Epistemic Norms

The following combination of views seems a natural one to me, but (in my admittedly limited excursions into the field) I have not yet seen it defended:

1. True epistemically normative statements are made true by objective facts.
2. These facts are naturalistically respectable facts, having to do with e.g. the truth of the subject’s beliefs, or (better) their probable truth given her evidential position.
3. This does not mean epistemically normative statements are equivalent in meaning/sense to statements which are explicitly about truth of beliefs or their probable truth.

Cf. the corresponding view in ethics, that true moral statements are made true by objective natural facts about e.g. the maximization of utility, but do not have the same meaning/sense as statements which are explicitly about the maximization of utility.

Naturalism about epistemic norms often seems to get equated with the Quinean project; other options are not mentioned. I’d be interested to know if 1-3 are defended in some area of the literature that I haven’t yet managed to stumble upon!


Epistemic Norms — 8 Comments

  1. I accept the following as a conceptual truth about belief:

    A belief that p is correct if and only if p is true.

    The claim that a belief is correct is a normative claim, whereas the claim that a belief is true is a descriptive claim. Since I favor an expressivist treatment of nromative claims, I deny that the two statements are equivalent in meaning. So I think I hold a position of the type you describe.

    However, in the literature in epistemology about epistemic normative statements that I am familiar with the distinction is rarely drawn between the question what an epistemic statement means and the question what makes an epistemic statement true, so it is very difficult to determine whether the position that you describe is advocated by anyone.

  2. Thanks for the comment Nishi. (Is there something of yours I could read about the expressivist view?)

    You write:
    ‘Since I favor an expressivist treatment of nromative claims … I think I hold a position of the type you describe.’

    As an expressivist, do you accept my claim 1? I guess one might, for minimalist/quasi-realist reasons …

    Also, one could, I suppose, accept a biconditional like yours without thinking the fact described on the RHS truthmakes that described on the LHS (given certain views about truthmaking).

  3. Here is a link to a paper that describes the expressivist view I favor. Much more work needs to be done to fully articulate such a view though.


    You are right that it is only under a deflationary interpretation of the truth predicate that an expressivist can accept 1. And come to think of it, I’m not sure that I want to tie myself to a deflationary view of the truth predicate. So maybe it is best for the expressivist to take your suggestion and deny that the RHS of the biconditional gives the truth maker of the LFS. But it is open to a non-expressivist to accept the biconditional and to accept 1-3.

  4. Hi Jenkins,

    It seems to me that, at least to the non-contextualist, epistemically normative statements are made true by the satisfaction of a suitable supervenience relation. If by ‘epistemically normative statements’ (ENS) you mean statements like

    (i) S is justified in believing that p

    then I think Feldman (‘Epistemology’ (2003)), Alston (1989), Van Cleve (‘Foundationalism, Epistemic Principles, and the Cartesian Circle’ (1979)), Sosa (1991), and others, all claim that (i) is made true by non-epistemic facts about S and/or her environment.

    Since you claim that ‘This [the fact that (ENS) can be made true] does not mean [that] [(ENS)] are equivalent in meaning/sense to statements which are explicitly about truth of beliefs or their probable truth’, I believe you are not interested in statements like the ones extensively discussed by Pollock (‘Contemporary Theories of Knowledge’ (1999))

    (ii) A belief is epistemically permissible if and only if the epistemic agent believes it to be highly probable

  5. Carrie — I’m totally sympathetic to your general project, which is to distinguish certain broadly speaking reductive relations (such that the relation that you indicate by saying that certain statements are “made true” by certain facts) from relations of meaning equivalence and the like.

    However, I have to admit that I really don’t know what is meant by “naturalistically respectable facts”, and I also really don’t know what is meant by saying a statement is “made true” by a certain fact. The philosophers who like appealing to these locutions have in my opinion never clearly explained which of many different notions in the general neighbourboods it is that they really have in mind.

    Perhaps I shouldn’t whinge too much about this, since I strongly suspect that however your claim (2) is elucicated, I would end up resolutely rejecting to it. This is because I am strongly inclined to think that epistemically normative statements (like ‘S knows p’ or ‘S is justified in believing q’) express irreducibly normative facts. (Indeed, I am even inclined to think that the property of truth, at least when it is intelligibly predicated of speech acts or mental events, is also an essentially normative property — not an austerely logical property as the the deflationist typically believes; and the property of truth that is intelligibly predicated of sentences and propositions is then at least in a weak sense a normative property as well, since it is an essential feature of sentences that their utterance can be used to effect speech acts, and of propositions that they can be the objects of mental events like judgments etc.)

  6. As the topic Carrie is proposing is a very interesting one I would like to exchange some references with you all.
    As far as I can see, the epistemologist that more strongly made the point that every epistemogical theory must account for the fact that epistemically normative statements are made true by non-normative facts is Ernest Sosa, in his article ‘The Raft and the Pyramid’ (1980). Sosa’s discussion of epistemic supervenience in that article has one main target: subjectivist accounts of justification such as Keith Lehrer’s. Coherence theories of justification that work with a tighter connection between truth and justification than subjectivist accounts do can, according to Sosa, respect the supervenience constraint on how justification is generated.
    Lehrer’s book ‘Self – Trust: A Study of Reason, Knowledge, and Autonomy’ (1999) has a chapter (chapter 3) entirely devoted to denying the objections to his theory that are based on supervenience considerations. In that particular chapter, Lehrer presents an argument against the thesis that normative properties of the epistemic type are completely determined by natural properties. Lehrer calls his argument for non-supervenience ‘The Independence Thesis’. As far as I know, the more recent publication by Sosa that deals with Lehrer’s argument for the independence of the epistemic in relation to the non-epistemic can be found in Erik J. Olsson’s ‘The Epistemology of Keith Lehrer’ (2002). The title of Sosa’s article is ‘Epistemology: does it depends on independence?’. The book also includes a reply by Lehrer to the objections Sosa presents to him.
    This takes me to Ralph’s point against (2) in the original post

    […] I am strongly inclined to think that epistemically normative statements (like ‘S knows p’ or ‘S is justified in believing q’) express irreducibly normative facts.

    What do you mean by ‘irreducibly normative facts’? Is it something similar to Lehrer’s ‘Independence thesis’? One consequence of Lehrer’s attack on supervenience is a coherence theory of justification that endorse something like the following generative principle of epistemic justification to the belief (acceptance) that p.

    (CO) If the acceptance system A of S contains (CO) and, if p coheres with A, then S is justified in accepting p

    Ralph, would you like to endorse a principle such as (CO)? As you can see, by endorsing (CO) you will have theoretical room to argue that epistemically normative statements are not reducible to natural facts: (CO), an epistemic property, can determine another epistemic property, namely, epistemic justification.

    Thank you very much for your discussion!!

  7. Note:
    Erik J. Olsson’s ‘The Epistemology of Keith Lehrer’ was published in (2003)and not in (2002) as I wrote before

  8. Carrie, GP Baker and PMS Hacker suggest, in their book _Frege: Logical Excavations_ (p. 43), that Frege held a view much like the one you have in mind. According to them, the normativism put forward in the introduction to _The Basic Laws_ conceives of logical laws as anankastic rules which “rest on objective logical relations between judgments”. Now, insofar as truth can’t sensibly be predicated of rules, the plausibility of any such view depends on one’s providing some satisfactory account of the “grounding” of rules on facts (or of their grounding on true descriptions of the facts). As far as I know, Frege didn’t do anything in that regard. (And maybe he actually thought that logical laws are themselves descriptions of logical facts and wasn’t troubled by a normative/descriptive duality, though I think their interpretation is intriguing _and charitable_, the one to be preferred.) Baker and Hacker don’t do anything on his behalf either — as they are concerned with advocating some form of the anthropologistic relativism that Frege opposed. For a discussion of alternatives, see chapter 21 of Crispin Wright’s _Wittgenstein on the Foundations of Mathematics_.

    For a sympathetic (but very brief and ultimately noncommittal) discussion of the view favored by Ralph above (comment # 5, third paragraph), you may want to see Richard Feldman’s “Authoritarian Epistemology” (1995, reprinted in Conee and Feldman, _Evidentialism_), where Keith Lehrer, Stewart Cohen and Richard Fumerton are credited with endorsing some form of the view that there are brute epistemic facts.

Leave a Reply

Your email address will not be published. Required fields are marked *