Miriam Schoenfield’s “Permission to Believe: Why Permissivism Is True and What It Tells Us About Irrelevant Influences on Belief”

  1. Introduction

Shoenfield begins by presenting a concern regarding belief. Oftentimes, our most deeply held beliefs are caused by seemingly irrelevant influences, such as the community in which we grew up, the school we attended, or the immediate people in our lives. These cases are called “IF cases.” Shoenfield claims we need not worry in many such cases. First, Schoenfield argues for the truth of permissivism. Second, Shoenfield argues that, if we take permissivism to be true, the concern raised by irrelevant influences on belief is, in many cases, unwarranted. Third, Shoenfield raises an issue with her view and responds to it. Finally, Shoenfield turns to the issue of peer disagreement and examines how permissivism sorts out issues related to the topic.

  1. What is permissivism and why is it true?

Shoenfield defines permissivism as “the claim that sometimes, there is more than one rational response to a given body of evidence” (194). There are non-permissive cases in which there is only one rational way to respond to a body of evidence, and permissive cases in which there is more than one rational way to respond to a body of evidence. Shoenfield argues that, starting with a non-permissive case, if we learn that it is an IF case, a significant reduction in confidence may be warranted. In a permissive case, however, finding out such a case is an IF case need not worry us. Shoenfield’s main argument goes as follows:

P1. Permissivism is true.

P2. If permissivism is true, the view that you should reduce confidence significantly in permissive irrelevant influence cases is unmotivated.

P3. If the view that you should reduce confidence significantly in permissive irrelevant influence cases is unmotivated, you don’t need to reduce confidence in such cases.

C1. You don’t need to reduce confidence significantly on the basis of irrelevant influences in permissive cases. (P1-P3).

To argue for the truth of P1, Shoenfield first introduces a countering view:

UNIQUENESS: For any body of evidence E, and proposition P, there is only one doxastic attitude to take towards P that is consistent with being rational and having evidence E.

Permissivism is simply the denial of UNIQUENESS. In other words, permissivism states that there are at least some cases in which there is more than one rational response to a body of evidence E. Shoenfield provides two arguments to motivate permissivism, one intuitive and one theoretical.

  1. Intuitive Motivations

Intuitively, it seems reasonable people can disagree when confronted with the same body of evidence. This extends beyond scientific contexts. For example, it seems jury members may rationally arrive at different conclusions given the same body of evidence. Some argue that everyone’s unique experiences constitute different bodies of evidence which may influence what is rational to believe. If this is true, we can maintain belief in UNIQUENESS while accepting that different jurors can rationally come to different conclusions. Shoenfield shows it seems very unlikely that all cases work this way. For example, paleontologists’ unique life experiences don’t seem to make substantive differences in the debate between what killed the dinosaurs. A further concern is, if we assign such importance to life experiences, this calls into question the trustworthiness of experts’ opinions in their given fields.

  1. Theoretical Motivations

Shoenfield argues that many plausible theories of justification require the truth of permissivism including coherentism, conservatism, and subjective Bayesianism. Furthermore, some theories of justification that reject permissivism contain unfortunate metaphysical commitments. For example, consider the Bayesian who thinks in terms of degrees of belief. According to UNIQUENESS, the Bayesian must believe in a unique real number between 0 and 1 that measures the appropriate credence to have in a proposition. Furthermore, she must believe that there exists some general principles that can objectively determine such a number. Given a proposition such as “there exist more than three hundred elephants,” this seems like an undesirable feature of UNIQUENESS. Rather, it seems like some principle is necessary to explain what is unreasonable about a given position, rather than some brute fact.

Question/Comment: Are you persuaded to believe in permissivism given these intuitive and theoretical motivations? It seems to me that the intuitive motivations are much stronger than the theoretical ones. Nevertheless, it seems to me like there is a rational position which denies these arguments. So is permissivism itself a permissive case? Note that Schoenfield’s entire argument hinges on the truth of permissivism.

2.1 Problems with permissivism

First, Shoenfield notes an important aspect of her permissivism. That is, “what one ought to believe depends, in part, on what epistemic standards one has” (199). Shoenfield defines a set of epistemic standards as “a function from bodies of evidence to doxastic states which the agent takes to be truth conducive” (199).

Question/Comment: Shoenfield’s epistemic standards are like stances/values in the sense that these standards precede beliefs and thus do much to determine them. Furthermore, Shoenfield argues that, according to permissivism, there are multiple permissible standards. Do you see any benefits/drawbacks to defining standards in this way?

With this talk of epistemic standards in our back pocket, we are ready to look at the objections levied against permissivism.

  1. The evidence pointing problem (Sosa, White)

A seemingly problematic aspect of permissivism is that it is permissible, given E, to believe p and permissible, given E, to believe ~p. Surely, we do not want to accept that the evidence supports both p and ~p. This idea rests, however, on a faulty assumption. Shoenfield’s permissivist does not think that the evidence “dial” points both to p and ~p simultaneously. Rather, there are multiple evidence “dials” corresponding to different permissible ways of weighing the evidence, due to different epistemic standards. Thus, it is not the case that given one set of epistemic standards, one is permitted to believe p and ~p.

  1. A cluster of worries: arbitrariness (Christensen, Feldman, White)

If one thinks that, given E, it is reasonable to both believe p and ~p, it seems having one belief, rather than the other, is arbitrary. According to Roger White, we can think of belief forming as taking a pill. If we take pill #1, we believe p. If we take pill #2, we believe ~p. Either we can look at evidence and come to whatever conclusion we come to, or we can take a randomly selected pill and come to our conclusion that way. If one is a permissivist, it seems both methods are equally likely to lead to truth, both being rational. Standards can help ease our worries here. Recall that one’s epistemic standards are thought to be more truth conducive to the individual than other standards. Taking the pill could either result in a belief conflicting with one’s standards, or result in changing one’s standards themselves. Both of these instances look undesirable to the permissivist, so she should deny the pill. Importantly, regardless of whether one is a permissivist, one cannot give independent reasons for weighing evidence in a certain way.

  1. How permissivism bears on irrelevant factor cases (defense of P2)

Now that we can assume permissivism is true, Shoenfield turns to P2.

P2. If permissivism is true, the view that you should reduce confidence significantly in permissive irrelevant influence cases is unmotivated.

Shoenfield introduces two hypotheses which motivate decreasing confidence.

RATIONAL INDEPENDENCE: Suppose that independently of your reasoning about p, you reasonably think the following: “were I to reason to the conclusion that p in my present circumstances, there is a significant chance my belief would not be rational.” Then, if you find yourself believing p on the basis of your reasoning, you should significantly reduce confidence in that belief.

TRUTH INDEPENDENCE: Suppose that independently of your reasoning about p, you reasonably think the following: “were I to reason to the conclusion that p in my present circumstances, there is a significant chance my belief would not be true.” Then, if you find yourself believing in p on the basis of your reasoning, you should significantly reduce confidence in that belief.

Schoenfield then presents a practical case COMMUNITY in which growing up in a religious community leads one to rationally believe in the existence of God. Had this individual grown up in a different place, she would have come to believe otherwise given the same body of evidence. Taking this to be a permissive case, Schoenfield argues that one should not decrease confidence in one’s belief.

P3. The best motivation for reducing confidence in permissive irrelevant influence cases requires TRUTH INDEPENDENCE.

P4. TRUTH INDEPENDENCE says to decrease confidence in all permissive cases (even when there are no irrelevant influences!).

P5. If you have to give up your belief in all permissive cases, there are no permissive cases, (definition of permissivism)

P6. TRUTH INDEPENDENCE is inconsistent with permissivism. (P4, P5)

C2. If permissivism is true, the view that you should reduce confidence in permissive irrelevant influence cases is unmotivated. (P3, P6)

P3 is true because if RATIONAL INDEPENDENCE is true, it is permissible to maintain belief, whereas if TRUTH INDEPENDENCE is true, you must give up your belief. Thus the worry about irrelevant influences, such as in COMMUNITY, only arises with TRUTH INDEPENDENCE. If Schoenfield can show that TRUTH INDEPENDENCE is inconsistent with permissivism, we need not worry about irrelevant influences in permissive cases.

TRUTH INDEPENDENCE is inconsistent with permissivism because it requires one to give up belief in permissive cases all the time. Consider the following example without irrelevant influences:

A caveman considers the arguments for and against the existence of God. He comes to believe in the existence of God (G), but also recognizes one could rationally reject G given the same evidence. TRUTH INDEPENDENCE will require him to give up his belief, as he must consider the likelihood of being right independent of his reasoning about the existence of God. Doing so leads to the conclusion that he is not likely to be right. So, if we accept TRUTH INDEPENDENCE, there can be no permissive cases, because each time we are forced to reject our belief absent independent justification. Since permissivism says there can be such cases, TRUTH INDEPENDENCE is inconsistent with permissivism. “What TRUTH INDEPENDENCE demands is exactly what the permissivist cannot provide: an independent reason for thinking it likely that her beliefs, in permissive cases, are true” (208).

  1. A problem with permissivism

Schoenfield notes that it is odd to argue on the one hand that a permissivist should not be willing to take a belief pill, but if one does take that pill (or grow up in a certain community), there is no reason to abandon that belief. The idea that we need not reduce confidence in permissive cases seems to conflict with REFLECTION.

REFLECTION: If you know that, in the future, you will rationally, without loss of information, have doxastic attitude A towards p, you ought to now have doxastic attitude A towards p.

Schoenfield gets around this issue by revising REFLECTION.

PERMISSIVE REFLECTION: If you know that, in the future, you will rationally, without loss of information, rationally have doxastic attitude A towards p, and your future self has the same standards of reasoning as your current self, you ought to now have doxastic attitude A towards p.

This move allows the permissivist to deny believing p or ~p, because the future self may have epistemic standards that conflict with the current self.

  1. Disagreement

Schoenfield’s permissivism can shed light on the debate surrounding peer disagreement. Views that argue a decrease in confidence in light of peer disagreement are motivated by the same independent reasoning principle that motivates a decrease in confidence for IF cases. Independently of one’s reasoning, one cannot assign a high probability of being right. This can lead to the spineless accusation because we would too often have to reject our beliefs. But according to permissivism, this is not enough to decrease confidence (you should instead decrease confidence in light of a high probability of being irrational). Thus decreasing confidence in IF cases hinges on whether or not the case is permissive. So, there will be some scenarios where we should decrease confidence and some where we should not.

5 thoughts on “Miriam Schoenfield’s “Permission to Believe: Why Permissivism Is True and What It Tells Us About Irrelevant Influences on Belief”

  1. Peter Huggins

    Phin, thanks for the great summary! I thought that the article was really interesting, but some of the ideas give me pause. This article seems very applicable in our current environment.

    Like Rebecca, I’m interested in what causes a case to qualify as permissive. However, I have a different question for you.

    My question is this: very similar to stances or values, Schoenfield talks about epistemic standards, and she defines them in terms of truth conduciveness. This seems rather odd when you think about the fact that schoenfield rejects Truth Independence (TI). How can ones standards need to center around being true if doubting our belief is likely true need not concern us?

  2. Nam Nguyen

    Thanks so much for the summary. Like many of the students here, I also found the article and the ideas presented to be extremely interesting despite having many reservations about those ideas.

    I think part of my reservations come from a disconnect with the internal intuitive and external theoretical motivations for permissivism. Though I understand that permissivism is not meant to be applied to every belief and any proposition, it seems that cases where Permissivism would be allowed are incredibly rare. That is to say cases wherein permissivism would make sense are far and away the exception and not the rule. In cases of a jury for example, there is an objective fact to be known regarding whether or not a crime was committed. If presented with evidence where it would be rational to cast a vote for either conviction or exoneration, then it seems to me that the evidence itself is not complete. If the evidence was complete, it should reflect the state of affairs during the time of incident which would give the jurors a vote that could never be wrong. By Schoenfields’ light, the jurors in a case where the evidence was incomplete and indecisive would vote based on their epistemic standards. Like with previous discussions, it seems to me that these standards are themselves debatable. It should, in my understanding, be entirely possible for there to be very few/ only one “permissible” standard to be held. From my perspective, the entire concept of law rests on this idea. If it were possible to cast a vote as the jury in a case where the evidence was inconclusive for either side, it would be rational for multiple cases of the same incident to be judged entirely differently. Someone committing a crime in the same circumstances could, in Schoenfields’ view, rationally and fairly get freedom. At the same time, another person with that exact same circumstance could rationally and fairly get convicted. This seems to violate the basic conceptions of fairness and accountability that we associate with any justice system. This intuitive example was something I held intense reservations about.

    I was also a bit confused regarding truth independence and rational independence. These two strike me as epistemic standards. If that were the case and Schoenfields’ argument regarding independent argument between epistemic standards to be impossible, then it seems to me that one could never move between a preference for truth independence to rational independence. If this is the case, then anyone who initially has a reservation with Schoenfields’ perspective could not help to disagree. At the same time, it would seem then that Schoenfields could never be convincing and there can be no logical pull towards rational independence. What seems to result is a sad state of affairs where there are a variety of unfalsifiable, which seems to be inherently suspect (though I suppose this can be regarded as an epistemic standard).

  3. Rebecca Amen

    Thank you for the summary, Phin! I found this article to be extremely interesting.

    I’m interested in what attributes of a case, exactly, cause it to qualify as permissive. It seems that this is key to the argument and to a lot of questions that arise while examining the argument. My issue with the claims that permissiveness is true and that truth independence and permissiveness are incompatible, making truth independence false, is that they both seem to depend on what kind of case we are discussing. But the entire argument hinges on the truth of these claims, making the argument have a conditional sort of truth. With the example of opinions about the dinosaur extinction, it is quite clear that peripheral, “irrelevant” influences would not taint the reasoning of a rational, well-versed paleontologist. But epistemic standards or inclination to believe or integrate new evidence may differ among paleontologists, such that two experts would come to different—and perhaps, equally rational, given the limited and interpretable evidence available and their, as far as they can see, equally rational standards—conclusions. So, would this case of the paleontologist really qualify as a permissive or non-permissive case? It depends on how whole/reliable/objective/solid you think the body of knowledge and methodology relating to paleontology is. It seems that sciences concerned with less interpretable evidence, or “harder” evidence, would be more disposed to produce non-permissive cases, or cases in which Schoenfield’s argument would fundamentally not apply because some principle of uniqueness would be in operation. Conversely, cases where there is little to no incontrovertible evidence (such as those that debate the existence of God) seem to epitomize what Schoenfield is discussing and invite the application of her argument. In those cases, I am in agreement with Schoenfield that irrelevant influences are not cause for concern because either conclusion is not, in terms of pure reason, preferable to the other; it is a case in which you can choose to ignore those influences because fundamentally, given the limits of human knowledge, neither conclusion can be proven true. I can see how truth independence would be fully incompatible with permissivism in these cases because truth independence seems to be diametrically opposed to permissivism, and permissivism does seem to do the work to explain why opposing beliefs can be equally rational. In the case of an argument for the existence of God, say, I would not be inclined to say that either side of the debate is irrational; there is simply no uninterpretable evidence to substantiate the claims of either side. But again, that doesn’t mean that truth independence is false; it just means that it is not applicable in permissive cases. That leads to my question: How do we decide just which cases are permissive and which are not? This seems to be quite an important question because this is how we make a judgment about rationality. Not asking for a clean line or some solid set of criteria here, as of course, there wouldn’t be any. But on a case-by-case basis, how do we go about doing that?

  4. Kenzo Okazaki

    Thanks Phin!

    I suppose I’m still stuck on this rationality issue as it is presented on page 202. Here, Sally supposedly thinks that her standards are more truth conductive than others, yet other standards may be just as rational. It would seem to me that from Sally’s point of view it would be irrational to believe in any standards that are not maximally truth conductive. I think we would all share the intuition that it would be strange to rationally endorse a standard that is not going to get us to truth. If this is the case and Sally also thinks that her standards are most conductive to truth, then it would be odd to say from her perspective that other standards are just as rational. This seems to point to a problem that I had from last week’s class about what the idea of collaborative epistemology looks like. It seems that we still end up having the same problems with rationality and that they are just pushed back a little.

    On another note in response to one of your questions Phin, I was not as convinced by the intuitive arguments for permissivism. It seems to me that if available evidence encapsulates everything (as illustrated by the paleontology example) you would end up with the same person. That is, if someone were to have all of the same experiences as me they would probably end up agreeing with everything I say. This is especially true with a determinist view of the issue. As for the paleontology example in full, I do not feel that the analysis is compelling. Perhaps the experiences that Schoenfield points out (the far field ones) are not relevant, but that does not deny that a palentologist’s view of the dinosaur extinction would be very similar to that of another if they had consulted all of the same kind of evidence. It seems to assume that the palentologist would have studied all the relevant material. In my mind, it is not enough that books and articles are “the kind of evidence that can be easily shared.”

  5. Farhan Islam

    Great write-up, Phin!
    I thought that this paper was really interesting but couldn’t help but feel like I was having the rug pulled from under me. The meat of the argument seems to proceed as follow.

    1. Permissivism is true
    2. Truth Independence (TI) runs counter to permissivism
    3. Therefore, TI is false
    4. Therefore, permissive IF cases are not a cause of belief revision

    This argument patterns is rather fishy to me because it doesn’t give me a particularly good reason to belief that TI is false and rather makes me think that the intuitiveness of TI suggests that permissivism is false.

    As I see it, TI is the intuition that if we think that our belief is likely to be false, we should believe it less. This supposedly conflicts with permissivism because if we can accept that it is rational to believe two opposing propositions, we must think that there is a significant likelihood that whichever one we believe is false and therefore we should not have high confidence in it. Doesn’t this view presuppose non-permissivism? This seems to assume that what “rational” means is that our credence in P is the same as the probability of P. But that would assume that there is some such single rational probability of P.

    I would have expected a permissivist to tell this story as:
    “It is rational to believe 100% that P is true even if you think that it could be rational for, by some other set of standards, someone to believe that ~P is 100% true.”
    If this is what permissivism is, then I don’t see the contention with TI. Perhaps my confusion comes from not being able to grasp how one could reason about the likelihood about their belief in P being true independent of their reasoning about whether or not P.

    On top of this, I am a bit puzzled by the weight of this conclusion. If the final conclusion is that “If you think that it is rational to believe P or ~P and find out that your reason for picking P instead of it’s negation is something that is not the truth of P, then you don’t need to revise your position” that seems like it needs a clearer story about what makes P permissive. What I mean by this is that IF cases only seems to be of great concern in non-permissive cases. Take COMMUNITY. Presumably the theist does not view G as a permissive case unless they have already come to terms with how and why their community shaped their belief.
    Perhaps this means that for someone to view a case as permissive is for them to believe that it there are irrelevant factors at play that can still be part of rationality. This seems like the right way of thinking about it because any permissive IF case I can think of is one where the irrelevant factors are the factors that determine one’s epistemic standards, which presumably we already take to be independent of the truth of P.

Leave a Reply