INTRODUCTION
Suppose S believes p and then learns that someone she justifiably believes is an epistemic peer believes ~p. Should S withhold her belief, or at least decrease her confidence, that p? Conciliationists tend to say “yes”, and steadfasters will more often say “no”.Footnote 1 My description thus far does not precisely draw a line between conciliationists and steadfasters. How might we better divide the two camps? David Christensen believes that a deciding factor between them is whether they accept the following principle:
Independence “In evaluating the epistemic credentials of another's expressed belief about P, in order to determine how (or whether) to modify my own belief about P, I should do so in a way that doesn't rely on the reasoning behind my initial belief about P.” (Christensen Reference Christensen2011: 1–2)
According to Christensen, the conciliationist will agree with Independence, and the steadfaster will reject it.Footnote 2 This is, in Christensen's words, “the fundamental theoretical difference between the two camps” (1).Footnote 3
Although I agree with the general spirit of Independence, I believe that it requires revision. A central aim of this paper, therefore, is to develop a principle that improves upon the original Independence. In section 1, I both explain Independence and also present a revised, more widely applicable, version of it. In section 2, I present a series of counterexamples to both versions of Independence, and then formulate and endorse a third version of Independence that avoids those counterexamples. Lastly, in section 3, I show how this third version of Independence reveals two new ways one may remain steadfast in the face of two real life disagreements: one about God's existence and one about moral realism.
1. INDEPENDENCE AND THE FIRST REVISION
Christensen, a conciliationist, finds Independence to be plausible because it states a norm that disallows believers from too easily demoting the epistemic credentials of another's belief. He writes of Independence,
[I]t's intended to prevent blatantly question-begging dismissals of the evidence provided by the disagreement of others. It attempts to capture what would be wrong with a P-believer saying, e.g., “Well, so-and-so disagrees with me about P. But since P is true, she's wrong about P. So however reliable she may generally be, I needn't take her disagreement about P as any reason at all to question my belief.” (2011: 2)
To illustrate this point, consider his famous restaurant case:
Mental Math
After a nice restaurant meal, my friend and I decide to tip 20% and split the check, rounding up to the nearest dollar. As we have done many times, we do the math in our heads. We have long and equally good track records at this (in the cases where we've disagreed, checking with a calculator has shown us right equally frequently); and I have no reason (such as those involving alertness or tiredness or differential consumption of coffee or wine) for suspecting one of us to be especially good, or bad, at the current reasoning task. I come up with $43; but then my friend announces that she got $45. (Christensen Reference Christensen2011: 2)Footnote 4
Intuitively, Christensen ought to withhold belief, or at least lower his confidence, that the tip is $43. Independence is in accord with this intuition because it implies that he may not use his initial calculation in favor of his belief that the tip is $43 to discount the epistemic credentials of his friend's belief that the tip is $45. Instead, he may only use dispute-independent reasons to evaluate the epistemic credentials of his friend's belief, such as whether she had too much to drink or is sleep-deprived.
These considerations make Independence seem plausible. However, many objections have been given to Independence.Footnote 5 Christensen (Reference Christensen2011) and Matheson (Reference Matheson2015) have recently replied to many of these objections, and it is beyond the scope of this paper to delve into the very interesting and nuanced issues raised in those discussions. Instead, in the following section, I will present completely new counterexamples to Independence.
Before doing this, it will be valuable to present a slightly revised version of Independence, which has greater applicability. Consider the following well-known Quad case, by Richard Feldman (Reference Feldman and Hetherington2006):
Suppose that you and I are standing by the window looking out on the quad. We think we have comparable vision and we know each other to be honest. I seem to see what looks to me like a person in a blue coat in the middle of the quad. (Assume that this is not something odd.) I believe that a person with a blue coat is standing on the quad. Meanwhile, you seem to see nothing of the kind there. You think that no one is standing in the middle of the quad. We disagree. In isolation – before we talk to each other – each of us believes reasonably. But suppose we talk about what we see and we reach full disclosure. At that point, we each know that something weird is going on, but we have no idea which of us has the problem (424).
Let us add to Feldman's Quad case that he decides to reason in the following way. “Well, there is a person in a blue coat. So, there is something wrong with the way you formed your belief that there is no such person.”
Notice that Independence does not forbid this way of demoting the epistemic credentials of your belief. This is because Feldman is not relying on the reasoning behind his belief in order to assess the epistemic credentials of your belief. And this is simply because there was no reasoning behind his belief! He noninferentially formed the belief on the basis of his perceptual evidence. This is especially troublesome because Christensen formulated Independence in order “to capture what would be wrong with a P-believer saying, e.g., ‘Well, so-and-so disagrees with me about P. But since P is true, she's wrong about P’.” But this example shows that Independence does not even rule this reasoning out. It would be valuable to have a principle like Independence that not only forbids the use of the reasoning behind one's belief, but the use of the belief itself.
I therefore propose the following extension of Independence,
Independence1: In evaluating the epistemic credentials of another's expressed belief about P, in order to determine how (or whether) to modify my own belief about P, I should do so in a way that neither relies on the reasoning for my initial belief about P nor relies on my belief about P itself.
Independence1 thereby has wider applicability than Independence and can apply to Feldman's Quad case and cases like it. Furthermore, since Independence1 is so much in the spirit of Independence, it is likely that anybody who accepts the latter will also accept the former.
2. THE COUNTEREXAMPLES TO INDEPENDENCE AND THE SECOND REVISION
Most of my counterexamples will apply to both Independence and Independence1. The last one, Counterexample 5, will apply only to Independence1. Here is the first one:
Counterexample 1
Someone I know to be a reliable testifier tells me,
(P) Peggy is at the party
and
(Q) Quinn's beliefs about Peggy's location are unreliably formed.
I have no other special information about Quinn. I come to have a justified belief that P&Q. I infer via conjunction elimination that P. Quinn then comes up to me and tells me that ~P. So, I believe P and Quinn believes ~P. I infer Q from my belief that P&Q, and then give a low evaluation of the epistemic credentials of Quinn's belief that ~P on the basis of my belief that Q.
In evaluating the epistemic credentials of Quinn's expressed belief about P, I relied on a premise, P&Q, which was also a premise in the reasoning behind my initial belief about P.Footnote 6 Yet it seems that my reasoning was permissible and that I may remain steadfast in my belief that P. This is therefore a counterexample to Independence.
One might object that my example makes unrealistic assumptions about human psychology. If I were to learn P&Q from a reliable testifier, I would first learn P, then Q, and then finally, P&Q. Even if the reliable testifier utters “P&Q” very quickly, my human mind will first come to believe P and then come to believe Q. But in that case, I would be inferring neither P nor Q from P&Q.
My first response to this objection is that it still seems possible (in a broadly logical or metaphysical sense) that there be intelligent, nonhuman creatures that could reason as in Counterexample 1; that is enough for it to serve as a counterexample. My second response is that we can tell the story in such a way that it is clearer that I come to believe P&Q all at once. Suppose the reliable testifier first tells me to consider P&Q. After a moment, she says, “And that conjunction is true!” Immediately, it seems, I will come to believe the conjunction all at once.Footnote 7
My third response is to formulate a new counterexample to Independence that even more clearly avoids the objection. Recall that ‘P’ denotes the proposition that Peggy is at the party and ‘Q’ denotes the proposition that Quinn's beliefs about Peggy's location are unreliably formed. Here is the new counterexample:
Counterexample 2
Someone I know to be a reliable testifier tells me,
If ~(P&Q), then Sally is at the store.
I have no other special information about Quinn. I come to justifiedly believe this conditional. I immediately go to the store and see that Sally is not there. I infer P&Q via modus tollens. I then infer via conjunction elimination that P. Quinn comes up to me and tells me that ~P. So, I believe that P and Quinn believes that ~P. The rest of the story proceeds as in Counterexample 1.
In this case, it is clear that I learn P&Q all in one fell swoop. As in Counterexample 1, it seems that I may give a low evaluation of the epistemic credentials of Quinn's belief that ~P and remain steadfast in my belief that P.
I draw the following lesson from these counterexamples:
Lesson 1: If S's reasoning for P includes a justifiedly believed premise that makes a claim about the epistemic credentials of T's belief about P, then S may use that premise to evaluate the epistemic credentials of T's belief about P.
In Counterexamples 1 and 2, P&Q is a premise in my reasoning for P, and it makes a claim about the epistemic credentials of Quinn's belief about P.Footnote 8 According to Lesson 1, it is thereby permissible to use that premise to evaluate the epistemic credentials of Quinn's belief about P.
Lesson 1 specifies one sufficient condition for the permissibility of using a premise in one's reasoning for P to evaluate the epistemic credentials of another's belief about P. Are there any other such conditions? I will present two more counterexamples that support the existence of another condition. Consider,
Counterexample 3
Someone I know to be a reliable testifier tells me,
(P) Peggy is at the party
and
(R) If Sally is at the store, then Q. (Reminder: ‘Q’ denotes the proposition that Quinn's beliefs about Peggy's location are unreliably formed.)
I have no other special information about Quinn. I come to have a justified belief that P&R. I infer via conjunction elimination that P. Quinn then comes up to me and tells me that ~P. So, I believe that P and Quinn believes that ~P. I then run to the store and see that Sally is there. I use modus ponens to infer Q. I give a low evaluation of the epistemic credentials of Quinn's belief that ~P on the basis of my belief that Q.
Notice that in this case, the premise that I used in my reasoning for P, the conjunction P&R, is not itself a premise that makes a claim about the epistemic credentials of Quinn's belief.Footnote 9 Its second conjunct, R, is merely a conditional. Hence, unlike in Counterexamples 1 and 2, neither of the premises make a claim about the epistemic status of the other person's belief. However, I could still use P&R to validly infer a proposition, Q, which does make a claim about the epistemic credentials of Quinn's belief. Furthermore, as in the other cases, it seems that my reasoning is permissible, and so this is another counterexample to Independence.
Some readers have expressed suspicion about my reliance on conjunctions in Counterexamples 1–3.Footnote 10 I do not think there is, but to assuage those worries, here is a counterexample to Independence that does not rely on conjunctions:
Counterexample 4
Someone I know to be a reliable testifier tells me,
(U) If Sally is at the store, then P. (Reminder: ‘P’ denotes the proposition that Peggy is at the party.)
(V) If Sally is at the store, then Q. (Reminder: ‘Q’ denotes the proposition that Quinn's beliefs about Peggy's location are unreliably formed.)
I come to justifiedly believe both U and V. I run over and see that Sally is at the store. I use modus ponens to infer P. Quinn then comes up to me and tells me that ~P. So, I believe that P and Quinn believes that ~P. I then use modus ponens to infer Q. I give a low evaluation of the epistemic credentials of Quinn's belief that ~P on the basis of my belief that Q.
In Counterexample 4, the proposition that Sally is at the store is both a premise in my reasoning for P and also a premise in my reasoning for Q. On the basis of my belief that Q, I give a low evaluation of the epistemic credentials of Quinn's belief about P. As in the other cases, it seems that my reasoning is permissible. Furthermore, this counterexample to Independence does not rely on a conjunction.
One might draw the following lesson from Counterexamples 3 and 4:
Lesson 2: If there is a valid inference from a justifiedly believed premise in S's reasoning for P (along with other propositions S justifiedly believes) to a proposition that makes a claim about the epistemic credentials of T's belief about P, then S may use that premise in support of her evaluation of the epistemic credentials of T's belief.
However, Lesson 2 faces the following objection.Footnote 11 Suppose that in the Mental Math case, Christensen knew the premises in his reasoning for his belief that the tip is $43. Furthermore, suppose that there is a valid inference from those premises to the proposition that the tip is $43. Then there is also a valid inference from those premises to the proposition that the tip is not $45, and hence, to the proposition that his dinner partner's belief is false. Now, note that while some epistemic properties do not entail truth (e.g., being formed epistemically blamelessly and being formed reliably), other epistemic properties do.Footnote 12 For example, knowledge entails truth. A number of epistemologists have argued that warrant entails truth (where ‘warrant’ is stipulatively defined as that which turns true belief into knowledge).Footnote 13 Supposing that it does, there is a valid inference from premises in his reasoning to the proposition that his friend's belief is false, and is therefore unwarranted or not an instance of knowledge. Hence, there is a valid inference from premises in his reasoning to a claim about the epistemic credentials of his friend's belief. Lesson 2 permits reasoning in accord with this valid inference.
But the Mental Math case is supposed to be a paradigm example in which he may not use those premises to evaluate the epistemic credentials of his friend's belief! Since I find this objection to be persuasive, I propose the following lesson to replace Lesson 2.
Lesson 3: If there is a valid inference from a justifiedly believed premise in S's reasoning for P (along with other propositions S justifiedly believes) to a proposition that makes a claim about the non-truth-entailing epistemic credentials of T's belief about P, then S may use that premise in support of her evaluation of the non-truth-entailing epistemic credentials of T's belief.
Lesson 3 avoids the objection to Lesson 2 by restricting itself to non-truth-entailing epistemic credentials.Footnote 14
My final counterexample only applies to Independence1 and not to Independence. As a reminder, Independence1 states
Independence 1 : In evaluating the epistemic credentials of another's expressed belief about P, in order to determine how (or whether) to modify my own belief about P, I should do so in a way that neither relies on the reasoning for my initial belief about P nor relies on my belief about P itself.
Now consider,
Counterexample 5
Someone I know to be a reliable testifier tells me,
(P) Peggy is at the party.
(W) If P, then Q. (Reminder: ‘Q’ denotes the proposition that Quinn's beliefs about Peggy's location are unreliably formed.)
I have no other special information about Quinn. I come to justifiedly believe both P and W. Quinn then comes up to me and tells me that ~P. So, I believe that P and Quinn believes that ~P. I then use modus ponens to infer Q from P and W. I give a low evaluation of the epistemic credentials of Quinn's belief that ~P on the basis of my belief that Q.
The use of variables in this case allows for its concise statement of it but also may interfere with its eliciting a clear intuition. So, I will state it again without the variables. On the basis of known, reliable testimony, I learn both that Peggy is at the party, and that if she's at the party, then Quinn is unreliable about Peggy's location. After I hear this testimony, but before I infer Quinn's unreliability, I see Quinn walking up to me, reporting to me that Peggy is not at the party. I then finish my inference and conclude that Quinn's unreliable. It seems both that this reasoning is permissible and also that I may remain steadfast in my belief that Peggy is at the party on the basis of my low evaluation of the epistemic credentials of Quinn.
Notice that in this example, in evaluating the epistemic credentials of Quinn's expressed belief about P, I relied on the belief that P itself! This is what makes this case apply specifically to Independence1. I now propose
Lesson 4 : If, independent of the disagreement, S justifiedly believes P, and there is a valid inference from P (along with other propositions S justifiedly believes) to a proposition that makes a claim about the non-truth-entailing epistemic credentials of T's belief about P; then S may use the belief that P in support of her evaluation of the non-truth-entailing epistemic credentials of T's belief.
Lesson 4 states a condition that allows for the person to continue to justifiedly believe that P even in light of the disagreement.
Let us take stock. I have provided five counterexamples and drawn three lessons (Lessons 1, 3, and 4) from those counterexamples. In light of that discussion, I propose the following principle, which is identical to Independence1, except for the new conditions (a), (b), and (c):
Independence 2 : In evaluating the epistemic credentials of another's expressed belief about P, in order to determine how (or whether) to modify my own belief about P, I should do so in a way that neither relies on the reasoning I used to form my initial belief about P nor relies on my belief about P itself, unless
(a) A justifiedly believed premise in S's reasoning about P makes a claim about the epistemic credentials of the other's expressed belief about P or
(b) There is a valid inference from a justifiedly believed premise in S's reasoning about P (along with other propositions S justifiedly believes) to a proposition that makes a claim about the non-truth-entailing epistemic credentials of T's belief about P or
(c) Independent of the disagreement, S justifiedly believes that p, and there is a valid inference from S's belief about P (along with other propositions S justifiedly believes) to a proposition that makes a claim about the non-truth-entailing epistemic credentials of T's belief about P.
(a) is drawn from Lesson 1, (b) is drawn from Lesson 3, and (c) is drawn from Lesson 4. Independence2, like Independence1, is an improvement on the original Independence because it has a wider range of application. Furthermore, it is an improvement to both Independence1 and Independence because it avoids all counterexamples 1–5.Footnote 15
3. APPLICATIONS OF THE LESSONS
3.1 Set Up
In the following, I will show that Lesson 4 has interesting, real world applications, specifically, cases where condition (c) in Independence2 is met. Lesson 4, you will recall, states,
Lesson 4: If, independent of the disagreement, S justifiedly believes P, and there is a valid inference from P (along with other propositions S justifiedly believes) to a proposition that makes a claim about the non-truth-entailing epistemic credentials of T's belief about P; then S may use the belief that P in support of her evaluation of the non-truth-entailing epistemic credentials of T's belief.
The following is supported by Lesson 4:
Lesson 4a: If, independent of the disagreement, S justifiedly believes P, and S also justifiedly believes <if P, then Q>, where Q is a proposition that makes a claim about the non-truth-entailing epistemic credentials of T's belief about P; then S may use the belief that P in support of her evaluation of the non-truth-entailing epistemic credentials of T's belief.
For an illustration, the believer in Counterexample 5 reasons according to Lesson 4a to evaluate the epistemic credentials of Quinn's belief.
Can Lesson 4a be applied to real life? In most cases, we do not justifiedly believe any such <if P, then Q>. Put another way, in most cases, we do not justifiedly believe a conditional proposition linking the proposition believed to some relevant proposition about the non-truth-entailing epistemic credentials of the other person's belief. In my counterexamples, I relied heavily on reliable testifiers to provide the needed justified beliefs, but their insertion is what made the examples unrealistic. In real life, we rarely have such reliable testifiers. So, are there any real life cases in which a person justifiedly believes a conditional proposition linking the proposition believed to some relevant proposition about the non-truth-entailing epistemic credentials of the other's belief?
3.2 Application to Disagreement about Theism
I argue that there are such cases. Consider Frederick, a theist, and Bertrand, an atheist. Each justifiedly holds his belief about God's existence. They then meet and find that they disagree. Imagine that Bertrand believes,
(1) If God does not exist, then Frederick formed his theistic belief in an epistemically defective way.
Returning to Lesson 4a, P is the proposition that God does not exist; Q is the proposition that Frederick formed his belief in an epistemically defective way; and (1) is the <if P, then Q>, the conditional proposition linking the proposition Bertrand believes to a proposition about the non-truth-entailing epistemic credentials of Frederick's belief.
But is Bertrand's belief in (1) justified? How might it be justified? Bertrand could employ the following premises in favor of (1):
(2) If God does not exist, then theists who form belief in God noninferentially do so in an epistemically defective way.
(3) Frederick is a theist who forms belief in God noninferentially.
But what is Bertrand's justification for (2) and (3)? We can easily stipulate that (3) is true and that Bertrand knows that it is true. But how would Bertrand get justification for (2)? Nothing is wrong with noninferential belief formation in itself; paradigm cases of justified belief, such as perceptual or introspective belief, are noninferential. And why think that God's existence has anything to do with whether the theist's noninferentially formed belief in God is epistemically defective?
Ironically, Bertrand can find help in the work of Alvin Plantinga. In the relevant section of Plantinga's work that Bertrand is reading, Plantinga is writing specifically about noninferentially formed theistic belief. He argues that “if theistic belief is false, it is not produced by cognitive processes successfully aimed at truth” (Plantinga Reference Plantinga2000: 187). He supports this claim with the following:
If there is no such person as God, of course, then there is no such thing as a sensus divinitatis [a faculty or set of faculties designed to form true beliefs about God]; and what truth-aimed faculty would be such that it is working at the limit of its ability in producing the belief that there is such a person as God, if that latter belief is false? It is exceedingly hard to think of decent candidates (187).
Plantinga then goes on to give some candidates of non-truth-aimed processes that might be responsible for belief in God, on the supposition that God does not exist:
You may think we human beings are the product of blind evolutionary forces; you may think that there is no God and that we are part of a godless universe. Then you will be inclined to accept the sort of view according to which belief in God is an illusion of some sort, properly traced to wishful thinking or some other cognitive mechanism not aimed at the truth (Freud) or to a sort of disease or dysfunction on the part of the individual or society (Marx). (191)
Bertrand uses these considerations to justify (2). He then infers (1) from (2) and (3).
My goal is not to argue that Plantinga's reasoning is successful or that (2) is true. That would require more than the brief summary of Plantinga's argument that I have given here.Footnote 16 I only need to point out that it is plausible that Bertrand could come to justifiedly believe (2) and (3), and thereby (1), along the lines I've mentioned. And it seems that Bertrand could justifiedly believe these propositions on the basis of Plantinga's argument in the same way that any person comes to justifiedly believe some proposition on the basis of an argument in the philosophical literature. Bertrand could then follow Lesson 4a and also not violate Independence2 in his evaluation of the epistemic credentials of Frederick's theistic belief because of condition (c).
Is my example of Bertrand and Frederick just a toy example that is not found in real life? I do not think so. It seems that many atheists find themselves in a position like Bertrand's. First, there are many atheists who justifiedly believe a proposition similar to (3), that is, they justifiedly believe that a theist noninferentially believes in God. This is partly so because it has become common for theists in philosophy to say that their belief in God is properly basic (noninferentially justified or warranted).Footnote 17 And both philosophical and nonphilosophical theists will more colloquially say things like, “I just know that there's a God”, or they will report having directly perceived God, often by saying that they felt God's presence.Footnote 18 It is reasonable to think that some of these people believe that God exists noninferentially.
Second, there are many atheists who justifiedly believe a proposition like (2). They might have directly read Freud, Marx, or some cognitive scientists of religion arguing that theistic belief is produced by an epistemically flawed process.Footnote 19 They might have reflected on the works of these thinkers and wondered, like Plantinga, “But if God is the ultimate creator and designer of humanity, then it's not so clear that that process wasn't actually God's truth-aimed design plan. On the other hand, if God does not exist, then yes, that process is likely to be epistemically defective.” They would then conclude (2). Note, once again, that (2) need not be true; it only must be that there are atheists who justifiedly believe it.
My last reason for thinking that there are many atheists who are in a position like Bertrand's is simply that I know many atheists who seem to be in this position! You, the reader, might even be such an atheist. So, we have good reasons to think that the case of Bertrand is not unrealistic. There are atheists who come to justifiedly believe a proposition like (1) on the basis of (2) and a proposition like (3). I conclude that Lesson 4a has an interesting, real life application. Bertrand and some atheists meet the exception clause (c) in Independence2. Lastly, I'll note that it seems that Bertrand would be justified in maintaining his atheistic belief in the face of disagreement.
Can Frederick argue in a similar way? He would need the following proposition:
(4) If God exists, then Bertrand formed his atheistic belief in an epistemically defective way.
And Frederick might support (4) like Bertrand supported (1):
(5) If God exists, then atheists who form belief in atheism noninferentially do so in an epistemically defective way.
(6) Bertrand forms belief in atheism noninferentially.
Interestingly, Frederick will receive no direct aid from Plantinga, who never argued for anything like (5). Plantinga (Reference Plantinga2000: 186–91) does argue that if God exists, then the theist's belief in God has certain valuable epistemic properties, and that is because the belief would be formed by a God-designed cognitive faculty, a sensus divinitatis, that is successfully aimed at truth. But Plantinga does not directly talk about the implications of theism on the atheist's belief.
Formulating (5) and (6) like (2) and (3) would leave us with an example that is not likely to have much real life application. While there are theists who form their belief noninferentially (and will report that they have done so), it is far less common to hear atheists say that their belief in God's nonexistence is properly basic, and rarely will an atheist report that she just knows that atheism is true. So, we could still stipulate that Frederick believes (6), but the example would then have little real world application since most theists don't have good justification to believe that they know atheists who form their belief noninferentially.
We might try to alter (5) and (6). To make it work, we would have to specify another method of belief formation. Frederick would have to justifiedly believe both that <Bertrand forms his atheistic belief in way X>, and also that <if God exists, then forming atheistic belief in way X is epistemically defective>. Furthermore, this case of Frederick and Bertrand would have to not be too unusual, or we would lose the whole point of this discussion, namely, finding real life applications of Lesson 4a. I think that it would be difficult to find a suitable candidate for X. Many atheists will say that their atheistic belief is a result of standard philosophical reasoning (via, say, the problem of evil), or inference to the best explanation, or whatever reasoning is involved when one comes to believe there are no unicorns. It would take a lot of work to argue that these methods are epistemically defective on the supposition that God exists. To be clear, I am not saying that Frederick does not have interesting moves to make here or that he (and other theists) must withhold belief in God in the face of disagreement from atheists. I am only saying that Lesson 4a does not straightforwardly apply to Frederick as it did to Bertrand.
At this point, it will be useful to note what made the application of Lesson 4a possible. In the case of Bertrand's reasoning about Frederick, there is a decent case to be made that atheism has interesting implications for what sorts of belief forming processes are responsible for theistic belief. Plantinga (Reference Plantinga2000) argued that the de facto question (whether God exists) is connected to the de jure question (whether theistic belief is rational or warranted). The very same reasoning would make it plausible that the de facto question is connected to the de paribus question (the question of whether someone is a peer). Bertrand's negative answer to the de facto question about God's existence made plausible his negative answer to the de paribus question about Frederick. Are there other cases where the de facto question and the de paribus question are so connected? In the following section, I will show that there is such a connection in debates about moral realism.
3.3 Application to Disagreement About Moral RealismFootnote 20
Let ‘moral realism’ denote the view that there are some stance-independent moral truths. Consider David, a moral realist, and Sharon, a moral antirealist. Each justifiedly holds his or her belief about moral realism. They then meet and find that they disagree. Imagine that Sharon believes,
(7) If moral realism is false, then David formed his belief in moral realism in an epistemically defective way.
How might Sharon justify her belief? She could employ the following premises:
(8) If moral realism is false, then moral realists who form belief in moral realism on the basis of moral intuitions do so in an epistemically defective way.
(9) David formed his belief in moral realism on the basis of moral intuitions.
Sharon might believe (7) on the basis of (8) and (9).
But what is Sharon's justification for (8) and (9)? We can easily stipulate that (9) is true and that Sharon justifiedly believes that it is true. But how would Sharon get justification for (8)? Perhaps Sharon has read work on evolutionary debunking arguments (or ‘EDAs’).Footnote 21 Here is Katia Vavova's (Reference Vavova2015) recent summary of their general form:
The thought is that just as evolutionary forces shaped our eyes and ears, so they shaped our moral beliefs – or, at least, early, proto-versions of those beliefs. But evolutionary forces select for survival, not moral truth. So, if realism is true, and the evolutionary story is roughly thus, then evolution has pushed our moral beliefs in directions having nothing to do with the attitude-independent moral truths. If this is right, then realism entails moral skepticism. (104)
Here is a way to state one part of this argument. Given the truth of moral realism, it is unlikely that the evolutionary processes that shaped our moral-intuition-producing dispositions are reliable. Some objectors to EDAs have argued that it is likely that evolutionary processes would shape us to have reliable moral-intuition-producing dispositions.Footnote 22
But Sharon can successfully side step these objections in her reasoning for (8). Perhaps, if moral realism is true, then evolutionary processes shaped our moral-intuition-producing dispositions so that they would be reliable. But if moral realism is false, then of course it is likely that our moral-intuition-producing dispositions are unreliable! Sharon can then go on to believe (7) on the basis of (8) and (9). Note that I only need to argue that Sharon could come to justifiedly believe (8) and (9), and thereby (7), along the lines I've mentioned. Sharon could then follow Lesson 4a and also not violate Independence2 in her evaluation of the epistemic credentials of David's belief in moral realism.
Is this example realistic? First, many antirealists justifiedly believe a proposition similar to (9), that is, they justifiedly believe that someone believes in moral realism on the basis of their moral intuitions. Second, although it is not very obvious whether many antirealists explicitly believe (8), it is clear that antirealists could easily come to justifiedly believe it. Hence, even if no antirealist currently believes (8), an antirealist's being in a position to justifiedly believe (8) and (9) – and hence, (7) – puts them in a position to use Independence2 and Lesson 4a to give a low evaluation of the epistemic status of many moral realists.Footnote 23
So, we can apply Lesson 4a to disagreements about theism and moral realism. It remains to be said whether there are other applications of Lesson 4a, cases where a person is justified in believing a relevant conditional linking the de facto claim to the de paribus claim. It is also unclear whether there are other applications of Lessons 1, 3, and 4. It would be interesting if there were, and that is where this paper can point to further research.
4. CONCLUSION
In conclusion, I have argued that the Independence principle has a number of counterexamples. I then presented a new version of Independence, Independence2, that both expands the scope of Independence and also avoids the counterexamples I raised in this paper. Lastly, I applied a lesson from the paper and showed a new way that atheists and antirealists can remain steadfast in the face of disagreement.Footnote 24
