Epistemic justification and the ignorance excuse

Abstract

One of the most common excuses is ignorance. Ignorance does not always excuse, however, for sometimes ignorance is culpable. One of the most natural ways to think of the difference between exculpating and culpable ignorance is in terms of justification; that is, one’s ignorance is exculpating only if it is justified and one’s ignorance is culpable only if it not justified (call this the justification thesis). Rosen (J Phil 105(10):591–610, 2008) explores this idea by first offering a brief account of justification, and then two cases that he claims are counter examples to the justification thesis. The aim of this paper is to defend the justification thesis against Rosen’s two cases. The argument will proceed in the following way. First, I clarify a few things about the nature of culpable ignorance generally and why the justification thesis is so intuitive. I then present Rosen’s purported counterexamples. Once this is done, I argue that Rosen misses an important view of justification in the epistemology literature that I call the pragmatic view. I present a general picture of the pragmatic view, and explain how it fits naturally with our practices of criticizing people’s beliefs, including claims of culpable ignorance. Finally, I address Rosen’s cases arguing that, if the pragmatic view is right, then Rosen’s cases are not counterexamples to the justification thesis.

1 I

Ordinarily we think people are blameworthy for the wrongs they commit. Unless, that is, they have an excuse. One of the most common excuses is a claim of ignorance; when accused of wrongdoing people often defend themselves by making a claim that amounts to “I didn’t know!” When we accept such a defense that is because it shows that the person under suspicion isn’t blameworthy after all. Ignorance does not always excuse, however. Sometimes we refuse to accept a plea of ignorance because we think that the ignorance was culpable and culpable ignorance is no excuse.

To bring out the difference between exculpating ignorance and culpable ignorance consider two doctors. The first doctor makes sure to ask the patient about any allergies, and checks the patient’s medical records and charts before prescribing medication. The second doctor fails to interview the patient appropriately, and does not bother to check the patient’s medical history or chart before prescribing medication. Imagine that, in both cases, the patient was allergic to the medication prescribed; that there was no indication of this allergy in the patient’s testimony or medical history; and that all other factors are held fixed. Both doctors are ignorant of the patient’s allergy, but this ignorance only serves as an excuse in the former case. The reason often given for why culpably ignorant people are not excused is that they should have known. However, the latter doctor shows this isn’t quite right. Had the latter doctor done her due diligence she would be excused, just as the former doctor is. Even so, she still wouldn’t have known about her patient’s allergy (just as the former doctor didn’t know). So then, it isn’t whether she should have known that matters. Rather, the problem with our latter doctor is that she failed to take appropriate measures to avoid her ignorance.¹

This normative evaluation of a person’s ignorance suggests a connection between epistemology and moral philosophy (Austin 1956; Zimmerman 1997, 2008; Rosen 2008). Indeed, it appears that what we are doing when we claim a person’s ignorance is culpable is evaluating a person’s epistemic standing. Our first doctor has done the best she could to gather and evaluate the available evidence. As such, her epistemic status is impeccable. The same is not true of our latter doctor. Both are ignorant, but the former doctor’s ignorance is not open to criticism, where the latter doctor’s ignorance is. The question of whether a person’s belief (or more properly, doxastic state) is open to criticism is a matter of epistemic justification. That is, when a person’s belief is justified it is not open to criticism. When a person’s belief is not justified, it is open to criticism.² Since a claim of culpability is a form of criticism, an intuitive conclusion to draw is something like the following two-part thesis.

The Justification Thesis

• (1) One is culpable despite one’s ignorance only if that ignorance is not justified, and

• (2) One’s ignorance exculpates only if it is justified.

The justification thesis isn’t just highly intuitive. It also matches up well with our ordinary practice. Consider our two doctors again. One very natural description of our judgments about the cases is that the reason our first doctor is excused is that her ignorance is justified. The reason our second doctor is not excused is that her ignorance is not justified. Whether this is right will, of course, depend on our theory of justification. Nevertheless, the intuitive appeal and natural fit with our practices make the justification thesis attractive.

Despite its intuitive appeal, Rosen (2008) challenges the legitimacy of the justification thesis. After presenting a sense of justification that he takes to represent well established views in epistemology, Rosen constructs two cases that he claims show that both parts of the justification thesis are false.

The aim of this paper is twofold. The first aim is to present a plausible, robust account of the justification thesis. The second aim is to defend the account against Rosen’s cases. The paper will proceed in the following way. First, I clarify a few things about the nature of culpable ignorance generally and why the justification thesis is worth exploring. I then present Rosen’s two cases in detail and explain the challenge they pose to the justification thesis. Once this is done, drawing on work in epistemology, I describe a view of justification that I call the pragmatic view, and explain how it fits with our practices of criticizing people’s beliefs, which includes claims of culpable ignorance. This should establish the plausibility of both the pragmatic view and the justification thesis. Finally, I address Rosen’s cases arguing that, if we understand justification in terms of the pragmatic view, then Rosen’s cases are not counterexamples to the justification thesis.

2 II

As Rosen (2003) points out, ordinary moral assessment operates with a presumption of responsibility. That is, we ordinarily begin from the assumption that a person is responsible for her actions, and we do not take ourselves to be obligated to perform further inquiry before making such an assumption unless given a reason to do so. Suppose we are told that Sally has done something we ordinarily consider wrong. Sally has promised to meet Susan downtown at four o’clock, and never showed up. Without being told more, we presume that Sally is to blame for standing Susan up. However, suppose we discover that Sally’s watch was set to the wrong time; that Sally got in a small fender bender on the way and was stuck waiting for the police; or that Sally was only 12 years old and her parents refused to grant her permission. These considerations may dispel the presumption of responsibility, and we may conclude that, while breaking a promise is wrong, Sally is not to blame for her broken promise in this case.

A class of considerations, like those just mentioned, that defeat the standing presumption of responsibility are known as excuses (Rosen 2003; Austin 1956). One such consideration mentioned in Sally’s case above was that Sally’s watch was set to the wrong time. Given that she had no special reason to suspect that her watch was wrong, Sally did not know what time it was. Ignorance sometimes excuses.

Nevertheless, standing presumptions of responsibility are not easily set aside. We may, for example, find it hard to believe that a person like Sally really did lack any special reason to suspect her watch was set to the wrong time. After all, we might think, if Sally is anything like the rest of us, then she is all but inundated with devices that tell her the time, so she really should have known the time, even if her watch was off. If she should have known the time, then her ignorance is no excuse! In other words, Sally’s ignorance is culpable, and culpable ignorance does not excuse.³

There are a few things to briefly touch on before we continue. First, it bears mentioning that considerations that undermine presumptions of responsibility are typically divided into two types: the control condition and the epistemic condition (Fischer and Raviza 1998). My concern is only with the epistemic condition, and as such, following Smith (1983), in what follows I assume that the agents under consideration have the relevant control necessary for moral responsibility and concentrate on the agent’s ignorance, for cases where the only question left open is whether the epistemic condition is met will most clearly tell us the difference between exculpating and culpable ignorance. Second, there are different forms of ignorance, and not all of them count as exculpating (Peels 2010, 2014).⁴ Rosen (2008) argues that the only relevant sort of ignorance is the absence of true belief. Further, some philosophers draw a distinction between ordinary factual (non-moral) ignorance and moral ignorance and argue that moral ignorance does not exculpate.⁵ Here, I will limit my discussion to non-moral ignorance of the sort with which Rosen was concerned—a lack of true belief.

Finally, I shall accept that culpable ignorance is a matter of derivative responsibility. What I mean by that is, one is blameworthy for an action done in ignorance only if one is blameworthy for one’s ignorance. But one is blameworthy for one’s ignorance only if one is blameworthy for some past action that resulted in the ignorance in question.⁶ Consider the following influential passage from Smith (1983, p. 547):

…notice that the relevant cases all involve a sequence of acts: an initial act, in which the agent fails to improve (or positively impairs) his cognitive position, and a subsequent act in which he does wrong because of his resulting ignorance.⁷

Smith refers to the initial act as a “benighting act” and to the final act for which we hold the agent responsible as the “unwitting wrongful act.”⁸ Benighting acts, per Smith, must also be acts for which the agent is responsible.

Smith’s claim needs refinement, however. One may improve one’s cognitive position and yet the improvement may have no effect on whether one commits the resultant unwitting wrongful act. Sally, for example, is not only ignorant of what time it is when she fails to keep her promise to meet Susan, she is also unaware that her watch has stopped working. Suppose Sally discovers that her watch is not working. Sally has improved her cognitive position—she is less ignorant than she was before—but while her overall epistemic position is better, she’s in no better an epistemic position vis a vis knowing the time. And the fact that it is time to meet Susan is one of the facts that makes Sally’s action wrong. We therefore need to restrict our discussion to ignorance of facts relevant to what makes the action under consideration wrong.

Gideon Rosen has given an account of culpable ignorance that builds off Smith and takes these things into consideration, arriving at the following package of claims taken from Rosen (2008):

• (a) X does A from ignorance when X is unaware of each of the features of his act that makes it wrong.⁹

• (b) When X does A from ignorance, X is culpable for A only if X is culpable for the ignorance from which he acts.

• (c) X is culpable for his ignorance when it is the foreseeable upshot of culpable negligence in the management of his opinion.

• (d) X is negligent in the management of his opinion when X fails to take some required precaution against ignorance or error.

Given these claims, it is easy to see why Rosen thinks that justification is the best candidate for an epistemological concept that might explain the difference between exculpating and culpable ignorance. Indeed, one might think that descriptions like “negligent in the management of one’s opinion” and “failing to take some required precaution against error” are just alternative ways of saying one’s ignorance is not justified. This is at least a prima facie reason to think Rosen’s package of claims are captured by the justification thesis. However, Rosen offers two cases he claims show that both parts of the justification thesis are false; namely that one’s ignorance can be culpable despite being justified, and that one’s ignorance can be exculpating despite lacking appropriate justification. I describe these cases in detail in the next section.

3 III

There are two key points to take away from the discussion in the previous section. First, one’s culpability for one’s ignorance is derived from one’s culpability for the acts that resulted in one’s ignorance. Second, one’s culpability for bringing about one’s ignorance is a matter of negligently failing to manage one’s beliefs given one’s circumstances and appropriate considerations of the future.

Rosen (2008) considers whether a notion of justification captures these features. He says:

In the clearest cases of blameless action from ignorance, the agent’s ignorance is not just real: it is justified. It may sound odd to speak of justified ignorance but the idea is straightforward. We often have good and sufficient reason for believing falsehoods or for suspending judgment on some issue, thereby failing to believe a truth. When one fails to believe a true proposition in either of these ways, one’s ignorance is justified or reasonable.

We might then consider two questions: Does justified ignorance of the wrong-making features of one’s actions always excuse bad conduct? And does ignorance excuse bad conduct only when it is justified? (Rosen 2008, p. 599. Emphasis in the original.)

To answer these two questions, Rosen considers two senses of justification. The first is a strict sense of justification where one is justified just in case one’s belief is appropriately supported by the evidence of which one is aware. Rosen immediately points out that this sense of justification is too strong, noting that there are cases where a person’s evidence gathering is inadequate. In such a case, one’s belief may be supported by the evidence one has, but it is not in line with additional evidence that a person negligently failed to gather. He then offers a second sense of justification which includes standards for evidence gathering procedures and doxastic maintenance. In this sense, ignorance that is not justified is just ignorance that results from a failure to meet these standards. I will have more to say about these two senses of justification in Sect. 4. With this second sense of justification in hand, Rosen then goes on to give two cases that he claims are counterexamples to both parts of the justification thesis.

Let me describe these in the order of the questions posed in the passage above. To answer the question “does justified ignorance of the wrong making features of one’s action always excuse?” Rosen offers the following case.

Suppose Dr. Feldman consults a chart that would lead any competent doctor to conclude that her patient is diabetic, but that Feldman fails to draw this conclusion and harms her patient as a result. Feldman acts from ignorance in our sense: she is unaware that her conduct has the features that make it wrong. Moreover, her ignorance is not justified. Given her evidence and background knowledge, the only reasonable thing for her to think is that the treatment will harm her patient. And yet it does not follow that she is morally culpable for the harm she causes. Suppose that Feldman fails to draw the obvious conclusion only because her blood pressure medication has temporarily addled her judgment. If the effects of the medication are severe enough and in no way foreseeable, we should conclude that Feldman is not responsible for her mistake or for the harm it causes (Rosen 2008, p. 601).

Rosen’s claim is that, given that she has evidence before her that would lead any competent doctor to conclude that the patient is diabetic, Dr. Feldman’s ignorance of her patient’s diabetes is not justified. Nevertheless, because the effects of the medication—the ultimate cause of her ignorance—are not foreseeable, Dr. Feldman is clearly not culpable for her ignorance. If this is right, then the case of Dr. Feldman shows that ignorance sometimes excuses even though it is not justified.

In answer to the second question, “does ignorance excuse bad action only when it is justified?” Rosen offers the following case.

Goldberg has been charged with investigating financial impropriety at Acme Corp. The evidence so far points to Himmelfarb as the guilty party, and it is predictable that things will go badly for him if he is named. Goldberg conducts a scrupulous inquiry, which ultimately sustains the case against Himmelfarb. At this point, Goldberg has done everything reasonable investigators ought to do, and his evidence amply supports the conclusion that Himmelfarb is a crook. In response to this evidence he forms a belief to this effect and accuses Himmelfarb of embezzlement. In fact, Himmelfarb has been framed by a malicious colleague. This means that in filing his report, Goldberg accuses an innocent person of a crime, and that is wrong. But if there is nothing more to the story this is a straightforward case of blameless wrongdoing from excusable ignorance. Moreover, if we are asked to say why Goldberg’s ignorance is not his fault, it would be natural to reply that Goldberg’s ignorance is excusable precisely because it is epistemically justified in the broad sense.

Suppose however, that there is more to the story. Suppose that Goldberg has promised Himmelfarb’s mother that he would spend an extra day on the investigation digging more deeply into the evidence, despite the fact that he (and she) had no reason to believe further inquiry would reveal anything new. Since he was in a position to keep his promise without great cost, Goldberg was morally obliged to put in the extra time. But he did not. Out of moral indifference or an eagerness to move on to other projects, Goldberg made up his mind without going through the file one more time. We may add that if he had kept his promise, he would have discovered the fraud and exonerated Himmelfarb (Rosen 2008, p. 601).

According to Rosen, Goldberg’s ignorance is justified. He has adhered to all the normal standards applicable to anyone in his shoes, and he has drawn a conclusion that is well supported by the evidence. To bolster this Rosen offers this further consideration.

Imagine that Goldberg is a member of a team whose other member has made no promise to Mrs. Himmelfarb. The partner’s belief in Himmelfarb’s guilt is epistemically justified by any standard. Goldberg has the same information produced in just the same way, and he has reasoned to just the same conclusion. Moreover, the stakes are the same for both: the costs of error are the same; the epistemically relevant costs and benefits of pursuing the inquiry further are the same, and so on. This strongly suggests the status of their beliefs should be the same, and hence that Goldberg’s belief in Himmelfarb’s guilt should be reckoned epistemically justified in the same broad sense. (Rosen 2008, p. 601)

If Goldberg’s partner is justified in his ignorance (as he clearly is), and Goldberg is on the same epistemic footing as his partner, then, Rosen claims, Goldberg must also be justified in his ignorance. Nevertheless, Goldberg is blameworthy, not merely for breaking his promise, but also for his action of accusing an innocent man. If this is true, then Goldberg’s ignorance is both justified and culpable.

In what follows I argue that Rosen does not account for all the possible considerations that factor into our ordinary concept of justification. To do this I examine three views of justification found in the epistemology literature. The first two views match up well with the two senses of justification Rosen considers. The third view, which I call the pragmatic view, builds off the first two, adding to them a further, pragmatic consideration that fits very naturally with our practices of criticizing beliefs—practices that are central to the concept of culpable ignorance.¹⁰

4 IV

Suppose Sally unwittingly misses her date with Susan because her watch is set to the wrong time. When we refuse to accept Sally’s ignorance of the time as an excuse, it is because we think she should have known the time. One way to understand the claim that she should have known is that, while Sally has some reason to believe her watch, the total set of considerations at her disposal do not support her false belief. It is very natural to think of this situation in terms of Sally’s justification; Sally’s ignorance (false belief) is not justified and that is why it is culpable.

Treating justification as a normative, deontic notion in this way is not a new idea. The epistemological literature has a lively, ongoing debate regarding the ethics of belief, in which one topic up for discussion is the deontic character of justification. Can we, for example, criticize people for the beliefs they hold in the same sorts of ways that we criticize people for their actions? Are beliefs the appropriate target for normative judgments?¹¹ Even those who answer such questions in the negative admit that justification is most naturally thought of as a deontic notion.¹² That is, the most natural understanding of justification is that when one has a belief that is not justified one shouldn’t believe it, and this sense of “should” leaves one open to criticism for that belief. Conversely, if one’s belief is justified then it is at least permissible for one to believe it. This natural understanding of justification is the one that concerns the justification thesis, and so it is the one we will be working with for the remainder of the paper.¹³ For if the claim of culpable ignorance is, at bottom, the claim that one should not have been ignorant, and we are basing a judgement of blameworthiness on such a claim, then we are, by extension, criticizing one for one’s doxastic attitudes. The first question we should ask in an inquiry into the nature of culpable ignorance, then, is the very same question we ask in an inquiry into the nature of justification: when is one open to criticism for what one believes?

A good place to start when attempting to answer this question is where Rosen does; with the traditional epistemological approach. He says, “In one sense of the phrase, a belief is justified when the subject’s total evidence constitutes a good and sufficient reason for it.” (Rosen 2008, p. 600). As I understand Rosen, this is representative of a strict form of the standard deontic view of justification in epistemology known as evidentialism.

Evidentialism is the thesis that what one is justified in believing, or more properly, the doxastic attitude one is justified in holding, is that which is best supported by one’s evidence.¹⁴ According to this general picture, Sally, for example, should hold the doxastic attitude (e.g., belief, disbelief, or suspension of judgment) best supported by her current total evidence—the evidence of which she is aware. To see this let us expand the case of Sally.

Case 1:

Sally has promised Susan she would be downtown at four o’clock for dinner. As the time approaches, Sally continually checks her watch to make sure she has time to get ready. Prior to dinner Sally is out running errands, and she visits an outdoor flea market as her last stop before meeting with Susan. Leaving her cellphone in the car, she begins sifting through the various odds and ends, though she is careful to check her watch regularly, trusting it to tell her the correct time. Unfortunately, Sally’s watch is wrong. Since she is at an outdoor flea market, there are no other clocks around, and thus little opportunity for her to realize that her watch is wrong. Based on what her watch tells her, Sally returns to her car with what she believes is thirty minutes to spare. When she arrives at the meeting place she learns that she is an hour late and Susan has already left. Sally immediately contacts Susan to explain. Upon hearing Sally’s story Susan excuses her, thinking that Sally didn’t really mean to break her promise; she just had bad information. Her watch is usually very reliable, as most peoples are, and it was her only means of keeping track of the time. It was just a dinner, after all. They can always try again next week. On this basis, Susan realizes that she shouldn’t blame Sally for missing dinner. It takes a minute for Susan to calm down, but when she does, she realizes Sally really does have a good excuse.

Here we can see how evidentialism is supposed to work. Sally falsely believed she had plenty of time because the evidence provided by her watch—the evidence she has—tells her as much. Since Sally’s false belief is the one best supported by her evidence, according to evidentialism it’s the belief she ought to have, and is therefore not open to criticism.

On the face of it, evidentialism is an attractive view. However, as Rosen (2008) points out, this sense of justification will do us no good in answering questions about culpability, for it is easy to construct cases where one’s belief is well supported by the evidence one currently has, but nevertheless one is open to criticism for that belief because one has been negligent. Consider this contrasting version of Sally’s case:

Case 2:

As in case 1, Sally has promised Susan she would meet her for dinner at four o’clock. Prior to dinner Sally is out running errands. This time, however, Sally is out looking for a new clock for her living room. In the clock store Sally continues to check her watch to be sure that she is not late, trusting it to tell her the correct time. Distracted by her errands and picking out the right clock for her living room, Sally fails to notice that all the clocks in the store say the same thing and they all say something different than her watch. Additionally, she fails to notice one of the clocks has a sign under it that reads “atomic time”. Sally forms the false belief that she has plenty of time before dinner (because her watch tells her so), and after purchasing a new clock for her living room decides to stop at home and set it up before heading to dinner. That stop makes her an hour late, though she believes herself to be on time. Upon arriving Sally finds that Susan has already left. Sally immediately contacts Susan to explain what happened.

Sally is blameworthy in this case, but not because her belief goes against the evidence she has; indeed, Sally has the exact same evidence in case 1 and case 2. Rather, the reason seems to be that Sally had many opportunities to improve her evidence. As Smith (1983) says, Sally has culpably failed to improve her cognitive (epistemic) situation. The problem with evidentialism is that it demands only that the agent have beliefs that are appropriate per her current evidence. However, in practice we tend to criticize people’s beliefs based on certain norms of doxastic maintenance, or what Rosen refers to as the “management of one’s opinion” (Rosen 2008, pp. 600–601). A doctor, for example, who fails to check a patient’s chart may form the opinion that the best treatment to prescribe for someone with the patient’s symptoms is the very medicine to which that patient is allergic.¹⁵ According to evidentialism, since the doctor’s current evidence does not include the fact that the chart says the patient is allergic, the doctor’s false belief is the one she ought to have; the belief she has is the one best supported by her current evidence. But, as Rosen points out, this is a highly counterintuitive result. As a doctor, she has an obligation to meet a certain standard that includes checking a patient’s chart before prescribing treatment. In other words, she shouldn’t form judgments that will be used as the basis for prescribing treatment before checking the patient’s chart, interviewing the patient, etc. If she has a belief she shouldn’t, at least when it is the result of a failure to meet a certain standard of evidence gathering, then such a belief is unjustified.¹⁶

The evidentialist may insist that doxastic justification is a uniquely epistemic enterprise. As such, additional obligations, while perhaps virtuous and sensible, are not relevant to whether a belief is justified. A person can have justified beliefs that they should not hold for non-epistemic reasons. It may, for example, be prudentially, or even morally, required of a doctor that she meet a certain standard before diagnosing patients, but this is not an epistemic requirement. According to evidentialism, a doctor’s ignorance of her patient’s allergy is justified so long as her ignorance is warranted given the evidence of which she is aware.¹⁷

DeRose (2000) has responded to this objection. He says, “Certainly there is a very good sense of ought in which [our doctor] ought not to believe that [her patient was not allergic to the prescribed medication]. This is the sense that my intuition responded to. And this sense, I think, has a good claim to the title of ‘epistemic.’”¹⁸

To put the point succinctly, the strict evidentialist insists that only the evidence a person currently has matters to whether that person’s beliefs are justified because only the evidence she has can be appropriately described as epistemic (Feldman 2000; Feldman and Conee 1985). DeRose’s point is that this demand runs contrary to our intuitions. We think that our negligent doctor shouldn’t be ignorant of her patient’s allergy, and the reason we think this is that her ignorance is the foreseeable result of a failure to meet a standard of evidence gathering. There is no reason to think that factors related to belief formation, especially those that pertain to one’s evidence and the standards it must meet, are not appropriately thought of as epistemic.¹⁹

So then, per Rosen (2008) and DeRose (2000), we need to buttress evidentialism with a more expansive sense of justification that includes reasonable standards of doxastic maintenance that match our intuitions regarding cases like Case 2 and our negligent doctor.²⁰ Rosen describes this broader sense of justification like this:

An opinion is justified in the broad sense if (a) the subject has complied with every pertinent procedural requirement in developing and assessing his evidence and (b) his opinion is justified in the narrow (evidentialist) sense. (Rosen 2008, p. 601. Emphasis in the original. Parenthetical added for clarification.)

It is this sense of justification at which Rosen’s purported counterexamples are directed. However, there is still something important missing from even this sense of justification.

As Rosen points out, these procedural requirements vary greatly with context (ibid). When it comes to beliefs regarding the diagnoses of her patients, a doctor has an obligation to gather a significant degree of evidence in very specific ways. In other contexts, the standards are not so strict. If our doctor watches the news and sees that the weather report predicts nice weather, and then uses this information as the basis for setting a tee time for a golf game, then we would hardly criticize her. When she defends herself by saying, “The weatherman said it would be clear all day,” refusing to accept her excuse would be excessively arduous.²¹ If procedural requirements vary from case to case as Rosen suggests, our notion of justification should include an explanation for these changing requirements.²²

Notably, some influential epistemologists have suggested that one’s pragmatic circumstances play a role in determining the context upon which these procedural requirements depend (Stanley 2005; Hawthorne 2004; Fantl and McGrath 2009). A doctor who fails to check a patient’s chart puts herself in a position to be blamed. One reason is because she fails to meet certain procedural obligations in her evidence gathering that are dictated by her job as a doctor—she ought to check the patient’s chart. There are, however, two further reasons. First, the belief has practical relevance; the doctor will use that belief as a reason to act—it will be the basis for prescribing treatment. Second, the belief also has a high degree of significance; the patient’s well-being, even her life, is at risk. In short, there is a lot riding on the doctor making the correct diagnosis. These two factors—practical relevance and degree of significance—combine to dictate a belief’s practical significance. Because being wrong has immense practical significance, the doctor has an obligation to take reasonable steps to mitigate the possibility that she is wrong. Both her job, and the practical significance of the resultant belief—how much it matters that she is wrong—are part of what constitute her pragmatic circumstances (Stanley 2005, chapter 5; Fantl and McGrath 2009, chapters 1 and 4).

Call the thesis that one’s pragmatic circumstances play a role in justification the pragmatic view.²³ To see the appeal of the pragmatic view consider a third variation of Sally’s case.

Case 3:

Sally has promised Susan she would meet her for dinner at four o’clock. This time Sally knows that the dinner date is very important. It is Susan’s last night in town before she moves away and she only has an hour to spare. They do not know when they might see each other again, and both Sally and Susan want to make sure they see each other one final time. Sally has a few things she wants to get out of the way before dinner so she decides to stop off at the bank and, if she has time, the post office. While inside the bank she checks her watch to be sure she has time to go to the post office. Sensitive of how important the dinner is, Sally is sure to continue to remind herself of how much time she has by continuing to check her watch. Seeing that her watch says she has a little over an hour before dinner, she believes she has plenty of time to go to the post office. There is a clock on the bank’s sign that says that Sally only has a few minutes to get to dinner, though Sally, in her anxiousness to get to the post office, doesn’t see it. When she arrives at the post office there is a long line, so she checks her watch again. Because it tells her she has just over an hour before dinner she decides to wait in line. The post office has a clock behind the counter in plain view, and another on the wall adjacent to the line, both of which tell her that she only has a few minutes to get to dinner, but Sally’s mind is busy with nostalgic thoughts about her friend Susan, and doesn’t notice either clock. It takes her nearly an hour in the post office, and she heads to dinner not realizing she has already missed Susan. Sally immediately contacts Susan to explain. Susan’s anger isn’t quelled by Sally’s attempted excuse and she responds by saying, “Sally, you knew how much this dinner meant to me. I’ve been to that bank, there’s a clock right outside that flashes the time in giant lights! There are two clocks in in the post office as well! If this were any other day I’d not be bothered, but today is important! You should have made sure!”²⁴

Here, Susan, while refusing to accept Sally’s excuse, admits that “on any other day” Sally would be off the hook, but in virtue of the significance of the dinner date Sally had an obligation to put in a bit more effort. Susan is not being unreasonable in demanding that, in virtue of how important it was that she be there, Sally at least make use of immediately available opportunities to keep track of the time.²⁵

This brings out what was claimed just before case 3. A doctor needs a great deal of evidence from reliable, generally accepted sources before a diagnosis should be accepted. But she need not consult multiple weather reports before making a tee-time for her golf game. One reason for this is that the stakes are significantly higher when the doctor forms a belief about her patient than when the doctor forms a belief about the weather. All of this is to say that the epistemic standard one is required to meet is at least partly determined by how much it matters if one is wrong (Fantl and McGrath 2009; Stanley 2005; Hawthorne 2004). It matters a great deal if a doctor misdiagnoses a patient. It hardly matters at all if she gets rained out of a round of golf.

The same is true of Sally. The reason Sally is blameworthy in cases 2 and 3, but not in case 1, is that in case 2, while the practical significance of her ignorance was not particularly high, the ease with which she could improve her cognitive situation makes us hesitate to excuse her; the evidence at her disposal was so obvious and simple to acquire Sally really should have done so. In case 3, Sally was not in such an obvious position to improve her cognitive situation. In fact, were it not so important that she get the time right, she would be excused. Unfortunately, it is important, and even Sally recognizes and appreciates this fact. As such, it is reasonable to expect that Sally make more of an effort to be certain about what time it is, especially since doing so does not require her to go very far out of her way (though more so than in case 2).

What these contrasting cases signify is that whether one’s beliefs are open to criticism is a function of three factors: (1) Whether they are formed based on good evidence, (2) whether the belief meets some minimal standard of doxastic maintenance, and (3) the practical significance of the resultant belief. The relationship between these things is straightforward. One is justified when the degree of certainty provided by one’s evidence meets the minimum standard set by the practical significance of the belief being formed. Put another way, the standards of certainty one’s current evidence must meet before a belief based on such evidence is justified is determined in part by how much it matters if the belief is wrong.

5 V

Readers may be quick to point out that while the pragmatic view may be a good way of describing how we are open to criticism for our beliefs in terms of justification, the question of culpable ignorance isn’t about when one is open to criticism for one’s beliefs per se. It seems that one may be open to criticism for any false belief that leads to action if it is not justified, but the class of beliefs under consideration in culpability are only those that lead one to perform a morally bad act. How does one get to moral criticism for one’s actions that are a result of one’s ignorance from being open to criticism for one’s ignorance on justificatory grounds?

Certainly, it is true that the pragmatic view is a view about when all of one’s beliefs are open to criticism vis-a-vis a lack of justification, but the contrasting cases of Sally are meant to show that the criteria for justification also underwrite the moral criticism we aim at actions that are the result of our ignorance. The point Rosen (2008) makes about the foreseeable upshots of one’s ignorance will help make this clearer.

Some of the foreseeable, practically significant upshots of our ignorance are moral in nature. When our ignorance has the foreseeable upshot of resulting in an action for which we may be (morally) blamed, that consideration is typically a salient one. It is plausible that when the possibility that one might be ignorant will foreseeably lead one to commit an act that has moral significance, then the steps one is reasonably required to take to avoid such ignorance may be morally significant as well.

To be sure, ignorance of a certain fact does not always have moral significance. Instead of meeting Susan, Sally’s ignorance may cause her to miss her favorite television program. This is hardly morally significant. She may still leave herself open to criticism if she could easily discover the time (like in case 2), but we do not blame her (in the moral sense) because she did nothing (morally) wrong. Even so, ignorance of this very same fact is what leads her to do something (morally) wrong in cases 2 and 3. The difference in the cases is in whether Sally’s ignorance was of a fact pertinent to what makes her action wrong.

Rosen (2008) states the point like this.

There is a sense in which we always act from ignorance. Whenever X poisons Y there are countless features of his act of which the poisoner is unaware. He may not know how the poison works, or that it was manufactured in Alabama; and for the most part such ignorance is completely irrelevant to his culpability. The first order of business, then, is to distinguish potentially exculpatory ignorance from the ignorance that makes no difference. When an act is wrong, it is wrong for a reason (or several reasons). These reasons are the features of the act in virtue of which it is wrong: the features that would figure in a complete and correct answer to the question, “What is it about this act that makes it wrong?” The straightforward idea is that potentially exculpatory ignorance is always ignorance of the “wrong-making” features of one’s actions. (p. 593. Emphasis in the original).

In our cases, Sally is breaking her promise to meet Susan. The features of her action that make it wrong are those that pertain to her failure to keep her promise; namely that it is now four o’clock and Sally should be at dinner. Sally is ignorant of these facts; she falsely believes that it is not yet four o’clock and she does not know she should be at dinner right now.²⁶ Since she is ignorant of the features of her act that make it wrong, we must then ask whether she is culpable for that ignorance.

My contention is that the reasons we give to explain why Sally’s ignorance is culpable will be the same reasons the pragmatic view claims her ignorance is not justified. Sally’s ignorance is justified in case 1 because she had no access to alternative sources of information, and the practical significance of being late is low; Sally has done all that can be reasonably required of her. Intuitively, Sally is excused for acting wrongly on the basis of her ignorance for the very same reason; she has done all that can reasonably be required of her.

In case 2, even though the practical significance of her tardiness is the same as in case 1, Sally had easy access to many sources that could tell her the time. As such, on the pragmatic view her ignorance is not justified. Intuitively, Sally’s ignorance is culpable for the same reason. It is reasonable to demand that Sally pay attention to any one of the dozens of clocks she is already looking at.²⁷

In case 3, on the pragmatic view, given how important it was that she be on time, Sally should have made certain she knew what time it was. Since Sally failed to do something her pragmatic circumstances dictate she should have done, her ignorance is not justified. Her ignorance is also culpable on the same grounds. Susan is not being unreasonable when she blames Sally for being late despite Sally’s ignorance, and she may cite the fact that Sally should have made certain of the time as evidence that her blame is appropriate.

What this all suggests is that, if we understand justification in terms of the pragmatic view, then the question of whether one’s ignorance is culpable really comes down to two questions: (1) was the action caused by the ignorance in question one we would ordinarily consider wrong? and (2) if so, was the ignorance that caused this action justified? If the answer to the first question is yes, and the answer to the second question is no, then one’s ignorance is culpable.

6 VI

Now that we have a plausible view of justification in hand, we are finally in a position to evaluate Rosen’s purported counterexamples to the justification thesis given in Sect. 3. Let me respond to the counterexamples he offers in turn.

Recall Rosen’s first case wherein Dr. Feldman, whose judgment is compromised due to medication, fails to draw the conclusion that her patient is diabetic. Rosen claims that Dr. Feldman’s ignorance is not justified because she has access to information that would lead any reasonable doctor to conclude the patient is diabetic (she has looked at the patient’s medical history and by hypothesis it indicates diabetes). Nevertheless, because it is an unforeseeable consequence of her medication her ignorance regarding her patient’s diabetes is exculpating. If this is right, then Dr. Feldman’s case represents a counterexample to the first clause of the justification thesis which states that one is excused on the basis of ignorance only if one’s ignorance is justified.

First, it is not clear that Dr. Feldman’s ignorance is not justified. Per the pragmatic view, Dr. Feldman’s beliefs are justified if they are appropriate given the evidence she has, the ease of access to available evidence, and her pragmatic circumstances. The question is whether it would be reasonable to expect that a person impaired by blood pressure medication notice that the chart contains information that would lead any non-impaired doctor to the conclusion that the patient is diabetic. It is not obvious that we can (or should) hold Dr. Feldman to this standard. If we know that an impairment caused by her blood pressure medication interfered with her ability to draw the right conclusion from her evidence, then we might think that the resultant belief was the one she ought to have, given her impairment. After all, we might say, any reasonable doctor in her circumstances would have believed the same thing. However, on the pragmatic view, one’s circumstances play a role in determining whether one’s belief is justified.

The question of whether Dr. Feldman’s belief is justified should be this. Given that she is impaired, is it reasonable to expect her to believe that her patient has diabetes? The answer to this question will depend on the answer to the question of whether she knew she was impaired in the first place. If she took her blood pressure medication knowing full well that it may cause impairments in her judgment, then, surely, the resultant false belief is not justified. She knew that forming false beliefs was a likely occurrence and so she is therefore required to take extra measures to ensure she does not form false beliefs. However, if she did not know that the blood pressure medication would cause her to be impaired, then the resultant false belief may well be justified. The reason for this is that justification is derivative in the same way as responsibility. Let me explain.

Consider common cases of derivative responsibility, say, driving while impaired. If a person drives while impaired because she had a few drinks, then we think she is responsible for any accident she may cause despite her impairment. The natural explanation for this is that she is responsible for becoming impaired, and so will be responsible for what happens while she is impaired. However, if, unbeknownst to her, someone laced her soda with a tasteless drug that caused her to become impaired without noticing, then intuitively she is not responsible for driving while impaired, and hence not responsible for any resultant accident.

Justification works similarly. If one is aware that one’s epistemic faculties are impaired, then the resultant belief is not justified. If one is not aware that the evidence one has, or the methods one uses to gather and evaluate that evidence are faulty, then one has no reason to doubt them. If one bases one’s judgment on grounds that, for all one knows, are adequate and trustworthy, then the resultant belief is justified.

In Rosen’s case Dr. Feldman’s impairments are “severe enough and in no way foreseeable”. But this is just to claim that Dr. Feldman is justified in believing that her medication would not cause her to make poor future judgments. This belief is false, of course, but it is also the explanatory locus of Dr. Feldman’s excuse. Since she is justified in the belief that her medication will not addle her judgment, she has no reason to doubt that her judgement will get her to the correct diagnosis. Accordingly, Dr. Feldman has done all that can reasonably be required of her on the pragmatic view, hence her belief that her patient does not have diabetes is, in fact, justified. The case of Dr. Feldman is therefore not a counterexample to the first clause of the justification thesis that one is excused on the basis of one’s ignorance only when that ignorance is justified.

There is one more case to consider. Recall the case of Goldberg. Goldberg is a detective assigned to a case of embezzlement. Because of a clever frame job, Goldberg accuses an innocent man named Himmelfarb. In the case, Goldberg has performed all the procedures that his job as a detective requires (evidenced by the fact that his partner has also performed these procedural requirements and drawn the same conclusion). However, Goldberg makes a promise to Himmelfarb’s mother that he will spend an extra day digging into the evidence more deeply before making his accusation. Goldberg breaks his promise, and accuses Himmelfarb anyway. Assuming Goldberg would have discovered Himmelfarb’s innocence, Rosen’s claim is that Goldberg’s ignorance is culpable; he is blameworthy, not merely of breaking a promise, but also of accusing an innocent man.

As far as I can tell, the reason Rosen thinks Goldberg is blameworthy derives from his promise to Himmelfarb’s mother. This promise imparted on him a further obligation, which, while not epistemic in nature (p. 603), is still what Rosen calls a “procedural requirement” in the maintenance of his beliefs. In other words, according to Rosen, Goldberg’s belief is epistemically justified but he still should not believe it. He says, “the active side of epistemic life is constrained by a range of obligations, only some of which are recognizably epistemic.” (p. 603). This is brought out by Rosen’s further addition to the case. Recall that Goldberg’s partner is in the exact same epistemic position as Goldberg. Both have completed all the requirements mandated by their job as detectives before making an accusation. Rosen also adds the claim that “the stakes are the same for both: the costs of error are the same; the epistemically relevant costs and benefits of pursuing the inquiry further are the same, and so on” (Rosen 2008, p. 601). If Rosen is right, this indeed strongly suggests that the epistemic status of Goldberg and his partner’s beliefs should be the same; if Goldberg’s partner’s belief is justified, so too is Goldberg’s.

It bears mentioning before we proceed that not everyone shares Rosen’s intuition that Goldberg is blameworthy for accusing an innocent man. Indeed, most will probably think Goldberg is blameworthy for breaking his promise but not for accusing an innocent man. In which case, no defense of the justification thesis is needed. Goldberg is excused from accusing an innocent man because his ignorance is justified (just as his partner’s is). However, let us grant Rosen’s intuition for the sake of argument. There are two points I want to make in response.

The first point is to answer the following challenge. If the justification thesis is true, and Rosen’s claim that Goldberg is blameworthy is true, then it follows that Goldberg’s ignorance of Himmelfarb’s innocence is not justified. However, Goldberg’s partner’s ignorance of Himmelfarb’s innocence is clearly justified. If they have all of the same evidence and come to the same conclusion, how can Golberg’s partner’s ignorance be justified but not Goldberg’s? Fortunately, the pragmatic view has the resources to answer this question.

It is open for advocates of the justification thesis to challenge Rosen’s stipulation that Goldberg and his partner have the same stakes. Indeed, on the pragmatic view, Goldberg and his partner’s stakes are not the same because the costs of error are not the same. Goldberg’s partner has met the obligations that his (the partner’s) pragmatic circumstances required of him, and so he has addressed the risk of being wrong to an adequate degree. However, Goldberg’s promise ups the stakes for him. In virtue of his promise, there is now a cost of error for Goldberg that his partner does not share; Goldberg is not only answerable to his superiors as a detective, but also to Himmelfarb’s mother. Goldberg’s promise changes his pragmatic situation thereby changing the minimum standard Goldberg must meet. If he is to keep his promise, Goldberg must meet a higher standard than just those required of an ordinary detective before he does any accusing. His partner is under no such obligation.

Even so, Rosen allows that Goldberg and his partner are being held to different standards. The claim isn’t that they are being held to different standards (indeed they must be, if one is blameworthy and the other isn’t) the claim is that the difference between Goldberg and his partner is not “recognizably epistemic”. This brings us to my second point. If there is an obligation related to what Rosen refers to as one’s “epistemic life” and that obligation pertains to whether one ought to believe something, then it is hard to see why Rosen says such an obligation is “not recognizably epistemic.” Consider that Rosen’s account of justification (in the broad sense) is similar to the pragmatic view. Recall from Sect. 3 that he suggests there is a “broad sense of justification” that includes the requirement that a subject comply “with every pertinent procedural requirement in developing and assessing his evidence…” and he further grants that these procedural requirements “vary massively from case to case” (Rosen 2008, p. 601), which is just what the pragmatic view explains. But if Goldberg has failed to comply with a “pertinent procedural obligation” then it is hard to see why, even on Rosen’s account, Goldberg is epistemically justified, and yet still blameworthy.

To make this clear, consider the following question: what has Goldberg promised to do? According to the case, Goldberg has promised to “spend an extra day going over the case and digging more deeply into the evidence.” In other words, Goldberg is explicitly promising to do something more than would ordinarily be required of him before drawing a conclusion about Himmelfarb’s guilt, or at least, before using his belief as the basis for accusing Himmelfarb.²⁸ To put the point another way, what does Goldberg have to do to keep his promise? Evaluate the evidence in more thorough ways, think about the case more carefully, attempt to find additional evidence, consider possibilities that wouldn’t normally need to be considered (perhaps because they are exceptionally unlikely), and so on. But these are all things that even Rosen admits play a role in justification (Rosen 2008, p. 601). In fact, it is difficult to see what ways Goldberg can keep his promise that are not recognizably epistemic. Therefore, certainly on the pragmatic view, and apparently on Rosen’s own view as well, Goldberg’s case is not a counterexample to the justification thesis.

7 VII

In this paper, I set out with two aims. The first was to present a plausible version of the justification thesis. To accomplish this, I described what I take to be a plausible view of justification, namely the pragmatic view. I then explained how the pragmatic view lends plausibility to the justification thesis. The second aim was to defend this version of the justification thesis against Rosen’s two cases. With the pragmatic view in mind, I explained how the justification thesis can accommodate Rosen’s cases. Coupled together with its intuitive appeal, the pragmatic view’s ability to defend against Rosen’s cases gives us strong reason to accept the following conditional claim: at least if we understand justification in terms of the pragmatic view, then one is excused on the basis of ignorance only if one’s ignorance is justified, and one is culpable despite one’s ignorance only if one’s ignorance is not justified. One thing worth noting is that I have not argued that the truth of the justification thesis is dependent upon the truth of the pragmatic theory, though, it certainly speaks in favor of the pragmatic theory that it is connects our epistemic practices and our moral practices in such a natural way. In this paper I have set out only to test one plausible theory against Rosen’s examples, and it seems to work. I think it is certainly worth exploring whether other theories can do the job just as well.²⁹

One final thing I would like to note is that, if the justification thesis is right, it raises a number of further questions about the distinction between moral justification and excuses. According to Austin (1956), when an agent performs an action that appears wrong, the explanation of her action can render blame inappropriate either by showing that, despite appearances, the action was not wrong—the action was justified—or some other considerations render blame inappropriate—the agent has an excuse. On the justification thesis, ignorance only renders blame inappropriate when that ignorance is justified. It might be asked, then, why we should think ignorance counts as an excuse rather than serving to justify action. Some theorists have argued that a justified belief that one ought to perform some action guarantees that one is not required to do something else (Littlejohn 2012; Gibbons 2013). Others (Way and Whiting 2016) have argued that justified beliefs about what one ought to do must be true.³⁰ On these views, it looks like one’s justification for one’s ignorance also justifies one’s action done in/from that ignorance. But ordinary parlance treats ignorance as an excuse. Which is it? My sense is that an adequate answer to this question will need to explore whether the distinction between justified and excused action is as substantial as Austin thought. Unfortunately, due to space limitations and the depth of the issue, I cannot hope to say anything even remotely satisfying to address these questions here. I bring them up in closing only to acknowledge that, even if what I have said here is correct, there is still a lot of work to be done.