RSS

Category Archives: Philosophers

Evidence of Evidence is Evidence, or is it?

04 Jul

Branden Fitelson (forthcoming) provides counterexamples to Richard Feldman’s principle that Evidence of Evidence is Evidence (EEE). Here’s the principle in its initial (naïve) form:

(EEE1) If E (non-conclusively) supports the claim that (some subject) S possesses evidence which supports p, then E supports p. (Fitelson forthcoming: 1).

Fitelson’s counterexamples to (EEE) work by presupposing the “positive relevance” (i.e., increase-in-probability) notion of evidential support. In footnote 6 he indicates a more substantive principle of evidential support might be wielded in defending (EEE). In this post I want to explore this possibility, specifically in relation to the notion of propositional justification. Consider the following principle of propositional justification:

S is justified in believing that p iff S’s total evidence sufficiently supports p (Neta 2007: 197).

Though there are many issues that could be raised with this formulation of propositional justification, let’s see if a less demanding iteration of the principle could be used to resist Fitelson’s counterexamples to (EEE). Neta’s principle suggests the following notion of evidential support:

(1) E (evidentially) supports p iff S’s total evidence includes E and S’s total evidence (necessarily) supports p.

The counterexample to (EEE1) involves drawing a card c at random from a deck. All the evidence we are given regarding c is as follows:

(E1) c is a black card.

(E2) c is the ace of spades.

(p) c is an ace.

Imagine a guy named John knows what card c is, and the evidence above constitutes all the facts about the case. This means the following is the case:

(2) E1 supports the claim that John possesses evidence (E2) which supports p.

Positive relevance creates a problem for (EEE1) because (E1) doesn’t raise the probability of (p). (E1) alone is probabilistically irrelevant to (p); so, even though (E1) supports (E2), the second conjunct in (EEE1) is false (i.e., E1 doesn’t support p).

How does the counterexample fare under principle (1) instead of positive relevance? John’s total evidence includes (E1), and John’s total evidence (E1 and E2) necessarily supports (p). (E1) alone doesn’t necessarily support (p), but it also doesn’t support (not-p), and when coupled with (E2) it does necessarily support (p). In fact, (E2) entails (p). John’s total evidence might not sufficiently support (p), but his total evidence does necessarily do so. The next iteration of (EEE) runs as follows:

(EEE2) If E1 supports the claim that S possesses evidence E2 which supports p, then the conjunction of E1 and E2 supports p (Fitelson forthcoming: 2).

This seems like the defense I just gave for (EEE1), assuming (1). Didn’t I just claim the conjunction of (E1) and (E2) supports (p)? If so, then, assuming evidential support principle (1), it looks like the next counterexample will sink (EEE2). However, I think (EEE2) and (1) escape unscathed. Fitelson’s counterexample to (EEE2) is about a guy named Joe:

(E1) Joe has a full head of white hair.

(E2) Joe is over 35 years of age.

(p) Joe is bald.

The example works given the positive relevance principle because the conjunction of (E1 and E2) fails to raise the probability of (p). Being over 35 years of age might raise the probability that one is bald, but having a full head of white hair doesn’t raise the probability one is bald. (E1) supports (not-p), so the conjunction of (E1) and (E2) refutes (p). What about in relation to principle (1) instead of positive relevance?

(E1) is part of the total evidence. But, does the total evidence support (p)? More specifically, is the notion of total evidence equivalent to the conjunction of all the (relevant) evidence? I would argue it is possible to equate the total evidence with the conjunction of all relevant evidence, as I did in defending (EEE1) against a counterexample, but it is not necessarily the case that the total evidence must be regarded as all of the evidence conjoined. There is a probabilistic consideration in favor of this point.

The probabilistic consideration is that the total evidential support for (p) is not determined by simply conjoining the individual probabilities that (E1) and (E2) afford (p). Returning to the card counterexample, the probability that the card is a black card (E1) supports the claim the card is an ace (p) is the probability that: if the card is black, then the card is an ace. Only the ace of clubs and the ace of spades satisfy this condition, so the probability is 2/52 (approx. 4%).

The probability (E2: the card is the ace of spades) supports the claim the card is an ace (p) is 1, as the entailment takes on the maximum value. The fact that (E2) entails (p) means Pr(p|E2) = 1. The amount that the total evidence supports (p) is not simply the product of the probabilities of (p) given (E2) and (p) given (E1). This would yield a probability of Pr(p|2/52 x 52/52) = 104/2704 or about 4%.

A  better estimate is attained through subtraction of the two probabilities. This measures the degree to which one bit of evidence lessens the impact of another bit of evidence on the target proposition. This better approximates the total impact of (E1 and E2) on (p). This makes it the case that Pr(p|52/52 – 2/52) = 50/52 or 96%. The actual probability, though, would be 1 or 100% because the probability the card is black (E1) is swamped by the probability the card is the ace of spades (E2) in relation to the probability the card is an ace (p). As such, (E1) can be disregarded. Again, the correct probabilistic impact of the total evidence on the target proposition is not determined by a conjunction of all the evidence. Let’s apply this back to the second counterexample about Joe’s hair or lack-thereof:

Pr(p|E1) = 0

Pr(p|E2) = .20 (estimate of men over 35 who are bald)

Conjoining the total evidence, again, doesn’t refute the target proposition. Suppose John knows the age of Joe and (2) is true. That is, (E1) supports the claim that John possesses evidence (E2) that supports (p). Does the conjunction of (E1 and E2) refute (p), as Fitelson urges? Multiplying the probabilities, as given above, would yield a probability of 0. It would indeed refute (p). However, this case is dissimilar from the card drawing case because it uses vague terms. Being “bald” is not defined as having “no hair”. Most men who are “bald” still have some hair on their head. Being bald is defined in relation to “male pattern baldness.” This is a progressive condition and, much like the term “a heap”, is wrought with vagueness. The fact that Joe has a full head of white hair (E1) and Joe is over 35 years of age (p) doesn’t make it the case that there is zero probability that Joe is bald. Joe may have hair loss as a result of male pattern baldness, yet to a casual observer (like John) he may appear to have a full head of hair. This is especially the case for someone who has all white hair because the threshold for counting as having a full head of hair is plausibly lower when all of one’s hair is white. Due to vagueness in the terms in (E1) and (p), the fact that Joe is over 35 years of age (E2) is not swamped by (E1).

A better probability estimate is attained by (i) allowing Pr(p|E1) = .05 as a correction on boundary vagueness in the terms, then (ii) subtracting the probabilities to yield the impact of the total evidence on the target proposition: Pr(p|.20 – .05) = .15 or 15%. The total evidence still supports (i.e., does not outright refute) (p) even though the total evidence makes it more likely that (not-p) than (p) is the case.

References

Fitelson, Branden (forthcoming). “Evidence of Evidence is not (Necessarily) Evidence.” Analysis.

Neta, Ram (2007). “Propositional Justification, Evidence, and the Cost of Error.” Philosophical Issues.

 
Leave a comment

Posted by on July 4, 2011 in Branden Fitelson, Evidence, Probability

 

Contrastive Bayesianism

09 Feb

One way to argue against Elliot Sober’s Contrastive Empiricism is to claim that the approach violates an important principle of confirmation theory. Branden Fitelson (2010) has done this in a recent paper by providing a counterexample to Sober’s Law of Likelihood (LL) for ‘favoring’ relations. According to (LL) evidence E favors hypothesis H1 over hypothesis H2 if and only if H1 confers greater probability on E than H2 does. You compare the likelihoods of alternative hypotheses. Fitelson argues against (LL) by proposing a counterexample. The counterexample includes the following facts:

  • E: The card is a spade.
  • H1: The card is the ace of spades.
  • H2: The card is black.

This makes the probabilities run as follows: Pr(E|H1) = 1 and Pr(E|H2) = 1/2. So, on (LL), E favors H1 over H2. However, this seems wrong. The truth of E guarantees (or necessitates) the truth of H2, but the truth of E doesn’t guarantee the truth of H1. This makes E conclusive evidence for H2 but not H1. So, it seems E favors H2 over H1 despite the likelihoods pointing the other direction. For Fitelson this suggests the following confirmation principle, called Conclusive Evidence (CE), a principle which (LL) violates:

  • (CE) If E constitutes conclusive evidence for H2, but E constitutes less than conclusive evidence for H1, then E favors H2 over H1.

In email correspondence with Fitelson I argued against (CE) by claiming that it causes probability and entailment to come apart. In this post I want to propose a different idea.

Carnap (1962) proposes two independent confirmation measures. The first is confirmation as firmness. Confirmation(f) tracks both truth and entailment. The second is confirmation as increase in firmness. Confirmation(i) tracks entailment in addition to an intuitive relevance. I think this suggests a Relevance (R) principle:

  • (R) If E constitutes conclusive evidence for H2, but E constitutes less than conclusive evidence for H1, then E is more relevant to H2 than H1.

If Fitelson’s use of ‘conclusive evidence’ is to work with regard to confirms(i) it should work with regard to relevance. In the card-drawing counterexample is the fact that the card is a spade (E) more relevant to the fact that the card is black (H2), or is (E) more relevant to the fact that the card is the ace of spades (H1)?

For one thing, relevance cannot be assimilated into entailment. This is because confirms(i) tracks entailment AS WELL AS relevance, according to Fitelson. If E is conclusive evidence for H, then Pr(H|E) = 1. In the counterexample, E is conclusive evidence for H2, as E establishes H2 with certainty, yet Pr(E|H2) = 1/2 and Pr(E|H1) = 1. This means H1 entails E, but H2 doesn’t. However, H2 seems more relevant to E than H1 even though H1 entails E. That the card is black is more relevant to the claim that the card is a spade than that the card is the ace of spades is relevant to the claim that the card is a spade. The latter claim is redundant whereas the former claim delimits, and adds to, the semantic content on the table. “The card is black, and by the way, the card is a spade not a club.” On this account, relevance provides non-redundant semantic information in helping the inquirer delimit the space of logical possibilities.

By contrast, it is true that if the card is a spade, then the card is black. But this is running things in the opposite direction from what I stated above. This makes it the case that the card is a spade (E) is more relevant to the claim that the card is the ace of spades (H1) than (E) is relevant to the claim that the card is black (H2). In this direction the color of the card in terms of meaning is already bound-up in the card being a spade. A spade in a standard deck is by definition a black card. My basic idea is that (intuitive) relevance is a measure of the degree to which something doesn’t participate in semantic redundancy.

In conclusion, I claimed (R) is false. Intuitive relevance, when it comes to conclusive evidence, works in the opposite direction of probabilistic strength and entailment. This means (CE), which implies (R) in order to qualify under a confirmation(i) measure, is also false. If confirmation(i) only tracks entailments, and relevance is dropped from the measure, then confirmation(i) can be assimilated into confirmation(f). This is because confirmation(i) involves necessary preservation of truth. If confirmation(i) isn’t different than confirmation(f) by virtue of tracking a type of relevance, then confirmation(i) is merely tracking truth-preservation. That is, confirmation(i) is merely tracking truth and, as such, is not independent from confirmation(f).

 
1 Comment

Posted by on February 9, 2011 in Bayesianism, Branden Fitelson, Confirmation Theory

 

Tags:

Post 4: The Factivity Problem Remains Dissolved

20 Jan

This post concludes coverage of the factivity debate. Brueckner and Buford (2010) fire the last shot. They claim that Baumann’s response, which I covered in post 3, only works because the time-indexing of knowledge-attribution sentences was dropped. Let me explain.

As you may recall, the debate is over premise (3) in the Factivity Problem. The question is: Can Mary know in her demanding context a sentence about Frank’s epistemic status (i.e., that he knows that Mary has hands) is true in Frank’s less demanding context? According to Brueckner and Buford, Mary would have to know that she has hands in order to know whether the sentence about Frank’s epistemic status is true. Given Mary’s evidence and her demanding context, it’s not possible for Mary to know that she has hands; thus, it’s not possible for Mary to know the sentence about Frank’s epistemic status in his context is true.

Baumann claims this response doesn’t work because it requires Mary to have prior knowledge that she has hands independent from, and prior to, her knowing that Frank knows that she has hands. This is false because testimonial knowledge can be given and attained, even within a demanding context. This knowledge allows, for instance, Wiles to know that Fermat’s Last Theorem is true by reading about his results in a newspaper. Wiles had no prior knowledge of the results of Fermat’s Theorem. By analogy, Mary can “read” knowledge of Frank’s epistemic status off Ann’s testimony. Ann has better evidence of the truth of the sentence about what Frank knows about Mary having hands.

Enter: Brueckner and Buford’s final response. The problem stems from a mischaracterization of the time sequence. At time (t) Mary’s evidence about her having hands doesn’t qualify to meet the demands put on knowledge in that context. Baumann loses the time-indexing and proposes another time (t’) at which Mary meets Ann. At (t’) Mary gains testimonial knowledge and comes to know that Ann’s utterance of the sentence ‘“Frank knows that Mary has hands” is true in O’ is true in D. This is because of (X):

  • (X) ‘Ann knows that Mary has hands’ is true in D.

However, contrary to Baumann’s insistence, the time shift from (t) to (t’) doesn’t hinge on a principle like (Prior). Mary doesn’t have prior, independent knowledge that the sentence is true. At (t) she has no such knowledge, but at (t’) she does have such knowledge. Brueckner and Buford are fine with this type of gain in testimonial knowledge. Their key point is that Mary’s epistemic state has altered from (t) to (t’). There’s nothing wrong with Mary failing to know that she has hands at (t) and Mary knowing that she has hands at (t’). Focusing on (t) Brueckner and Buford’s response to the Factivity Problem still stands.

Tracking this debate has been interesting. Regarding the debate itself, I would say Brueckner and Buford’s challenge is still on the table. It seems the Factivity Problem is not a problem for Contextualism. The burden of proof is on those who want to apply the problem to contextualism and then fend off the problem. Otherwise, the Factivity problem seems like a pseudo-problem. It dissolves upon closer inspection.

There might be different avenues for the contextualist to pursue in showing the problem is real and that the problem can be handled. The ‘B vs. B&B’ debate focuses around a single premise in the Factivity argument because it is viewed as, “By far the most promising way of attacking the view that there is such a factivity problem for contextualism” (Baumann 2010: 84). Yet, different avenues remain open to the contextualist. She could attack closure, factivity or disquotation principles. I wonder what the prospects are for focusing on the factivity claim in the argument ?  Listen to how strongly Baumann expresses the pull of the factivity claim:

It certainly seems weird if not crazy to deny the factivity of knowledge. Whatever knowledge is, it is factive. Nothing is a concept of knowledge in a broad sense if what it is a concept of isn’t factive. The solution to the factivity problem proposed here will therefore not deny factivity (Baumann 2008: 584).

The quote above strikes my ear as argument by brute force. Baumann doesn’t think denying the factivity claim is promising. He doesn’t argue against it because “obviously” knowledge is factive. But, couldn’t a contextualist deny factivity on fallibilist grounds (e.g., Stewart Cohen 1988)? Why does Baumann think attacking factivity is so “obviously” unfruitful? Why must it always be the case that if someone knows something then what they know is the case?

I close on a methodological note. It’s interesting to track a philosophical debate because there’s a narrowing effect that often occurs. For example, the Factivity Problem originally addressed contextualism and SSI. However, SSI dropped out of the picture never to return. I have found this true in my own work. I recently was involved in a debate that focused around a single issue. I kept trying to explain and argue a small point. After failing to make headway I conceded the point to my interlocutor. Ironically, conceding the smaller point allowed me to argue for a larger point more effectively. This is why it’s beneficial to frequently zoom out and locate the contested point in the bigger picture. For the Factivity Debate to continue the discussants need to let go of premise (3) and look at other lines of debate, including forgotten aspects of the bigger picture.

 
Leave a comment

Posted by on January 20, 2011 in Anthony Brueckner, Contextualism, Factivity, Methodology, Peter Baumann

 

Post 3: Baumann Argues the Factivity Problem is Real

07 Jan

This post continues following the dialectic in the factivity problem for contextualism. Now our attention turns to Baumann (2010) in the journal Analysis.

In post 2 we saw Brueckner and Buford (2009) dissolve the factivity problem. They did this by claiming step (3) in the factivity reductio is false. Step (3) is ‘Mary knows that “Frank knows that Mary has hands” is true in O’ is true in D. Brueckner and Buford claimed Mary can’t know that Frank knows that she has hands. The truth of (3) requires the truth of:

  • (6) ‘Mary know that she has hands’ is true in D.

On contextualism (6) is false, and because the truth of (6) requires the truth of (3) this makes (3) false as well. As such, for Brueckner and Buford, the factivity problem doesn’t apply to contextualism. Baumann (2010) tries to locate Brueckner and Buford’s argument for the requirement principle—that the truth of (3) requires the truth of (6). He is not able to find such an argument. Instead Baumann claims that Brueckner and Buford are relying on a stronger principle than the requirement principle. Baumann lists the Priority Principle (Prior) as follows:

  • (Prior) If B knows that A knows that p, then B has antecedent knowledge that p independently from and prior to the knowledge that A knows that p (p. 86).

After pinning (Prior) on Brueckner and Buford, Baumann raises a counterexample to (Prior). The counterexample involves Paul reading in a newspaper that Wiles proved that Fermat’s conjecture is true. From doing this Paul could have come to know that Fermat’s conjecture is true. Arriving at this knowledge didn’t require that Paul had prior knowledge that Fermat’s conjecture is true. If this were the case then only mathematicians familiar with the truth of the conjecture could have come to know that the conjecture was true through reading the report that Wiles proved Fermat’s conjecture. Similarly, testimonial knowledge (even in demanding context D) can be given such that Mary could learn from Ann that (1) [i.e., ‘Frank knows that Mary has hands’ is true in O].

Baumann’s counterexample works based on transmission of knowledge within a context that is demanding but not completely skeptical. Baumann takes this a step further and says that, even within context D, Ann might have better evidence than Mary. This evidence might make it true that ‘Ann knows that Mary has hands’ is true in D. As Baumann says, “Mary can thus gain testimonial knowledge about Frank’s epistemic state concerning the proposition that she, Mary, has hands” (p. 86). This would make (3) true [i.e., (3) ‘Mary knows that (1)’, is true in D].

While I’m sympathetic to Baumann’s line of reasoning about transmission of testimonial knowledge I’m not clear how Ann’s evidence escapes the (Prior) principle? I agree that Mary can learn about Frank’s epistemic status from Ann regarding the proposition that she has hands, but how does Ann learn about Frank’s epistemic status from within D without entering into an infinite regress of testimonial justification? At some point, someone must have known that p independently from and prior to the knowledge that S knows that p. But this requires that someone to know that p, which from within context D is false. This is precisely what makes the context demanding: Mary can’t know that she has hands; she can’t directly know that p from within D.

 
Leave a comment

Posted by on January 7, 2011 in Anthony Brueckner, Contextualism, Factivity

 

Post 2: Brueckner and Buford’s Solution to the Factivity Problem

14 Dec

I’m back from a brief hiatus from blogging. In this post I continue writing on the factivity problem for epistemic contextualism and subject sensitive invariantism (SSI). In the first post I outlined the factivity problem for contextualism as articulated by Peter Baumann (2008). I also discussed Baumann’s solution to the problem. Enter Anthony Brueckner and Christopher Buford (2009).

Brueckner and Buford propose an alternative solution to the factivity problem. They extend the factivity problem to SSI. As a result of this extension, their solution to the problem is uniform in that it applies to both contextualism and SSI. According to Brueckner and Buford the factivity problem is only an apparent problem, not a genuine one; whereas, for Baumann the factivity problem is a genuine problem.

Recall that the factivity problem results from combining a factivity claim with a standard closure principle. The factivity claim (F) is:

  • (F) ‘S knows that p‘ (as uttered in a context) is true → p

The closure principle (Clos) is:

  • (Clos) ['S knows that p' is true in context C and 'S knows that p → q is true in C] → ‘A knows that q‘ is true in C.

The reductio involves Mary and Frank. I will not rehearse the entire reductio here (see the first post for schematic details). However, I will explain the last part of the reductio because it involves the claim around which the controversy centers (i.e., claim 3).

Mary is in a demanding context (D) while Frank is in an ordinary context (O). A contextualist will want truth-value of utterances (sentences) to vary with the context of utterance. If Mary utters ‘Mary knows that she has hands’ in context D this sentence is not true in D. If Frank utters the same sentence in less-demanding context O, then ‘Frank knows that Mary has hands’ is true in O. Mary in context D can learn from Frank and come to utter claim 3:

  • (3) ‘Mary knows that “Frank knows that Mary has hands” is true in O’ is true in D.

Using the factivity of knowledge claim (F) we can claim that:

  • (4) ‘Frank knows that Mary has hands’ is true in O → Mary has hands.

Because Mary understands the factivity claim this leads to (5):

  • (5) ‘Mary knows that (4)’ is true in D.

Combining (Clos), (3) and (5) results in:

  • (6) ‘Mary knows that she has hands’ is true in D.

This is a problem for contextualism because it contradicts contextualism’s commitment that truth-value varies with context (i.e., that ‘Mary knows that she has hands’ is not true in D). Brueckner and Buford generate a parallel factivity problem for SSI. According to SSI the truth-value of an utterance in a context varies with the subject’s interests and stakes. For one subject it matters a lot whether φ is the case while for another subject not much depends on whether φ is the case. The SSI version of the reductio contains something like controversial claim (3). Both versions of the factivity problem share this feature: a subject who does not have knowledge that φ can correctly attribute that knowledge to another another subject. Brueckner and Buford dissolve the factivity problem for contextualism and SSI by denying this is the case. As they claim, “the theories are not committed to the possibility of such asymmetrical knowledge attribution” (2009: 434).

Brueckner and Buford hold that (3) is not true. That is, contextualists are not committed to claiming that Mary can know that ‘Frank knows that Mary has hands’. Recall that to know whether ‘Frank knows that Mary has hands’ is true in O Mary must know whether Mary has hands (i.e., must know whether p is the case). However, anyone in D, including Mary, cannot know that they have hands. So, a condition for the truth of ‘Frank knows that Mary has hands’ is true in O fails to be satisfied from within D. The contextualist can, however, claim that the truth of ‘Frank knows that Mary has hands’ is satisfied up to the truth condition. But, in order for ‘Mary knows that “Frank knows that Mary has hands” is true in O’ is true in D to be the case it requires the truth of ‘Mary knows that she has hands’ is true in D. Because the last statement is false for the contextualist, (3) is false as well. So, the factivity problem does not go through. It’s not really a problem for the contextualist. A similar claim is made about the factivity problem for SSI.

 
Leave a comment

Posted by on December 14, 2010 in Anthony Brueckner, Contextualism, Factivity, Peter Baumann, Subject Sensitive Invariantism

 

Index of Posts on Williamson’s “Evidence” Chapter

19 Jul

For ease of reference, below is an index of my posts on Williamson’s “Evidence in Philosophy” chapter in The Philosophy of Philosophy. Williamson’s chapter 7 is broken down into sections. I commented and analyzed each section in the chapter.

 
Leave a comment

Posted by on July 19, 2010 in Epistemology, Evidence, Index of Posts, Methodology, Timothy Williamson

 

The Factivity and Contextualism Debate

09 Jul

It’s that time again. I’m going to provide explanation and analysis of a slice of philosophical literature. Previously I reviewed the “Evidence” chapter in Williamson’s The Philosophy of Philosophy, a series of posts I’ll index soon. This time I’m going to play analyst and referee in a debate pitting Peter Baumann vs. Anthony Brueckner and Christopher Buford. Hopefully my review of this literature will serve you by giving you a synopsis of the debate. In addition, this series of posts will help me better understand contexualism and subject-sensitive invariantism–two hot topics in epistemology. This debate might also link-up with my previous post on factivity in a surprising way. We’ll see.

While I cannot guarantee the rate at which I’ll post I can guarantee that I’ll go in order. I discovered the debate by looking at the Brueckner and Buford (2010) paper in the current issue of Analysis. This led me to trace the debate backwards, which started with Baumann’s PPR paper “Contextualism and the Factivity Problem.”

 
Leave a comment

Posted by on July 9, 2010 in Anthony Brueckner, Contextualism, Factivity, Peter Baumann

 

Links: Reasons Workshop and Most-cited Philosophers

14 May

There are a couple of pointers I’d like to provide. The first is a workshop at Leeds University on theoretical and practical reasons June 24-25. Click here for details. Second, Eric Schwitzgebel over at The Splintered Mind is doing some interesting work using the Stanford Encyclopedia of Philosophy (SEP). One interesting set of results is a list of the 200 most-cited authors in the (SEP). Click here to access that list.

 
Leave a comment

Posted by on May 14, 2010 in Conferences, Philosophers

 

Reasons and Evidence: The Provide/Consist Distinction

07 May

Recently, while looking in Knowledge and Its Limits, I came across an interesting distinction. The more I thought about the distinction it started to make sense of a topic I’m currently researching—the relationship between epistemic reasons and evidence.

The distinction is: provide vs. consist. Williamson uses this distinction to mention an objection to his view that all evidence is propositional. For Williamson even perceptual experiences, which are often regarded as non-propositional evidence, consist of propositions. An objector might claim: “Experiences provide evidence; they do not consist of propositions” (197, italics mine). However, only propositions we grasp can be used in confirmation, inference to the best explanation, and choice between rival hypotheses. Even though words fail to completely capture perceptual experience it does not mean evidence is non-propositional. Instead, experience makes propositions e1…en count as evidence for a hypothesis h. Having an experience bestows the status of evidence on propositions. As such, evidence is inextricably linked to (and mediated by) propositions. Thus, experience consists of propositions.

It is possible to ask the same thing of epistemic reasons and evidence: Are they inextricably linked? There are defenders of two theses concerning this question:

  • Inseparable:  Where you find one you find the other (i.e., reasons and evidence serve the same function, appear under the same analysis, or are constitutionally equivalent).
    • For every proposition p, if p is a reason R then p is evidence E.
  • Separable: Reasons and evidence come apart (i.e., in some scenarios you have reasons but no evidence, and vice versa).
    • There is some p such that p is an R but not an E.

Now I’ll relate this to the provide/consist distinction. One way of arguing for Inseparable is by claiming ‘having’ evidence for the truth of p ‘provides’ you with an epistemic reason for believing p. In response one might argue for Separable by showing evidence for p doesn’t always generate a reason to believe p. A strategy to counter this move is to claim the evidence for p is not really (good) evidence for p. What is taken as evidence for p doesn’t ‘consist’ of evidence; it doesn’t have the status of evidence because its status is undercut by other pieces of evidence.  These moves have the following assumptions:

  • Pro-Inseparable: If you have good evidence for p, then you have a good epistemic reason for believing p.
  • Pro-Separable: Rejects the assumption endorsed by Pro-Inseparable.

Pro-Inseparable claims ‘providing’ sanctions ‘believing’. ‘Believing’ connects to ‘consisting’ in that one is ‘believing’ appropriately if that believing is based on that which has the status (consists) of good evidence (reasons). By transitivity ‘providing’ sanctions ‘consisting’.  A way to argue for Pro-Separable is to show you can have good evidence for p without that evidence grounding a good reason for believing p because the reason is not based on the evidence. As a result, it’s not the case where you have good evidence you always have a good epistemic reason. The epistemic reason needs to be appropriately linked to the evidence to result in ‘believing’ in a way that’s sufficient for the belief to be justified. Simply claiming the two entities are inextricably linked (i.e., where you find one you find the other) doesn’t secure this connection. There’s another way of putting this point.

‘Providing’ focuses on the function of evidence or how it’s used in an argument. Williamson takes this line by arguing evidence ‘is’ (consists) only in so far as it ‘functions’ (provides). For evidence (experience) to play its evidentiary role within an argument it must be propositional. Because experience functionally provides evidence for hypotheses, and evidence must be propositionally grasped in order to be used, experience consists of propositions. Williamson argues for ‘consists’ by way of ‘provides’. The problem with this is that something may ‘consist’ without ‘providing’. I can be justified in believing that p even if no agent has engaged in the activity (function) of justification. This is because p’s status as evidence justifies believing in p in a way that doesn’t depend on anyone having used it in argumentation. It doesn’t require that the evidence is possessed, grasped, or used. That it can only ‘function’ a certain way if it ‘is’ a certain way simply shows that ‘function’ (provide) depends on ‘status’ (consist). It doesn’t show that ‘consist’ can be derived from ‘provide’ when it comes to evidence.

 
Leave a comment

Posted by on May 7, 2010 in Epistemic Reasons, Epistemology, Evidence, Timothy Williamson

 

Epistemological Dogmatism about Evidence

27 Jan

In “Defeating the Dogma of Defeasibility” Ram Neta argues against all positions defeasible. Neta’s paper, which is found in Williamson on Knowledge[1], is a comprehensive argument against the alleged “defeasibility” of knowledge. Neta endorses an often neglected stance on knowledge: that knowledge is not capable of being defeated by future evidence. While I cannot cover Neta’s entire argument in a blog post I will get his position on the table, explain Williamson’s response to one aspect of Neta’s thesis, explain how Neta, willingly, embraces epistemological dogmatism, and then raise a problem with Neta’s account. According to Neta knowledge is indefeasible justified true belief (IJTB):

  • S knows that p = S has a justified, true belief that p, and there is no true proposition e such that the conjunction of e and S’s actual evidence set E does not constitute a justification for S to believe that p (169).

Neta defends the IJTB account of knowledge by showing it can properly handle a battery of cases where non-knowledge gets counted as knowledge. I’ll leave those details to the reader and fast-forward to a part in Neta’s paper that Williamson takes issue with.

Are there counterexamples against IJTB? One counterexample is found in Williamson (2000: 219). This counterexample involves a person putting one black and one red ball into a bag (e), making 10K draws and getting a red ball every time (e’), resulting in rationally doubting whether a black ball was really put in the bag and not just a red ball made to look black by a trick of lighting. The initial knowledge is defeated by future evidence. Thus, knowledge seems defeasible.

But, not so fast says Neta. There are many ways the additional evidence (e’) can interact with the person’s beliefs: (i) the belief that p is true is lost, (ii) some of the initial evidence for p is lost, (iii) being able to reasonably form the belief that p on the basis of one’s evidence is lost. A person might lose knowledge that p in any of these ways. This is not a problem for IJTB because, “IJTB says nothing about what would happen to our epistemic subject if she were to gain an additional bit of evidence. It says that, if S knows that p, then, for any true proposition e’, the conjunction of our subject’s actual evidence set with e’ constitutes a justification for our subject to believe that p.” (171) For Neta, infallibility entails indefeasibility because, “Knowledge…is belief that is properly based on infallible evidence (indeed, on evidence that can be known—perhaps upon reflection alone—to be infallible)” (354). This means that a subject can know p only based on infallible evidence for p.

Williamson points out a fallacy in Neta’s argument for the idea that S knows that p only when it’s based on infallible evidence for p. It’s possible to deduce, according to Neta, from the assumption that S knows p on the basis of e that e is not misleading evidence for p. This makes a disjunction hold: either the subject can know the evidence is not misleading based on some further, independent, evidence e’ or e is infallible evidence for p. According to Neta the first disjunct leads to a regress because it’s possible to know that e & e’ are not misleading with regard to p, and this can be known based on further, independent, evidence e”, and so on. Williamson counters this assumption by saying it only shows that all cases of the first disjunct cannot be true, but it is possible that some cases are true. The result is that only some of the time the second disjunct is true, namely, in cases when the first disjunct is false. This does not show that the second disjunct is true all the time because, as Neta argues, the first disjunct is always false because it generates an infinite regress. According to Williamson the following is quite possible:

S knows that e is not misleading with respect to p on the basis of evidence e’ distinct from e; e is fallible evidence for p; S does not know that e & e’ is not misleading with respect to p on the basis of evidence e” distinct from e & e’; e & e’ is infallible evidence for p. (354).

So, it’s possible for e to be fallible evidence for p, yet for S to know p on the basis of e (i.e. when e is conjoined with e’). Williamson launches his second wave of attack against the assumption of entailment between infallibility and indefeasibility. If the evidence is infallible, then it’s the case that if S knows p on the basis of e, then p is true; possessing the evidence guarantees the truth of the belief. However, according to both Neta and Williamson’s views, the subject may not (and need not) be in a position to access or reason to the truth of the belief. So, even knowing that e is infallible evidence for p, it’s not clear that this entails indefeasible justification for p. Simply knowing that e is infallible does not entail that, “the conjunction of e with anything should constitute a justification for S to believe p” (355). It seems Williamson missed Neta’s disclaimer about justification, as Neta confides:

[T]here may be examples in which S is justified in believing that p on the basis of evidence that can be expanded into something that is not a justification for S to believe that p—justification itself may be defeasible. But knowledge is not defeasible, according to the IJTB theory (180).

Neta leaves justification as an outstanding, perhaps defeasible, position to develop. Williamson’s comments do bring out a worry with Neta’s account. Neta is committed to epistemic dogmatism. In fact, he embraces dogmatism. As long as an agent maintains her current evidence for her knowledge that p, future evidence will never justify disbelief in p. Being dogmatic about belief that p is OK because there is no epistemic cost, reasons Neta. Holding p as a settled belief and continuing to believe p in the face of new evidence, as long as one does not lose one’s current evidence, allows one to rationally continue knowing that p. What does it mean to lose evidence? If a true proposition in one’s evidence set becomes false it falls out of the evidence set. Beliefs once true can become false in light of new evidence (i.e. true evidence may falsify other true evidence once added to the evidence set). This means one must account for future evidence because it can cause one to lose one’s current evidence for the belief, and may, as a result, justify disbelief in p. Unless Neta embraces a view that factivity is absolute (i.e. once true, always true), it seems a cost of Neta’s view is that it sanctions a conflict: retaining evidence allows one to keep belief irrespective of future evidence, yet one can lose one’s evidence in the face of future evidence and so future evidence  can significantly impact one’s current evidence. It seems: future evidence is no big deal with respect to one’s current evidence and yet a big deal with respect to one’s current evidence. Which is it?

[1] All page references and quotes are from Williamson on Knowledge.

 
Leave a comment

Posted by on January 27, 2010 in Dogmatism, Epistemology, Knowledge, Ram Neta, Timothy Williamson

 
← Older posts
 
Follow

Get every new post delivered to your Inbox.

Join 409 other followers