The Pragmatic Dimension To Knowledge

By Ciprian Barnutiu, 9002294 Phil. 264, Prof. Brian van den Broek November 23, 2009

In The Pragmatic Dimension to Knowledge, Fred Dretske confronts the skeptical argument to knowledge, namely that knowledge is an ideal that we never truly attain, even when we think we do. Our justifications cannot be absolutes because we can compare them and they can be more or less adequate. But propositional knowledge is an absolute because you cannot know a proposition more or less than me. Knowing implies a certain degree of justification that qualifies your belief as knowledge. Justification is the qualifier for knowledge, meaning that we need to reach a certain threshold of justification to say we have knowledge of something. This means that we can be more certain in knowing a proposition by being more justified, but we cannot know a fact better just because we have more justification. There is a limit to how justified we need to be in order to know something. Dretske picks up on Peter Unger’s idea of absolute knowledge in order to illustrate how we use the concept of absolutes to build a pragmatic dimension to knowledge. We conclude from Under’s analysis that knowledge is absolute because the threshold of justification for a person to have knowledge is absolute. It exposes the irregularities of our world. The skeptic position argues that since knowledge is absolute, and since we can never be absolutely certain, we have no knowledge. However, Dretske departs from Unger at this point.

1

They agree that knowledge is absolute, but Dretske argues that we need not side with Unger’s skeptical conclusions. We should not conclude, as the skeptics do, that since we cannot eliminate all possible justification we do not know anything. Dretske arrives at a conception of knowledge as relationally absolute by offering an alternative source of absoluteness. For Unger, the absoluteness comes from a certainty that one knows, whereas Dretske regards justification for knowing as the source of absoluteness [Dretske, 541]. He is right to make this distinction because if an absolute is something devoid of a certain thing, than we must decide what that thing is. After all, if there were no thing we were looking to exclude from the concept, than it would cease to be an ideal. We have to ask what the concept is devoid of? Once we do, it becomes evident that what I consider a broken computer is not a broken computer to a computer salvager. This computer is devoid of its expected capabilities (e.g., powering on) but the hardware found inside the computer, is not broken because it can be recycled or reused. Thus, the absoluteness of broken (and any other absolute term) depends on the intention of the person using it, not on the certainty required for knowledge. So it is appropriate to say that everyday absolute concepts are mediated by a kind of specific relevancy. However, skeptics make the claim that everything is relevant and nothing is irrelevant. Dretske does not disprove this, because the best he can do is to say that the possibilities that a skeptic sees to be relevant, are simply too remote to be considered as such [Dretske, 543]. Dismissing unlikely alternatives is possible, and acceptable, because as I noted earlier, knowledge has shown itself to be governed both by the absolute and the relevant. This ideas of the relevant allows us to dismiss the skeptics alternatives in the same way the salvage specialist would dismisses the notion that the computer is broken (e.g., just

2

because it doesn’t power on when its plugged in). Dretske claims that the dismissal of some alternative is “a half conscious attempt to exhibit the relationally absolute character of our cognitive concept. [Dretske, 543]” Therefore, any skeptical argument that denies this claim, Dretske follows, is a misunderstanding of what needs to be justifiably excluded in order to know. The relevancy set, as a subset of the contrasting set, is a constantly changing set of alternatives that needs to be conclusively proved wrong in order to make claims to possession of knowledge. This set of alternatives cannot be the same in every situation even if the knowledge one has is the same [Dretske, 545]. In contrast to the relevancy set stands the contrasting set, which is the set of all possible alternatives. For the skeptic, CS=RS, because relevancy does not mediate possible alternatives. Skeptics may be true on a theoretical level, but Dretske is looking for a pragmatic way to consider knowledge. It is a way to apply absolute concepts to everyday cases. The only problem Dretske faces is now is putting forth a criterion for membership in the relevancy set. If Dretske cannot do this, than his idea of relevance cannot stand the arguments of the skeptic. The language we use helps us discriminate what is relevant or not. In articulating a proposition, we can have different intentions. That is, we can focus on different aspects of the proposition to indicate a different range of relevant alternatives. The proposition that you are reading an essay right now can require different relevancy sets. By the use of contrastive focusing we would either need to eliminate the possibility that you are eating the essay right now, or that you are reading the cover-page of the essay right now. This contrasting can be continued until we have multiple focuses on the proposition and consequently, until we have different relevancy sets. This is a compelling example of

3

how we can separate relevant from irrelevant alternatives in our struggle to know. Another way of determining membership of the RS is by taking for granted that something is known. This is achieved by our choice of subject term [Dretske, 547]. In claiming that I saw a car run over some person, I am taking epistemic responsibility for claiming to know that the thing being run over was in fact a person. But by begging the question, I am saying that “person” in this sense is included in the relevancy set. We also take for granted the way in which we come to acquire knowledge. Again, we take epistemic responsibility for our source of knowledge, but the premise is that if these “channels” of information are in fact reliable, than we can claim to have knowledge. The “taken for granted” source is therefore included in the relevancy set, with the caveat that if I do not eliminate all the relevant alternatives (which would prove that the source is faulty) than I cannot claim to have knowledge. The importance of what is known could also predict the relevancy set. That is, if the stakes are really high we must take a greater consideration of relevant alternatives; otherwise we cannot claim to know. Dretske disagrees with this idea, but not because it does not have any appeal. Indeed, he admits that it does, but I believe he also shows why it is wrong. In the auto gage/nuclear reactor case, there is some reason to doubt this criterion for the RS. I agree with Dretske that there is no reason we cannot say to know the coolant level by looking at the transplanted auto gage. In principle, a measuring tool, assuming it functions in a different context, remains a measuring tool. However, as Dretske notes, our concern in this ‘unusual’ example is not whether the gage is accurate. But rater, we should be cautious of when we stop knowing that the auto gage is accurate. We have to remember that in our premise the auto gage can tell the coolant level,

4

therefore we need only to refute the possibility that the gauge does not tell the right coolant level, not whether it will break down. Finally, we have to consider what counts as an irrelevant possibility. We have to ask ourselves what makes an alternative relevant or irrelevant. The pragmatic way wants to make the distinction between what is regarded as a relevant possibility and what is objectively possible in an every day situation. We see in the bird-watcher example that if the grebe is not able to migrate to North America, than we need not include it as a relevant alternative to the bird-watcher’s claim that the bird is a Gadwall. In other words, there needs to be an objective reason for excluding a possibility. If a rumor is spread that there are grebes in the same area as the bird-watcher, we are compelled to say that since the birdwatcher can’t disprove the alternative, he does not know that he saw a Gadwall, even if he thinks he does. This discussion has shed light on the relational absolute theory of knowledge. It shows that if we look at the relevant implications of knowledge we can (and do) create relations between our truth claims and our daily experience of the world. I can narrow down the relevancy to say that I know the fridge is empty of food, but I cannot say that the bird I see is not a grebe (if indeed, grebes exist). Therefore, Dretske’s argument is very useful for explaining the applicability of absolutes to every day situations, but it does not disprove the global skeptic, in fact, it agrees that there are no absolute justifications if there is no subset to the contrasting set. But Dretske shows us that our use of absolutes can actually give us knowledge. We can use absolutes to consistently agree in our claims to know what are the relevant implications for knowledge. It is a theory of knowledge that works on the local level of skepticism.

5