Oct 29: Philosophical analysis and the a priori
|
|
To do for next class - 1) Read the remainder of chp. 2 and chp 3. 2) Try exercise 2. on page 67. |
To do for next class - 1) Read chp. 4; 2) Second essay due. |
Introduction
Do you recall from last day the difference between the criteriological analysis and a definitional one? What do they have in common? They are both forms of conceptual analyses of simple terms. However, a definitional analysis conveys the meaning of the concept by relating to other concepts. A criteriological analysis provides a list of criteria (in particular non-evaluative criteria for evaluative concepts) which, when met allow for correct application of the concept. Can you provide some examples of non-evaluative and evaluative concepts?
Why would we want to provide a list of criteria for the application of an evaluative concept in nonevaluative terms? There is extremely strong intuition here that the justification (or moral status of an action) cannot be a brute fact which is unexplainable. Rather, there must be certain characteristics which only justified beliefs share that make them justified. In other words, philosophers feel it is their job to determine what the characteristics of a given belief are that provide it a particular epistemic status.
One way of re-phrasing this intuition is to adopt the claim that evaluative properties supervene on non-evaluative properties. This means that normative properties are determined by, depend on, or are a consequence of non-normative properties. A good example of the supervenience relation is that between shape and physical composition. We would say here that shape supervenes on physical composition. Note that this is somewhat of a one way relationship. We can't change the shape of a thing without changing something about its physical composition (spatial relations) but we can change physical composition without changing the shape of a thing. The supervenience doctrine is also used in philosophy of mind can you guess how? (The mental supervenes on the physical).
In epistemology the doctrine becomes: a beliefs epistemic status supervenes on its non-normative properties. So if I believe there are three people in this room, what kinds of non-normative properties will determine the epistemic status of this belief? (Its being a perceptual belief, its being a belief formed in a well lit room, it being about a medium-sized object, it being at such and such a location at such and such a time).
Weak and strong supervenience
Does anyone recall the difference between strong and weak supervenience (notice that these are metaphysical doctrines... why)? It comes down to a question of entailment. On strong supervenience, the nonevaluative properties entail the evaluative properties. In the case of weak supervenience does he non-normative properties determined the normative ones but not to the same degree. That is, weak supervenience ensures a consistent mapping from non-normative to normative properties but does not ensure what that mapping is. Steup explains this difference by noting the introduction of a modal operator in strong supervenience (see p. 33). Under strong supervenience then the given non-epistemic properties determined that a belief is justified or unjustified, it does not simply make it justified if similar ones are justified.
What are the reasons that philosophers offer for accepting this intuition? Well, denying the weak supervenience thesis seems to make justification somewhat arbitrary. This is because nearly identical situations could have opposite belief generating results. However, we must be careful Powell on p. 34 Steup it is describing the similarity of the situations. If we think back to Taylor is examples of trying to explain a nonnecessary cause, we might see a pattern. In particular, we can ask the question "how do we determine what are relevant differences?" If the answer to our question simply states that "relevant" means relevant to believe formation Steup has simply begged the question. What do you think of Steup parrying of the second attack?
In the case of the strong supervenience thesis Steup presents three claims about possible worlds which must be viable if we deny it. These possible worlds seem very strange and show exactly how arbitrary the linkage between non-normative and normative properties would become on the weak supervenience thesis alone. Steup argues for the strong supervenience thesis from analogy to moral properties. However is not clear how absurd such results for possible worlds are considering that the possible worlds themselves are somewhat difficult to imagine. (I'm not sure why time can be admitted as a epistemically relevant property in weak supervenience but not in strong supervenience as Steup seems to suggest at the top of p. 36)
Beyond supervenience
Say that we agree that epistemic status supervenes on non-normative properties. We have yet to identify what these properties are. The two choices Steup considers by that there is one non-evaluative property which determines justification or that there are many such properties (monistic and pluralistic theories). Are you familiar with utilitarianism and prima facie duties? (Maximize nation of utility determines completely what is right; some set of duties been fulfilled determine rightness with a certain context sensitivity). More closely related to epistemology is the very unreliable listen. Like utilitarianism it claims that there is one property that determines the evaluative status of a claim. In this case it is the reliability of a cognitive process producing a belief. In contrast, Chisholm has produced a list of epistemic principles which must be satisfied. Something only counts as a principal if it is supposed to express entailment and does not contain any normative terms in its antecedent. These kinds of principles are difficult produce an even more difficult to agree upon. The most noticeable difference between these kinds of theories is that the pluralistic theory provides only a sufficient set of conditions for justification and does not address necessity, unlike the monistic theory. Of course, we could claim that our list was necessary though Chisholm does not. In pursuing either a pluralistic the monistic theory it seems to depend on how wedded we are to the importance of unified theories.
Chisholm's pluralistic theory
What is an important consequence out of supervenience claims being metaphysical? It means, of course, that it is possible that we can never know what the non-evaluative properties are on which the evaluative ones supervene.
What kind of change can we make to the claim: "if a belief is grounded in perception it is then justified" to make it at least plausible? What is the problem with a kind of change? (It is probably evaluative). How can we weaken the consequent of that claim to render it nonevaluative? We can introduce the notion of prima facie justification. This means that something is justified "at first glance" but might not be, depending on a particular context. The seems to be a very similar distinction to that mentioned last day between evidential and factual justification. And, it might be as difficult to evaluate. Can you give an example of prima facie justification? How does this new principal of prima facie justification compare to the principle which includes a defeater. They are logically equivalent. We are headed down the road towards epistemological pessimism.
What the such pessimism amount to? Essentially, it admits that the metaphysical problem is one which we cannot solve. That is, we are not in a position to determine those non-normative properties on which normative properties strongly supervene. All that we are then allowed to do is be an epistemic judge of sorts. In other words, we must judge each individual case of the basis of what we believe are relevant data available to a given individual. In the case of duties, when general principles about duties come into conflict as they inevitably will, we must judge them individually as there is no general principle to help us decide which will be right in which cases. In epistemology, the conflict comes from bodies of evidence rather than principles of duty. What do you think of epistemological pessimism? Is it plausible and it so why?
A priori knowledge
A priori knowledge is nothing new (pun intended). We encountered it in our definitions of rationalism and empiricism. The recall how it distinguished the two? We also encountered a priori and knowledge in our discussion of kinds of knowledge. The particular term "a priori" and its opposite "a posteriori" were introduced by Kant. Examples of both are easy to come by. Classify the following: "I am 26" "two plus 24 is 26" "I am either here or not here" " the air orbits the sun", etc. What are these kinds of knowledge prior or posterior to? (Experience). A priori knowledge then, is knowledge which is independent of experience. What must we explain about this definition? (independence and experience).
What we mean by experience? (Sense experience; i.e. stimulation through one of the five senses). However, this might not be experience enough. Why not? (Memory and introspection). What were your intellectual feelings in considering the two sentences Steup presents on p. 48? Perhaps he felt compelled. In this experience is enough for some to consider any proposition which elicits it a priori. If that is true then we can't include these kinds experiences in our definition of experience. So by stipulation Steup excludes such experiences while including the other seven. What were they again?
What we mean by independence? In the obvious sense Steup notes all concepts needed to understand language are dependent on experience. However, by independence we mean that they need not be used to provide justification to a particular proposition. So it is independent of experience with respect to the property of being justified. To make this clearer we can, after Steup, identify two different kinds of dependence. One is causal dependence (if x had not happen y would not have happened). The other is a semantic dependence (x makes y mean what it does).
Necessity again
In trying to understand what apriority is, many appeal to the notion of necessity (logical necessity). If a proposition is necessarily true then simply thinking about it (i.e. having no other experiences) is enough to recognize its truth and to justify it. Any proposition like those on p. 50, which has this property will be a priori... or so would seem. But as Steup points out belief in a necessary proposition is not sufficient for a priori justification. Why? For two reasons: a. Because my belief could be a lucky guess and b. Because believing a necessary truth might be justified a posteriori. An example of the second case is my believing in the church Turing thesis. I've never seen a proof that many people are trust of told me it is true. What is an objection to this kind of claim for justification? It is that we are only justified in believing that the church Turing thesis history not believing the contents of the church Turing thesis. What the thing as Steup's argument for the Goldbach conjecture? I'm not to sure. How does d4 p. 50 to avoid the problems of d3? (And it stipulates a manner in which something is believed and it rules out a posteriori justification).
Notice how important the word necessary is in this context. We must distinguish between "is true" and "is necessarily true". The second of these is a modal claim. However, when considering claims like "whatever is read is colored" it seems we are a priori justified in believing his proposition whether or not we consider its necessity. In this case, something like d4 is too narrow. This definition also asserts that a priori justification is not fallible. What does that mean? It means that the by definition of the word "grasp", for one to truly grasp something you cannot be mistaken in your belief that that something is necessarily true. In other words, grasping and mistakenly believing something are inconsistent. That see then if we can construct an example which shows the fallibility of a priori justification. We can recall here the Sorites paradoxes that we examined in the metaphysics section. In with a what was our standard example? (A bald man) vagueness and polarization seem to be playing important role in the arguments Steup makes for the fallibility of a priori justification. How can resort this out?
So far Steup has identified for criteria which an adequate
definition of a priori justification must satisfy. What are they?
1. Constraints must be placed on the manner in which p
is believed
2. A posteriori justification for necessary truths must be possible
3. We must allow for the possibility of being justified a priori
in believing a proposition to be true without believing it to
be necessarily true
4. A priori justification must be fallible
How do these criteria lineup of the different kinds of a priori knowledge we have seen?
A few more ways of defining a priori
Consider this: we must only understand an a priori proposition in order for us to realize that it is true. This is clearly a very narrow definition of a priori. However, if we have a model of knowledge which parallels that of logic, we could consider such a definition to accord us our axioms. Note that a standard logical system consists of axioms for which there is no proof and theorems for which there are proofs that can be derived from the axioms. Appling this to a priori knowledge, we would have a priori knowledge for which there was no proof (axioms) and a priori knowledge which was derived directly from those axioms which did not have proof. This is essentially the picture Steup paints on p. 55 and 56. How does this kind of picture avoid the problems of the previous to definitions of a priori? (First, what were the problems? The fallibility, necessity, a posteriori justification, the manner of believing) What new problems could arise for this definition? In short, it is possible for people to disagree on axioms. Nevertheless, the those axioms which some accept and others reject must be knowable a priori if they are to function as axioms.
Another way of defining a priori returns us to Kant's original definitions. However, those original definitions only told us what the a priori was not (the seven forms of experience) not what it is. So, an attempt a tell us what it is has been formulated by the likes of Alvin Plantinga. In particular, he identifies intellectual attraction as being the hallmark of a priori propositions. Steup uses this to identify purely intellectual experiences which can be either mediate or immediate. Of course such experiences could be defeated by a posteriori experiences as Steup notes. So wrapping these things together he presents us with the following definition:
S is justified a priori in believing that p if and only if S has an undefeated purely intellectual experience that p is true
Other than the objection Steup discusses can you think of others? (What was the one he discusses?) In what sense is this a definition of a priori if it depends on a posteriori considerations?
The analytic/synthetic distinction
The analytic synthetic distinction is one which was first introduced by Kant. There since been seriously challenged by Quine in his paper "Two dogmas of empiricism", a very important paper which Steup for some reason ignores. In any case, the distinction has a long history which parallels that of empiricism and rationalism. Simply, empiricists claim that all a priori propositions are "simply" propositions about the truths of language and logic, i.e. they are all tautological. In contrast, rationalists insist that a priori proposition state facts about the world. In other words, there things we can learn about the universe without using any of our (7) senses. The real argument comes down to whether or not all a priori propositions are analytic. Analytic propositions are those which state truths of language and logic i.e. truths of definition. The synthetic propositions are those which are constructed out of experience and perhaps logic or language as well i.e. they are constructive.
Kant himself defined analyticity as: a proposition is analytic if and only if it is a proposition whose predicate is conceptually contained in its subject. So, "all bachelors are unmarried" would be paradigmatically analytic. Whereas, "all bachelors are lonely" would not be even if true. What is an obvious problem with this definition? (What of logical truths like the law excluded middle?). Either any other problems? (What does conceptual containment mean? What about the sentence "whatever is read is colored"?)
The great logician Frege has another definition of analyticity which is more general than Kant's. He claims that a proposition is analytic if and only if it is a truth of logic or reducible to a truth of logic by substitution. Logical truths are valid schema in their simplest form (note on p. 61 the examples Steup provides are redundant). A schema is simply an expression with place holders for propositions. Valid schema are true for any propositions put in those place holders. Steup presents objections to this definition by examining the sentence "whatever is red is colored". He notes that you cannot perform the appropriate substitutions on the sentence to get in the form of the logical truths. Is this true? Can we really not define red? Might there be other ways than what Steup attempts? The second more interesting objection by rationalists is that of realizing that the notion of a priori itself has not been adequately handled. In other words, this definition does not explain the status of the logical truths themselves.
Let us considered a third way of defining analyticity. In this case we want to define analyticity in terms of sentences (not propositions). So consider the definition that a proposition is analytic if and only if it is true solely in virtue of its meaning. Note that probably most sentences are true in virtue of two properties: meaning and fact. So the sentence "in there are three people in this room" is true because of what the words mean i.e. what they refer to an and also because it agrees with the fact that there are three people in this room. So analytic sentences then, should be true regardless of fact. The recall the rationalist reply to this definition? Steup claims that there are in fact to things which make analytic sentences and like "whatever is red is colored" true. What are those two things? What is it that doesn't seem well-defined enough to be able to arbitrate this objection? (meaning.) what is the discussion so far tell us about Steup's position? Kant was the first to identify synthetic a priori knowledge.
Skepticism about the a priori
Of course, we have so far been assuming that there is such a thing is a priori knowledge. Do we have good reasons for thinking so? Against the a priori skeptic Steup gives us an interesting argument. Were the fair any premises which you thought he did not defend adequately? So, suppose that you accept his argument what is it that is devastating about it for a skeptic? The problem is this: the skeptic must have given argument whose conclusion follows from its premises or not. If not you can ignore his argument. If so, he is assuming you can recognize its validity and thus assuming a priori knowledge. So you can ignore it or it is self-defeating. So the a priorists if given the premise that grasping a belief is a priori win the day... not surprisingly. Obviously there is more to fall back on the Steup gives to the empiricists. He has built validity and a priority together, which isn't obviously the only way to do things. But, as he suggests empiricists to have a lot of work to do though some believe they are succeeding.
To do for next class
1) Read chp. 4; 2) Second essay due.