II. MARCO TEÓRICO
2.2. Antecedentes de Investigación
The discussion in the previous section suggests the following view:
Advisor’s Evidence. For any context of epistemic advicecA, ‘S should haveD
to P’ is true atcA iff adoptingDto P accurately reflects how much the advisor’s
evidence supports P.
This gets us the right results for all the discussed cases so far since Kima’s evidence supports believing that it was Omar, meaning that (5) turns out true on Advisor’s Evidence. It also fits with the theoretical consideration that it is absurd, given the aim of advice, for advisors not to look at their own evidence when giving advice.12
12The idea that the practical ‘should’ is in some contexts sensitive to the speaker’s, i.e., the
advisor’s, evidence has been discussed in the literature. See Kolodny and MacFarlane (2010), Bj¨ornsson and Finlay (2010), Dowell (2012, 2013).
However, Advisor’s Evidencefaces a problem. Consider the following case:
Police-Reverse. Lester and Kima both know that it was either Omar or Marlo who killed Bob. Kima was told by her generally very reliable informant Bubbles that Marlo was at the other end of town at the time of the murder. Lester knows that Bubbles told this to Kima, but Lester also knows that Bubbles is mistaken to think that he has seen Marlo at the other end of town. Lester has not shared this information with Kima. Kima walks into the room, where Lester is pondering about the case.
(5) Kima: You should believe that it was Omar.
(5) seems to be false here. One of the reasons is that following it would make Lester adopt a belief that it is not supported by his own, better evidence. A parallel example from practical advice should make this even clearer. Imagine you want to go from St Andrews to Edinburgh by the cheapest mode of transport. You know that this usually would be by bus, but that the bus company doubled the price today because of the Edinburgh Festival. Not knowing about this, I tell you: ‘You should take the bus’. If advice was correct if it was true in light of the advisor’s evidence, my utterance would be true. Contrary to this, my advice appears to be clearly false.
Furthermore, it’s certainly not the goal of advisees to have a doxastic attitude that is supported by their advisor’s evidence, but not supported by their own information-state, where this information-state is better. It follows that Kima’s information-state inPolice-Reverse is not the one Kima should be talking to in order to serve the pragmatic purpose of her utterance. However, since she is apparently not aware that Lester has better evidence than her, she seems to reasonably intend to speak relative to her own evidence. Police-Reverse
could therefore be the first case we’re encountering in which the information-state suggested by the pragmatic purpose and the one the advisor reasonably intends to speak to come apart. If it is the speaker’s reasonable intention that determines which is the contextually provided information-state and Kima really intends to speak to her own information-state, then (5) inPolice-Reverseis actually true. I think we can save our intuition that (5) is false inPolice-Reverseand avoid
having to decide between reasonable speaker intentions and pragmatic purpose as determining the information-state. As I argue in the next subsection, the contextually provided information-state is the collective information-state, the information-state that pools the advisor’s and the advisee’s evidence. On this account, (5) is false in Police-Reverse. Moreover, the account is compatible with Kima’s intention determining the context’s information-state since Kima might just as well be intending to speak to this collective information-state when uttering (5) in Police-Reverse. Since she falsely believes that Lester has no further relevant evidence on whether it was Omar, she falsely thinks that the collective information-state supports believing that it was Omar and therefore utters (5).
Before turning to my preferred account, I’d like to discuss one possible modifica- tion of Advisor’s Evidence. On this version of it, the advisory information-state is the information-state of an advisor who is better informed than the advisee. Kima’s information-state in Police-Reverse couldn’t be the information-state that, according to this account, (5) in Police-Reverse is relativized to, since Kima is not better informed than Lester. However, what would the advisory information-state be instead? If there was none, (5) would have no content. But if (5) is false, it must have content. In conclusion, the refined version of Advisor’s Evidence must also tell us what the advisory information-state is where the advisor is not better informed. Here is such an account:
Better-Informed Evidence. For any context of epistemic advicecA, ‘Sshould
have D to P’ is true at cA iff adopting D to P accurately reflects how much i
supports P, whereiis the information-state of the advisor if the advisor is better informed and that of the advisee S if the advisee is better informed.
so far. It is not without difficulties either, though. The following case shows this:
Police-Equals. Kima only knows that either Marlo or Omar killed Bob. Lester only knows that either Omar or Avon Barksdale killed Bob. Kima says to Lester:
(8) Kima: You should suspend judgement on whether it was Omar who killed Bob.
The first problem is that Better-Informed Evidencedoes not determine an advisory information-state for (8). Kima’s and Lester’s evidence are equally good, so neither information-state is better than the other. We could fixBetter- Informed Evidenceby stipulating that in cases of ties either the advisor’s or the advisee’s information-state is the advising one. One problem with this is that I’m a bit at loss to see which of the one to choose. Choosing either seems arbitrary.
Second, if we chose either of them, (8) turns out true. Both information- states support suspending judgement about whether it was Omar since on either information-state it could as well have been someone else (Marlo or Avon Barks- dale). However, there’s good reason to believe that (8) is false. It is bad advice. Kima seems not to be interested in what Lester’s evidence concerning the case is. To give advice to someone without first checking whether they might have evidence that could put one in an epistemically better position to give advice is arguably wrong.
Third,Better-Informed Evidence requires making precise what makes an information-state ‘better’ than another one. In Police-Equals, it is easy to measure the relative quality of the two information-states. Intuitively, they are equally informative on the issue of who killed Bob. But the more complex, that is, the more like in the real world, the involved information-states get, the harder it will be to make comparisons. I’m not saying that this is impossible. But providing this adds to the complexity of our account of correct epistemic advice. If we could give a simpler model, this would be preferable.