Nov 19: Coherentism and the great debate
|
|
|
To do for next class - 1) Read chp. 6, pp. 125-140 and chp. 7; 2) Try question 5 of the exercises (State and defend your position). |
To do for next class - 1) Read chps. 8 & 9; 2) Give specific examples to explain the generality problem; 3) Precisely what is it that an epistemic nonnaturalist asserts?. |
Lehrer's coherentism continued
Do you recall from last day the two main ingredients of Lehrer's theory? (Coherence through competition; level ascent to determine trustworthiness of perceptual experience). What basic problem does Steup point out about level ascent? He notes that, similar to modern foundationalism, people just don't perform level ascent (as they do not form beliefs about appearances). Thus, most beliefs become unjustified and we fall into skepticism. How can Lehrer avoid this problem? He can claim that we need only have acceptances about level ascent, not beliefs. Don't forget, an acceptance is simply the set of all true propositions of the form "the agent accepts the belief at some time". (It seems to me the distinction between standing and occurent beliefs might solve the problem for foundationalists as well). So, we accept the plethora of things we do not reflect on. This is important also for being able to beat out competitors (the first condition of Lehrer's theory). There are probably innumerable competitors most of which are not considered. A person need only act as if they were false and need not prove it. Steup claims that there is in trouble with this move because he has asserted that acceptances need not be beliefs, which he denied earlier. However, given the occurent/standing distinction for beliefs I think Lehrer is safe.
Steup continues to show that foundationalism and Lehrer's theory are not inconsistent. However, this demonstration again relies on the distinction between acceptances and beliefs (in particular, the claim that acceptances are not a subclass of beliefs). Since some acceptances are not beliefs and foundationalism requires for a basic belief that it justified by something other than a belief, the theories are consistent. If, however, acceptances are a subclass of beliefs then there is no such thing as a belief which is not justified by something of its subclass so the theories are inconsistent as Lehrer suspects.
BonJour's coherentist theory of justification
BonJour holds a somewhat stronger version of a level ascent type of coherence. In essence, he thinks in order for someone to be justified in believing something they must belief the reasons for thinking that thing is justified. In other words, they must have meta-beliefs, or beliefs about beliefs. So, meta- justification is for him justification at the level of meta- beliefs. Can you think have an example of meta- justification? (The chair is blue, is a perceptual belief. Perceptual beliefs are likely to be true. Therefore the chair is blue. Or the more detailed example on pp. 130-131) In what way are we performing level ascent? What is it that this theory is interested in explaining? (Something like why basic beliefs are justified, which is why they cohere with background beliefs).
The reason BonJour holds this level ascent view has to do with epistemic responsibility. Why? (I can only be epistemically responsible if I "know why" I believe what I believe, this is like the voluntariness principle perhaps). The particular sense in which we "know why" for BonJour is to have "beliefs about why". His view is decidedly internalists, as "cognitive accessibility" is the main ingredients. However, internalists like Steup don't think that the only "cognitively accessible" things are beliefs. In particular, perception and memory will be things we do not have beliefs about but which support justification.
What are some criticisms of his view? (Most obvious, that it is psychologically unrealistic. First, people don't seem to form these beliefs. Second, people must belief something to cohere with other beliefs but, Steup argues, most mortals don't have a concept of coherence. Now, Steup rightly notes that this is an empirical question but continues to claim that it is "don't let ordinary people would bother considering such matters" which seems to be equally empirical. Examples from talk out loud paradigm of complicated thought processes.) BonJour avoid some criticisms, he notes, by distinguishing approximate an ideal justification claiming we only achieve approximate justification. For this reason, Steup argues BonJour has turned into the skeptics he is trying to argue against.
The isolation objection
What are the two versions of the isolation objection? (1. Detachment from reality: a belief system remains constant and equally justified although the world in which is situated varies dramatically. 2. Alternative systems: even though the world is the same, two different systems can be equally justified). Steup illustrates this with the aardvarks in the enclosure. How? What is the coherentist response? (That in changing world's, your perceptions will into juicy to believe things that our incoherent with mistaken belief). What is the critics response? (This is not enough. Though it is unlikely, it is possible that the systems coherence could remain preserved and absurdly imply the mistaken belief is justified). What of the mad scientist example and Steup's conclusions from it? We could, of course, deny something other than coherentism in the example. A parallel example for the alternative systems case can be constructed. What do both of these cases rely on? The deep intuition that we can have perceptions without generating beliefs. Isn't this just the intuition which divides foundationalism from coherentism?
So, can we have perceptions without perceptual beliefs? How can we decide on this question? Is there an experiment we might do? If not, how can we decide? Steup rightly notes that this example brings into focus what is at the heart the foundationalism coherentism controversy: namely, the legitimacy of non-doxastic justification.
The great debate
The foundationalist coherentist debate is over precisely this question. What is Davidson's arguments in favor of coherentism? What is the one thing we can view when evaluating ourselves? ("Get outside our skins"). In other words, we can't become an observer of a belief/experience conflict since we are the subject of both. However, the foundationalist argue that this is applicable only to the active justified are beliefs, not to determining whether a belief is justified. So, when we provide reasons for something, obviously we must appeal to other things we believe. However this is not all there is to justification. In fact, we have characterized the main role of epistemologists in analyzing justification as determining what nonevaluative features give a belief the evaluative property of justification. That is, we are looking for what it is about a belief that gives it the property of being justified. This is distinct from arguing for the justification of a particular belief. Just because we must appeal to beliefs to argue for the justification of a perceptual experience doesn't mean that those further beliefs are the reason for its being justified, or so the foundationalist would argue.
Steup outlines a parallel situation with consequentialists. Do you recall what it was? (Consequentialism: an act's moral status is determined by its consequences.) What is in that both of these examples have in common? It is a distinction between what consequences there really were, and those we know. What is strange, however, is that the once we know seem to be the internalist ones. Whether ones there really are seen metaphysical, in some sense externalist. Though this parallel is a weak one. Why?
Another argument against foundationalism
BonJour argues directly against non-doxastic justification. To do so, he poses the following dilemma: non-doxastic experiences are states of awareness or not. If they are they can justified only if justified. If not the can't be justified and can't justify. The foundationalist's have to choose either of these horns? Which one is more likely? If so, how will they argue against the dilemma? The first horn is deemed unacceptable by Steup, but it is not clear to me why. In any case, he casts the foundationalist tasks as one of justifying the second horn. In other words, they must show how a non-belief which does not have justification can give justification.
What is the foundationalist's reply? The reply is simply that coherentism it is in as much trouble here as foundationalism is. In other words, the fundamental assertion of coherentism is that the justification of a system depends on its coherence. How can we justified that? It seems that we can't. We must take it as a foundation for our theory of belief systems. Or similarly, coherence is a non-doxastic source of justification for coherentists. What can coherentist do about this? They can mistaken of as BonJour does and make the "doxastic move". That is, claim that the justification is not by virtue of coherence but rather by virtue of the subjects belief about the coherence of the system.
What are two problems with this response? (1. It seems to generate an infinite regress. 2. It is at odds with supervenience.) can you think of any other moves a coherentist might make? Where does the infinite regress come from? How is it different from the other infinite regress we considered in coherentism? (This is a regress of levels, that was a regress in levels). Can we solve this regress the way we solved that one? Why/why not? How can we use the infinite hierarchy of beliefs to construct an argument for skepticism? (Detailed examination of the infinite level regress which is a necessary consequence of BonJour's theory). What if we say justification through beliefs about coherence is not necessary? (That opens the door for non-doxastic justification again).
What you think of Steup's criticism of BonJour is theory. Recall first that BonJour tries to argue that only the kinds of beliefs which foundationalist take to the basic are subject to his schema. That is only such beliefs have meta-beliefs formed about them. Steup calls these non-inferential beliefs. Inferential beliefs, on the other hand, don't have a regress problem since they can be justified by other beliefs which are not "the level up". Steup's criticism, then, is that BonJour is theory makes non-inferential beliefs unjustified since using the schema results in regress, whereas inferential beliefs are more likely to be justified. Is there any problem with Steup's critique? (If Steup really accepts BonJour his claim that non-basic beliefs can be justified by other beliefs of the same level he has solved his regress problem in both cases so Steup's argument against BonJour fails).
Supervenience revisited
The recall the distinction between weak strong supervenience? Weak supervenience is the assertion that any two beliefs which share non-normative properties share an epistemic status. Strong supervenience is the claim that any beliefs which have non-normative properties that make them justified in one possible worlds makes them justified in all possible worlds. With this distinction mind, Steup argues that it is mysterious how coherentism can be reconciled with supervenience. Do you recall the problem? Suppose you are a coherentist who endorses the doxastic move. How will you answer the following questions: why is my belief that there is a blue chair before me justified? (Because it coheres with your other beliefs, in particular, my beliefs that it is a perceptual experience). Why does this not satisfy supervenience? (Because it explains belief justification in terms of belief justification both of which are evaluative). Well then, says the foundationalist, is epistemic justification non-normative or not? There is a dilemma here. If the coherentist answers yes then the admits coherence as non-doxastically justified. If the coherentist answers no then they are denying supervenience... which is bad, why? (Because they cannot offer explanations, see page 34).
To try to clarify the discussion, Steup outlines to principles about justification on p. 151. The first is about the justification of any two beliefs (a belief which justifies another must be justified) and second is about the justification any two things (similar but more general than above). Steup notes that we can reject the second while still holding the first of these principles. And insinuates, then, that BonJour another coherentist's have confused these improperly. Steup continues to claim that if coherentist except non-doxastic justification they have no grounds for claiming that coherence is the only element of a justificatory theory which have that status. Why not, he asks, admit introspection memory and perception? However, I would like to emphasize that admitting non-doxastic justification does not force us to be foundationalists.
The recall how status analysis of Lehrer's theory progresses? Recall the two conditions of Lehrer's theory: one all competitors and except inserting; two, the subject accepts that their trustworthy source of information under the circumstances. How does Lehrer avoid the infinite regress problem that would arise from two? (Global trustworthiness). What is the problem with a solution? (We have a separate theory for better beliefs, or acceptances, and first order beliefs.) Why can't the trustworthiness principles day today? (Because it only seems to apply to half of our beliefs (meta-beliefs), for no good reason).
Rather than rehash Steup's final verdict, what is yours? Which is more reasonable, coherentism or foundationalism? And why.
To do for next class
1) Read chps. 8 & 9; 2) Give specific examples to explain the generality problem; 3) Precisely what is it that an epistemic nonnaturalist asserts?.