Does Knowledge Entail Belief or Exclude Belief?

A reader who says he is drawn to the view that knowledge excludes belief comments:

I am taking a philosophy class now that takes for granted that knowledge entails belief. My sense is that most philosophers now think that that condition is obvious and settled. They tend to dispute what "justification" means, or add more conditions to the Justified True Belief formula.

That knowledge is justified true belief is a piece of epistemological boilerplate that has its origin in Plato's Theaetetus.  The JTB analysis is extremely plausible.  It is first of all self-evident that there is no false knowledge.  So, necessarily, if S knows that p, then 'p' is true.  It also seems obvious that one can have a true belief without having knowledge.  Suppose I believe that at this very moment Peter (who is 60 miles away) is teaching a class on the philosophy of science, and suppose it is true that at this very moment he is teaching such a class; it doesn't follow that I know that he is teaching such a class.    Knowledge requires justification, whatever exactly that is.  Finally, if S knows that p, how can it fail to be the case that S believes that p?  It may seem obvious that knowledge entails belief.  Necessarily, whatever I know I believe, though not conversely.

So I agree with my reader that most philosophers now think that the belief condition is "obvious and settled."  But most academic philosophers are fashionistas: they follow the trends, stick to what's 'cool,'  and turn up their noses at what they deem politically incorrect.  And they read only the 'approved' journals and books.  I pronounce my 'anathema' upon them.  In any case it is not obvious that knowledge entails belief.

The Case for Saying that Knowledge Excludes Belief

Why not say this:  Necessarily, if S knows that p, then it is not the case that S believes that p?

One cannot understand belief except in relation to other mental states. So let's consider how believing and knowing are related, taking both as propositional attitudes.  They are obviously different, and yet they share a common element. Suppose we say that what is common to S's knowing that p and S's believing that p is S's acceptance of p. I cannot (occurrently) believe that Oswald acted alone unless I accept the proposition that Oswald acted alone, and I cannot (occurrently) know that he acted alone with accepting the very same proposition. To accept, of course, is to accept-as-true. It is equally obvious that what is accepted-as-true might not be true. Those who accept that the earth is flat accept-as-true what is false. Now one could analyze 'S knows that p' as follows:

a) S unconditionally accepts-as-true p
b) p is true
c) S is justified in accepting-as-true p.

This is modeled on, but diverges from, the standard justified-true-belief (JTB) analysis of 'know' the locus classicus of which is Plato's Theaetetus.

And one could perhaps analyze 'S believes that p' as follows:

a) S unconditionally accepts-as-true p
d) S does not know that p.

These analyses accommodate the fact that there is something common to believing and knowing, but without identifying this common factor as belief. The common factor is acceptance. A reason for not identifying the common element as belief is that, in ordinary language, knowledge excludes belief. Thus if I ask you whether you believe that p, you might respond, 'I don't believe it, I know it!' Do I believe the sun is shining? No, I know the sun is shining. Do I know that I will be alive tomorrow? No, but I believe it. That is, I give my firm intellectual assent to the proposition despite its not being evident to me. Roughly, belief is firm intellectual assent in the absence of compelling evidence.

Surely this is what we mean by belief in those cases that clearly count as belief. Lenny the liberal, for example, believes that anthropogenic global warming is taking place and is a dire environmental threat. Lenny doesn't know these two putative facts; he believes them: he unconditionally accepts, he firmly assents to, the two propositions in the absence of compelling evidence. And it seems clear that an element of will is involved in our boy's belief since the evidence does not compel his intellectual assent. He decides to believe what he believes. His believing is in the control of his will. This does not mean that he can believe anything he wants to believe. It means that a 'voluntative surplus' must be superadded to his evidence to bring about the formation of his belief. Without the voluntative superaddition, he would simply sit staring at his evidence, so to speak. There would be no belief and no impetus to action. Beliefs typically spill over into actions. But there would not be even a potential 'spill over' unless there were a decision on Lenny's part to go beyond his evidence by superadding to it his firm intellectual assent.

"But aren't you just using 'believes' in an idiosyncratic way?"

It is arguably the other way around. Someone who says he believes that the sun is shining when he sees that it is shining is using 'believes' in an idiosyncratic way.  He is using 'believes' in a theory-laden way, the theory being the JTB analysis of 'knows.'

"But then isn't this just a terminological quibble? You want to substitute 'accepts' or 'accepts-as-true' for 'believes' in the standard JTB analysis of 'knows' and you want to reserve 'believes' for those cases in which there is unconditional acceptance but not knowledge."

The question is not merely terminological. There is an occurrent mental state in which one accepts unconditionally propositions that are not evident. It doesn't matter whether we call this 'belief' or something else.  But calling it 'belief' comports well with ordinary language.

Let me now elaborate upon this account of belief, or, if you insist, of Aquinian-Pieperian belief.

1. Belief is a form of acceptance or intellectual assent. To believe that p is to accept *p*, and to disbelieve that p is to reject *p*. One may also do neither by abstaining from both acceptance and rejection.  (Asterisks around a sentence make of the sentence a name of the Fregean proposition expressed by the sentence.)

2. If acceptance is the genus, then knowing, believing, and supposing are species thereof. In knowing and believing the acceptance is unconditional whereas in supposing it is conditional. It follows that believing is not common to believing and knowing as on the JTB analysis. To think otherwise is to confuse the genus (acceptance) with one of its species (belief).

                                                           Genus:  Acceptance-as-true

[Species 1:  Knowledge                          Species 2:  Belief]                      [Species 3:  Supposal]

                     Unconditional Acceptance                                                   Conditional Acceptance

 

3. What distinguishes believing and knowing is that the believer qua believer does not know, and the knower qua knower does not believe. Both, however, accept.  What I just wrote appears objectionably circular. It may seem to boil down to this: what distinguishes believing and knowing is that they are distinct!  We can lay the specter of the circle by specifying the specific difference.

If believing and knowing are species of the genus acceptance, what is the specific difference whereby the one is distinguished from the other? Believing that p and knowing that p are not distinguished by the common propositional content, p. Nor are they distinguished by their both being modes of unconditional acceptance. Can we say that they differ in that the evidence is compelling in the case of knowing but less than compelling in the case of believing? That is true, but then the difference would seem to be one of degree and not of kind. But if knowing and believing are two species of the same genus, then we have a difference in kind.  Perhaps we can say that knowledge is evident acceptance while belief is non-evident acceptance.  Or perhaps the difference is that belief is based on another's testimony whereas knowledge is not.  Let's explore the latter suggestion.

4. It is essential to belief that it involve both a proposition (the content believed) and a person, the one whose testimony one trusts when one gains access to the truth via belief. To believe is to unconditionally accept a proposition on the basis of testimony. If so, then there are two reasons why it makes no sense to speak of perceptual beliefs. First, what I sense-perceive to be the case, I know to be the case, and therefore, by #3 above, I do not believe to be the case. Second, what I sense-perceive to be the case I know directly without need of testimony.

On this approach, the difference between believing and knowing is that believing is based on testimony whereas knowing is not. Suppose that p is true and that my access to *p*'s truth is via the testimony of a credible witness W. Then I have belief but not knowledge. W, we may assume, knows whereof he speaks. For example, he saw Jones stab Smith. W has knowledge but not belief.

Comments

[Account Deleted]

Bill,

Very interesting analysis. A couple of thoughts.

1) You note that in ordinary language one might quite intelligibly say regarding some proposition p:

(*)I don't believe that p; I know that p.

You take this to suggest that knowledge *excludes* belief and, therefore, JTB's thesis that the former is included in the later is false.

However, I think that you might be misinterpretation the common sense evidence to which you appeal. One who says (*) does not mean that they know that p, but do not believe it. Rather, (*) is a shorthand for something like the following:

(**) I do not *merely believe* that p; I know that p.

Thus, they are not denying belief while affirming knowledge. Rather they are denying *mere belief* and affirming something stronger; i.e., knowledge.

2) I am not convinced that you have offered compelling reasons to think that 'accept-as-true' is a common genus of both knowledge and belief and, hence, in some sense conceptually more fundamental. My reservation has to do with the fact that acceptance is a pragmatic notion and, hence, is often relativized to a specific purpose.

For instance, we may agree that John knows (and, hence, believes) that 2+3=5, but he may accept as a premise the contradictory proposition *it is not the case that 2+3=5* for the sake of an argument or an indirect proof. Hence, knowledge does not entail acceptance, at least for a specific purpose.

If we were to accept your anti-JTB approach outlined above, we would turn both knowledge and belief into pragmatic concepts. Are you willing to accept such a consequence? I do not!

[Bill Vallicella]

Thanks for the comments, Peter.

As for your first point, the lingusitic evidence could be read both your way and mine. I wasn't arguing from ordinary language; I was making the point that the view I sketched is consistent with ordinary language, not ruled out by it, though, if you are right, not ruled in by it either.

There is an interesting issue here. People say things like 'Religion is practice, not doctrine' when they should say 'Religion is not mere doctrine, but practice too.' The issue would be whether (*) fits the same schema.

I don't see that acceptance is a pragmatic notion. To accept a proposition is to affirm it; to reject a proposition is to deny it. What's pragmatic about that?

[Account Deleted]

Bill,

I do not think that the linguistic evidence about (*) can be read the manner you suggest. In every case that I can think about a speaker would not intend to deny belief, while affirming knowledge; rather the phrase is used to affirm something stronger than merely believing.

"I don't see that acceptance is a pragmatic notion."

If acceptance is not merely a synonym for belief, then so far as it is distinguished from it, it seems to me to be pragmatic. I gave an example regarding knowing 2+3=5, but accepting the opposite for a given purpose. The same can be said about belief.

"To accept a proposition is to affirm it; to reject a proposition is to deny it."

But then acceptance/rejection are simply synonyms for believe/disbelieve and, hence, cannot be said to be conceptually prior to belief. So I claim this:

(a) Either your notion of acceptance is synonymous with the notion of belief or it is not;

(b) if it is synonymous, then it cannot be a deeper conceptual notion;

(c) if it is not synonymous, then it is a pragmatic notion;

therefore

(d) It cannot serve the purpose of both being a common component to belief and knowledge as well as preserving the non-contextual conception of knowledge proposed by JTB.

[Bill Vallicella]

Peter,

If you asked me whether my father is dead, and I said, 'I believe he is,' your would take that to mean that I did not know that he is. So this is a clear case in ordinary language in which believing that p entails not knowing that p.

What do you mean by pragmatic?

[Account Deleted]

Bill,

Your father-example does not show that believing that p alone *entails* not knowing that p. Rather it shows that believing that p is not a sufficient condition for either knowing that p or for not knowing that p.

Hence, if I hear you say 'I believe my father is dead', I cannot infer anything regarding whether you know he is dead or do not know. Therefore, even in ordinary language I would not take such an assertion to tell me anything about what you know. That is exactly what the JTB view entails.

I will respond in a separate post about acceptance versus belief and my claim that when the two are not viewed as synonymous, then the former is a *pragmatic* notion.

[John]

“Your father-example does not show that believing that p alone *entails* not knowing that p. Rather it shows that believing that p is not a sufficient condition for either knowing that p or for not knowing that p.”

Peter,

If to 'know' p entails having experiential knowledge of p, then Bill believing p entails Bill not having experiential knowledge of p AND someone (other than Bill) does know p. In other words, unless someone (other than oneself) knows p, one cannot truly be said to believe p because we don’t really believe p, we believe the person who claims to have experiential knowledge of p, that is, the one who knows p.

For example, suppose my sister, who has been attending to my father through an illness, calls and tells me (p) ‘Dad is dead’. Presumably, she ‘knows’ p and I believe her, because of which I completely assent to the truth of p. On the other hand, suppose my father just went out for a walk several years ago and no one ‘knows’ what happened to him. In such an instance, I don’t think I can really say I believe p. I can assume p is true or make an informed inference that p is true, but absent a witness with experiential knowledge of p, the assent to truth isn’t complete as it was when I believe p on account of my sister’s testimony. For instance, if X were to tell me he swears he saw my father at the mall, such testimony would not overturn my belief that p is true based on my sister’s testimony, whereas it would challenge my assumption of the truth of p in the case of my father just wandering off.

So to me, belief in p, while maintaining the same complete assent to the truth in p as knowing p, excludes having experiential knowledge of p while mandating someone knows p. In other words, we can’t claim to believe p lest someone claims to know p, nor can we claim to believe p if we have experiential knowledge of p. So unless someone told Bill p, he can’t be said to believe p AND if Bill has experiential knowledge of p, he can’t be said to believe p.

-John