3.1 What Assertion is Good for 42
3.1.4 Overriding Functions 53
A trait/artefact/act can have several e-‐functions simultaneously, even several e-‐functions of different types; take, for instance, the functions served by food for humans. One important such function will surely be a biological one, a nourishment function. Plausibly enough, though, on top of this, food also serves an aesthetic function for us, that of generating pleasant gustatory experiences. Now, normally, the aesthetic function complements the nourishing function. It serves, as it were, the greater goal of survival, by increasing the probability of us ingesting nourishing substances. This need not be the case, however; there can be situations where the two functions come in conflict, at which point the more stringent requirement will take precedence and dictate what’s the all-‐things-‐considered good to observe. Think about a case where I am on a deserted island and all I can eat in order to stay alive are my boots; surely, against my aesthetic well being, that is exactly what I should do, all-‐things-‐considered.
Similarly, it is highly plausible that the epistemic function is merely one of the many functions served by assertion. Surely, when I tell you that the weather is nice while sitting with you in a café, generating testimonial knowledge in you with regard to meteorological states of affairs is hardly among my chief concerns. In this case, assertion rather serves a social bonding function.
One of the most important functions of assertion, as with action in general, will be a prudential one, serving our survival. Again, just like in the case of food, the epistemic function will, in most cases, complement this prudential function. Generating testimonial knowledge in one’s hearer with regard to an imminent threat, or about the whereabouts of resources are paradigm cases. However, again, this
need not be the case. For instance, even if one knows that one’s boss is bald, it may not be polite, prudent, or relevant to point this out to him ((Jessica Brown 2010, 550), henceforth, BALD)): surely, here, the prudential function comes in conflict with the epistemic one and overrides it.
Here is, then, my proposal: both Classical Invariantism and KNA hold. In virtue of the function of assertion of generating knowledge in hearers, one is in a good enough epistemic position to make an epistemically proper assertion that p if and only if one knows that p. In turn, where whether one knows (or ‘knows’) that p is insensitive to practical matters. Crucially, on the account defended here, what KNA claims is mere epistemic propriety. The standards for all-‐things-‐ considered propriety of assertion – as for action in general – will vary with context, while standards for both epistemically proper assertion and knowledge will remain fixed. The way to ascertain whether the requirements at work in one case or another are genuinely epistemic requirements is by looking at the function that is plausibly being served.48 An epistemic function will be associated with an epistemic norm.
To see how this works, consider Williamson’s TRAIN case:
TRAIN: Suppose that I, knowing that it is urgent for you to get to your destination, shout “That is your train” upon seeing a train approach the station, although I don’t know that is the case; I merely believe there is some chance that is the case, and I think you should check it out. (Williamson (2000)).
According to the present proposal, the assertion in TRAIN is epistemically defective, but all-‐things-‐considered proper, due to prudential considerations stepping in and requiring a lower degree of warrant (Override2). In view of the epistemic function of generating testimonial knowledge, this assertion is not permissible. However, in view of the prudential function of raising your chances at catching the train, the assertion in TRAIN is perfectly fine. Given that the prudential constraint overrides the epistemic constraint in this case (as in most), by lowering the degree of necessary warrant, the assertion is all-‐things-‐ considered proper, which is what triggers our relevant intuitions.
48 At this point, the reader might worry that the functionalist account defended here,
in virtue of being a variety of epistemic consequentialism, suffers from the classical problems associated with this view. Fortunately, this is not the case: in a nutshell, that is because while functionalism is, indeed, a variety of consequentialism, crucially, it is a rule consequentialism. The classical cases against consequentialism, however, arguably, only affect act consequentialism. For extensive discussion, see section #3.2.3 below
Similarly, in ASPIRIN2, strictly epistemically speaking, if you have memorial knowledge that you have aspirin at home, you are permitted to assert. You are in a position to make an assertion that, in normal conditions, will fulfil its epistemic function of generating testimonial knowledge in your hearer. However, prudential constraints referring to the tragic consequences your assertion might have override the epistemic constraint, raise the necessary degree of warrant, and make the assertion all-‐things-‐considered improper. As such, again, it is all-‐things-‐considered propriety that varies with context, not epistemic propriety.49
3.2 Objections and Replies
Before I continue, in what follows, I would like to take a brief look at (and try to answer!) what I take to be the main objections that come to mind when presented with my account. I will try to do justice to three importantly different directions of possible discontent: contextualism and pragmatic encroachment, anti-‐functionalism and truth-‐centred functionalism.
3.2.1 Objection #1: A Contextualist Rejoinder
On the view defended here, the contextualist data is explained by normative overriding: while epistemically fine, it is prudentially impermissible for DeRose to assert ‘The bank will be open tomorrow’ in the high stakes scenario. Again, this claim draws its plausibility from what I have dubbed Value Individuation of normative constraints: look at the goods associated with the norm! In the bank cases, of course, practical goods are at stake when it comes to variation in assertability; therefore, on the account defended here, unassertability is triggered by the prudential norm corresponding to the prudential function of assertion, overriding the epistemic function.
Recall, however, that, according to DeRose, not only can he not assert ‘The bank will be open’ or the corresponding ‘I know that the bank will be open’ in high stakes cases, but it is even appropriate for him to deny knowledge to himself. Now, what does the functionalist picture have to say about this?
49 Note that it is not in my intention to suggest that practical concerns will always
override epistemic ones (or aesthetic ones, for that matter). The claim made by this chapter (and needed here) is weaker than that: practical requirements will, on occasion, do so. I gather that this is a fairly plausible thought.