The concept of propositional knowledge -- knowledge that one has through holding a justified belief in a proposition that states a fact -- is a foundational one in epistemology (for example, it is what the SEP entry on epistemology leads with.) There also seems to be wide agreement that there are other forms of knowledge, acquired by direct sensory experience. Sometimes, however, I see the latter conflated with the former: knowledge acquired by direct sensory perception is considered propositional, even when no-one, not even the person having this knowledge, is able to state the proposition that this person allegedly believes.

To be clear, the literature has many examples where a person might not recognize that a given proposition is semantically the same as, or follows from, something they do believe in, and it also seems possible, in the manner of blindsight, that a pathology might leave a person unable to articulate, or even recognize, a proposition that they nevertheless believe, but I do not think that is the case in the examples I am thinking of.

For example, I recently came across (thanks, Conifold!) a defense of the Knowledge Argument that included this passage [1]:

Perhaps Churchland is working with a narrow conception of propositional knowledge in which such knowledge is necessarily quasi-linguistic (or necessarily symbolic). However, proponents (and most opponents) of the knowledge argument use the term "propositional knowledge" in a broader sense, e.g., for the narrowing down of possibilities, whether or not this involves language or symbols.

The emphasis is mine, and I simply have no sense for what this broader sense is, and no idea what the narrowing down of possibilities could mean here.

I realize that once one has gained "what it is like" knowledge through experience, one can, for the first time, understand, formulate, and hold a justified belief in, certain propositions relating to that experience (such as "the color of the sky looks like this one you say is blue"), but it does not seem possible that the "what it is like" knowledge itself is a belief in such a proposition, as a justified belief in any of these propositions depends on already knowing what "it" is like.

I am aware of the doctrine of intellectualism, but it is presented there as a minority view; more to the point, the arguments for it seem mostly to be counter-arguments to anti-inteIlectualism that do not hold up if one is careful about the distinctions between, for example, knowing-how and know-how (knowing a recipe for a soufflé is propositional know-how, but being able to execute it requires more, including the knowing-how of sensorimotor skills.) The remainder strike me as taking the definition of propositional knowledge and presupposing that it must be so for all knowledge -- "if not that, then what?", as it says near the start of the article.

Perhaps my question can be best summarized thus: when someone says they have propositional knowledge of something, is it unreasonable to expect them to be able to state the proposition that they have a justified belief in?

  1. Alter, Torin. (2014). Churchland on Arguments Against Physicalism (section 5.3) 10.1007/978-94-007-6001-1_5.
  • I would say the terminology you are using is vague and ambiguous. Propositions are defined a specific way. They are not justified true beliefs. If one is not aware of a truth value of a statement why call it prepositional knowledge. The term scientific knowledge is more what you likely mean when someone claims x is TRUE in the world. There is also OBJECTIVE knowledge when I make a claim x is forever true. The way you use the term proposition you do not understand the concept. Propositions are concepts or ideas in the mind. They are not sentences but are expressed by declarative sentences.
    – Logikal
    Aug 28, 2019 at 21:08
  • @Logikal "[Propositions] are not justified true beliefs" - it is not clear to me how you could think I am saying this, but if you could point out actual occurrences of this error, I will make corrections... "Propositions are concepts or ideas in the mind. They are not sentences but are expressed by declarative sentences": Either usage is acceptable, depending on context. In this case, my use of 'proposition' is consistent with its use in the articles I quote (if you disagree, please be specific.)
    – A Raybould
    Aug 28, 2019 at 22:17
  • What you are expressing has nothing to do with propositions but how human beings come to obtain knowledge. So I dont see why you are authors are even mentioning propositions because what I see is a mixture muddle of using proposition in more than one context. One is an objective truth context & the other is about sensory knowledge. You can have propositions about both fine but to throw around the phrase propositional knowledge without declaring in what respect is confusing. What is being expressed is how do we know the value of propositions. For the most part how normal folk use knowledge.
    – Logikal
    Aug 28, 2019 at 22:31
  • It is NOT clear in the context you provide what is meant by propositional KNOWLEDGE. The what constitutes as knowledge is not directly stated but there is a high chance they mean SCIENTIFIC KNOWLEDGE aka sensory experience. That is, to claim propositional knowledge is something you were aware scientifically or just using made up semantics. That seems to be the vibe or direction it is going in. But a proposition has a truth value. There is no believing in a proposition as it can only be true or false if meaningful. We can be wrong about a proposition being true. The proposition is just false.
    – Logikal
    Aug 28, 2019 at 22:38
  • 1
    I do not think that "what it is like" itself being propositional is the point. Jackson's argument is rather that, whatever it is, it induces a gain in knowledge, and that could be propositional (or not). The "narrowing down of possibilities" is reminiscent of Hintikka's thesis that epistemic growth amounts to ruling out possible "worlds", see Jago, Logical Information and Epistemic Space. One can certainly have non-propositional means (intuition, skill, know-how) for doing that. Mental models of Johnson-Laird are analog, not symbolic.
    – Conifold
    Aug 28, 2019 at 23:01

4 Answers 4


This is not a definitive answer, but too long for a comment. The OP quote has a footnote listing the "proponents (and most opponents) of the knowledge argument" who take propositional knowledge "in a broader sense". Among the references are Lycan, who is classified by SEP under The New Knowledge/Old Fact View on Mary. According to this view, "what it is for an organism to acquire and possess a phenomenal concept can be fully described in broadly physical terms", but "a subject can acquire and possess phenomenal concepts only if it has or has had experiences of the relevant phenomenal kind".

So can a propositional, in this sense, belief be put into words? This is reminiscent of the Kant's conundrum. On the one hand, a priori propositions can be justified a priori. On the other, concepts in them can only be acquired through experience. In this case, we have description in place of justification, but, to quote SEP again:

"Physical concepts and phenomenal concepts are cognitively independent: it is impossible to see a priori that something that falls under a physical concept of a particular phenomenal character also falls under the corresponding phenomenal concept of that phenomenal character".

It sounds very much like one, at least, can not apply the relevant concepts based on linguistic description alone. In a way, the "linguistic" side of belief ("physical concepts") is not the whole of belief.

Another reference is Stanley and Williamson's Knowing How. Stanley and Williamson are the leading current champions of intellectualism, i.e. to them knowledge-how is a species of knowledge-that. The linked paper is a long defense of this thesis. However, this may not matter that much because of what Alter calls "the propositional knowledge claim: phenomenal knowledge is at least in part propositional". This is what they have to say (p.427):

"According to the first theory, the contemporary Russellian theory, propositions are ordered sequences of properties and objects. According to the second theory, the Fregean theory, propositions contain modes of presentations of properties and objects, rather than the properties and objects themselves. Finally, according to the third standard theory, verbs such as 'believes' and 'knows' express three-place relations between persons, Russellian propositions, and ways of thinking of Russellian propositions.

Our view can be stated in any of these three frameworks. For clarity's sake, however, we shall take propositions to be Russellian, as in the first and third of these theories. The propositions that concern us [those of knowledge-how] will contain ways of engaging in actions. To be more precise, we shall take ways to be properties of token events."

This sounds more puttable into words wholesale. But Stanley and Williamson go on to distinguish between the "demonstrative mode of presentation" and the "practical mode of presentation" (of a way of riding a bicycle), with the latter involving "possession of certain complex dispositions". And "it is simply a feature of certain kinds of propositional knowledge that possession of it is related in complex ways to dispositional states". Since they themselves draw a parallel between that and "first-person mode of presentation", again it sounds a lot like the "demonstrative mode" alone, which is the verbalizable part, presumably, does not suffice.

  • I have to admit that there is much I do not follow here, but I get the impression that you and @jobermark broadly agree - is that fair to say?
    – A Raybould
    Aug 30, 2019 at 3:22
  • @ARaybould I think the issue is deeper than simply having beliefs one can not express, for lack of vocabulary or insight. These phenomenal concepts or first person modes of presentation are not something suited for such an expression. They are not mystically ineffable, even mundane, it is just that language is a limited medium that has to open into reception and action to function, beyond transcribing words into more words. Churchland draws a bright line between the two types of concepts and calls equivocation, but "proponents and most opponents" see a fusion there.
    – Conifold
    Aug 30, 2019 at 7:43
  • Alter also references his own Know‐How, Ability, and the Ability Hypothesis, where he argues against the Lewis-Nemirow identification of what-it-is-like with a kind of ability, to contrast it to propositional knowledge. That is in the same spirit as Churchland's bright line (although he does not equate the phenomenal to ability), hence Alter's "phenomenal knowledge is at least in part propositional" counter.
    – Conifold
    Aug 30, 2019 at 7:51
  • 'Ineffable' was probably the wrong word. Nevertheless, there seems to be a very pragmatic distinction between the sort of knowledge that can be learned from books, and the rest. Churchland suggests that the physiology of learning is different between the two cases. Maybe this should be the basis for a different question.
    – A Raybould
    Aug 31, 2019 at 11:54
  • @ARaybould But would a pragmatic distinction suffice? If all knowledge is in part propositional and in part phenomenal (and/or ability), with one or the other emphasized for pragmatic purposes (say, style of learning), then knowing what it is like to see red and "discursive scientific knowledge" are no different in principle. They are just extreme cases of the same phenomenon. Churchland needs a more robust distinction for his objection, I think. One could say that even book knowledge is not knowledge without a link to some acquaintance and/or use, at least for writing more books.
    – Conifold
    Sep 1, 2019 at 7:52

It is reasonable to point out that one cannot take a position on this question until one has decided what 'belief' means. A pragmatic definition of a belief is that one believes a proposition if one acts as though it were true, or is aware of a violation when one fails to do so (experiencing surprise, fear or confusion.)

In that case, then, yes, one can have propositional knowledge without realizing that one has it, and in such a case, it is unreasonable to assume that you can state it. People entirely without language clearly know that if you duck when someone throws something at you, it is less likely to hit you. Very young children know this, many before they can articulate anything but desires. You can tell because they deploy this strategy.

It seems that children are born with a certain fund of propositional knowledge. Contrary to the notions of earlier theorists like Piaget, a child does not at some point 'develop' the notion of object constancy. Instead, they develop the ability to actively deploy it. Prior to that period, they indirectly express confusion when object constancy seems to be violated. That means, using the pragmatic definition of belief, that they have that propositional knowledge from much earlier in life, most likely from birth.

Maybe we need some motivation for this definition of belief.

From the POV, that provides the definition here, behavior is what matters and language is irrelevant. The name 'propositional' is not about what is or is not statable. It is about what kinds of behavioral patterns propositional logical operations apply, and what kinds it doesn't. Propositional logic applies to justifiable beliefs -- the possibility of justification via rules has to be there, and a specific form of 'true' has to matter. There are certainly other kinds of knowledge, but why insist it is their form that sorts them, and not their associated behavior?

There are certainly behaviors that cannot be combined (processes), or that when combined do not admit the construct of negation (most fear responses), or that are heuristic in nature and do not keep form when combining (e.g. tribal politics), or that one can otherwise derive productive value out of without acting as if they are 'true' in the propositional sense.

The making of bread is not a proposition, it cannot be 'or'-ed or 'and'-ed with other processes based primarily on their instantaneous state, the way a decision can. The kind of combinations possible are varied, but completely different. You cannot act like your bread recipe is true, only as if it is good.

Tactical actions embedded in the immediate physical environment tend to combine rationally according to some informal physics with definite rules. So ducking is worth considering to be a response to something propositional -- I and that moving object cannot properly occupy the same space at the same time.

Emotional manipulations don't -- If I keep mentioning how nice your mother is, you may in fact viscerally respond to the fact II am judging her at all, even though I have purposely negated the attack -- this is Freud's 'primary process' which is prior to propositional logic. They seem like beliefs, but they are not true in a way that makes their negations false. I can love and hate your mother, and it might mean I consciously tell myself she is nice.

Likewise, if you are a voter who 'believes' in small government and elects people who bloat the deficit with defense contracts, the two things do not constitute a contradictiion, or surprise you at all. Because they are not actually beliefs in the sense that one acts as if they are true, whether they are stated that way or not. They are instead encoded tribal affiliation, and identies have internal contradictions, so they follow a different kind of logic from propositional logic.

Why not decide whether knowledge is propositional based on how we observe it works, instead of creating a fiction of 'how the mind uses it' that forces language into places it doesn't belong, and which we cannot validate or really describe?

  • Let me see if I follow... When, say, we see someone duck a missile, it is useful to regard the situation as if they had a belief in some inexpressible or currently unknown proposition? As I have been slowly reading Stanley and Williamson's "Knowing How" paper (from @Conifold), I have been wondering if what they have done is to show that this is how language (specifically, English) copes with the ineffable character of what some people call non-propositional knowledge.
    – A Raybould
    Aug 30, 2019 at 3:20
  • You have missed the point, which is that propositional form does not make something a belief. The ability to behave as if it is true (in a specific way) does. And that is what determines whether the rules of propositional logic apply to it or not. I have edite in a motivation for this point of view. It is a point of view, different from the one you advance, but it is not some kind of front or lie which the tone of your last sentence implies. It is not a cheat, it is a different framing.
    – user9166
    Aug 30, 2019 at 13:36
  • And while you are criticizing my notion of what a belief is, you still haven't given one. So don't be too dismissive.
    – user9166
    Aug 30, 2019 at 14:32
  • I did not intend to be at all critical. In fact, I think you are making some very good points, and I am trying to learn from them.
    – A Raybould
    Aug 31, 2019 at 11:38
  • In this view, would how we come by our beliefs and acquire knowledge be within the scope of how it works? I would think it very reasonable for it to be so.
    – A Raybould
    Sep 4, 2019 at 21:10

However, proponents (and most opponents) of the knowledge argument use the term "propositional knowledge" in a broader sense, e.g., for the narrowing down of possibilities, whether or not this involves language or symbols.

Ignorance may be regarded as the incapacity to exclude possibilities. Knowledge may similarly be thought of as the capacity to exclude possibilities as being understood somehow as impossible. Thus, knowledge, in the broadest sense conceivable, is the narrowing down of possibilities by the exclusion of impossibilities.

Obviously, this capability can only be based on the subject's beliefs unless the definition of this "narrowing down" became circular. Beliefs may be thought of as the subjective and only justification for the subject in this operation of the narrowing down of possibilities. If I believe he is dead, I will exclude a range of possibilities, for example that the two of us could again discuss what knowledge is.

No belief is of course quite enough in itself. The narrowing down of possibilities should somehow have at least some degree of operational effectiveness. The narrowing down by the police of the possibilities that the suspect has in terms of hiding places will amount to knowledge only if it is effective in increasing the probability that the suspect be caught.

I take these sorts of theorising of knowledge as metaphysical desperation whereby we try to conjure the rabbit of knowledge to come out of the proverbial hat of our ignorance.

This notion of the "narrowing down of possibilities" and its implication that knowledge is properly operational effectiveness can only transmogrify knowledge from something subjects have--i.e. something subjective subjects have--to something agents do, that is, something agents do objectively.

If CIA agents are effective in tracking down some terrorists to their hiding place, it would have to be because they, somehow, know something. Doing an effective tracking down would be what knowledge consists of. It would be just what agents do, and not only CIA agents. Thus, subjects have subjective beliefs and act on them, but only the operational effectiveness of what the agent does would constitute knowledge.

Claims to knowledge, in this perspective, would now have to be seen as something again agents do. To believe agents really know what they are doing, we need only look at what they actually do. Knowledge as fitness of the agent to its environment.

But then what kind of agents? Humans? Why only humans? Ants are very effective agents. But, is one solitary ant effective or is the the collective of the ant-hill which is the real effective agent? And then, is not the true agent in effect nature?

But the general drive to define knowledge as some sort of objective operational effectiveness doesn't care for metaphysical conundrums. The agent does, therefore the agent knows. To each agent its share of total knowledge.

Sounds like we end up with some sort of omniscient nature. An interesting twist.

  • "Ignorance may be regarded as the incapacity to exclude possibilities" i get what you mean, but you shouldn't begin answers on philosophy.stack with new definitions. -1
    – user38026
    Aug 30, 2019 at 14:21
  • @another_name It is not a definition. It is an explanation. You are free to buy it or not. The rest of my piece is inferred from that and as such it is at best only as good as this first explanation. Aug 31, 2019 at 10:19

Can one have propositional knowledge without knowing the corresponding proposition?

Depends on what you mean.

  1. We can, I think, have propositional knowledge (non propositional knowledge cannot be expressed in propositions) that is not expressed in propositions. Surely I don't need to know that this is blue, to know this blue (Which is no proposition)?

There may be disagreement on this, I'm not sure. e.g. this from a Finnish University page says that it is useless to claim that someone could justify a true proposition but won't. If right, then surely propositional knowledge must consist of a proposition that we've justifed.

Crisholm (Theory of knowledge, 1966) has speculated... that the subject can have adequate evidense, even if s/he doesn't know that s/he knows. However, his own formulation produces same problems, as he himself states: some propositions can be evident to the subject, and still the subject doesn't know that they are true... Crisholm's speculations are so senseless and self-contradictory that they are not even worth of disarguing here.

  1. Can we have propositional knowledge (believe that this is blue) that is not propositional knowledge (not believing that this is blue)? Not in any robust sense.
  • i think i misunderstood...
    – user38026
    Aug 31, 2019 at 20:47

You must log in to answer this question.

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