Disclaimer: These are notes which I took for myself to help me study for an exam. They were prepared over the course of about 48 hours and are known to contain some errors. Use at your own risk. 1. Words are human products, and consequently, they are sometimes not made as well as they should be. Consider the word, 'know'; perhaps all that the skeptical scenarios and other philosophical conundrums that have swirled around the topic of knowledge show is that the word 'know' has been designed badly, and should be replaced with another one. Discuss. Is this a question about the way we ordinarily use 'know' being different than how philosophers use the word? If so, then I agree that the term 'know' is misleading. Philosophers are taking a word that already has a meaning and overloading it with a special epistemological meaning. Biperspetivalism. And by using the same word there is the possibility of confusion because different kinds of knowledge may be incorrectly used interchangeably. (Austin accuses Ayer of similar things.) But what of substance would changing the word really show? What we want it to make a distinction, not a language change; it's just that language is how we present ideas. But changing know with one other word? What would that change? A rose by any other name... http://www.princeton.edu/~jimpryor/courses/epist/notes/twosenses.html ---- 2. Some might say, "There is admittedly some difference between 'the first person perspective' and 'the third person perspective'. But the difference is without consequence for epistemology. There is nothing we can know from the first person perspective which is unknowable from the third. There is no significant immunity from error which the first person perspective enjoys over the third. It is true that it is easier to gain warrant for certain claims (like the claim that one is in pain) from the first person perspective. But it is likewise easier to gain warrant for certain other claims (like the claim that the Antarctic is icy) from the 'South Pole perspective'. No one thinks that this makes a pilgrimage to Antarctica mandatory for budding G. E. Moores." Of what significance is the first person perspective to the theory of knowledge? The difference between first and third person does seem to be of consequence. We might be able to know everything third person (through sensory perception, testimony, etc) that we can know first person (either by showing that both are extremely reliable sources of knowledge or that neither are), but the method by which we acquire the knowledge is different. It's also not clear to me that we can know as much through the third person as through the first person perspective. We may not be immune from error in either case, but first person "knowledge" seems a more likely candidate for knowledge, and the question about knowing seems to me to be rich enough to support a spectrum of kinds of knowledge rather than just two alternatives (either you know it or you don't). Drawing on the South Pole example, wouldn't you be more likely to believe there is ice if you saw it than if someone told you (people often lie/mislead/are uninformed)? There are degrees of reliability in testimony. Wouldn't we believe the testimony of someone who had been to the South Pole more than the testimony of someone who hadn't? Testimony seems, in general, less certain to be reliable than the first person perspective. http://www.philosophypages.com/ph/moor.htm Ayer Phenomenalism, 1946-47 ---- 3. Some philosophers think that 'know' is a word which shifts in its standards according to the context in which it is used. Discuss this claim, and the possibility of using it as tool for undermining skeptical positions about knowledge. Is this a question about biperspectivalism or contextualism? I definitely think the word 'know' shifts. At the very least, there are two senses of the word know (biperspetivalism). This is discussed by Ayer in Foundations and by Austin (discussing Ayer) in Sense and Sensibilia (chapter 2). But there are likely far more meanings of know depending on context (contextualism). Locke concedes to the skeptic but says nobody is so mad as to doubt that fire is hot, etc. (See question 1.) There are probably more than just two standards for 'know'. I think 'know' does act as a kind of indexical, at least if we are trying to describe how we use the word 'know'. I think our standards for knowledge really do shift, and it's not just that we are willing to use the word incorrectly. The standards of knowledge that are useful (and/or worth wanting) is context-dependent rather than some abstract philosophical notion. It also might be that we use relevant alternatives (which are context-sensitive) as knowledge criteria. The multiple meanings of 'know' can be used to undermine the skeptic because one can concede that the skeptic is right (in his context), while holding that skepticism is no real threat because the skeptic's standards of knowledge are too high and/or are not relevant in ordinary contexts. There are different kinds of relevant alternative theories: Contextualism is the theory that the word "know" is context-sensitive. Much as we could truly say "The piglet is small" (when compared with full-grown pigs) and "The piglet is large" (when compared with mice), we can truly say of something in one context that we know it, but not in another context. (The meaning of "small"/"know" is relative to a standard.) Understood in this way, the question "How much evidence does it take for a true belief to count as knowledge?" has different answers depending on the context. In the skeptical context, the bar for knowledge is very high. But the contextualist can reply that even if there isn't knowledge in those contexts, there is knowledge in other important contexts. Dretske's relevant alternatives theory: You see a bird that has all the markings of a Gadwell duck (and really is one). Do you know it's a G duck? That depends on whether there are any relevant alternatives (such as that it might be a grebe). Similarly, when you see a zebra in a zoo. Is there a possibility that it's a painted mule? This seems to deny closure. You know the animal is a zebra. You know that the animal being a zebra is incompatible with it being a mule. You don't know the animal is not a mule. How is this possible? Because knowledge is not closed in the general case (it can be closed for relevant alternatives). And this can be seen as a response to skepticism because the skeptics argument relies on closure to demonstrate the inconsistency of three individually plausible claims: 1) If we know any proposition p, we know the skeptic's hypothesis (that we don't know anything) is false. 2) We don't know the skeptic's hypothesis is false. 3) We do know (the truth of) some propositions. So the skeptic says one of the claims has to go; in particular, 1) if we know any proposition p, we know the skeptic's hypothesis is false, and 2) we don't know that the skeptic's hypothesis is false leads to we don't know the truth of any propositions, which is the denial of 3. The contextualist has a response to this argument: Deny closure (Nozick, Dretske). Knowledge is relative (context-sensitive) so the inference to 3 is unjustified. (Alternatively, deny the justification of inference between statements in different contexts.) http://pantheon.yale.edu/~kd47/responding.htm http://www.princeton.edu/~jimpryor/courses/epist/notes/twosenses.html http://www.princeton.edu/~jimpryor/courses/epist/notes/stroud2.html http://www.princeton.edu/~jimpryor/courses/epist/notes/rel-alterns.html http://www.utm.edu/research/iep/s/skepcont.htm http://www.iep.utm.edu/c/contextu.htm DeRose in Alston, 109 Williams, p2+, 129+, etc ---- 4. It is sometimes held that skepticism is true (or at least irrefutable), that it involves the same kind of epistemic assessment that we carry out in our everyday lives, and that it has no practical import. Is this position consistent, given that our ordinary epistemic assessments do normally have practical import? If not, which of the claims should be abandoned? I think skepticism doesn't involve the standards of knowledge that we use in our everyday lives; the standards for the knowledge we use successfully (with predictive reliability) are remarkably low. So skepticism has no (direct) practical import. Or skepticism may involve the same kind of epistemic assessment, but it doesn't have any practical import because we do believe we have knowledge (and even if we don't have knowledge, the important thing is that we believe we do and act accordingly). It may be that we have no choice but to believe, that we believe because our beliefs seem reliable, that we believe because it provides a better explanation than skepticism, etc. Blackwell, Skepticism Williams, ??? ---- 5. What are some of the main examples of Gettier problems? Is there a plausible way of handling them? Does it matter if there isn't? Prior to Gettier's 1963 paper, the common account of knowledge was that knowledge is justified true belief. (Subject S knows proposition P if and only if S believes P is true, P is true, and S is justified in believing that P is true.) Gettier problems are counterexamples to the JTB account of knowledge. They are cases where there's belief, truth, and justification, but where we don't want to say there was knowledge. 1) Smith and Jones apply for a job. Smith has a justified belief that Jones will get the job. Smith also has a justified belief that Jones has 10 coins in his pocket. So Smith can can justifiable conclude (by transitivity of identity) that the man who gets the job has 10 coins in his pocket. It turns out that Smith gets the job and that Smith also has 10 coins in his pocket. So Smith's belief was true, but not knowledge. 2) Smith has a justified belief that Jones owns a Ford. Smith concludes (by disjunction) that Jones owns a Ford or Brown is in Barcelona (without having any knowledge about the location of Brown). Well, Jones doesn't own a Ford, but Brown really is in Barcelona. So the belief is true, but not justified. These seem to be cases where having a JTB is accidentally, and this doesn't capture the idea that knowledge is more than epistemic luck. (Knowing something isn't have a lucky guess; it's having certainty.) More counterexamples: Williams' candle: You conclude there is a candle in front of you based on your belief that you see a candle in front of you. There is a candle in front of you, but it's hidden from view by a mirror so you're seeing a candle behind you. Goldman's barn: You see what looks like a barn and so you belive you are justified in believing you see a barn. But you find out you are in fake barn country (with barn look-alikes) and call your claim to knowledge into question. It turns out that the barn you were looking at was real, but in the context you don't seem to know it. Lehrer's theft: You are justified in believing a boy stole a book because you saw him. When you tell his mother she tells you the boy was out of town and it was his twin. But she's lying and the boy did steal the book. Some say there's a flaw in these examples: some say these are examples of knowledge (of some very fallible sort); others say that the examples are not in fact JTB. I think the latter claim has some appeal; in particular, I'm not sure if the beliefs in the examples are justified (Meyers, Stern, Armstrong). It's important to realize though that Gettier's use of justified is not like Descartes' demand for certainty; it's like Chisholm's adequate evidence and Ayer's right to be sure; according to Gettier it is possible for a person to be justified in believing something which is false. Gettier is addressing not whether one's beliefs are (objectively) justified, but whether one is (subjectively) justified in believing. It is important to realize that the Gettier counterexamples rely on this kind of subjective not-certain justification. They answer not "what is justified belief?" but "what are we justified in belief?" (or "what should we believe")? Most believe the Gettier counterexamples are correct in their aim (they demonstrate instances that JTB, of the sort we use in ordinary life, is not sufficient for knowledge), and most believe that it is important to have an analysis of knowledge in terms of sufficient conditions. Responses are roughly divided into two kinds: those which add another requirement to JTB (JTB++) and those which (more radically) abandon JTB entirely in favor of something else. Proposing JTB++ solutions: The first two counterexamples relied on false premises. So maybe for knowledge there must be JTB and all of the premises must be true. (Something like the argument must be sound: the inferences must be valid and the premises must be valid.) But this seems bizarre because it implies we could know something without knowing that we know it (because we didn't know the premises were true). Also, it turns out that this is not strong enough to explain the barn example (where what we are justified in believing varies from context to context). Nozick's subjunctive account: knowledge is TB where the subject could not have had the belief were it to have been false. Goldman causal theory of justification (also process reliabilism): for a subject's belief to be justified that belief must be causally connected (indirectly caused) by the truth of the belief. (Additionally the subject must be able to demonstrate the causal chain.) (Say more about Goldman!) Lehrer's defeasibility condition: knowledge is JTB and undefeated. There are no (ultimate) defeaters which if known would defeat the subject's justification. This seems to me a slightly strange response, but it shares some features with what I will offer as one of my own suggestions. Certainty: knowledge is certain JTB. This seems to prevent us from counterexamples. But I suspect it also discredits many things we'd like to call knowledge (lapses into skepticism). Similarly with the proposal that to know P is to have a JTB that P and to have a JTB that the belief in P is a JTB. My first suggestion: JTB and all of the premises must themselves be JTBs. This is a lot like JTB and premises must be T. Does it help any (with the barn example)? Not really if I mean the premises must be subjectively JTB. The problem seems to be that I can be subjectively J in something for which I am not objectively J. My second suggestion (similar to Pollock): both subjectively JTB and objectively JTB. My intution is that the pre-Gettier JTB notion fails because it allows one to be justified in holding a false belief. That's subjectively justified. It seems to me that one cannot be objectively justified in holding a false belief. (If one could know all the facts, one would no longer believe it; that is, objectively justified is what it would be for an omniscient being to be subjectively justified.) The concern with this is that it open to the question of how we (being non-omniscient) could know if something were objectively justified. And this is the worry with externalism. I think there are plausible ways of handling the Gettier problem and these rely on objective justification. http://en2.wikipedia.org/wiki/Gettier_problem http://www.princeton.edu/~jimpryor/courses/epist/notes/gettier.html file:///Users/sara/Desktop/%20epistemology/Sheffield/04%20The%20Gettier%20Problem http://www.simons-rock.edu/~sara/index.php?topic=philosophy/gettier Gettier in Moser, p237-268 Blackwell, Gettier Williams, ??? ---- 6. What are the doctrines of foundationalism and coherentism in epistemology? What are the best arguments for and against each one? Which has the stronger support? Is there any third alternative? Foundationalism and coherentism are both positions about the structure of knowledge. Both attempt to explain how the claim "Every justified statement is justified by a statement" can be true without turning into an infinite regress. (Note: Neither are positions which make the claim that there are any justified statements; they just say, if there are any, they look like this ...) The regress problem and possible solutions. Foundationalism answers that there are some basic foundational self-justifying beliefs. A is justified by B is justified by C ... by Z which is justified by Z itself. Coherentism says that no basic beliefs are needed; beliefs are justified by cohering with a system of beliefs. Sosa's metaphor is that of a pyramid (supporting structure) and a Neurath's raft (which can be rebuilt at sea). (Other representations of coherentism might include an arch or a fully connected graph or a crossworld puzzle.) Foundationalism is supposed to be the more intutive approach (how deduction works), but I think coherentism does a better job of sounding like how we come to formulate and accept scientific theories. One problem for foundationalism (BonJour) is having to explain how self-justification counts as justification and how something can be immediately justified. It seems impossible to explain how there can be a justification for something basic without impugning its status as basic. Another problem is identifying what the basic beliefs are. What are candidates for basic (self-justifying) beliefs? Chisholm says statements about one's own beliefs and psychological attitudes (I believe that X, I think that X, It appears to me that X). Also a priori truths like (I think) logical inference (that A and A=>B implies B). Another problem for foundationalism (an important one for Alston) is explaining how mediately justified beliefs are justified by the immediately justified ones. Another problem for foundationalism is the doxastic assent argument (which I don't really understand, see coherentism in Sheffield). And yet another problem has to do with its reliance on mental states which don't give us direct contact with reality (see Sosa in Alcoff, p 190+). Lots of problems with coherentism (mostly from BonJour): One problem is having to explain how this holistic approach is free from a vicious threat of circularity (A is based on B is based on ... Z is based on A). Chisholm, BonJour, and others quickly reject this. Inferential justification (for the coherentist) is not essentially linear in character; this objection misunderstands justification; beliefs are justified by being inferentially related to other beliefs in the overall context of a coherent system; justification is systematic and holistic. It's not that one belief depends on another belief, but that it coherences with a system of beliefs. Another problem for coherentism is that it seems there could be several possible systems of coherent beliefs with no system privileged over the others. This might be possible, but it's not as likely as it first seems. Foundationalists offering this argument against coherentism might be thinking that a coherent system of beliefs is just one which is logically consistent. But it's more than that; it has to do with hanging together. More formally, it looks like coherence requires: logical consistency, probabilistic consistency, large number of inferential connections between components of beliefs, lack of unconnected subsystems of beliefs, explanatory relations between members, lack of anomalies. Another problem for coherentism is having to explain how an internal subjective notion of justification correspond to the external objective reality/truth. If coherence is the sole basis for empirical justification, it seems that a system of beliefs might be justified (might be empirical knowledge) without having anything to do with the world it purports to describe. Epistemic justification seems to require input from the world outside the system of beliefs, but coherentism says that a belief is justified solely on its coherence with other beliefs in the system. I think (but I'm B.S.ing here) the answer to this is perhaps that it is coherence with a system, but not a system made up of just beliefs (the outside world is part of the system). Another problem is that our theory should have a connection between justification and the goal of truth, justification must be truth-conducive), but it's not clear how coherentism provides this. (I'd say more about these two problems, but I don't really understand them. See BonJour in Greco p 127 and in Alcoff p 226.) Could foundationalism and coherentism be combined? I think so. How about a foundational account of justification for rational beliefs and a coherentist account for empirical ones? Or perhaps coherentism plays a role (either negative or positive) in foundationalism. There's a theory (associated with Haack) called foundherentism. Foundationalism and coherentism actually have much in common. Both are looking for a simple theory that explains how epistemic justification supervenes on the nonepistemic (Sosa in Alcoff, p 200). Both are (I think) considered to be forms of justification internalism: justification depends on grounds that are internal to the knowing subject. But there's another approach -- externalism. Many philosophers seem to be scared by externalism, but I'm not quite sure what about it is scary. I think what's supposed to be scary about is that basic beliefs may be justified even when the person for whom they are basic is not in cognitive possession of the appropriate justifying argument. But I think this seems right. Armstrong: what makes a true non-inferential belief a case of knowledge is some natural relation which holds between the belief-state and the situation which makes the belief true. BonJour: it seems as though many empirical beliefs are justified not by other beliefs but by sensory and introspective experience. There are many externalist approaches, one of which BonJour calls foundationalism (although not the same kind of foundationalism I've discussed so far). According to this foundationalism, not all beliefs are justified on the basis of other beliefs; some are justified on the basis of states of affairs which are immediately apprehended (or perhaps experience, sensations, etc). One form of externalism (associated with Goldman) is a process reliabilism according to which beliefs are justified not on the basis of other beliefs but by being the products of reliable processes. Another alternative to foundationalism and coherentism is contextualism (a sort of relativism), according to which there is no single structure for justification; what counts as justification varies from context to context. http://www.aare.edu.au/97pap/parks060.htm http://www.simons-rock.edu/~sara/index.php?topic=philosophy/epistemologyexam/structure http://www.princeton.edu/~jimpryor/courses/epist/notes/regress.html http://www.princeton.edu/~jimpryor/courses/epist/notes/bonjour1.html http://en.wikipedia.org/wiki/The_regress_argument_in_epistemology http://en.wikipedia.org/wiki/Foundationalism http://en2.wikipedia.org/wiki/Coherentism file:///Users/sara/Desktop/%20epistemology/Sheffield/09Foundationalism file:///Users/sara/Desktop/%20epistemology/Sheffield/10Coherentism Blackwell, coherentism, foundationalism, externalism BonJour in Greco, p 117+ Audi in Foundationalism-Coherentism, p 143+ Part 3 in Alcoff, p 165-248 ---- 7. What could count as evidence for the existence of innate knowledge? What is this question asking? Is it asking what could count as knowledge for the existence/possibility of any innate knowledge? Or what could count as evidence for a particular innate belief being knowledge? Or what could count as evidence for a particular piece of knowledge being innate (or being justified innately)? Is the question asking for an example of innate knowledge? Or a reason why the existence of innate knowledge is necessary? I guess it's worth figuring out what innate knowledge is first. Innate ideas are defined as either 1) ideas consciously present to the mind prior to sense experience or 2) ideas which we have an innate disposition to form. Are there any such things? Locke and many others reject the existence of innate ideas. But many philosphers (including Descartes) embrace them. Why do philosophers think there are such things? Well, there are some things we think we know that we seem to know immediately, intuitively, directly through introspection and/or prior to experience. Innate ideas can be invoked to account for our recogition of certain truths without recourse to experiential verification. So what are some candidate examples? Some claim that many religious and moral claims are innate (e.g., God exists, murder is wrong). These examples may trigger some warning alarms. Why? Because if someone tells me "murder is wrong" I want to ask how they know that murder is wrong. And when the reply is "I know murder is wrong innately", I won't feel enlightened. More generally, it seems like the answer to what would count as evidence for a particular belief being innate knowledge? Nothing, besides itself. Innate knowledge is self-evident. We know it because we know it. Innate knowledge are precisely those things which we just know, no further explanation or evidence is needed or even possible. They are truths that are necessary. I've just suggested something fairly important. I've made the claim that for any piece of innate knowledge, we have no evidence that it is knowledge. Is this really true? It seems this follows directly from an externalist view of justification, but I'm not sure about an internalist view of justification. Suppose for now that for any piece of innate knowledge, we have no evidence. How then could we know that any innate knowledge exists? Well, it's not always necessary to demonstrate an example of something to know that it exists. (I know there's a number which is the largest prime number less than 2^1234567; I have no idea what it is, but I know that some such number exists.) So even if we can't produce evidence for the existence of any particular piece of innate knowledge, that doesn't mean that there is no innate knowledge. How then could we go about going finding evidence for innate knowledge? If we could demonstrate that the existence of knowledge is necessary for (or necessarily follows from) something which we know, then that would count as evidence for the existence of innate knowledge. Why might we need innate knowledge and what might some examples be? Mathematical truths perhaps. The rules of logical inference (deductive proof). http://faculty.washington.edu/himma/phil3604/trans002.htm http://www.lclark.edu/~rebeccac/lockei.html Blackwell, innate ideas ---- 8. What, if any, are the differences in goals, methods, and background assumptions between philosophical theory of knowledge and empirical psychology of cognitive processes? What, if anything, can either enterprise contribute to the other? ---- 9. Does science show that we never directly perceive physical objects? What does "directly perceive" mean? Well, what does indirectly perceive mean? From Austin: In some cases it's obvious we aren't immediately perceiving physical objects: When you see the president on TV you aren't seeing the president immediately/directly; you are seeing an image which represents the president. When you see something in a mirror or through a telescope you aren't seeing it directly. When you see blips on a radar screen representing enemy ships, you aren't seeing the ships directly. These are just some of the many different ways that objects can be indirectly perceived. So what does directly perceive mean? It means that the immediate objects of our perception are the physical objects themselves. So a related question is "What are the immediate objects of our perception?" The answer to which, according to representative realists (such as idealists), is sense-data or some similar mental thing. This seems initially plausible. Just as we indirectly see things on TV, we might really be indirectly perceiving all external objects. When we say "I see the chair" it might just be shorthand for "I see .. mental image .. shape .. color .. which represents a chair." Quite a few philosophers find this objectionable because it posits strange, private representational entities. Reid, for instance, argued that if representationalism is correct, we must accept either skepticism or phenomenalism, both of which are absurd; he concluded that representationalism is wrong and instead that there immediate/direct objects of perception are external objects. Those who oppose representational realists are direct realists, and they believe that the immediate objects of perception are not something mental. They may be the physical objects themselves (as Reid suggests). But they need not be: according to an Aristotelian view, they may be the non-mental perceptual properties of the physical objects. What's the difference between the representationalist view and Aristotle's view? Whether the immediate objects of perception are mental or non-mental. If the immediate objects of perception are mental, we have to explain what they are, how they are caused, how we know that we can reliably infer (properties of) physical objects from them, etc. In the Aristotelian case (e.g., Armstrong's view) things are simpler. Physical objects are perceived by (and only by) direct perception of certain properties belonging to the objects. We perceive that something is red by perceiving that it refects or emits light of a particular wave length. This is not some private mental entity, but an objective property of the physical object itself. There might be a worry here that we never perceive objects directly. We don't perceive colors; we perceive frequencies. Similarly, we perceive sound waves (rather than sounds) and particular shaped molecules rather than smells or tastes. And this is taken by some to mean we aren't directly perceiving the objects. But I think we are. I think that directly perceiving an object supervenes on directly perceiving its properties. I think this is exactly what we mean by (want from) direct perception. Surely we don't mean that to directly perceive an object the object itself we must be in direct mental/cognitive contact with the object (ouch! the book hit my brain). And so science, rather than telling us that we perceive properties rather than objects, can tell us how it is that perceiving properties is perceiving objects. So I don't think we should worry about science showing us that we don't directly perceive. Instead, I think we should worry about theories that answer the question "How do we perceive external objects?" with the simple answer "Directly." Because this sounds like a useless answer to me. It doesn't tell us how we perceive. This is something science can tell us, both by explaining the objective properties of the external objects we perceive (to borrow Locke's terms, by reducing secondary qualities to primary qualities) and by explaining the mechanisms (sensory, cognitive) by which we perceive. Science can (or at least likely can) explain the process of perception. I guess this might be seen as a threat to direct perception because we aren't perceiving directly, but rather through a process. But I think having a perceptual process is consistent with direct perception. The fact that we perceive through a complex neurophysical process does not mean perception is indirect; it merely establishes the method by which direct awareness is secured. (Much as a direct airplane flight doesn't mean one immediately and inexplicably arrives at the destination.) There's another way in which science may seem to show that we do not directly perceive objects. Assuming that "we" means the mind/brain, it is clear that we do not directly perceive physical objects. When we have a visual perception experience, brain does not do the seeing; our eyes see something and translate that and send that sensory data along to our brains. In this way, perception involves a link (via sensory organs) between the mind/brain and that which is external. But I don't think this is any cause for worrying about indirect perception. Perception is a process, and there is no particular point (eyes, brain) at which it occurs. http://www.iep.utm.edu/p/perception.htm http://en.wikipedia.org/wiki/Philosophy_of_perception http://www.simons-rock.edu/~sara/philosophy/epistemologyexam/austin.txt journal articles Blackwell, direct realism ---- 10. Can a case be made out that Descartes was a "naturalized epistemologist"? Why or why not? The Meditations are pretty rational (as opposed to empirical). He doesn't seem to be looking to science. But Descartes wrote a lot of other things and did a lot of science, and science was part of philosophy. So I think he isn't a naturalized epistemologist by today's standards, but perhaps by the analogous ones of his time. But no, I don't think Descartes was a naturalized epistemologist. He was doing science as part of philosophy rather than philosophy as part of science. ---- 11. "Part of one's epistemic duty is to reflect critically upon one's beliefs, and such critical reflection precludes believing things to which one has, to one's knowledge, no reliable means of epistemic access." What is at stake in the view expressed here? Is the view correct? How would we explain our knowledge of mathematical objects on such a view? ---- 12. C. D. Broad called the unsatisfactory state of Hume's problem of induction "the scandal of western philosophy". What is Hume's problem? Why is (or isn't) its state unsatisfactory? Inductive reasoning is not deductive reasoning. Inductive reasoning involves concluding a generalization from a number of instances (ampliative). Here are some examples of inductive reasoning: I've observed thousands of tomatoes and every single one of them had seeds; therefore, I conclude by inductive reasoning, that all tomatoes have seeds. Every morning the sun rises (okay, it appears to rise); the sun will rise tomorrow. These sounds like ordinary (non-philosophical) things to say. I do expect that the sun will rise tomorrow, and the reason I believe it will do so is because it has reliably done so in the past. But is this belief justified? Isn't it possible that the sun won't rise tomorrow? It does seem possible (though very improbable). (Russell gives an example of a turkey which observes that it is fed at the same time every day so it expects to be fed at that time every day. Then Christmas morning comes and the farmer kills it.) So can I really be said to know that the sun will rise tomorrow? No, because induction is fallible. But it sounds like I can make a similar, but weaker, claim: I know that the sun will very probably rise tomorrow. Is even this claim justified? More generally can I be considered rationally justified in holding as very probably beliefs which are the conclusions of inductive generalizations? From an ordinary standpoint, accepting inductive inferences as yielding probable beliefs seems rational; inductive inferences do give us the right results. But wait, isn't this itself an inductive generalization? Why do we think inductive inferences give us the right results? Because they have given us the right results in the observable cases. This is begging the question; that is, if we assume the principle of induction, the principle of induction is true. Hume considers this problem (or rather a subset of this problem at length). When all observed As are B, is the conclusion that all As are B rational? His answer is no; inductive inferences are the result of an a-rational process, a belief in constancy and habit. (That we observe frequent repitition is the cause of our believing uniformity, but it does not show anything about knowledge.) Hume challenges his dissenters to present a demonstration that the principle of induction is justified. Hume says that such a demonstration would have to either accept the principle of induction a priori or rely on either empirical reasoning (in which case it falls prey to circularity in assuming the principle). This seems to leave us in something of a bind. We do rely on inductive inferences, and if this is unjustified, it seems that we must be unjustified in holding a lot of our beliefs. In particular, Bacon claimed that science is based on induction. If so, this renders all our beliefs in science unjustified. (Popper provides one answer to Bacon; science relies on falsification, not induction.) Of course, induction might actually be defensible. Arguments by inductivists in defense of induction either try to show how induction can be inductively justified without the threat of circularity (Skyrms) or that a priori justification of induction is possible (Russell, BonJour). (This second approach makes some sense to me; it seems that we have knowledge of the inductive hypotheses prior to/independently of experience.) Russell's argument is something like this: We are justified in beliving the inductive hypothesis: that a sufficient number of cases of association will make the probablility of a fresh association very high. This principle is incapable of being proved by experience (without begging the question). And it's incapable of being proved by deduction (there's no deductive inference rule about generalizing from instances). Yet our conduct is unhesitatingly based on the assumption that the associations which have worked in the past will likely work in the future. Our belief in the inductive principle is firmly rooted in us. So why not take the inductive principle as a self-evident inference rule, just as the laws of identity, contradiction, and the excluded middle (and the deductive inference rules) are? We don't deduce that If "A" and "if A then B" then B; that's simply what deduction is. We know this rules independently of experience -- a priori. (Many reject this response because they find a priori justification problematic.) Additionally, some concede that Hume is right that inductive knowledge cannot be justified in the sense of showing that the conclusion of an inference is likely to be true, but that there is still some form of justification for our belief in induction. This justification may be in either the form of ordinary language (Strawson) or pragmatic reasons (Reichenbach). Strawson's argument goes something like this: The demand to produce deductive-like justification for the inductive inference is illegitimate because induction is fundamentally different from deduction and the standards for justifying induction are different than the standards for justifying deduction. (I think this amounts to saying that inductive claims can not be reduced to deductive claims; or, that not all knowledge must come from deductive arguments.) So the way to justify the inductive inference is by the stands for justifying induction and according to these the inductive inference is obviously justified. But Strawson doesn't fully address Hume's problem because he says that whether inductive conclusions will continue to be true is contingent and may turn out to be false. As for pragmatic reasons, induction isn't so much a form of inference as a wager. Reichenbach claims that if there are any truths to be known about generalizations/regularities, they will be known through induction, so we should accept the inductive hypothesis. The method of finding truths is as follows: we should accept inductive hypotheses as posits provisionally and make corrections as we get more information. (Popper says something similar.) I think induction, whether philosophers want to say it is justified or not, is something which can play an important role and which we do in fact rely on in ordinary life. (Not only do we rely on the belief that the future will be like the past, but this turns out to be a good survival skill for us most of the time.) So there's no real problem with induction which is relevant outside of philosophical circles. http://en.wikipedia.org/wiki/Problem_of_induction http://marr.bsee.swin.edu.au/~dtl/het704/lecture3/logfals/logfals.html http://www.simons-rock.edu/~sara/index.php?topic=philosophy/epistemologyexam/induction Blackwell, problem of induction, Hume Russell, The Problems of Philosophy, p60