Philosophy 5550

Seminar 1 - Skepticism - Part 1

Michael Huemer on Skepticism

Chapter II. The Lure of Radical Skepticism

    Mike Huemer defines radical skepticism as follows: “Philosophical skeptics hold that one cannot know anything about the external world” (SVP, p. 7).

But this needs to be interpreted as saying that, according to the radical skeptic, one cannot even know that there is an external world, since he later goes on to say, “. . . the skeptic maintains that we do not know there is an external world.” (SVP, p. 7)

Comment

There is a less radical form of skepticism, according to which, while one may be able to know that there is an objective world outside of one's mind, one cannot know what it is like, since one cannot know, for example, whether it consists of mind-independent objects, or, instead, of a Berkeleian God and other immaterial minds.

****************************************************************************

    Mike Huemer, in contrast to a number of epistemologists, takes skeptical arguments seriously in one respect, since he says that it is important to understand skeptical arguments, and to know precisely “where they go wrong.” (8)

    At the same time, however, he says, “the skeptic's position seems extravagant.” (8)

Comment

Would Mike Huemer also say that the more limited skepticism mentioned above also “seems extravagant"? In view of his later discussion, in chapter III, it seems clear that he would.

But suppose, for example, that one's parents had been followers of the philosopher Berkeley. Would one be inclined to think that the idea that one cannot know whether the external world consists of mind-independent objects, or, instead, of a Berkeleian God, seem extravagant to one?

****************************************************************************

    Mike Huemer now proceeds to set out four arguments for skepticism, the first of which is as follows:

1. The Infinite Regress Argument

Mike Huemer offers the following summary of this first argument:

"1. In order to know something, I must have a good reason for believing it.

2. Any chain of reasons must have one of the following structures: Either

    (a) it is an infinite series,

    (b) it is circular, or

    (c) it begins with a belief for which there are no further reasons. But,

3. I cannot have an infinitely long chain of reasons for any of my beliefs.

4. Circular reasoning cannot produce knowledge.

5. Nor can I gain knowledge by structure 2c, for

     (a) I would not know my starting beliefs to be true (from 1) and

(b) I cannot gain knowledge by deriving it from assumptions that I do not know to be true.

6. Therefore I cannot know anything."                    (SVP, p. 9-10)

**************

    Mike Huemer points out that most philosophers think that the error in this argument lies in the first premise, and they would say that, contrary to that premise, there are self-evident, or foundational propositions, where a proposition is foundational if one “can know it to be true without having a reason for it . . ..” (SVP, p. 10)

Comments

(1) Notice that Mike Huemer defines foundational propositions as ones that one can know without having any reason for thinking that the proposition is true.

(2) Since it seems to me that it is justified belief, and not knowledge, that is crucial, I believe that it is preferable to define foundational beliefs as ones that one can be justified in accepting in the absence of any supporting evidence for the beliefs in question.

(3) So one does not need to be a foundationalist in Mike's strong sense to answer the above argument.

(4) In addition, if one formulates things in terms of justified beliefs, one need not even claim that the foundational beliefs – understood as the starting points for justification – must be beliefs that are justified, since one can argue that justified beliefs can be justified on the basis of beliefs that are not justified – and, a fortiori, are not known.

(5) How can this be? Suppose that one is slightly inclined to believe that someone told one that p was the case. Given that one is only slightly inclined to believe it, the probability that it is the case may be very small, in which case one will certainly not know that someone did so testify. But suppose that one is also inclined to believe that someone else also told one that p was the case. Given a sufficiently large number of beliefs of this sort, it might very well be extremely likely that someone or other told one that p was the case, and this in turn, if the person is reliable, might make it likely that p was the case. So justified beliefs can rest upon unjustified beliefs!

(6) Given some accounts of knowledge – such as Michael Clark's, or Roderick Chisholm's, or my own view, which is intermediate between Clark's analysis and Chisholm's analysis – it will then also be possible to argue that knowledge can rest upon what, if one thinks in terms of the believing/not believing/suspending belief model of belief will be ‘unjustified’ beliefs.

(7) But things can be expressed instead, and much more happily, in terms of the degrees-of-assent model of belief, and one can then say that justified beliefs, and also knowledge, may rest upon beliefs whose epistemic probability is less than one half, and that are assented to at the appropriate, less-than-one-half level.

****************************************************************************

    Mike Huemer next points out that the skeptic has a response – namely, he or she can contend that one needs to be able to provide a way of distinguishing between propositions that are foundational – such as the proposition that 2 = 2, or the proposition that one is conscious – from propositions that are not – such as the proposition that there is a twelve-headed purple dragon on Venus.

    Mike Huemer next presents a BonJour-style argument, which can be summarized as follows:

(1) If proposition A is foundational, while proposition B is not, then it must be because there is some feature, F, that A has, but B lacks.

(2) If I am not aware of feature F, I cannot know (or be justified in believing) that proposition A is foundational.

(3) If I am aware that proposition A has feature F, then that fact – that proposition A has feature F – provides a reason for A, and so proposition A is not foundational.

(4) Hence it is logically impossible for there to be foundational propositions.

Comment

One is aware, first of all, of certain states of affairs, and, secondly, of the fact that those states of affairs make certain propositions true. But this is not to have evidence for the proposition in question, since there are no propositions that are distinct from the proposition in question that provide evidence for it.

****************************************************************************

    This brings us to the second of the skeptical arguments that Mike Huemer considers:

2. The Problem of the Criterion

Mike Huemer illustrates this second argument in terms of a community that uses the method of the “Magic Eight-Ball", and that, when this method is challenged, uses that method to determine whether the method itself is reliable.

Mike Huemer offers the following summary of this skeptical argument:

"1. All my beliefs are formed by some method.

2. I am justified in accepting a belief formed by method M only if I am first justified in believing method M is reliable.

3. I do not have an infinite series of belief-forming methods.

4. Thus, all of my beliefs must rest on beliefs formed by methods whose reliability has not first been established. (from 1 and 3)

5. Therefore none of my beliefs are justified. (from 2 and 4)"        (SVP, p. 13)

Comments

1. Because a belief may be either non-inferentially justified or inferentially justified, this argument poses two challenges:

(1) How does one justify any method that one uses in forming basic beliefs about contingent states of affairs?

(2) How does one justify any methods of reasoning that one uses – either deductive or inductive?

2. Suppose that one says that the belief that p is justified because one is directly aware of a state of affairs S and also directly aware that state of affairs S is a truthmaker for p. If the two claims to direct awareness are true, then it follows deductively that proposition p is true. Is the ‘Problem of the Criterion” argument challenging, then, the claim that one is directly aware of some states of affairs? Or is it asking for a justification of the analytically true claim that if the two claims to direct awareness are true, then it follows deductively that proposition p is true?

****************************************************************************

    The third skeptical argument that Mike Huemer considers is as follows:

3. How Can You Get outside Your Head?

This argument has the following overall structure:

(1) One cannot have any knowledge of a world external to one's mind unless either direct realism is true or indirect realism is true.

(2) Direct realism is false.

(3) Indirect realism is false.

(4) Therefore, one cannot have any knowledge of a world external to one's mind.

How are (2) and (3) to be established? In the case of (2), Mike Huemer mentions the sort of argument alluded to in the first chapter, which he summarizes as follows:

"1. As your focus shifts to the background, the fingerlike thing you are seeing splits in two.

2.   No physical object splits in two at this time.

3.   Therefore, the thing you are seeing is not a physical object."    (SVP, p. 14)

Comments

(1) In a footnote, Mike Huemer points out that terminology differs among indirect realists, and that some do not speak of seeing mental images.

(2) Mike Huemer introduces the technical expression “sense data” to refer to the “mental items that allegedly exist whenever we exercise any of the five senses.”

(3) The term “allegedly” suggests that he does not think there are any such items – as emerges later, in Chapter VII of his book.

****************************************************************************

The next question is how (3) is to be established. Here Mike Huemer sets out Hume's famous argument. This he summarizes as follows:

"1. In order to have knowledge of the physical world, we must be able to know that our sense data are caused by physical objects.

2. In order to know that A causes B, one must have experience of A and B.

3. We have no experience of physical objects.

4. Therefore, we do not know that physical objects cause our sense data. (from 2, 3)

5. Therefore, we have no knowledge of the physical world. (from 1, 4)"    (SVP, p. 16)

Comments

(1) If Hume's argument were sound, it would also show that one cannot have justified beliefs about very small things, such as protons, neutrons, electrons, quarks, and so on.

(2) Hume’s argument is open to an immediate objection, namely, that Hume is overlooking what is variously called ‘abduction', ‘the method of hypothesis', ‘hypothetico-deductive method', and ‘inference to the best explanation'.

(3) This objection shows that Hume's argument, as it stands, is unsuccessful, provided that one can then go on to argue that, contrary to what Bas van Fraassen has claimed, inference to the best explanation is a defensible principle of inductive inference. This, however, is not a trivial matter.

(4) Moreover, even if one succeeds in showing that inference to the best explanation is sound, that does not yet provide one with an answer to external, physical world skepticism, since there is certainly at least one alternative, skeptical explanation that must be ruled out – namely, Berkeley's hypothesis that the only things that the world contains are immaterial minds.

****************************************************************************

    The fourth skeptical argument that Mike Huemer considers is as follows:

4. The Brain in a Vat

Mike Huemer summarizes this argument as follows:

"1. Your sensory experiences are the only evidence you have for propositions about the external world.

2. The BIV scenario predicts that you would be having the same sort of sensory experiences as you are actually having.

3. Therefore, the sensory experiences that you are actually having are not evidence that the BIV scenario isn't true. (from 2)

4. Therefore, you have no evidence that the BIV scenario isn't true. (from 1, 3)

5. Therefore, you don't know that you are not a BIV. (from 4)

6. Therefore, you do not know anything about the external world. (from 5)"                                             (SVP, p. 17)

Comments

(1) The crucial assumption in this formulation of the brain-in-a-vat argument is that if you don't have evidence against a theory, you are not justified in rejecting it.

(2) But this assumption is clearly false, since the a priori probability of a theory may be very low, and so you may be justified in rejecting it even though you have no evidence against it.

(3) Here is a different way of formulating the brain in the vat argument, which avoids that false assumption:

1. Your sensory experiences are the only evidence you have for propositions about the external world.

2. The BIV scenario predicts that you would be having the same sort of sensory experiences as you are actually having.

3. When two theories predict precisely the same sensory experiences for you, you cannot be justified in assigning a higher probability to the one than to the other.

4. Therefore, the sensory experiences that you are actually having cannot make it the case that the thesis that you are perceiving an external world is more likely to be true than the thesis that you are a BIV. (From 3)

5. If p and q are two incompatible propositions, and p is no more likely than q, then you are not justified in believing p.

6. Therefore, you are not justified in believing that that you are not a BIV. (From 4, 5)

7. Therefore, you don't know that you are not a BIV. (From 6)

8. Therefore, you do not know anything about the external world. (From 7)

***********************************

(4) This alternative formulation involves a different erroneous premise, namely, the assumption that when two theories entail the occurrence of precisely the same sensory experiences for you, you cannot be justified in assigning a higher probability to one than to the other.

(5) The reason this assumption is false is that the two theories that generate precisely the same experiential predictions need not have the same a priori probability. (This can be shown by taking some proposition with a very low a priori probability that entails no experiential predictions – such as that there are exactly 999,999 causally inaccessible spatiotemporal worlds, and conjoining that proposition with any proposition that does have experiential predictions.)

****************************************************************************

This concludes what Mike Huemer regards as the four most important skeptical arguments. He now goes on to consider different forms of skepticism.

5. Skepticism and Common Sense

    First of all, Mike Huemer offers the following, more general definition of skepticism than that offered earlier in Chapter 1:

“I define skepticism as any philosophical theory that challenges a significant class of common sense beliefs.”                 (SVP, p. 18)

    Mike Huemer then defines “common sense beliefs” in terms of the following three characteristics:

"i.      They are accepted by almost everyone (except some philosophers and some madmen) regardless of what culture or time period one belongs to.

ii.     They tend to be taken for granted in ordinary life . . . .

iii.    If a person believes a contrary to one of these propositions, then it is a sign of insanity."                        (SVP, p. 18)

Comments

(1) Notice that given this definition of “common sense beliefs”, something cannot be a common sense belief unless it is accepted by almost everyone at every “time period”. That means that the fact that a belief has been accepted by everyone up until a certain time does not suffice to make it a common sense belief, and to know (or be justified in believing) that something is a common sense belief one must know (or be justified in believing) that it will be accepted by almost everyone at every time in the future. But how can one know, or be justified in believing, that this is the case?

(2) To believe something that is contrary to the ordinary belief that there is a spatiotemporal world containing physical objects is not treated as a sign of insanity when, as in the case of Berkeley, arguments are offered in support of the opposing view.

(3) It is important to note that acting in ways that would be rational if a certain belief were true, while rejecting that belief, need not be irrational in any way, provided there is some other belief that it is rational to accept, and that makes precisely the same actions rational. So consider, for example, Bas van Fraassen's view, which he defended in his book The Scientific Image, and elsewhere, that while we are not justified in believing that certain scientific theories are true, we are justified in believing that those scientific theories are empirically adequate.   Or consider the corresponding view with regard to the proposition that there are physical objects.

In short, in the case of physical objects, one may act in the way that it would be rational to act if there were physical objects if, rather than believing that there are physical objects, one believes instead that the theory of the existence of physical objects is experientially adequate, either because there really are such objects, or else because Berkeley's theory of reality is correct.

(4) As I shall discuss in more detail in connection with the G. E. Moore response to skepticism, it is crucial whether the common sense beliefs in question are inferential. For if they are inferential, the correctness of the inference is very much something that one may change one's mind about, given philosophical considerations. For example, if ordinary beliefs about physical objects involve, as a matter of fact, an inference to the best explanation of one’s sensory states, then consideration of alternative theories – such as Berkeley's – may lead to one's forming a much lower estimate of the plausibility of the original beliefs.

****************************************************************************

    Mike Huemer goes on to discuss the second of the three characteristics listed above, pointing out that philosophical skeptics continue to act, in ordinary life, as if the beliefs that they claim to be unjustified were true.

    Next, Mike Huemer explains what he means by a challenge to P: Q is a challenge to P if P and Q are not rationally cotenable.

    Mike Huemer argues – very plausibly, it seems to me – that his definition of skepticism is preferable to standard definitions – such as the definition of skepticism as “the view that no one can know anything.” For one thing, it applies to skepticism about justified belief as well as skepticism about knowledge. For another, skeptics typically grant that one has some knowledge.

6. Skepticism and Internal Justification

    Given the above definition of skepticism, skeptical challenges can both be directed against different sets of beliefs, and can take the following different forms:

(1) Certain beliefs are not true.

(2) Certain beliefs are not knowledge.

(3) Certain beliefs are not justified.

    Mike Huemer says that it is the last of these three sorts of claims that he wants to label “radical skepticism”. He also says that the radical skeptic that he wishes to respond to is not saying merely that certain beliefs are not all that highly justified:

“The radical skeptic I will be confronting holds that (some significant class of) common sense beliefs are not at all justified, which is to say: there is no reason to believe that they are true; it is no more rational to think they are true than to think they are false.”                             (SVP, p. 20)

    Mike Huemer's final point in this section is that radical skepticism is concerned with the internalist sense of justification, with what is justified from one's own point of view. This sense of justification is a deontological notion: it is concerned with what beliefs one should have.

    Some philosophers think that it turns out that there is no way of showing that there are beliefs that one should have, and they then often focus upon the axiological notion of what beliefs it would be good to have. Often such philosophers speak of an externalist sense of justification, or they speak about what beliefs are “warranted". But “warrant” also sounds like a deontological term, since to say that a belief is warranted sounds as if it is saying that a belief is justified. I think it is better to speak instead about what beliefs it is desirable to have.