Oct 22: Knowledge and justification
|
|
|
To do for next class - 1) Read pp. 21-30 2) Exercise 3 on p.19. |
To do for next class - 1) Read the remainder of chp. 2 and chp 3. 2) Try exercise 2. on page 67. |
Introduction
As we discussed last week in epistemology is concerned with a number of different kinds of questions, including: 1. What is knowledge? 2. What do we know? 3. How can we know anything (skepticism)? Etc. these questions are all the theoretical questions and. That is, they are interested in conceptual analysis, identification of criteria for counting as knowledge, and theorizing about justification. However, epistemology is also concerned with practical questions. Which, in general, means they are interested in constructing or identifying methodologies which will lead us to acquire knowledge and to avoid error. We will focus on the more theoretical aspects of epistemology. However, it is often obvious from the theory what kinds of practical implications it will have (e.g. coherentism). The traditional account of knowledge is that it is justified true belief. Clearly, and each of the concepts of justification, truth, and belief will play an important role in our understanding of what knowledge is given that we accept this definition.
Note: Keep in mind that as we discussed knowledge we are referring only to propositional knowledge. As discussed last day this is different from non-propositional knowledge (pictures, sounds, kinesthetic sensation), procedural knowledge (knowing how to ride a bicycle), and knowledge by acquaintance (in French connaitre)
The JTB account
Steup presents us with the standard definition of knowledge as:
S knows that p if and only if (i) p is true; (ii) S believes that p; and (iii) S is justified in believing that p.
Each of these three conditions is the central to knowledge under the traditional account. The first condition, the truth condition, means that a proposition can only be knowledge if it is not false. Therefore, my belief that the Pillsbury dough boy is in this room is not knowledge since it is false. What does this tell us about some scientific theories, perhaps the once we hold now? (Might we have a polarization problem here?)
The second condition, the belief condition, simply says that I must believe something in order to have knowledge of it. In other words, if I explicitly disbelieve, or have no opinion about the proposition "the Pillsbury dough boy is in this room" I cannot know that proposition; that is, I should not form the proposition "I know that the Pillsbury dough boy is in this room." The know I'm a form the proposition "I know that the Pillsbury dough boy is not in this room."
The last condition is, not surprisingly, referred to as the justification condition and was first suggested by Plato in the Theatetus. He noted a distinction between opinions with reasons and opinions without reasons. Knowledge, on this definition, must be opinions with reasons. Why would this condition be suggested? Consider it wasn't. If it wasn't, any randomly chosen proposition about the future that happen to turn out to be true would count as knowledge now (this is the case whether or not a proposition is about the future). And, admitting random guesses as knowledge would destroy the utility of the term.
Truth
Do you recall the three theories of truth mentioned by Steup?
1. The correspondence theory: the belief that p is true if and only if it corresponds with the fact that p. For example, "snow is white" is true because of external facts in particular that snow is white. What two central concepts are left undefined by this definition? (Correspondence and facts)
2. Verificationism: the belief that p is true if and only if it is an instance of (idealized) rational acceptability. For example "snow is white" is true if I can show you that snow is white by following a set of agreed-upon rules for determining the truth of statements: e.g. deduce it. This theory disallows the truth of statements which we cannot establish as true. It also dismisses the truth of logically possible propositions.
3. Pragmatism: the belief that p is true if it only if it is useful. That is true beliefs must be a good basis for action. This theory disallows the possibility that actions based on true belief can lead to disaster while actions based on false assumptions may, but pure chance, produce the most useful of results.
The defeat of verification and pragmatism lead Steup to suggest
following two criteria (though he denies this) for the relation
between truth and justification:
1. It is possible that p is true although there is no one
who has a justified belief that p.
2. It is possible for us to be completely justified in believing
that p although p is false.
What do you think of these two criteria? Are they reasonable?
All of these theories essentially accept that "snow is white" is true if it only if snow is white. But, they also add the necessity of some other property (such as correspondence to reality, verifability, or pragmatics). The number of philosophers have rejected any need to identify a further property.
4. Deflationary theory: the truth predicate is not for describing propositions but for constructing a certain kind of generalization. E.g. Newton's theory is true at low speeds (expresses an infinite conjunction of his laws and all of their applications).
Belief
You may have heard of beliefs as being identified with propositional attitudes. This is because they are seen as attitudes (opinions, ideas, relations) towards propositions. In particular, the attitude that the proposition is false, true, or neither. Believing that p is false is the same as this believing p. note that the propositional attitude includes both in the attitude and a proposition. Steup notes that we cannot help to take one of these three positions. Could we run into a polarity problem here? However, he does go on to say that beliefs can be held with various degrees of intensity. Which seems incomplete contrast with the claim that they must be one of these three (we should only have a continuum from disbelief to belief).
The recall the difference between occurent and standing beliefs? Any given example of an occurrent belief that is not standing? (The belief that we are in this room). One that is standing? (The belief that the sun will rise tomorrow).
Justification
Only those beliefs which are justified are considered knowledge. What do we call an unjustified belief which happens to be true? A lucky guess. This is why many people don't believe in horroscopes despite the fact they may be occasionally right. Steup defines a lucky guess as being occasion on which the guesses true, we believe the guess, and we have no evidence for our beliefs. Note that this lucky guess is to be distinguished from a lucky truth. Can you remember the difference? A lucky truth is when the truth of proposition is not likely with respect to certain relevant facts. So, if I'm expecting a phone call from a friend of mine named John, and a friend of mine named John (a different one) decides to call, and I believe that the current ringing phone is caused by a friend of mine named John, then my belief a be true, my belief a be justified and it will be a belief. Nevertheless, it is a lucky truth because it was not likely that the friend of mine the would happen to be calling me would be John. Why are lucky truths important? (The Gettier problem which will discuss).
When discussing justification, as when discussing any process, who must be careful to distinguish both its product and its activity. The activity consists of supplying evidence, reasons, or explaining your belief to convince others. Whereas the product of such an activity is a belief with a property of being justified. Steup claims that a belief can have the property of being justified even if you have not engaged in the activity of justifying its. What you think of that? More worrying perhaps is his claim that it is possible to be unable to justify a belief though it is in fact justified. The problem with this is, of course, that it leaves open the ability of anyone to claim that they have a justified belief they just haven't figured out how to justify yet.
Evidence
Obviously evidence and justification are closely related. Can you recall any the relations? 1. We are justified in believing something if it only if it fits with our evidence. 2. Or more strictly we are justified in believing something if and only if we have adequate evidence for it and believe it because of that evidence.
Traditionally, philosophers have used for sources of evidence: perception, introspection, reason, and memory. Do all of these strike you as reasonable sources of evidence? Can you think of other sources of evidence? (Steup suggests a reliable authority though this is reducible to the other four sources). Note also that admitting reason as a source of evidence (and also of justify beliefs on p. 12) seems to move us onto the side of the rationalists. Note also that using reason as a to derive justified beliefs assumes a principal of transmission of certainty.
In around 1960, Quine introduce the idea of a web of belief. He suggested that the propositions at the edge of the web were more susceptible to revision than those in the center. It is not clear that they are more justified, as Steup suggests, but they are harder to revise and, given a coherence theory of justification, they would then seem more justified. What sorts of things do you think would be at the center? (Logical principles; a belief in God; a belief in material objects). Steup seems to suggest that more important than this location in a Web, is the conclusiveness of the evidence for the belief. Conclusive evidence seems to be only that evidence which provides us with complete certainty -- everything else is nonconclusive (not inconclusive, though I'm not so sure, I think we are supposed to simply reserve this word for usage in normal language).
Defeasibility
What is defeasibility? It is the ability of our justification for believing a proposition to be annulled. We have all experienced this. It is also called being wrong. Can you remember the sample from the book? (Blue paper with blue light) a similar situation occurs when leaves following on our tent causes us to think it is raining. Steup spells out evidence to defeat on p. 13. There are two important ways of satisfying this definition. The first is for us to acquire evidence that contradicts our current belief in the second is for us to acquire evidence which undermines, but does not contradict, our belief. Do you recall what the example of the blue paper is? (Undermined). What about the example of our tent? (Contradiction). What the think of the sorts of examples provided by Steup and the piece of paper? Perhaps they are too rigid. Much defeasibility seems to me, anyway, to be partial -- not complete defeat. People seldom seem to give up beliefs as easily as they get them...
The Gettier problem
This is a perfect example of how to make a career out of a 5 page paper. In 1963 Gettier published a short paper showing line the three conditions of the standard account are not sufficient. He did this by constructing counter examples. Did everyone understand the example in the textbook? Was there anything that struck you as odd? Notice was being relied on here: logical expansion (the properties of the logical or), and the principle of justification under known entailment. If, for some reason, we believe more strongly in our definition of knowledge and then either of these we could use that argument as a reductio ad absurdum.
Consider a different example: A teacher has two students, Mr. Nogot and Mr. Havit, in her class. Mr. Nogot seems to be the proud owner of a Ferrari (a rare and expensive car). He says he owns one, drives one around, and has papers which state that the car he drives is his - but he does not actually own a Ferrari. The teacher, on the basis of this evidence, concludes that someone in her class owns a Ferrari. This is true enough, but only because Mr. Havit, who shows no signs of Ferrari ownership, secretly owns one. So, it seems that the three conditions (truth, justification, and belief) of knowledge have been met, but that there is no knowledge. This example, at least, does not rely on the particular interpretation of the logical operator or, but it does still rely on the principle of justification under known entailment. There is, indeed, a cottage industry of constructing such examples. They have in common an attempt to satisfy the three conditions of knowledge while violating our intuitions about what should count as knowledge.
In order to alleviate the problems of the type Gettier originated we must either had a fourth condition on knowledge or introduce some other appropriate clause for eliminating such examples.
Now, we can introduce the notion of a second kind of defeat; that of factual defeat. After Steup we can call the previous kind of defeasibility evidential. In what sense you think these Gettier type problems suffer from factual defeat? In each of these cases, there is some piece of evidence which S does not have access to yet which would defeat a given belief if S did. In the case of the Ferrari, if the teacher had known that Mr. Nogot was lying about his car she would not have concluded that someone in her class own a car but that piece of evidence was not available to her. Note that this is different from having a piece of evidence and not bringing it to bear. For a belief to be knowledge, it must be neither it justificationally nor factually defeated. However, it can remain a justified beliefs without being knowledge in the case that it is actually defeated. Steup names the first of these a epistemizing justification and the second non-epistemizing justification. This lead Steup to call factual defeat a defeat of epistemizing potential. Let's apply these distinctions to the Ferrari case.
We seem well on our way to solving the Gettier problem. Let's consider a solution. In particular consider the claim:
S's justification for p does not depend on any falsehood.
This solves the original problem of the Ferrari, that is the teacher's dependence on a false belief that Mr. Nogot owns a Ferrari. So the teacher's justification would not go through if we add this new claim to our definition of knowledge. However, if we remove the deduction from the Gettier problem can simply rely on our senses as a source of evidence we can still construct Gettier like examples. Steup provides us with the hologram cat. We can construct a parallel example in our tent. Perhaps it is raining, but a large tree a sheltering our tent and dropping leaves on us (notice how this example seems little more like deduction the hologram cat, yet they are perfectly parallel. It seems we might have to have a particular kind of theory of perception in order to make such Gettier examples work. It also depends on how broad our notion of deduction is).
Consider a more likely solution to the Gettier problem based on defeasibility. I should highlight Steup's note here that there is nothing like a consensus on a solution to this problem. In any case, the defeasibility solution introduces the constraint that there must be no proposition that factually defeat's our evidence for believing our belief. Perhaps you can see right away why such a solution has not been immediately embraced (all the possible facts?). So, it is clear that this worked in the case of the Ferrari. Similarly it works in the case of the cat and tent. Clearly then, proponents of this view propose to add a fourth constraint to the JTB account of knowledge. However what happens when we include this constraint? (The truth condition is redundant). Why is this? It is because when p is false the proposition not be always factually defeats our evidence.
Take up question 3 on p. 19 of the exercises.
Conceptual analysis
It is anything but uncontroversial to consider concepts the same as properties. However, they often considered universals and should, indeed, be distinguished from words (though not always from ideas -- especially in psychology). As well, it is wise to distinguish propositions and sentences. Many philosophers have claimed that there are no such things as propositions yet they tend to be assumed by many others. Propositions are those things which can be either true or false, like sentences, but they are abstract, unlike sentences. Notably, the truth of a sentence is dependent and secondary to the truth of propositions. Do you recall any of Steup's arguments for the importance of propositions? (Type/token distinction gives back abstracts, sentences are bad truth bearers, they are accessible just not perceptible).
Steup also takes as through a quick tour of necessity and possibility. Do you recall his distinctions? (The only new one is that all propositions except necessarily false propositions are possibly true)
Steup then uses the concepts of necessity and possibility to define equivalence and entailment. What works are these concepts doing for him in his definitions? It is not clear that there doing any, and many philosophers, particularly those influenced by Quine, with much prefer to drop possibility and necessity from these definitions. Note that equivalence is the same as mutual entailment. We can look at the logical structure of these using the conditional and biconditional. The kind of equivalence Steup is talking about seems to be a special case of standard equivalence, in particular necessary equivalence. Note that any talk of possible worlds is a dead giveaway for someone embracing the notions of necessity and possibility. Now we can see more clearly what work these concepts are doing in the definitions... they're speaking to a kind of intuition that there is a difference between analysis and synthesis (which Quine has also questioned).
Now to conceptual analysis. Since the time of Locke, there has been a distinction between complex and simple ideas (which are often equated with concepts). Simple ones, like a red, cannot be broken into parts. Complex ones, like Santa Claus, can be. Only complex concepts can be analyzed that is broken into parts which have meaning which constitute the meaning of the complex concept. Note the interesting relation between conceptual analysis and causal analysis. In both cases the explanans (or analysans) specify individually necessary jointly sufficient conditions. Steup has provide us with example of analyzing the concept of being a mother. Steup notes that an analysis must be a special kind of biconditional in order for to be a good one. That is it must be noncircular. What the think of his example five on p. 28? How can we identify criteria for the concept of illuminating?
Steup introduces in other kind of analysis for normative concepts (that is, concepts which have an evaluative dimension, such as good, bad, or ugly). Can you think of some normative concepts we have already encountered? Criteriological analysis simply states the criteria needed to correctly apply a given normative concept. Of course, these will be the most controversial since they are essentially an attempt to identify how we should use a word. But notably this kind analysis translates evaluative terms into non-evaluative terms. This is in fact one kind of analysis which only philosophers have ever attempted, and still attempt. What is the importance of this kind of analysis to epistemology? (In particular justification).
To do for next class
1) Read the remainder of chp. 2 and chp 3. 2) Try exercise 2. on page 67.