Against Evidential Minimalism

Abstract

Evidence is often taken to be “normative” for doxastic agents. What accounts for the normativity of evidence? According to the view that I'll call “evidential minimalism”, there is a close connection between strong evidence for the truth of p and a normative reason to believe p: evidence is either itself a normative reason for belief, or evidence gives rise to such a reason when certain other minimal conditions are met. In this paper, I argue against evidential minimalism. I will argue that there are cases where: (i) an individual S possesses strong evidence E for the truth of p at time t, (ii) all other minimal conditions for the normativity of E are met at t, (iii) S doesn't believe p at t, yet (iv) S isn't open to any form of criticism on account of (i)–(iii) at t. I will then formulate a plausible linking claim connecting openness to criticism and the existence of normative reasons for belief. The minimalist can either accept or reject this linking claim. I will argue that, either way, the minimalist view falters.

1. Introduction

Strong evidence for the truth of p is often taken to give rise to an epistemic reason to believe p. Evidence is thus commonly taken to be “normative” for doxastic agents; it is the kind of thing that we ought to be sensitive to when forming and maintaining attitudes such as belief, disbelief, and suspension of judgment.¹ But what accounts for the normativity of evidence? Is it the case that evidence itself is normative for belief? Or is the normativity of evidence accounted for in terms of some extra-evidential consideration(s)?

In the following paper, I will consider, and reject, the view that says evidence itself (perhaps when coupled with certain other “minimal” conditions) is normative for belief. I will call this view “evidential minimalism”. I will take the minimalist as offering us an answer to the following question: Under what conditions does strong evidence for the truth of p generate an epistemic reason for S to believe p? According to the minimalist's answer, there are either very few or even no conditions that need to be met, over and above the obtaining of strong evidence for the truth of p, in order for there to be an epistemic reason for S to believe p.²

For instance, according to one way of being a minimalist, the obtaining of strong evidence is itself enough to generate an epistemic reason for S to believe p. I'll call this version of minimalism “objectivism” since it doesn't require that S possess or otherwise attend to the evidence in order for it to give rise to an epistemic reason for S to believe p. On another way of being a minimalist, there are certain minimal conditions that need to be met – over and above the obtaining of strong evidence for the truth of p – in order for there to be an epistemic reason for S to believe p. For instance, perhaps S has to possess the evidence, or possess the evidence and consider the question of whether p. Why think of these latter conditions as “minimal”? The idea is that “possessing” evidence and “considering whether p” are conditions that can be met rather easily; once strong evidence for p is already on the table, a normative reason to believe p isn't too far off, so to speak. Any normal epistemic agent possesses all kinds of excellent evidence at any given time, and questions of the form “whether p” can be thrust upon our conscious awareness from without (e.g. by an annoying friend or co-worker).

Consider, for instance, the view put forward in Hofmann (2021: 667) according to which:

neither a bit of evidence nor the fact that it is evidence for a certain proposition is a normative fact, but it is still the case that evidence provides normative reason for belief. In this latter sense, then, evidence is normative.

Hofmann argues that, in the prospective case (where a subject S possesses strong evidence for p but doesn't yet believe p), certain conditions have to be met in order for there to be a normative claim upon S to believe p. For instance, it has to be the case that “the question whether p is activated for S and S has all the abilities needed for following [the evidence]”.³ Similarly, Kiesewetter (2017) argues that “If A has sufficient evidence for p, and A attends to p, then A is rationally required to believe p”.⁴ I will count these further conditions – possessing the evidence, “activating” the question of whether p, paying attention to p, etc. – as “minimal”. The rough idea is that evidence is still shouldering most of the normative weight. Even though there are certain conditions that have to be met, over and above the obtaining of the evidence, in order for there to be a normative reason for S to believe p, these conditions don't add too much to the picture. In particular, they don't seem to add anything which would threaten to usurp the central role that evidence is playing in explaining why S has a normative reason to believe p.⁵

The minimalist view thus comes in different shapes and sizes. While the minimalist/non-minimalist distinction will perhaps not always be clear-cut, we can provide a preliminary disjunctive characterization of the minimalist position.⁶ Evidential minimalists subscribe to either one of the following:

Strong evidence for the truth of p is itself an epistemic reason for S to believe p.

Or

Strong evidence for the truth of p gives rise to an epistemic reason for S to believe p when certain other minimal conditions are met (e.g. S possesses the evidence and considers whether p).

A non-minimalist about evidential normativity, by contrast, would hold that something more substantial needs to be on the table, over and above the obtaining of strong evidence for the truth of p and certain other minimal conditions, in order for there to be an epistemic reason for S to believe p. The recent literature on epistemic normativity provides a number of non-minimalist possibilities. For instance, Steglich-Petersen (2018) argues that evidence for p constitutes a normative reason for S to believe p only when S has some further (e.g. moral or prudential) reason to form true beliefs about p. Thus, the evidence by itself (even when coupled with certain minimal conditions) isn't enough to underwrite a normative reason for belief.⁷ Such a view is also implicit in Papineau (2013: 69) where it is argued that epistemic evaluations carry “no prescriptive force on [their] own, independent of some further value attaching to the aim of truth.” Similarly, Maguire and Woods (2020) argue that evidential considerations are not “genuinely” or “authoritatively” normative. According to Maguire and Woods (2020: 229), “all and only the practical reasons are the authoritative reasons”. Non-minimalism is thus well represented in the contemporary literature.⁸ These views are unified as “non-minimalist” insofar as they introduce conditions which do threaten to usurp the central role that minimalists seek to reserve for evidence when explaining why S has a normative reason to believe p.⁹

Motivation for non-minimalism is sometimes provided by considering the “justificational fecundity” of evidence; the fact that evidence potentially justifies an infinite number of beliefs.¹⁰ Consider, for instance, Whiting (2013: 130):

Suppose that I have evidence that the cakes are burning. Whatever provides this evidence provides evidence for an infinite number of other beliefs, such as that there are cakes, that the cakes are burning or that Tolstoy wrote Great Expectations, that if the cakes are burning then the cakes are burning, that there is more carbon dioxide in the atmosphere than there was before I made the cakes, that there is no dancing monkey singing the national anthem in the space the cakes occupy, and so on without end.

One could take such observations to suggest that, even if one has decisive evidence E for the truth of p, one still does not have a normative reason to believe p. It is at precisely this point, however, where the minimalist view stands to be bolstered by appealing to minimal conditions. For instance, it's less clear that Whiting would lack a normative reason to believe some of the claims in the above passage if, in addition to possessing the relevant evidence, certain questions were to be brought before his mind. One could argue that this so even in those cases where Whiting lacks practical reason to be interested in the relevant claims.

Plausible as this might sound, I will argue that it is mistaken: the minimalist view is false. I will not argue against minimalism by defending the comparative theoretical strengths of some version of non-minimalism. Rather, I will argue directly against the minimalist view itself. If my argument is successful, then some form of non-minimalism will be the way to go when it comes to understanding the normativity of evidence. Here I remain neutral on which form of non-minimalism should be preferred. The bulk of the present paper will be devoted to my argument against minimalism. However, before presenting the argument (section 3), I would like to say a few things regarding evidence and its possession (section 2).

2. Evidence and its possession

There are a number of debates concerning evidence in contemporary epistemology.¹¹ Since it is beyond the scope of the present paper to argue for one particular approach, I will simply flag how I'll be understanding the notion throughout. While I think there is much that could be said in favor of how I'll be understanding evidence, I won't be able to defend my approach here. Note, however, that while the following assumptions about evidence are contentious, the success of my argument against minimalism doesn't essentially depend upon them. In other words, similar arguments could be offered against minimalism that operate with alternative conceptions of evidence. Nevertheless, it will be helpful to introduce the following simplifying assumptions.

First, I will assume that evidence is propositional. This rules out as evidence experiential states that are not propositionally structured, as well as ordinary objects and artefacts, e.g. fossils or murder weapons.¹² What this rules in as evidence are propositional claims like “that the victim's blood is on the knife”. Second, I will assume that evidence is factive. In other words, in order for E to be evidence for the truth of p, E must be true. This rules out false propositions as evidence.¹³ Third, I will assume that our evidence can include true propositions about the external world. The propositional nature of evidence combined with factivity doesn't automatically guarantee this third assumption. For instance, one could hold that evidence consists of all and only true propositions about how things appear to us. I will assume a more expansive view of evidence according to which true propositions about the external world can be a part of one's evidence.¹⁴

I will also focus on evidence E for the truth of p of a certain strength. In order for E to be evidence for the truth of p, E must somehow indicate or make more likely the truth of p. It's possible, however, for some proposition E to only raise the likelihood of p ever so slightly. To simplify matters, I will focus on instances of evidential support that are very strong. Specifically, I will focus on instances where the support relation between E and p is such that one could come to know p on the basis of E.¹⁵ I will not assume that this support relation must amount to entailment; I take it that I can know things (say, on the basis of testimony or memory) without my evidence entailing the truth of the relevant propositions. However, entailing evidence would certainly suffice to qualify as “strong” evidence. Thus, for the purposes of the following discussion, we can say the following:

In order for E to be “strong” evidence for the truth of p:

1. (i) E must meet the constraints on evidence introduced above, i.e. E must be a true proposition (possibly about the external world); and

2. (ii) E must make the truth of p sufficiently likely; it must be possible for one to come to know p on the basis of E.

Focusing matters in this way simplifies the discussion in certain respects. For instance, if it can be shown that the minimalist view falters even when we restrict our attention to evidence that is “strong”, then it doesn't seem that the view holds much promise.

Let's turn now to evidence “possession”. Given the broadly externalist view of evidence presented above, it's possible for evidence to obtain without a person's being in possession of it. Also, recall that, according to one way of being a minimalist, strong evidence E for the truth of p isn't itself an epistemic reason for S to believe p. Rather, certain other “minimal” conditions must also be met, e.g. S must possess the evidence and consider whether p. What is it to “possess” evidence? Let's focus on the following example discussed in Kelly (2007, 2014): Koplik spots are evidence of measles. This is a true claim, and it reflects a discovery of medical science. Say that patient X presents with Koplik spots. Let's also say that the presence of Koplik spots makes the claim that the patient has measles sufficiently likely; one could come to know this claim on the basis of the evidence. We thus have a case of “strong” evidence, with the following evidence E and supported proposition p:

E: Patient X has Koplik spots.

p: Patient X has measles.

When it comes to the question of evidence “possession”, we can distinguish between three different types of cases. In the first type of case, a subject S has no contentful mental state (whether occurrent or non-occurrent) which represents E as true. This is a clear case in which a subject doesn't “possess” the evidence. For instance, perhaps the subject has no such mental state because they haven't been made aware of E; they haven't seen the patient, nobody has told them about the patient, they haven't read about the patient, etc. Next there's the case in which a subject S clearly does possess the evidence. Imagine a doctor who is attending to patient X. Say that the doctor is aware of the connection between Koplik spots and measles, is proficient at identifying the spots, and wants to know about the patient's medical condition. After attending to the patient and noticing the spots, the doctor clearly “possesses” the evidence.

Finally, there's an intermediate case. Imagine that there's a different doctor attending to the same patient. However, this doctor is ignorant of the connection between Koplik spots and measles. Even though this may be the case, the doctor can still attend to the patient's spots. It's not clear whether, in such a case, the doctor “possesses” evidence that the patient has measles. Here we can simply say that there's a very weak sense in which this doctor “possesses” the relevant evidence. The second type of case mentioned above can thus be considered evidence possession in a more robust sense; “robust” possession for short.

We can now give a rough gloss on evidence “possession”. The kind of possession that is relevant for the minimalist is evidence possession in the robust sense. We can thus say the following:

A subject S possesses strong evidence E for the truth of p in the “robust” sense at time t iff (i) S has, at time t, some contentful mental state (whether occurrent or non-occurrent) which represents E as true, and (ii) S is aware of E's evidential import vis-à-vis p at t.¹⁶

A few clarificatory remarks: In the example of robust possession above, the doctor had an occurrent mental state at time t, and was also, at time t, explicitly and consciously appreciating E's evidential import vis-à-vis p. While this will suffice for “robust” possession, it's not necessary. For instance, the mental state which represents E as true doesn't have to be occurrent. An individual could, for instance, simply know that E obtains; something which could be committed to memory and then “called up” if the individual were to consider the matter. Also, “awareness” of E's evidential import vis-à-vis p is an ability or capacity had by S; something which S can have, at some time t, even if she's not, at that time, explicitly or consciously appreciating E's evidential import vis-à-vis p. Thus, an individual S can, at time t, possess (in the “robust” sense) strong evidence E for the truth of p without attending to E at t and without believing p at t. From here on out when I speak of evidence “possession” I will have in mind possession in the robust sense.

3. An argument against minimalism

My argument against minimalism will occur in two stages. In the first step (section 3.1) I will establish the following claim:

There are cases where: (i) a subject S possess strong evidence E for the truth of p at time t, (ii) all other minimal conditions for the normativity of E are met at t, (iii) S doesn't believe p at t, yet (iv) S isn't open to any form of criticism on account of (i)–(iii) at time t.

The second step of the argument (section 3.2) will involve the following linking claim connecting openness to criticism with epistemic reasons for belief:

If [there are cases where: (i) a subject S possess strong evidence E for the truth of p at time t, (ii) all other minimal conditions for the normativity of E are met at t, (iii) S doesn't believe p at t, yet (iv) S isn't open to any form of criticism on account of (i)–(iii) at time t] then, in such cases, it's not the case that E is (or gives rise to) an epistemic reason for S to believe p at t.

I think there is some plausibility to this linking claim, and I will say more about that below. However, my argument won't depend upon the truth of this claim. Once the claim in step 1 is established, the minimalist is faced with the following choice: She can either accept or reject the linking claim. I will argue that, either way, the minimalist view falters.

Before proceeding to carry out the first step of the argument, I'd like to say something about condition (iv) “S isn't open to any form of criticism on account of (i)–(iii) at time t”. Specifically, I'd like to clarify what I mean by “criticism”. When it comes to “criticism” I mean to be somewhat inclusive. Thus, I do not restrict myself to blame. Some authors hold that there is a distinctly epistemic form of blame.¹⁷ If such a thing exists, then I would include it as a form of criticism. However, I don't mean to restrict myself to blame. Criticism might involve aretaic assessment of the person which falls short of blame. Also, there are ways of holding a person accountable (e.g. via sanctions of various kinds) which do not imply that the person manifests a character defect. What I do mean to rule out is something that we might call “mere appraisal”; simply assessing someone's thought or conduct vis-à-vis a standard where this has no real weight or significance in our actual practices. The thought is that, in order to count as “criticism”, the form of response at issue must somehow go beyond mere assessment vis-à-vis a standard. For instance, blaming someone is sometimes thought to involve certain reactive emotions or modification of various attitudes and expectations vis-à-vis the person,¹⁸ sanctioning someone might involve imposing certain punitive measures, etc. The question of whether there is a distinctly epistemic form of response that counts as “criticism” in this sense is an interesting one, and I will return to it briefly below. For now, I'll simply note that “criticism” must go beyond mere assessment in some way.

3.1. An argument against minimalism: step 1

The first step of my argument against minimalism will involve establishing that there are cases where the above conditions (i)–(iv) are met. Keeping in mind the account of evidence “possession” presented earlier, consider the following case:

BORED AT HOME: Katlyn is at home with nothing to do. She recently left her job in order to take up a better position at a new company. There's a one month interval in between her leaving the old job and beginning training for the new job. Katlyn doesn't have to move for her new job, and she's tied up all manner of loose ends, taken care of various errands and housekeeping, etc. She finds herself at home one Sunday afternoon and is simply bored with nothing to do.

As a typical epistemic agent, Katlyn finds herself in this situation with a large stock of evidence upon which she could draw in order to arrive at new knowledge. For instance, on this particular Sunday, Katlyn could sit down and start drawing out consequences from things that she already knows, e.g. she could try to figure out how many hours she's been alive, or how many haircuts she's received in her life so far. Also, while her house will certainly not be as lively as, say, a busy city street, there will still be quite a bit of perceptual information that she could attend to in order to acquire new knowledge, e.g. she could figure out exactly how many tiles there are on her bathroom floor. However, Katlyn of course does not do this. Nor do we expect her to. We wouldn't think any less of Katlyn for not doing this. In fact, we'd probably be somewhat taken aback if she did start doing this.

Imagine one of the questions I mentioned above: How many hours has Katlyn been alive? Say that Katlyn knows her own age (32), and she's quite talented at quick mental math. She could easily figure out that she's been alive for more than 200,000 hours, were she to sit down and think about it for a minute. Thus, at any arbitrary time t when Katlyn is at home, she will possess, in the “robust” sense, strong evidence E for truth of the following claim: that she has been alive for more than 200,000 hours. However, Katlyn is of course in no way criticizable for not believing this claim. At any arbitrary time t when Katlyn is bored at home not thinking about the question of how many hours she's been alive, she manifests no fault and opens herself up to no form of blame, sanction, or criticism for not believing this claim. Examples like this are easy to generate. Here's another thing that Katlyn knows: that her name is Katlyn. Katlyn also remembers the disjunction introduction rule from her symbolic logic class in college. Thus, at any arbitrary time t, Katlyn will possess, in the “robust” sense, strong evidence for the following disjunctive claim (DC):

DC: Either her name is Katlyn or aliens exist.

As well as:

DC*: Either her name is Katlyn or aliens exist or the moon is made of cheese.

And:

DC**: Either her name is Katlyn or aliens exist or the moon is made of cheese or Santa Claus is real.

And so on. These propositions are, quite clearly, a bunch of disjunctive junk. Once again, Katlyn manifests no fault and opens herself up to no form of blame, sanction, or criticism, for failing to believe them. Consider one further case:

LAST DONUT: Someone at work ate the last donut. The box of donuts was in the lunchroom. David knows an assortment of information which is such that, were he to reason through it properly, he could determine who ate the last donut. For instance, he knows that Sarah always eats lunch in the lunchroom, but only between 12 and 12:45, that the last donut was present until approximately 1:15, that Sue was off that particular day, that Jamie doesn't like donuts, etc. Say that Carl ate the last donut. Moreover, Carl wasn't violating office protocol in doing so. Actually, David himself brought the donuts and placed a note on the box that said “please eat me”.

Say that, at time t, David knows everything he needs to know in order to figure out who ate the last donut and that figuring this out would take some concentrated effort but wouldn't be exceedingly difficult; David is fully capable of figuring it out. However, David takes no interest in this; he wanted the donuts to be eaten. Similar to Katlyn, David is in no way criticizable, at t, for not believing that Carl ate the last donut; he manifests no fault and opens himself up to no form of blame, sanction, or criticism for not believing this claim.

The cases of BORED AT HOME and LAST DONUT perhaps suffice to establish the following claim:

There are cases where: (i) a subject S possesses strong evidence E for the truth of p at time t, (ii) S doesn't believe p at t, yet (iii) S isn't open to any form of criticism on account of (i)–(ii) at time t.

However, this was not the claim we were to establish. That claim was as follows:

There are cases where: (i) a subject S possess strong evidence E for the truth of p at time t, (ii) all other minimal conditions for the normativity of E are met at t, (iii) S doesn't believe p at t, yet (iv) S isn't open to any form of criticism on account of (i)–(iii) at time t.

A minimalist could happily grant that BORED AT HOME and LAST DONUT suffice to establish the first claim but not the second. This minimalist could argue that there are further minimal conditions for the normativity of evidence that aren't yet met in the relevant cases. Let us, then, introduce further minimal conditions for the normativity of evidence. As I mentioned earlier, these will include things like being prompted with questions of the form “whether p” and attending to the relevant evidence. It is important to note that there will be certain non-minimal conditions that we will want to exclude. Imagine, for instance, BORED AT HOME*. In BORED AT HOME* everything is the same except that Katlyn becomes genuinely interested in the question of how many hours she's been alive. The question is before her mind and she wants to figure out an answer to it. She then “activates” her mathematical abilities, engages in a process of reasoning, and arrives at an answer to the question. Katlyn is thus engaged in a certain goal-directed or purposive activity; she's engaged in active inquiry when it comes to a certain question or subject matter. We can also imagine David doing so in similarly adjusted LAST DONUT*. Engagement in such an activity would be a non-minimal condition; it is precisely the kind of thing that threatens to usurp the central role that minimalists wish to reserve for evidence in explaining why S has an epistemic reason to believe p. Consider, for instance, the following example adapted from a case discussed in Kelly (2003):

MOVIE SPOILER: Liz often sees newly released movies only after they've been in theaters for some time. Prior to her seeing some particular movie, she has no interest in forming true beliefs about its ending. In fact, she actively does not want to take on such beliefs. Liz is chatting with a few friends before class. One of them quickly changes topic and begins talking about a movie that Liz hasn't yet seen. Before Liz has a chance to warn her not to spoil the ending the friend blurts out the ending and Liz comes to form a true belief about its ending.

In his (2003), Kelly argues that our epistemic reasons for belief do not depend upon our adoption of certain aims or goals, including the goals that we take on when engaged in active inquiry vis-à-vis some topic or subject matter. Regarding the movie example, Kelly notes that, when it comes to Liz's epistemic reasons for belief, there would be no difference between MOVIE SPOILER and MOVIE INFORMER; in the latter case everything is the same except now Liz wants to know about the ending of the movie and she asks her friend to divulge this information. According to Kelly, the goals and desires had by Liz in MOVIE INFORMER are extraneous and irrelevant when it comes to explaining what her epistemic reason for belief are.

Thus, we shouldn't count something like “S is engaged in active inquiry vis-à-vis whether p” as a minimal condition for the normativity of evidence; this would threaten to obscure the central role that evidence is playing in the minimalist's explanation as to why S has an epistemic reason to believe p. Somewhat ironically, then, the normativity of evidence seems to be most clearly on display in cases in which a subject is not engaged in conscious deliberation or active inquiry vis-à-vis some subject matter or question. But what, then, would count as a minimal condition for the normativity of evidence? As I've already mentioned, simply prompting S with a question of the form “whether p” will perhaps be enough, as will getting S to attend to the relevant evidence.

Let's return again to BORED AT HOME and LAST DONUT. Say that Katlyn has a friend over and her friend turns to her and says “Hey, you're 32 and good with math, how many hours have you been alive?”. Imagine that, in response, Katlyn simply bats the question away, refusing to “activate” her mathematical abilities because the question is silly and pointless. She thus doesn't go through the requisite process of reasoning in order to arrive at an answer (one that she easily could go through). Once again, it doesn't seem that Katlyn is open to criticism on account of this. She manifests no fault and opens herself up to no form of blame, sanction, or criticism for responding in the way that she did. Similarly, in LAST DONUT, a coworker might approach David and ask “Who ate the last donut?”. Even though David is in possession of everything he needs in order to answer this question, and could answer the question after giving it a little thought, he is in no way criticizable for simply answering “I don't know”.

Thus, even if certain minimal conditions for the normativity of evidence are met, a subject can still fail to be appropriately subject to criticism on account of not believing the target proposition. I think this lends strong support for our claim that there are cases where conditions (i)–(iv) are met. However, one might remain unconvinced. One could argue that, in the modified version of BORED AT HOME (where Katlyn's friend asks her how many hours she's been alive), it is not the case that all of the minimal conditions for the normativity of evidence are met.¹⁹ For instance, let's say that Katlyn knows the following conjunctive proposition E:

E: She is 32 years old & there are 24 hours in a day & there are 365 days in a year.

Say that, at time t, Katlyn is attending to E and she is asked “How many hours have you been alive?”. Perhaps, at time t, there are certain minimal conditions for the normativity of E that remain unmet. For instance, perhaps Katlyn has to perform the relevant calculations and thereby come to consciously and vividly “see” the evidential connection that obtains between E and the target proposition. Consider the following analogy: Say that someone has to look around the corner in order to acquire some evidence E bearing on the question of whether p. Prior to looking around the corner, perhaps the evidence doesn't give rise to a normative claim regarding what the person should believe. One could argue that Katlyn's case is similar. In Katlyn's case, the relevant “act” is, of course, mental rather than physical. However, perhaps a similar point holds: prior to engaging in this act, the evidence E generates no normative claim concerning what Katlyn should believe.

The problem with this proposal is that there are cases where conditions (i)–(iv) are met and S is “consciously and vividly” appreciating the evidential connection that obtains between E and the target proposition at t.²⁰ Return again to the disjunctive claims we saw earlier in the case of Katlyn in BORED AT HOME. Say that Katlyn consciously acknowledges that she knows her name is Katlyn at time t. Say that she also consciously and vividly acknowledges that the following disjunctive claim (DC) follows from this known fact:

DC: Either her name is Katlyn or aliens exist.

If the question of “Whether or not DC?” is somehow brought to her attention, would Katlyn then be appropriately subjected to criticism were she to not believe DC? It's hard to imagine Katlyn being more “in touch” with the relevant evidence. Perhaps we might expect Katlyn to assent to DC in such a circumstance. However, if Katlyn could somehow avoid believing DC even after assenting to it (maybe immediately after assenting to DC, the claim simply evaporates from her conscious awareness), would she thereby be open to some form of blame, sanction, or criticism? Paakkunainen (2018: 135–6) suggests that she would be:

there is a clear sense in which, if S has excellent, undefeated epistemic reasons to believe that p, then S ought to believe that p: he epistemically ought to believe that p. At least, this is so if S also considers the reasons, q, and the question whether p … If S considers the question whether there is an even number of dust specks on his desk, and considers the excellent evidence for answering “yes,” then he epistemically ought to believe that there is an even number of dust specks on his desk. Likewise, we (well, I) would fault him for failing to form that belief if he considered the question and his evidence was clear. He's not merely lacking a psychological compulsion if he fails to form the belief in this instance; he's being a less-than-excellent epistemic agent.

If we apply Paakkunainen's considerations to the case of Katlyn and DC do we get the same result? In other words, would Katlyn be a “less-than-excellent epistemic agent” and an appropriate target of criticism is she were to fail to believe DC? My own reaction is that she would not. If Katlyn were to fail to believe DC she would not be appropriately subjected to any form of criticism. Nor would she be if she failed to believe DC* in similar circumstances:

DC*: Either her name is Katlyn or aliens exist or the moon is made of cheese.

Or DC**:

DC**: Either her name is Katlyn or aliens exist or the moon is made of cheese or Santa Claus is real.

Let us, however, explore this a little further. In order to do so, I suggest we consider a few recent attempts to pinpoint a distinctly epistemic form of assessment that counts as “criticism”. If there is such a form of assessment – and I'm inclined to think that there is²¹ – then we can ask about the conditions under which it is appropriately taken up towards a person. If Katlyn is appropriately subjected to some form of criticism on account of her failure to believe the disjunctive claims (as Paakkunainen's remarks suggest), then it seems that a distinctly epistemic form of criticism would perhaps be our best candidate. The recent literature on this issue is often framed in terms of the concept of blame ²². I will (somewhat hesitantly²³) follow suit.

Boult (2021a) identifies four recent proposals when it comes to the nature of epistemic blame. First, there is the emotion-based view.²⁴ According to this view, “epistemic blame is the manifestation of reactive attitudes such as indignation and resentment, directed towards a target as a result of the judgment that the target has (culpably) violated some epistemic norm” (Boult (2021a: 5). If epistemic blame is understood in this way, then Katlyn is clearly not appropriately subjected to it; she is not the appropriate target of resentment or indignation on account of her failing to believe pointless disjunctive claims. Responding to her in this way would be wildly overblown and uncalled for.

Second, there's the desire-based view.²⁵ According to the desire-based view, epistemic blame consists in a characteristic set of dispositions (e.g. to reproach, feel upset, and to verbally request reasons) associated with a certain belief-desire pair. The relevant belief-desire pair concerns “believing badly”, i.e. the belief-desire pair concerns “those whose beliefs violate epistemic norms without excuse” (Brown 2020: 399, emphasis added). On this account, epistemic blame is what occurs when a person's desire that someone else not culpably violate an epistemic norm (e.g. by believing badly) is frustrated. If we understand epistemic blame in this way, then Katlyn is not appropriately subject to it. First, Katlyn doesn't believe badly; her (alleged) epistemic infraction is that she failed to form certain beliefs. Second, people do not harbor the kind of desire that would be needed in order to make epistemic blame appropriate in the case of Katlyn. Our desires aren't frustrated when individuals fail to believe pointless and silly disjunctive claims.²⁶

Third, there's Boult's own relationship-based view of epistemic blame (Boult 2020, 2021c). According to Boult's view, epistemic blame consists in a kind of relationship-modification. ²⁷ Boult (2021a: 6) argues that, as members of epistemic communities, we stand in “epistemic relationships” with one another. When a person A epistemically blames B, A judges that B “has done something that falls short of the “normative ideal” of this relationship” and A modifies her attitudes and expectations vis-à-vis B accordingly. For instance, imagine that A finds out that B is dogmatic and biased when it comes to a certain topic or issue. In response, A might cease to trust B's words when it comes to this topic. If we understand epistemic blame in this way, then Katlyn is, once again, not appropriately subjected to this form of response. In not believing pointless disjunctive claims, Katlyn is not doing something which compromises her epistemic relationship with other people. Matters would perhaps be different if Katlyn had a general tendency to not believe the claims that are well supported by her evidence (say, out of dogmatism or wishful thinking). However, this is not what's going on in her case; Katlyn is simply avoiding taking on unnecessary and pointless beliefs. Note that her failure to form the relevant beliefs in this case might still express an underlying tendency or disposition, viz. to avoid believing pointless and trivial claims. However, as long as this tendency is only restricted to those truths that are pointless and trivial, then Katlyn is not open to any form of criticism on account of her being disposed in this way.²⁸, ²⁹

Finally, there's what Boult calls the agency-cultivation view of epistemic blame.³⁰ This is a broadly “forward-looking” account of epistemic blame which says that blame functions as a “vector” for agency cultivation. According to this view, blame-responses function so as to discourage certain forms of behavior. By internalizing the expectations and demands implicit in the blame-responses of members of our communities, we become responsive to certain salient features of our situations. In the epistemic case, the relevant blame-responses discourage certain forms of epistemic behavior and this, in turn, cultivates epistemic agency. If we understand epistemic blame in this way, Katlyn will, once again, not be appropriately subjected to this form of response. A “well-cultivated” epistemic agent is not one who will draw out the consequences of any piece of strong evidence she just so happens to possess, no matter how trivial or uninteresting. Katlyn can perform as she did while also being a fully competent and mature epistemic agent.³¹

Thus, I conclude that Katlyn is not appropriately subject to any form of criticism – even a distinctly epistemic form of criticism – on account of her not believing certain disjunctive claims. This is so even though Katlyn is being prompted with a certain question while also possessing (and consciously and vividly appreciating the evidential import of) strong evidence E for the truth of p.³² This concludes the first step of my argument. We have now established the following claim:

There are cases where: (i) a subject S possess strong evidence E for the truth of p at time t, (ii) all other minimal conditions for the normativity of E are met at t, (iii) S doesn't believe p at t, yet (iv) S isn't open to any form of criticism on account of (i)–(iii) at time t.

3.2. An argument against minimalism: step 2

The second step of my argument will involve the following claim which links openness to criticism with epistemic reasons for belief:

If [there are cases where: (i) a subject S possess strong evidence E for the truth of p at time t, (ii) all other minimal conditions for the normativity of E are met at t, (iii) S doesn't believe p at t, yet (iv) S isn't open to any form of criticism on account of (i)–(iii) at time t] then, in such cases, it's not the case that E is (or gives rise to) an epistemic reason for S to believe p at t.

While I think there is some plausibility to this linking claim, step 2 of the argument will not depend upon its truth.³³ Rather, I will argue that the minimalist can either accept or reject the linking claim. Either way, the minimalist view falters.

But why think there's any plausibility behind this linking claim in the first place? The driving idea behind this linking claim is that normative reasons for belief should somehow show up in our practices of interpersonal criticism and assessment. Their “normativity” should be revealed in the ways that we respond to individuals who either conform or fail to conform to them. If there is no trace of such reasons in our actual practices, then by what right can we call them “normative”? Consider, for instance, Schmidt (2021: 4–5) who makes a similar point about moral normativity:

the significance of a moral requirement will make it often – in absence of an excuse or exemption – appropriate to show resentment or indignation. These emotions are expressions of the normative significance we attach to the moral requirement because they are appropriate in face of its violation.

Similarly, if evidence is normative (as the minimalist contends), then we should expect this to be somehow expressed within our interpersonal practices of criticism, blame, and accountability. When it comes to the epistemic (as opposed to the moral), such expressions don't necessarily have to involve certain emotional responses such as resentment and indignation.³⁴ Nevertheless, if evidence constitutes genuinely normative reasons for belief, then we should expect there to be some kind of expression of its normative significance within our practices. This is what's motivating the attempt to link openness to criticism with epistemic reasons for belief.

Thus, the above linking claim perhaps has some initial plausibility. Let us move on now to consider the two options mentioned above for the minimalist: either accepting or denying the linking claim. Recall that evidential minimalists subscribe to either one of the following:

Strong evidence for the truth of p is itself an epistemic reason for S to believe p.

Or

Strong evidence for the truth of p gives rise to an epistemic reason for S to believe p when certain other minimal conditions are met (e.g. S possesses the evidence and considers whether p).

Consider a minimalist who accepts the linking claim. If the minimalist accepts the linking claim, and if the arguments presented in section 3.1 are on the right track, then she must abandon her position. This is because, as we have seen, there are cases where conditions (i)–(iv) are met. In other words, the antecedent of the linking claim is satisfied. However, this means that a minimalist who accepts the linking claim must infer the consequent, viz. that, in the relevant cases, it is not the case that evidence E is (or gives rise to) an epistemic reason for belief. However, this is incompatible with both of the claims in the disjunctive characterization of the minimalist position. Thus, it seems that the minimalist must deny the linking claim. I would like to consider three different routes the minimalist might pursue in denying the linking claim. As we will see, none of these routes will prove to be favorable for the minimalist.

First, the minimalist might reject the linking claim as too strong. Specifically, the minimalist could argue that there are cases where conditions (i)–(iv) in the antecedent are met, yet the relevant evidence still gives rise to an epistemic reason for belief. Imagine cases involving excuses. Say that a person is temporarily impaired, exhausted, or confused. Being in such a state can lead one to violate certain standards or norms. However, this can also excuse a person from criticism. Imagine a friend who is suffering from depression or who is stricken with grief. Say that, as a result, the friend fails to follow through on some promise that was made to you. Perhaps a norm has been violated here (a norm that the person had good reason to follow), but the person is plausibly excused from criticism given their situation. Thus, the excuse doesn't imply that there wasn't a reason for the person to follow through on the promise, it just implies that the person shouldn't be criticized for failing to do so.

Similarly, there might be cases where conditions (i)–(iv) in the antecedent of the linking claim are met partly in virtue of the fact that the person in question is in a state which excuses them from criticism. The minimalist could argue that there are still epistemic reasons for belief in such cases (just as there was still a reason for the person in the above case to follow through on the promise). The point to emphasize here is that, as long as the minimalist still accepts some general connection between normative reasons for belief and openness to criticism, then appealing to excuses won't save her view. We could even introduce a fifth condition into the antecedent of the linking claim: (v) S is not excused or exempted from criticism.³⁵ Return again to the case of Katlyn and the disjunctive claims. This is a case where an agent is clearly and vividly attending to some evidence E for the truth of p without also believing p and yet is not appropriately subjected to criticism on account of this. Moreover, there are no excuses or exemptions at play in this example. For instance, we can say that Katlyn is an ordinary adult who is competent, intelligent, under no duress or stress, isn't impaired, etc. Cases like this show us that the appeal to excuses won't save a minimalist who endorses some general connection between normative reasons for belief and openness to criticism. In such cases, there are no excuses, yet the person still isn't appropriately subjected to criticism.³⁶

Consider a second way a minimalist might try to reject the linking claim; a way which also regards the claim as too strong. A minimalist could reject the linking claim by saying that evidence only ever provides warrant for belief. For instance, in Buckley (2021) I discuss the notion of a “warranting” reason.³⁷ A warranting reason to believe p doesn't require a person to believe p. Rather, warranting reasons are reasons that it would be appropriate to base one's doxastic attitudes on, whether or not one actually forms the attitude.³⁸ Arguably, Katlyn still has a warranting reason to believe the disjunctive propositions even if she doesn't take on beliefs in them, and even if she's not open to any form of criticism in virtue of this. Moreover, one could say that “warranting” reasons are a kind of normative reason for belief. Thus, according to this minimalist response, the linking claim should be rejected since the antecedent can be satisfied (along with our additional clause (v) ruling out exemptions and excuses) while the consequent is false. In the case of Katlyn and the disjunctive claims, the antecedent is satisfied, however the evidence does give rise to an epistemic reason to believe p, viz. a warranting reason.

I'm actually inclined to think that this response is correct. However, the point to emphasize here is that it won't help the minimalist. In order to see why, we need to ask the following question: under what conditions would a person be warranted in believing p? Return again to BORED AT HOME and LAST DONUT. In the former case we imagined Katlyn being asked “How many hours have you been alive?”, in the latter case we imagined David being asked “Who ate the last donut?”. In both cases, the individuals possess strong evidence E for the truth of p, however they must carry out some mental act M (e.g. a process of calculation or reasoning) in order to “see” the evidential connection between E and some target proposition p. The point I'd like to make here is that, in such cases, the person will only have warrant for believing the target proposition if they actually engage in the mental act M. Prior to engaging in that act, neither Katyln nor David would be warranted in believing their respective target propositions. However, engagement in such acts will often be a non-minimal condition. For instance, these will be acts that one might carry out when engaged in active inquiry vis-à-vis some question or subject matter. As we saw earlier, engagement in active inquiry is a non-minimal condition. Thus, there will be many cases where strong evidence E for the truth of p only gives rise to a warranting reason to believe p when some non-minimal condition is met, and this is incompatible with the minimalist's position.

Finally, a minimalist might reject the linking claim by simply denying any inherent connection between normative reasons for belief and our practices of interpersonal criticism, accountability, and blame. Consider, for instance, the objectivist described earlier; someone who thinks that strong evidence just is a normative reason for belief. This minimalist might hold that claims regarding our ordinary practices of interpersonal criticism and assessment are neither here nor there when it comes to the nature of objective normative reality. If the minimalist takes this route, then the normativity of evidence becomes utterly mysterious. The forgoing discussion reveals that there are cases where conditions (i)–(iv) are met, the subject in question is consciously and vividly appreciating the evidential connection that obtains between E and p, and the subject is not exempted or excused from criticism. The case of Katlyn and the disjunctive claims is a case of this kind. The objectivist must persist in claiming that, nevertheless, in such cases the evidence E constitutes a normative reason for S to believe p. However, it's not clear what the objectivist could mean by “normative”. If the above conditions are met, then in what sense does S have a “normative” reason to believe p?

The objectivist could appeal here to the notion of a warranting reason. She could, moreover, attempt to avoid the points I made above regarding warranting reasons by appealing to an objective notion of warrant. Thus, just as there might be objective “oughts”, there might also be objective epistemic warrants. These would be (warranting) reasons to believe certain propositions which obtain for an agent regardless of whether or not the agent is properly apprised of them (e.g. by “possessing” them and appreciating their evidential import vis-à-vis some target proposition). However, this has the following unfavorable result: as I sit here typing, I have warranting reason to believe infinitely many propositions. Indeed, I have warranting reasons to believe things that I can't, given my psychological limitations, believe. Just imagine all of the pointless disjunctive claims that are entailed by the simple proposition “I am typing on my computer right now”. The objectivist described above has to say that, as I sit here typing, I have warranting reason to believe all of them, and countless others. It's entirely unclear how the objectivist can maintain this while also taking “warrant” to be a normative notion.

4. Conclusion

I have argued against a view that I have labeled “evidential minimalism”. According to minimalists, there is a close connection between strong evidence for the truth of p and a normative reason to believe p: evidence is either itself a normative reason for belief, or evidence gives rise to such a reason when certain other minimal conditions are met. My argument proceeded in two steps. In the first step, I established that there are cases where a subject S possesses strong evidence for p, all of the minimal conditions for the normativity of that evidence are met, yet S doesn't believe p and isn't open to any form of criticism on account of this. The second step involved a linking claim which connected openness to criticism with epistemic reasons for belief. I argued that the minimalist can either accept or reject the linking claim. Either way, the minimalist view falters. If my arguments against minimalism are successful, then some form of non-minimalism will be the way to go when it comes to understanding the normativity of evidence. Here I remain neutral on the question of which form of non-minimalism should be preferred.³⁹