The aim of this paper is to vindicate the Cartesian quest for certainty by arguing that to aim at certainty is a constitutive feature of cognition. My argument hinges on three observations concerning the nature of doubt and judgment: first, it is always possible to have a doubt as to whether p in so far as one takes the truth of p to be uncertain; second, in so far as one takes the truth of p to be certain, one is no longer able to genuinely wonder whether p is true; third, to ask the question whether p is to desire to receive a true answer. On this ground I clarify in what sense certainty is the aim of cognition. I then argue that in judging that p we commit ourselves to p’s being certain and that certainty is the constitutive norm of judgment. The paper as a whole provides a picture of the interplay between doubt and judgment that aims at vindicating the traditional insight that our ability to doubt testifies our aspiration to know with absolute certainty.
The notion of absolute certainty has been almost completely removed from the agenda of contemporary epistemologists. There is a plethora of reasons why this is so. First, certainty seems impossible to attain. Second, it seems that most of our judgments can be justified and amount to knowledge even if they fall short of being certain. Third, and more generally, it seems that the Cartesian aspiration to possess absolute certainty is not some sort of inescapable philosophical task, but rather the contingent product of particular intellectual contexts that have framed much of philosophical reflection on knowledge in medieval and modern times.
In this paper, I wish to counter this tendency by arguing that the desire to possess certain knowledge is a constitutive feature of our inquiring mind. It is not some sort of psychological or philosophical compulsion that can be extirpated through proper philosophical analysis. Nor is it the product of a particular and objectionable philosophical Zeitgeist. To possess absolute certainty – the sort of doubt-free knowledge Descartes was looking for in his emph Meditations – is the emph constitutive aim of cognition. The aim of this paper is therefore to vindicate the traditional quest for certainty – to be found in Descartes, but also in Husserl’s phenomenological project, among others – by arguing that it is in the very nature of our inquiring mind to aim at certainty.
My argument hinges on three claims concerning the nature of doubt and judgment. In section 2 I clarify what I mean by doubt by connecting it to questioning, viz., to doubt whether p is true is to raise the question whether p. Then I defend the main claims. First, section 3, it is always possible to raise the question whether p in so far as one takes the truth of p to be uncertain. Second, section 4, it is not possible to wonder whether p if one takes oneself to know the answer to that question with certainty. Third, section 5, to ask the question whether p is to desire to receive a true answer. On this ground, in section 7 I explain in what sense certainty is the aim of cognition, where cognition is understood as the conscious activity of asking questions and answering them by forming judgments based on evidential grounds, that is, grounds that speak in favour of the truth of the content to be judged. Before that, I explain in section 6 the methodology adopted in this paper and argue that phenomenology is more fundamental than linguistic analysis when we inquire about the metaphysics of our conscious mental reality.
In the rest of the paper I clarify this claim, draw some consequences, and respond to objections. In section 8 I consider the objection according to which certainty can’t be the aim of questioning as our questions can be answered by forming uncertain judgments. I reply that certainty is the commitment of judgment: in judging that p we commit ourselves to p’s being certain. The argument hinges once again on the nature of questioning, and crucially rely on the observation that if certainty is not the commitment of judgment then it is impossible to explain why it is unintelligible and impossible to judge that p and wonder whether p is true at the same time. In section 9 I argue that certainty is a constitutive norm of judgment – a judgment is correct only if certain – on the ground that certainty is both the aim of questioning and the commitment of judgment. Finally, in section 10 I consider an objection according to which some doubts are irrational and as a result in some cases to aim at certainty is irrational as well. I reply by considering several ways in which our doubts might be irrational and by arguing that these cases are compatible with the claim that certainty plays a threefold constitutive normative role for cognition: certainty is the aim of questioning, it is the commitment of judgement, and it is the norm of judgment. The paper as a whole provides a picture of the interplay between doubt and judgment that aims at vindicating the traditional insight that our ability to doubt testifies our aspiration to know with absolute certainty.
My argument crucially relies on the nature of questioning and the act of resolving one’s question through the formation of a judgment. Before presenting the argument, I shall clarify what I mean by questioning and by resolving one’s question.
By questioning I shall refer to the phenomenon of consciously asking the question whether p. The phenomenon I have in mind is when we ask yes/no questions like ‘Am I identical with my brain?’, ‘Is there anything I know with absolute certainty?’, ‘Will my friends win the game?’, and so on. These are what we typically describe as doubts, for doubts posit alternatives: either something is the case, or not. Yes/no questions are particular kinds of whether questions, that is, questions that present a finite number of direct answers. Whether questions come into two varieties: yes/no questions and questions that present two or more alternative direct answers other than yes and no (e.g., ‘Is red, green or blue your favourite colour?’). Whether questions in general should be distinguished to all those questions that do not present a definite set of alternative direct answers. These are questions like ‘Who am I?’, ‘What do I know?’, ‘Why do I care about truth?’.
All questions have in common a phenomenological core: to ask a question consists in being open-minded as to how things are, and it represents a sort of desire to know how things are. In asking the question ‘what am I?’ I want to know how things are with respect to my nature, and in asking the yes/no question ‘am I my brain?’ I want to know how things are with respect to the relationship between me and my brain. In both cases there is a desire to know something that one doesn’t know yet. I shall argue that the desire that our questions express is a desire to know the truth with absolute certainty. In what follows I shall concentrate on yes/no questions only, that is, on questions that ask whether a specific proposition p is true. However, the arguments I offer below are meant to apply mutatis mutandis to all kinds of questions, for they rely on the nature of questioning itself, regardless of its specific varieties.
2.2 Resolving a Question by Answering It
To resolve the question whether p is to end up in a position in which it is no longer intelligible, from the first-personal point of view of the inquirer, to have the question whether p. There are several ways in which we can resolve a question.
The canonical way of resolving a question is by answering it. When we ask the question whether p, we are looking for an answer to the effect that p is the case or not. Thus, to answer one’s question is either to come to judge that p, or to come to judge that not-p.
One might come to answer one’s question in various ways. The most direct way is to judge that p or to judge that not-p. There are some indirect ways. To illustrate, one way is to deny a presupposition that is displayed in the content of the question. If I ask ‘is my car in this street?’, but I am remembered that I have no car, then I have resolved my question by rejecting the presupposition that I have a car. Another way is to suppose that p and not-p is the case, or that neither p nor not-p is the case. To so judge resolves one’s question, for the judgment denies what seems to be presupposed when we ask whether p: that is, that either p or not-p is the case. For simplicity’s sake, in what follows I shall concentrate on direct ways of answering a yes/no questions by either judging that p or by judging that not-p.
2.3 Resolving a Question without Answering It
There are ways of resolving one’s question without answering it. One way is to have a definitive suspension of judgment about the truth-value of p. The kind of suspension of judgment I am speaking about here consists in forming a judgment about one’s grounds. There are different kinds of judgments about grounds that can be meaningfully said to be ways of suspending judgment about p. To illustrate, I can judge that right now my grounds are not sufficient to favour p over not-p while leaving it open whether it is in principle possible for me to have conclusive grounds for judging that p rather than judging that not-p. This is the standpoint of a temporary agnostic. A temporary agnostic hasn’t resolved one’s question, for it is still open to her the possibility of knowing the answer.
A question can be resolved by occupying the standpoint of a definitive agnostic, that is, by judging that it is impossible in principle to have grounds that favour one option over the other. By judging that it is in principle impossible to have grounds that favour p over not-p, the definitive agnostic ends up in a position in which it is no longer intelligible for her to keep searching for the truth on this topic. This way of resolving one’s question rejects a presupposition that is arguably concealed in the act of questioning: namely that it is possible to have grounds for favouring one option over the other. However, and this is the crucial point, to have a definitive suspension of judgment significantly differs from the other ways of resolving one’s question because there is also an important sense in which the question whether p is wide open for the definitive agnostic: the definitive agnostic doesn’t know whether p is true or not, and this is why she can keep asking the question whether p, even though the questioning attitude is no longer accompanied with the quest to discover its answer.
There might be other non-canonical ways of resolving one’s question. According to some views, we sometimes have the impression of asking a question, where in fact we are not. Some read Wittgenstein’s remarks in On Certainty as expressing the view that it is impossible to genuinely raise doubts about hinges, like the hinge that the external world exists. If this is true, then when we ask whether the external world exists, say, we might have the impression of being asking a question, but in fact we are not. To so conclude is to resolve the question because one judges that there was no question to begin with. This might represent another pertinent case (like suspending judgment) in which one resolves one’s question without answering it.
In what follows I will for the most part put on a side these non-canonical ways of resolving one’s question and will focus on cases in which one resolves a question by answering it directly, that is, either by judging that p, or by judging that not-p. This restriction is done for ease of exposition. However, the points I am going to make about this way of resolving a question applies to all kinds of questions and to all ways of resolving it. The key to understanding why certainty plays a normative role for cognition is to look at the nature of questioning and judgment. Since all forms of resolution of a question involve asking a question and forming a judgment, my arguments apply mutatis mutandis to all cases. Or so I will argue.
3 It is Possible to Wonder Whether p so Long as it is taken to be Uncertain Whether p is True
To possess certainty for judging that p involves at least the possession of conclusive grounds for p, that is, grounds that are incompatible with the possibility that p is false. So, if the truth of p is taken to be uncertain, then it is intelligible to think that p might be false. But if it is intelligible so to think, then it is still intelligible to ask whether p. If I am not certain whether the death of my body will lead to my death, I can intelligibly ask the question whether I will die when my body dies. In so far as I am not certain that p, there is a sense in which it is still an open question for me whether p is true or not. This sense is displayed in the fact that we are able to raise such a question when we realise that we lack certainty.
This claim – that it is possible to raise the question whether p so long as the truth of p is regarded as uncertain – is to be distinguished from the following claims. First of all, I am not claiming that one ought to raise the question whether p as soon as one becomes aware that p is not certain. I am only claiming that it is possible to raise such question so long as p is regarded as uncertain.
Moreover, I am not claiming that in judging that p is uncertain or that p might be false one is led by psychological necessity to form the question whether p. One might form such judgments and then turn her reflective gaze on other topics without thereby raising the question whether p.
I am not even claiming that it is possible to wonder whether p only if one is at the same time judging that p is uncertain or that p might be false. Most of the time we ask questions without first judging that we are uncertain about their answers.
Finally, I am not claiming, absurdly, that it is impossible to have uncertain judgments. Most of our judgments are at best fallibly grounded, that is, formed on the basis of grounds that are compatible with the proposition judged actually being false.
All I am claiming is that it is possible for the sort of cognitive agents that we are to raise the question whether p so long as we take it that we are not certain whether p is true.
4 It is Impossible to Wonder Whether p so Long as we take it that it is Certain Whether p is True
The second observation is that once we think we are certain that p, we can’t no longer wonder whether p. By taking yourself to possess certainty that p is the case, you understand that you can’t be wrong about the truth of p. But by wondering whether p you are thereby taking it as an open question whether p is true or not, and so you are taking it that p might as well be false. This is why you can’t at the same time comprehendingly judge that the truth of p is certain and doubt whether p is the case.
This claim needs two important qualifications. First, we should appreciate the distinction between seriously entertaining a doubt as to whether p is true and merely pretending to be doubting or voicing the doubt. Suppose you have thought carefully about (some version of) Descartes’ reasoning and you agree with him: the proposition that I am thinking is certain so long as it doubted or judged by someone. Now, even if we are certain that we are thinking now, we can still pretend to have a doubt about it by uttering the corresponding doubt out loud or in our foro interno. It is like uttering a sentence without genuinely asserting it. In this case we are saying something but we are not judging that the content said is true. My claim should then be restated by taking into account this distinction: once we think we are certain that p, we can’t no longer seriously wonder whether p.
Second, sometimes it seems that we can seriously entertain a doubt as to whether p even if in some sense we already take the truth of p to be certain. Consider again the case in which we are considering Descartes’ reasoning. We might keep seriously wondering whether I am thinking now with the underlying intent to fully understand Descartes’ point. In this case we are relying on our doubt in order to put ourselves in a position in which we can apprehend the ground that justifies with certainty the cogito. But the crucial point, for our purposes, is that in this case the doubt is silenced as soon as one becomes aware of the ground that make it certain that I am thinking now.
5 Questioning Aims at Truth
The third observation is that a question has an aim, or displays a kind of desire.
To wonder whether p is to want an answer. I ask whether p because I want to know whether p is true. This explains why we can’t ask whether p is true when we judge that p – because to judge that p is to be in a position where the aim or desire expressed by the corresponding question is taken to be already reached or satisfied. When I ask whether p is true, I want to know whether p is true, and to judge that p is to take it that p is true. In judging that p I take myself to possess the thing that I want when I ask whether p is true. Wondering whether p is true is being open-minded as to whether p is true, but to judge that p is to close one’s mind on the issue whether p is true.
The point can be further appreciated by looking at examples. It is useful to distinguish between cases where a question is directed to a new issue, and cases where a question is directed to an old issue – that is, respectively, a question about a proposition one has never entertained before, and a question about a proposition one has already entertained and with respect to which the judger has formed a belief.
Consider old issues first. Suppose we are considering a case where I firmly believe that p because in the past I wondered whether p and concluded upon some reflection that p is true. Having this belief that p involves at least having a disposition to judge that p when presented with p. Thus, by so believing that p I will judge that p as soon as I bring my mind to this issue, and as I am judging that p I am not wondering whether p is true. However, the question whether p might be prompted by the recognition that (or by the doubt that) the grounds that I have for p were misleading or bad. When I raise the question whether p I thereby stop judging that p.
Of course, doubting whether p is true might not be enough to remove my belief that p. The doubt might last a very short span of time, and I might come back to my previous belief – not only if I end up judging that my grounds were good enough, but also if I am lazy and I stop to ask the question and to push forward the inquiry about p. This of course happens in most cases since it is often painful to abandon one’s convictions. However, my point is that questioning and judgment are not co-tenable, even if questioning and belief might be co-tenable – when belief is understood as a dispositional mental attitude and is not confused with judgment, that is, the conscious act of endorsing p as true.
Consider the case of new issues. Here we might think of at least two different important cases: (1) the judger hasn’t previously formed a belief that p (this is why the issue is new), yet she has a disposition to judge that p when she is presented with p, for p is easily found to be true given the judger’s other previously formed beliefs; (2) the judger has no belief that p, nor does she have any disposition to judge that p when presented with p. Case (2) is straightforward: since the judger doesn’t have any opinion whatsoever about the issue she is wondering about, it is clear that she is not judging any of the candidate answers to her question. Case (1) is more subtle. Consider a judger in case (1) who asks: do I have more than one hundred hair? When she raises this question, she is not already judging that she has more than one hundred hair. Yet, since the truth of this proposition follows obviously from many other beliefs she possesses, she has a disposition to judge that she does indeed possess more than one hundred hair. Her disposition might help the judger to quickly resolve the question by judging that she has more than one hundred hair. As soon as she so judges, the issue is closed for her, and she is no longer wondering whether she has more than one hundred hair. But even if a disposition to judge might quickly remove one’s question, the point remains that when there is the question whether p, there is no judgment that p, and when there is the judgment that p, there is no question whether p.
If it were possible to judge that p and to raise the question whether p at the same time, then we would have an indication that a question doesn’t aim at being answered – for, if that were its aim, we would expect a question to disappear as soon as its aim is taken to be reached. But the two things are not co-tenable, and this impossibility is nicely explained precisely by the fact that the desire to get the truth that a question expresses is satisfied in the very act of answering it.
When we say that the aim of questioning is to have an answer we are not saying that its aim is to have any answer whatsoever. The aim is to have the right answer, and the right answer is one that is true. Of course, this is compatible with one’s judgment being false. Yet, one can’t answer one’s question unless one takes oneself to possess the true answer to one’s question, and being in a position in which one takes oneself to possess a true answer just is to have formed a judgment. The point is trivial, but it is worth pointing it out: answering a question is to form a judgment, and one can’t judge that p if one takes p to be false, for judging that p is to regard p as true. Another way of putting the point is to say that a question posits a true answer as its form of satisfaction.
6 Phenomenology and Linguistic Evidence
The claims defended so far are grounded on phenomenological considerations: we observe the dynamics of conscious cognition by being conscious cognizers and we observe that
we can doubt whether p is true if we take the truth of p to be uncertain;
we can’t seriously doubt whether p is true if we take the truth of p to be certain;
in asking whether p is true we want to know whether p is true;
we can’t seriously doubt whether p is true if we are judging that p.
The appeal to phenomenological grounds is apt when the claims we want to ground are about the nature of our conscious mind. However, there is a widespread tendency in contemporary debates to appeal to intuitions about the appropriateness of exchanges in our linguistic practice in order to reach conclusions about the nature of the mind. We might then want to provide further supports to the points just made by looking at the following putative linguistic data.
It seems odd to assert the following:
It is certain that it is raining, but I wonder whether it is raining
It is raining, but I wonder whether it is raining
The oddity is naturally explained by noticing that people are sensitive to the fact that (ii) it is unintelligible to raise a question while one thinks to know the answer with certainty (case A) and (iv) when one thinks to know the answer to that question (case B).
Yet, there is a way of speaking about belief that might seem to oppose claim (iv). Consider
I believe that it is raining, but I wonder whether it is raining
Is this odd? I think it obviously is if what one is doing here is just saying that p and that she is wondering about it. There is however a couple of cases in which one can speak felicitously using (C). These are cases in which one is conveying the information that she is inclined to think that p though she is not yet sure that p. It is a way of expressing the fact that one takes oneself to possess some good grounds for thinking that p is the case, while also communicating that one still isn’t sure about it and so wonder about it. This is entirely compatible with claim (iv), because this manner of speaking just shows that judging that there are good grounds for p is compatible with wondering whether p is true.
Another way of using belief that makes the above cases felicitous involves the report of one’s dispositional beliefs. So, consider a person who goes to the psychoanalyst and finds out that she believes that her husband is unfaithful. She finds out this belief of her by observing her behavior, the things she said to the psychoanalysts, and similar evidence. Yet, she is not fully endorsing her recalcitrant belief, and this is why she can happily say that though she believes that her husband is unfaithful, she is still wondering whether he really is.
I do not claim to have exhausted all the ways in which instances of (C) can be found felicitous, nor have I considered all manners of speaking about judgment, belief, question and doubt that might be taken to challenge (or support) claims (i)-(iv). However, even if it were proven that it is indeed felicitous to convey the information that, say, one is both judging that p and doubting whether p is the case, or that one is both taking p to be certain and wondering whether p is true, still I will maintain on phenomenological grounds that it is strictly speaking impossible to be in these mental states. Our ways of speaking about mental reality might rely on false presuppositions and the only way to know whether this is so is to look at the phenomenology of cognition and check whether the way in which cognition works cohere with the way in which we implicitly think cognition to work. At any rate, for the purposes of this paper, I shall stick to phenomenology, although we should of course integrate and further evaluate the claims defended here with evidence provided by linguistic analysis.
7 Questioning Aims at Certainty
Now we have all the resources to explain in what sense questioning aims at certainty. In asking a question we want to receive an answer (iii). By answering a question – that is, by judging – we loose the question, as it is evinced by the fact that judging and questioning are not co-tenable (iv). However, the question might be reopened if the subject takes her judgment to be uncertain (i). This shows that certainty is necessary in order to silence one’s doubt in a definitive fashion. Moreover, certainty is also sufficient for this purpose, for if the subject takes her judgment to be certain then she is no longer in a position to seriously doubt its truth (ii). The desire expressed by our questions and doubts are not satisfied if we know we lack certainty, and they are satisfied if we know we possess certainty.
It is important to appreciate that the way in which certainty and truth relate to the act of questioning is different. When I ask whether p, I am asking whether p is true. The question whether p is transparent to the question whether p is true. However, the question whether p is not transparent to the question whether p is certain. If I am asking whether p I am not thereby asking the question whether p is certain, as it is evinced by the fact that to answer positively to the former is not to answer positively to the latter. This is why truth and certainty should be regarded as the aims of questioning in two different ways. Truth is what we might call the internal aim of questioning, for it is internal to the understanding of what it takes to wonder whether p that one is wondering whether p is true. Certainty is not the internal aim of questioning, and yet in questioning we aim at certainty because unless we have certainty we can keep raising our questions, and only if we have certainty we can definitively satisfy the desire expressed by our questions.
The ways in which truth and certainty function as the aims and desires of questioning are coordinated. On the one hand, we desire certainty because we desire the truth. To possess certainty is to be in a position in which one is certain to possess the truth, and so to possess certainty is to be sure that the desire for the truth has been satisfied. On the other hand, we possess the truth only if we do not doubt it, and doubting is the expression of our recognition that we lack certainty. In asking a question we want to know the truth. To answer the question whether p by judging that p is to be in a perspective such that p is the case. This is the perspective in which the corresponding desire for the truth is taken to be satisfied. However, this satisfaction is lost as soon as a doubt is raised, and a doubt can be raised so long as one takes oneself to lack certainty. Thus, although we desire certainty because we desire the truth, our desire for the truth is definitively satisfied only if we take ourselves to possess certainty.
Summing up, there is nothing objectionable in the traditional quest for certainty. For this quest is the quest of cognition itself. For, again, if you wonder whether p is true and you know you lack certainty about the truth of p, then the desire for the truth that your question expresses is not satisfied and won’t be satisfied unless you take yourself to know the answer with certainty. You might stop to inquire about p if you are distracted or you loose interest in it. But so long as you care about whether p is true, your inquiry won’t be closed unless you take it that you possess certainty.
8 To Judge that p is to be Committed to p’s Being Certain
I have argued that questioning aims at certainty. This claim is prima facie puzzling. We can answer our questions without judging to possess certainty. In fact, most of our judgments are at best fallibly grounded. But in judging that p we do resolve our questions, albeit provisionally. How does this fact cohere with the claim that certainty is the aim of questioning?
A conservative reply will insist that we should just distinguish two notions of aim. Questioning aims at truth (what I previously described as the internal aim of questioning), and does not aim1 at certainty, and this explains why a fallible judgment is sufficient to answer one’s question. However, questioning aims2 at certainty, but does not aim2 at truth, for if the judgment is judged to be uncertain then the question is open again, and the question will be silenced in a definite fashion only if the answer is taken to be certain.
Although I think that this reply is sufficient, I think there is room for exploring a stronger reply. In what follows I shall argue for the view that to judge that p is to be committed to take it that p is certain and on this ground I will explain how certainty plays a normative role even in provisional or temporary resolutions of questions. On this view, every resolution of a doubt is a definitive one. For, from the perspective of the judger who judges that p, the world is such that p is the case, and by so judging the judger is excluding the possibility that p might actually be false, which is tantamount to be committed, albeit implicitly, to take it that it is certain that p.
In order to give prima facie plausibility to the claim that judgment is committed to certainty I will start with a defence of less controversial claims about the commitments of judgment. First, to judge that p is to be committed to there being good grounds for p. Suppose that to judge that p is not to be committed to the possession of such grounds. If this is so, then it seems that there must be nothing unintelligible or irrational in judging at the same time that p and that there are no good grounds for p. However, this is neither intelligible nor rational. There is something prima facie Moore-paradoxical in judging that p and that there are no good grounds for p. For, if there are no good grounds for p, then from the first personal point of view of the judger it is entirely arbitrary to regard p as true (as opposed to any other proposition incompatible with p). This provides evidence for taking it that in judging that p one is thereby committed to take it that there are good grounds for p.
The same point can be appreciated by looking at the interplay between questioning and judging. Judging that p is not co-tenable with asking whether p. However, judging that there are no good grounds for p is co-tenable with asking whether p. According to the picture that denies that judging that p commits one to judge that there are good grounds for p, it is neither unintelligible nor irrational both to judge that p and to judge that there are no grounds for p. But this can’t be right. For, in answering a question by judging that p I am thereby resolving the question whether p, whereas in judging that there are no grounds for p I am in a position in which it is perfectly intelligible and rational to wonder whether p is true. The view under evaluation lacks the resources to explain why it is irrational and even impossible for a judger to resolve her question whether p by judging that p while she at the same time judging that there are no good grounds for p and also wondering whether p is true.
Second, to judge that p is to be committed to p’s being known. Suppose that to judge that p does not commit the judger to judge that she knows that p. If there is no such commitment, then it seems that it must be entirely intelligible for a judger to hold that p and that she does not know that p. However, there is something prima facie Moore-paradoxical in holding that p and I don’t know that p. If I judge that I do not know that p, then I understand that so far as I know p might be false. But, again, there is something Moore-paradoxical in holding that p and that p might be false. When I judge that p I am excluding the possibility that p is not the case: for to judge that p is precisely to take it that p is the case.
To appreciate this point we should look at the connection between judgment and questioning. To ask whether p is not co-tenable with judging that p. However, it is entirely possible and indeed perfectly intelligible for a judger to ask whether p while judging that she does not know that p. These two mental acts aren’t just co-tenable. It is co-tenable to ask whether cats are animals while judging that ice creams are tasty. These two acts are co-tenable because they are completely unrelated. But the question whether p and the judgment that I don’t know whether p are not unrelated: the judgment that I don’t know whether p captures the ground that must be presupposed in order to make it intelligible to ask whether p is true. For in asking whether p is true I am desiring to know whether p is true, and there is no point in my desire if I already take it to possess what I want. Thus, there is an intimate connection between the judgment that I don’t know that p and the question whether p. But this shows that there is a problem for the suggestion that judging that p and judging that I do not know that p are co-tenable. For, I can’t ask whether p when I judge that p, but I can ask whether p when I judge that I don’t know that p. However, this can’t be right if the suggestion under evaluation is correct: for, according to that suggestion, when I judge that p I can also judge that I don’t know that p, and when I judge that I don’t know that p I can also wonder whether p; however, when I judge that p I can not also wonder whether p, and thus according to that suggestion it is possible both to wonder whether p and to refrain from wondering whether p, which is not. This provides evidence for thinking that in judging that p I am thereby committed to take it that I know that p. Moreover, this conclusion is coherent with both the fact that judging that p answers and resolves one’s question whether p, and the fact that asking the question whether p is to want to know whether p is true. For, to satisfy one’s question by judging that p is to be in a position in which one takes oneself to know that p.
The same reasoning can be extended in order to vindicate the claim that to judge that p is to be committed to p’s being certain. If judging that p is not to be committed to take it that p is certain, then it shouldn’t be impossible or unintelligible to judge that p while at the same time judging that p might be false (or that the truth of p is uncertain). However, and this is the crucial point, whereas judging that p is not co-tenable with questioning, judging that p might be false is co-tenable with questioning. Indeed, since it is unintelligible to judge that p is certain and to wonder whether p, to judge that p might be false (which is tantamount to take it that the truth of p is uncertain) is a condition for the very intelligibility of the question whether p. But this leads to nonsense: the judger cannot at the same time have resolved the question whether p, and yet still have this question unresolved. The judger can’t at the same time wonder whether p is true and answering the question whether p by judging that p. Thus, we can’t intelligibly regard the judgment that p as being co-tenable with the judgment that p might be false (or that the truth of p is uncertain). And this fact is what is captured by saying that to judge that p is to be committed to p’s being certain.
9 Certainty as the Constitutive Norm of Judgment
If the previous claims are correct, then we have the resources to hold that certainty plays a twofold constitutive normative role for cognition:
Certainty as the aim of questioning: the question whether p can be answered in a definitive fashion only if one takes oneself to possess certainty that p.
Certainty as the commitment of judgment: to judge that p is to be committed to take it that p is certain.
On this ground we can motivate the view that certainty is the constitutive norm of judgment.
Certainty-norm: to judge that p is correct only if it is certain that p.
In order to appreciate the motivation for the claim that certainty is the constitutive norm of judgment it is useful to consider an analogous plausible picture of the way in which truth plays a constitutive normative role for cognition.
It seems constitutive of judgment to be governed by the truth-norm, that is, by the norm according to which a judgment is correct only if it is true. For the sake of pursuing an analogy between the truth-norm and the certainty-norm, it is useful to consider two plausible motivations for the view that truth is the constitutive norm of judgment.
The first motivation appeals to the interplay between questioning and judging. In asking a question we are aiming to possess a true answer. In this way, a question posits truth as the standard of correctness for its answer. But we answer our questions by forming judgments. Therefore, our judgments are correct only if they are true. It is in the nature of questioning itself to posit truth as the standard of correctness for judgments.
The second motivation appeals to the committal nature of judgment itself. In judging that p we are taking a stance as to how things are. Moreover, in judging that p one is thereby committed to take it that to so judge is correct, and since a judgment is a commitment as to how things are, its correctness is to measured depending on whether it takes things as being as they really are. Therefore, it is part of the nature of judgement to posit truth as its own standard of correctness. And the commitment of judgment nicely fits with the fact that truth is the aim of questioning and that we answer our questions by forming judgments.
A parallel reasoning can be extended in order to provide two motivations for the claim that the certainty-norm is constitutive of judgment: one argument relies on the claim that certainty is the aim of questioning, and the other argument relies on the claim that certainty is the commitment of judgment. Taken together they offer a compelling picture of the relationship between questioning, judging, and certainty.
First, our questions are definitively answered only when the judgments that we form are taken to be certain. In this sense, it is as though it is in the nature of the dynamics of cognition itself that a judgment ought to be certain, for unless it is certain there is an important sense in which we haven’t reached what we wanted in asking our question.
Notice that it is not fit to speak of certainty as being the aim of judgment. The aim-talk makes sense in the case of questioning, for in questioning we want something that we don’t have yet – that is, an answer in the form of judgment. In judging, there is no thing that we want but we don’t have yet – judging that p is being in the state that was looked for in the state of question; it is, in a sense, to have already what one was looking for. Thus, judgment doesn’t aim at certainty. Rather, certainty sets the correctness standard, or the norm against which a judgment’s correctness is measured. If a judgment is taken to be uncertain, then it is hostage to a doubt, and when one raises the doubt the judgment is lost. It is as though an uncertain judgment were found wanting by cognition itself.
Second, judgment is committed to certainty. When I judge that p I am excluding that p might not be the case, for judging that p is to regard things as being such that p is the case. Thus, since in judging that p I am committing myself to take p as certain, it is a feature of judgment itself to posit certainty as its own norm. Again, this point is inferred from the fact that a judgment can be lost by raising a question when it is found uncertain. If certainty were not the standard of satisfactoriness for judgment, then we wouldn’t have an explanation why discovering that a judged p is uncertain will ipso facto put the judger in a position in which she looses her judgment that p. The fact that recognizing that p is uncertain makes it possible to wonder whether p and makes us loose the judgment that p highlights the fact that certainty is the standard that a judgment should respect in order to be found satisfactory by cognition itself.
The claim that the certainty-norm is constitutive of judgment can be seen as absurdly demanding. To remove this impression, it is important to keep in mind the following points.
First, the claim is not the implausible one according to which one ought to possess certainty about every proposition. The reason is that I do not care to know all truths, as I do not genuinely raise all questions. However, in so far as I do care about some issue – that is, in so far as I have questions about the issue – my judgments are correct only if certain, for only judgments that I take to be certain will remove my questions in a definitive fashion.
Second, and relatedly, the certainty-norm doesn’t make any demand whatsoever concerning the way in which one ought to conduct one’s own cognitive life. The certainty-norm doesn’t entail anything as to how one ought to act. Thus, it doesn’t entail that one ought to keep investigating about an issue unless the judgments formed are actually certain. These are questions for practical deliberation, and the certainty-norm doesn’t provide practical instructions. The certainty-norm merely highlights the fact that a judgment is in some fundamental sense incorrect if it is uncertain, as it is the mind itself that finds the judgment wanting as soon as it directs its gaze on it and takes it to be uncertain.
Third, and crucially, the claim is not that a judgment is subject to the certainty-norm only insofar as it is formed as an answer to one’s question, or insofar as it becomes targeted by one’s questions. The norm applies to all judgments, regardless of their aetiology. This might seem implausible. Sometimes we form judgments as a result of doxastic deliberation upon having raised questions, but often we form judgments quasi-automatically, without having first to form an explicit question to which our judgments are answers: perceptual judgments are often a case in point. It might seem exceedingly demanding to require that these judgments be certain in order to be correct. However, the claim should not look surprising. First of all, these judgments are lost as soon as the questioning gaze targets them. Moreover, if judging in general is committed to certainty then they do already display their own pretension to certainty and thus they are constitutively subjected to the certainty-norm. Compare again the certainty-norm with the truth-norm. We do not find surprising at all the claim that judgments that are formed unreflectively and quasi-automatically are incorrect if false. Their being formed unreflectively is not an obstacle to deem them as incorrect if false. Similarly, their being formed unreflectively should not be taken as an obstacle to their being incorrect if uncertain. To put it figuratively, any judgment belongs to the game of truth and certainty, even if it is not formed as a result of one’s desire to know the truth with certainty as expressed by one’s own questioning.
Finally, the certainty-norm captures (at least part of) the constitutive epistemic normative profile of cognition. There is then a dimension of epistemic evaluation that captures the constitutive norms of cognition itself. But epistemic evaluation might be broader, and it might feature standards of correctness that are not constitutive of cognition. Thus, accepting the existence of a constitutive certainty-norm is compatible with the existence of other epistemic standards according to which a judgment might be epistemically fine even if it falls short of being certain. To illustrate, a judgment might be fallibly justified – according to our best notion of internalist fallible justification – and thus in a sense it can be an epistemically good one even if it is in another sense an epistemically bad one as it doesn’t satisfy the constitutive norm of cognition. The same reasoning applies to externalist positive epistemic statuses, and to all epistemic statuses in general. Thus, to recognize that certainty is the constitutive norm of judgment should not jeopardize the other ways of evaluating the epistemic credentials of our judgments.
10 Irrational Doubts?
One might endorse the claim that certainty is the aim of questioning and the norm of judgment and yet downplay the normative role that certainty plays in our cognitive life by arguing that some of the doubts that we might have are irrational. On this ground, one might argue that when the question whether p is irrational then the corresponding aim to possess certainty is irrational as well. One might further argue that judgments that bear on the question whether p (paradigmatically, the judgment that p and the judgment that not-p) are not governed by the certainty-norm if it is irrational to wonder whether p. And if there are cases in which a judgment is correct (in the intended sense) even if it is uncertain, then certainty is not the constitutive norm of judgment.
In order to evaluate this line of objection I will distinguish a variety of cases in which a doubt might be deemed as irrational and I will claim that none of these cases can be used in order to downplay the constitutive normative role that certainty plays for cognition. We are going to explore cases that can be usefully classified according to the quality of the grounds for taking some doubt as irrational – that is, whether the grounds are certainty-conferring or not – and to senses in which the doubt is deemed as irrational – whether it is irrational for epistemic reasons or for non-epistemic ones. We can therefore envisage four kinds of cases:
certain ground for taking a doubt as epistemically irrational.
certain ground for taking a doubt as non-epistemically irrational.
uncertain ground for taking a doubt as epistemically irrational.
uncertain ground for taking a doubt as non-epistemically irrational.
10.1 Case (1a). To Wonder Whether p is Irrational because p is Known with Certainty
One obvious way in which the doubt whether p might be judged as epistemically irrational is when one judges that it is certain that p. This is a case in which one’s question has not only been answered – and thus resolved – but it has also been answered on grounds that one takes to be certainty-conferring – and thus the question has been definitively resolved. But the existence of such cases is hardly an objection to the claim that certainty is the norm of judgment – rather, these cases highlight the grounds for taking certainty to be the constitutive aim of questioning and the constitutive norm of judgment.
10.2 Case (2a). To Wonder Whether p is Irrational because p is True
Another way in which the doubt whether p might be taken as epistemically irrational is when one ends up judging that p and then argues that since p is true it is irrational to raise the question whether p. If the judger takes the truth of p to be certain, then this case is an instance of the previous kind of cases. If, however, the judger doesn’t take the truth of p as certain, then it is always possible for the judger to become aware that p is uncertain and thus to raise the question whether p. In that circumstance, when the question has been raised, the issue is open again and the doubt can’t be silenced as irrational by judging that p, for the truth of p is exactly what is targeted by the question.
10.3 Cases (1a) & (2a). To Wonder Whether p is Irrational because there can’t in Principle be Grounds that Favour p over not-p
As we saw in section 3, there are ways of resolving one’s question that do not amount to answering it. One such way is to suspend judgment about p. I distinguished between definitive and provisional suspension of judgment. When I suspend judgment provisionally, I am not deeming my question as irrational, for I am open to the possibility that there be grounds for answering my question.
The case of definitive suspension is trickier. One might object that in the case of a definitive suspension of judgment about whether p it is irrational to keep asking the question whether p precisely because one takes it that an answer can’t be found. However, this point doesn’t threaten the claim the certainty is the constitutive aim of questioning, nor the claim that certainty is the constitutive norm of judgment.
First, to reach a point where one suspends judgment in a definitive fashion is a way of honouring the aim of certainty. A provisional suspension of judgment is taken as a partial satisfaction of the aim of certainty, for one takes it that it is an open question whether it is possible for settling one’s mind one way or another. It is only by answering one’s question or by reaching a definitive suspension of judgment that one attains a point where no further moves are available as one takes it that everything has been done in order to reach one’s aim. To suspend judgment in a definitive fashion is thus a way to reach a point where one takes it that no satisfaction is possible, but it is also in another way to reach a point in which one is satisfied because she did all she could in order to honour the aim of questioning.
Second, it is not possible for a judger to (rationally) judge at the same time that p and that it is in principle impossible to have grounds that favour p over not-p, for by so definitively suspending judgment about p one is thereby committed to take it that she doesn’t possess grounds that favour p over not-p. So, in so far as one definitively suspends judgment about p, one is not judging that p, and thus, even if it is irrational to doubt whether p, in this case there is no judgment that p that fails to be constitutively governed by the certainty-norm. Moreover, if, ex hypothesis, the definitive agnostic were to judge that p, then she would no longer be a definitive agnostic, and thus the question whether p would no longer count as irrational on the ground that it is in principle impossible to have grounds that favour p over not-p. Therefore, the possibility of irrational doubts in the case of definitive suspensions of judgment does not threaten the claim that certainty is the constitutive norm of judgment, for in the case of irrational doubts there are no possible judgments that can count as correct and yet uncertain.
10.4 Cases (1a) & (2a). To Wonder Whether p is Irrational because there are Fallible Grounds to Judge that p
The most interesting and philosophically influential cases are those in which we judge that some doubt about p is epistemically irrational because we take it that even if we are not certain that p, still we do possess good, albeit fallible, grounds for judging that p. In fact, most contemporary theories of justification are fallibilist, and thus they take it that it is epistemically fine to hold a judgment even if one doesn’t possess certainty for it. A natural corollary of such views is that it is somehow epistemically misplaced to raise a doubt about a proposition if the judger has a fallible justification for it.
To evaluate this line of objection we must consider two versions of it. The first version is externalist in spirit and holds that even if from the first-personal point of view of the judger it is an open question whether p, still judging that p is justified for the judger, and thus the judger is wrong in doubting it, regardless of whether the judger takes judging that p as warranted. This externalist objection is beside the point: the inquiry conducted here concerns the constitutive norms of cognition as they are manifest in the phenomenology of cognition. Even if one’s judgment would be justified according to some externalist notion of justification, and even if as a result one’s doubt would be irrational according to a corresponding externalist notion of rationality, still the point remains that it is entirely fine to doubt whether p is true in so far as p is taken to be uncertain, for it is in the nature of cognition, as experienced from the first-personal standpoint, to aspire at certainty.
The second version takes seriously the first-personal perspective of the judger and insists that the judger herself might coherently deem her own doubt as irrational when she takes her judgment that p as fallibly warranted. However, this cognitive standpoint is unstable and defective. First of all, as the judger is doubting whether p is true, she is not judging that p, and hence judging that judging that p is warranted is pointless, as it is an open question for the judger whether p is true. Second, the judger might also raise a question concerning the epistemic status of the judgment that one’s judgment that p is warranted. If this judgment about warrant is put into question, then the judger loses the ground to judge that her initial doubt is irrational. And even if, ex hypothesis, the judger takes this judgment about warrant as certain, still the judger can coherently regard her judgment about warrant as certain while doubting whether p itself is true. This is because the possession of fallible grounds for p (or even the certainty to possess fallible grounds for p) is compatible with p actually being false. Thus, from the first personal perspective of the judger, there is no way in which a doubt about p can be taken as irrational on the mere ground that judging p is fallibly warranted.
10.5 Cases (1b) & (2b). To Wonder Whether p is Irrational for Non-Epistemic Reasons
There are then all sorts of cases in which we might judge that our doubts are irrational not so much for their epistemic properties, but for their non-epistemic ones. Some doubts might be irrational because they are useless or because they lead to suffering. However, this kind of non-epistemic irrationality doesn’t downplay the normative role of certainty. It might be prudentially irrational to raise some scientific and technological questions about atomic energy, say, if we know that answering them is likely to end up causing suffering to many people. However, in so far as one is asking these questions one is still aiming at certainty, and one’s answers are still answerable to the standard of certainty.
Other arguably frequent cases are those in which one judges that it is non-epistemically irrational to keep inquiring about some topic because it is enough to possess uncertain judgments about it. Thus, a judger might judge that she possesses fallible grounds for her everyday judgments, and even if she recognizes that her grounds are not certain, she might judge that it is pointless to keep inquiring about the issue, as she judges that this wouldn’t have any significant added value. Again, the existence of such cases represents no objection, as it leaves completely untouched the epistemic constitutive normative role that certainty plays for cognition.
I have attempted to vindicate the traditional quest for certainty by arguing that the aspiration to possess absolute certainty is a constitutive feature of our inquiring mind. To sum up, I have argued that certainty is the constitutive aim of cognition. More specifically, I have argued that certainty is the constitutive aim of questioning – in that we can doubt whether p is true if we take the truth of p to be uncertain and we can’t seriously doubt whether p is true if we take the truth of p to be certain – that certainty is the constitutive commitment of judgment – in that judging that p is to be committed to take it that p is certain – and that, on this ground, certainty can be regarded as the constitutive norm of judgment – that is, a judgment is correct only if certain.
If the quest for certainty is indeed inescapable for us, then certainty should be put back at the centre of our contemporary epistemological concerns.
An earlier draft of this paper has been presented in Bologna, in December 2018, at COGITO Research Center. I am grateful to all people in attendance for helpful comments and suggestions. Special thanks are due to Annalisa Coliva, Filippo Ferrari, Sebastiano Moruzzi, Nikolaj Pedersen, Tommaso Piazza, Daniele Sgaravatti, Robert Stern, Elena Tassoni, and Giorgio Volpe for their feedback on previous versions of this paper and on my PhD thesis Certainty and normativity from a phenomenological point of view, on which this paper is based. I am especially grateful to Franco Bertossa, whose work and teaching have inspired the ideas contained in this paper. Finally, I wish to thank the referees of this journal, whose comments and suggestions helped me to improve this paper significantly.
Beddor, B. 2020. “New Work for Certainty.” Philosophers’ Imprint 20 (8): 1–25.Search in Google Scholar
BonJour, L. 1985. The Structure of Empirical Knowledge. Cambridge, MA: Harvard University Press.Search in Google Scholar
BonJour, L. 2000. “Toward a Defense of Empirical Foundationalism.” In Resurrecting Old-Fashioned Foundationalism, edited by M. DePaul. Lantham, MD: Rowman & Littlefield.10.4324/9781003061038-34Search in Google Scholar
Cross, C., and F. Roelofsen. 2018. “Questions.” In The Stanford Encyclopedia of Philosophy (Fall 2020 Edition), edited by E. N. Zalta. https://plato.stanford.edu/archives/fall2020/entries/questions/.Search in Google Scholar
Fassio, D. 2015. “Belief, Aim of.” Internet Encyclopedia of Philosophy.Search in Google Scholar
Ferrari, F. 2018. “Normative Alethic Pluralism.” In Pluralisms in Truth and Logic, edited by N. Kellen, N. Pedersen, and J. Wyatt, 145–68. London, UK: Palgrave Macmillan.10.1007/978-3-319-98346-2_7Search in Google Scholar
Fumerton, R. 1995. Metaepistemology and Skepticism. Lanham, MA: Rowman and Littlefield.Search in Google Scholar
Kołakowski, L. 1975. Husserl and the Search for Certitude. St. Augustine’s Press.Search in Google Scholar
Railton, P. 1997. “On the Hypothetical and Non-hypothetical in Reasoning about Belief and Action.” In Ethics and Practical Reason, edited by G. Cullity, and B. N. Gaut, 53–79. Oxford: Oxford University Press.Search in Google Scholar
Velleman, D. J. 2000. The Possibility of Practical Reason. Oxford: Oxford University Press.Search in Google Scholar
Wittgenstein, L. 1969. On Certainty. New York: Harper & Row.Search in Google Scholar
Zanetti, L. 2018. “Certainty and Normativity from a Phenomenological Point of View.” (Dissertation Thesis), Alma Mater Studiorum Università di Bologna.Search in Google Scholar
© 2021 Luca Zanetti, published by De Gruyter, Berlin/Boston
This work is licensed under the Creative Commons Attribution 4.0 International License.