Compassionate Phenomenal Conservatism

Here, I explain why you should endorse phenomenal conservatism in epistemology.*

[* Based on “Compassionate Phenomenal Conservatism,” Philosophy & Phenomenological Research 74 (2007): 30-55.]

1. Side Notes

Terminological notes:

• The title of this paper alluded to the phrase “compassionate conservatism” that President Bush had made popular; the title hasn’t aged well. Compassion has nothing to do with the paper.

• “Phenomenal” comes from the Greek word “phainomenon”, meaning “appearance”. The “phenomenal conservative” view is so called because it has to do with preserving appearances.

Background: An earlier view in the literature is doxastic conservatism (which some people call “epistemic conservatism”), which holds that the mere fact that you currently believe P provides at least some (defeasible) justification for believing P. Richard Foley refuted this with the following example:

• S has almost but not quite enough evidence for P to justify believing P. S goes ahead and forms the belief anyway. According to doxastic conservatism, it looks like, as soon as S forms this irrational belief, it immediately becomes rational, since now, in addition to the original evidence, S has a new source of justification for P, namely, the fact that S believes it. This would push S over the threshold for justified belief.

So doxastic conservatism is not plausible. But maybe an analogous view with appearances in place of beliefs is more plausible.

2. Phenomenal Conservatism

This is the view:

PC: If it seems to you that P, and you have no grounds for doubting that appearance, then that gives you at least some degree of justification for believing P.

Comments:

• A “seeming” or “appearance” is a type of mental state that represents the world as being a certain way (it has propositional content, as we say).

• It is different from a belief, since you can refuse to believe the appearances, and that doesn’t necessarily make them go away (as when you see a perceptual illusion).

• But appearances normally cause beliefs (when you don’t have any reason to doubt them and you’re rational).

• Appearances aren’t the only things that can cause beliefs. E.g., a person sometimes believes P because they want to believe it; this is very different from believing it because it seems true.

• There are several species of appearances, including sensory experiences, (apparent) memories, and intuitions. Reasoning also relies on a special type of appearance called an inferential appearance (discussed in a later paper).

• PC doesn’t say appearances are always true; it says, basically, that appearances are to be presumed true until grounds for doubt appear.

• An “intuition”, btw, is a special type of appearance. It is an initial (non-inferential), intellectual appearance.

3. The Self-Defeat Argument

This is my favorite argument for PC (though it is not popular in the literature).

1. When we form beliefs in relevant cases (cases in which we are seeking the truth, cases that are initially plausible candidates for justified beliefs), our beliefs are always based on how things seem to us.

This is true in general, including when forming beliefs about epistemology, including when you’re forming beliefs about PC itself—your belief about the truth of PC will be based on what seems correct to you.

Beliefs based on reasoning are no exception to this. The arguments that count as at least prima facie serious arguments are those with premises that seem true to us and that seem to support the argument’s conclusion. Without some distinction between prima facie serious arguments and mere random strings of sentences followed by conclusions, there is no reasoning.

2. A belief is justified only if it is based on a valid source of justification.

This is widely accepted in epistemology. Even if you have good evidence for P, your belief in P won’t count as a justified belief unless you believe P because of that evidence. If you have some great evidence available but the actual reason you believe P is that your fortune-teller told you P, then you have an unjustified belief.

3. If ~PC, then appearances aren’t a valid source of justification.

If appearances are a source of justification, then it would seem that they should justify beliefs in the things that seem correct (rather than, say, things that seem wrong), and that they should do so when we lack grounds for doubting them. See further discussion below.

4. So if ~PC, then all beliefs are unjustified (including the belief in ~PC).

Follows from 1-3. So it’s self-defeating to deny PC, in that if you’re right, your own belief is unjustified.

4. Alternative Theories

Let’s look at some alternative theories of justification and discuss whether they avoid self-defeat.

4.1. Reliabilism

A reliabilist might say that only appearances that are in fact reliable confer justification, while other appearances do not. Maybe this avoids self-defeat as long as the reliabilist claims that the appearances his own view rests on are among the reliable ones.

Bear in mind that the self-defeat argument does not turn on whether one has justification for a belief. It turns on whether one’s belief is based on a valid source of justification, or whether one holds a belief because of the thing that provides justification for it.

On the reliabilist view, the actual reliability of a belief-forming method (regardless of whether the subject knows it is reliable) is what confers justification on a belief. But this actual reliability plays no role in explaining the subject’s belief, once we control for appearances. Reliability only influences your beliefs through affecting the appearances. E.g., more reliable information sources tend to cohere better with themselves, tend to better match what other people say, and tend to lead to more satisfied expectations, all of which affects the appearances that you will experience. However, once you control for that, reliability has no effect on your beliefs. In probability theory jargon, the appearances screen off reliability from your beliefs.

Thus, one can imagine pairs of cases in which you have the same appearances but different levels of reliability, and the subject would have the same beliefs. E.g., a brain in a vat has the same sensory appearances as you, but his are completely unreliable while yours are (I assume) reliable. The BIV (if equally rational as you are) would have the same sort of beliefs you have.

You can also imagine cases in which two belief sources have the same level of actual reliability but the subject has different appearances, so one source appears more reliable to the subject. In that case, the subject would form a stronger belief based on the one source than the other.

All of which shows that what really explains your beliefs is appearance facts, not reliability facts.

4.2. The Acquaintance Theory

Acquaintance theorists say that foundational justification is conferred by a kind of direct awareness (“acquaintance”) that we have with certain things, notably our own mental states and some abstract objects. (See Bertrand Russell and Richard Fumerton.) Acquaintance differs from appearance in that one logically cannot be acquainted with x unless x exists, whereas one can have appearances of things that don’t exist. E.g., if you’re hallucinating a pink hippo, then you have a hippo appearance, but you aren’t acquainted with a hippo. (Acquaintance theorists, by the way, generally deny that anyone can be acquainted with hippos or any other physical objects.)

You might think this view could avoid self-defeat by claiming to be itself based on acquaintance with the nature of justification, or something like that. But again remember that the question is whether the theory recognizes the actual basis of our beliefs as a source of justification. Are our beliefs based on acquaintance?

Parallel to the above argument about reliabilism, I propose that we consider pairs of cases that differ with respect to acquaintance but are the same with respect to appearance. So here’s an example.

Gottlob considers two axioms of set theory:

a. The Comprehension Axiom: For any well-formed predicate, there is a set containing all and only the things that satisfy that predicate.

b. The Axiom of Extensionality: Set A is identical to set B if and only if A and B have all the same members.

When he thinks about these, they both seem to him self-evident. He has the experience of “just seeing” that these things must be true. This isn’t purely made-up, by the way—in the late 19th century, most people who thought about sets in fact found (a) and (b) self-evident, they seemed to know (a) and (b) in the same way as they knew any other mathematical axioms, and they had no reason to doubt either.

Later, it turned out that (a), the Comprehension Axiom, is false. This is proved by Russell’s Paradox (consider the predicate “is not a member of itself”).

(b), however, the Axiom of Extensionality, is still accepted.

So no one was ever acquainted with a fact that made (a) true, since there is no such fact. All acquaintance theorists would agree with this. But they would also presumably say that we are acquainted with the fact that makes (b) true, and that is how we know it.

So here we have two propositions with respect to which the appearances are relevantly alike ((a) and (b) each seem equally obvious) but we have acquaintance with the truth of one of them and lack acquaintance with the truth of the other. That gives us a test case for whether beliefs are based on acquaintance. And the answer is that they are not, because in fact everyone who thought about (a) and (b) believed both with about equal confidence. The fact that one of them was actually an object of acquaintance and the other not made no psychological difference. All that mattered to people’s beliefs was the appearances: Gottlob accepted both because both seemed correct to him.

5. Conclusion

If you buy that, you can probably see how it goes for any other theory. Once we fix the appearances, I claim, nothing else affects our beliefs (in relevant cases, i.e., the cases that are initially plausible candidates for rational beliefs—though of course, you can have desires affecting beliefs when people are being irrational). That shows that appearances are the actual basis of our beliefs.

So any rival to Phenomenal Conservatism that identifies something other than appearances as the source of justification for our beliefs will imply (given the actual psychological facts) that our beliefs are not based on a valid source of justification, i.e., we do not believe what we do because of the justification that’s available for it.

And that, in turn, means that our beliefs would not actually be justified.

But it’s irrational to hold a belief while also thinking that that very belief is unjustified. So you should not endorse any rival theories to Phenomenal Conservatism.

By the way, this applies to skepticism just as much as to reliabilism and the acquaintance theory. The skeptic relies on his own intellectual appearances in coming to the conclusion that no beliefs are justified.