is also the most common, because it is interactional and we are both obsessively social animals and highly dependent upon environmental feedback for survival. Such beliefs are, in many ways, the only kind of truly new information. We do not know facts, we merely believe them, and we only believe them within a context that gives us feedback about them and their 'truth'. That context is a language game we enter into either in our internal monolog, or with others.

We seldom deal with this form of belief, as a belief, 'in the raw', however. Instead, once we conceive of something as a potential belief, we tend to internally model it in terms of the first form of belief by projecting it onto the second form, so that we can create abstract models and make predictions, or we do the opposite, and embrace it as the second form, acting with the risk of unplanned consequences so we can project them onto the first form as data about the outside world upon which to base an internal model.


Does 'to believe that P' imply 'to believe that P is true'?

If we rephrase this question to follow Jo Wehler's excellent suggestion, we get:

Does 'to believe that p' imply 'to believe that "p" is true'?

First, the answer to the rephrased question is "no", since to believe is to have a belief and one can have a belief that p without having a belief that "p" is true, in particular if one doesn't have a concept of truth, or indeed a concept of proposition.

However, the rephrased question may not be equivalent to the original question from its author's point of view.

The following rephrasing might be closer to Bernard Williams' original idea:

Does 'to believe that p' imply 'to believe that it is true that p'?

First, there is a sort of marked conceptual distance between "It rains" and "It is true that it rains". The latter wording suggests an abstraction on the part of the believer.

Second, although "It is raining" does imply "It is true that it is raining", we can believe the former without even having any thought corresponding to the latter.

The distinction, however, comes with the word "belief", which refers to a psychological state, not to any widely accepted logical concept, unlike for example the word "true".

A valid implication involving the word "belief" would have to include a formal and logically consistent definition of the term. Minimally, you could just posit the implication 'to believe that p' implies 'to believe that it is true that p'. That would help in making valid arguments about beliefs but, then, how would you justify your assumption that this implication is true?


No necessarily, You can directly believe that P is false. And here is a difference between what you believe and what you know, which therefore results in a fact that you can check and contrast to be generally true. Probably interchangeable terms but different logic meanings.

I believe that she is late - just an hypothesis

I know that she is late - a contrasted fact.

Well basically I “believe” that you are depicting what is known as the “Gettier Problem” so the answers to your questions were given by Edmund L. Gettier in his paper Edmund L. Gettier; Is Justified True Belief Knowledge?, Analysis, Volume 23, Issue 6, 1 June 1963, Pages 121–123, https://doi.org/10.1093/analys/23.6.121


I can't quote any of the famous philosophers, so I'll have to go forward as if I am one, forgive the vanity. If one is to say 'I believe P', it follows that one believes P to be true. Let's substitute P for 'she was late to class'. 'I believe she was late to class' and 'I believe it's true she was late to class' are identical statements.

In regards to belief and not proof, however, I would argue for the weaker argument as well. 'I believe I will win the lottery' is a belief a person might hold, even with the full understanding they will not. It's clearer to see with opinions that are not factual. 'I believe prisoner XYZ should be executed' does not translate directly to 'Prisoner XYZ should be executed'. To believe in a thing does not instantly make it thus, as all religions and world governments can attest.

So we seem to be at a contradiction. How can belief and truth not be equal here, and also be? Depends on the way that belief was derived. As far as we are able, if we can prove a thing, then to believe in it is to believe in the proof of the thing. 'I believe she was late to class, because she entered class after the bell' is both a belief and a provable statement. 'I believe red is the best color' is a belief, but an opinion. The first is a belief based on the truth, the other is an attempt to impress truth upon a pre-formed belief.

Thus, I would argue that 'I believe P' and 'I believe P to be true' can and cannot be identical, provided the direction in which they are being applied.

  • 1
    Welcome to Philosophy.SE! I would encourage you to check out the tour and look at the conventions around answers here. This is not a forum for expressing one's opinions. Jan 27, 2016 at 20:16
  • 1
    "To believe in a thing does not instantly make it thus, as all religions... can attest" - Certainly in my religion, we aspire to the opposite (believing in something because it is so, rather than believing in something is if that would make it so). This statement is (1) unsupported and (2) not helpful to your argument. Jan 27, 2016 at 20:19

You must log in to answer this question.

Not the answer you're looking for? Browse other questions tagged .