Download File

McDowell Credit and Confirmation Intervals

McDowell Credit and Confirmation Intervals

One of the most interesting points McDowell makes in his discussion of testimonial knowledge in “Knowledge by Hearsay” pertains to confirmation intervals—particularly—how it is that we can have knowledge of some proposition p at some time t occurring between confirmation intervals for p.

To borrow McDowell’s example, let p represent the proposition “George W. Bush is president.” Suppose further that you are a loyal watcher of the 6:00 p.m. news. Last night, when watching the news, you see that Bush is giving a speech in Minnesota. You don’t leave the house all day, and it is now 3:00 p.m., three hours before you will watch the news again (and, thus, have the opportunity to reaffirm your belief that Bush is President.)

Do you know, at 3:00, that Bush is president?

McDowell, recognizing that it would be counterintuitive to think that your knowledge that Bush is president “disappears” as the television screen turns black, thinks allows that you can know, at 3:00, that Bush is president.

Of course, if Bush had suddenly died in the time interval between the previous newscast and your forming the belief that he is president at 3:00, then you would not know that he is president. But, so long as he still is president, you do know.

McDowell thinks that, for a proposition such as “Bush is President”, belief in this proposition at some time within (for example) a day of the previous confirmation interval that it is so meets whatever criterion of doxastic responsibility on behalf of the agent that is needed for the agent’s factive state of remembering that Bush is president to constitute knowledge.

What McDowell doesn’t explain carefully enough, however, is how one would go about answering the following amended versions of the previous example:

Scenario 1: Amend the Bush-TV scenario so that the agent has no confirmation interval that Bush is president for 15 days, and then utters: “Bush is president”.

Scenario 2: Amend the Bush-TV scenario so that the agent has no confirmation interval that Bush is president for six months, and then utters: “Bush is president.”

What McDowell does say is that, at some point, your belief that “Bush is president” will require another confirmation interval (even if your reason for believing that he is president is that you ‘remember that’ he is president) if you are to continue to know that Bush is president. He likens this idea to the analogy of a plant, needing water every so often. What McDowell doesn’t say is (in his parlance) how often the plant needs watered, and perhaps more importantly, how we can come to know how we must water it.

Appealing to base intuition, I’d suspect that the agent does not know in Scenario 2, and ‘maybe’ knows in Scenario 1.

I have no intuition, though, as to how to go about assessing at what point your initial knowledge that Bush is president ‘runs out of water.’ This is because, as I see it, it’s not clear if it is relevant (for example) whether between your initial confirmation interval and (say) Scenario 1, there are death threats placed on Bush of which the agent is unaware. Would this affect the extent to which adequate doxastic responsibility would require a confirmation interval? If there are (suppose) bullets flying haphazardly through the air at Bush five minutes after I turn the TV off after watching him give an address, and none of the bullets hits Bush, can I still know that Bush is president (five minutes later?)

It’s just not clear how McDowell would go about addressing these sorts of questions.

At this point, I am drawing two conclusions:

1. McDowell has no good way of explaining how an agent can ever know that she knows any given proposition at some time t between confirmation intervals. This is because, to know that one knows such propositions, one must first know that one has been adequately doxastically responsible in attending to the needed confirmations. It’s not clear how, on McDowell’s view, knowledge of the latter would be within the reaches of one’s rational powers.
2. McDowell has no good way of explaining how an agent can ever be credit-worthy for knowing some proposition at a time t in between confirmation intervals. This is because, for the agent to be creditworthy for knowing some proposition p, the agent (and not brute luck) must be relevenatly instrumental in an explanation for why the agent knows p. On McDowell’s account, though, it seems that whether or not one has met the requirements of doxastic responsibility might well be a matter of luck.

I shall elaborate this last idea.

Suppose I bury a silver coin underneath a tree in my back yard. I see that the coin is there, and suppose further, that the next day, I remember that the coin is there. Suppose I never again check under the rock to see if it is there.

On McDowell’s view, there is some point in time tn which I no longer know that the coin is there, even if the coin is still there. Suppose that, seconds before tn I form the belief “the coin is there.” Seconds later, and infinitesimally after tn, I form the belief “the coin is there.” McDowell seems to rule that I know in the first case, but not in the second. If McDowell is to preserve the intuitive idea that knowers are credit worthy, then he must allow that my belief seconds before tn constitutes a creditworthy achievement, but not the belief I form seconds after tn. This consequence seems rather hard to swallow.

For a final thought: I’m not sure whether I am right to think the following or not, but here it goes: one might be able to argue that the McDowell doesn’t adequately satisfy the safety principle. The case that (might) work to illustrate this would be an amendment of the silver coin case.

Imagine that, five minutes after you plant the coin, it spontaneously (due to some quantum fluke) catches fire and burns to ash. Four minutes after you planted the coin, you form the belief that the coin is under the tree. You are (probably) doxastically responsible, and you remember that the coin was under the tree, however you could have easily been wrong. Why could you have easily been wrong? Because you could have just as easily formed your belief about the coin’s being there five or six minutes after planting it as you could have four minutes after planting it.

I’d be interested to get some feedback as to how a McDowellian could address some (or all) of these concerns.

Available link for download