So, I should withhold assent from almost all of my beliefs. (i.e. I should cease to maintain these as beliefs.)
This argument implies that for any current sense-based beliefs, e.g. my belief that there is a blue screen in front of me, there is a possibility that it is mistaken and so I should withhold assent from it.
However, there is gap at the last step of the argument, signaled by the blank space. The next to last premise (and sub-conclusion) speaks of doubt, the conclusion speaks of withholding assent. What connects them? It is a good heuristic for reconstructing arguments that every major term in the conclusion must be addressed in the premises, otherwise the premises won’t imply the conclusion. So what is the missing premise? Clearly, it is Descartes’ criterion:
I should withhold assent (belief) from any proposition where there is any possibility doubt.
13a. This seems to be an argument worth challenging: the reasoning appears plausible and the conclusion is radically at variance with common sense. This recognition is one stage of reading a difficult text, especially a philosophical one. To comprehend it, you need to recognize how it aligns with your own beliefs. This is especially so when it does not align with common sense, since common sense beliefs are ones that Descartes presumably shared, despite the historical distance between us. Surely he too did not start to question whether he recognized a friend on the street because at some time he misidentified someone?
For all of us, Descartes' reasoning does not at all fit with our practice. Right now I both believe that there is a blue screen in front of me, based on visual experience. I would hold on to that belief quite willingly allowing that sometimes beliefs based on visual experience have been mistaken. As we noted, Descartes himself, in his ordinary life, presumably thinks the same way. So why does he claim otherwise?
14. To answer this question, it is good to look for the underlying motivation or project. Descartes' project is to search for absolute knowledge and absolute knowledge implies the elimination of any possibility of error. If one wants to believe that p, then one must eliminate the truth of every proposition incompatible with p. In the case at hand, a crucial difference between us and Descartes is that he accepts, and we do not,
My senses having sometimes deceived me generates a possibility for doubt (or error) in any other sense-based belief.
So stated, we might like to agree with it, since we allow, as in the discussion of fallibilism, that it is possible that our sense-based beliefs are mistaken. But this is no reason not to believe any one. So to avoid a purely verbal dispute let's qualify Descartes' claim as one about 'genuine possibility' [epistemic or metaphysical?], where genuine possibilities of doubt (or error) do undermine (proper) belief. Descartes then affirms (and we, like Descartes in his everyday mood, deny):
My senses having sometimes deceived me generates a genuine possibility for doubt (or error) in any other sense-based belief (a doubt that excludes proper belief or knowledge).
How is Descartes going to convince us that his senses sometimes deceiving him is sufficient not to trust them, when we already accept his premises and flatly reject his conclusion? How is he going to persuade us that his senses sometimes deceiving him count as genuine, rather than spurious, reasons to doubt? Undoubtedly, he will allude to his search for absolute knowledge and how that search enjoins on him a demand for absolute certainty. As a consequence, any or virtually any possibility for doubt is a genuine one.
But an immediate rejoinder is this: What should I trust more my vision or my abstract reasonings? Would not the latter be guilty of ros? Put differently: Why shouldn't we treat the radical conclusion of Descartes' argument as a reductio of its premises, even if we cannot say exactly where he goes wrong? The current deliverances of my senses, that there is a blue screen in front of me, is based on vison, which has worked so successfully in the past and whose reliability we understand from science. Descartes argument is based on contentious principles and abstract reasonings. A facet of this worry about what to trust more is pressed by Descartes himself when he continues by qualifying the scope of the above argument:
Yet although the senses occasionally deceive us with respect to objects which are very small or in the distance, there are many other beliefs about which doubt is quite impossible, even though they are derived from the senses--for example, that I am here, sitting by the fire, wearing a winter dressing-gown, holding this piece of paper in my hands, and so on. (p.13)
Still, rather than seeking the comfort of Descartes own doubt, let us try to see if we can make better sense of Descartes skeptical argument in regard to perception, whose scope still remains wide e.g. my recognizing a friend across the quad.
13b. Here we continue to the next stage of reading a philosophical text (to understand the gap between a philsopher’s claims or conclusions and our common sense beliefs). Given that the conclusions defy our ordinary beliefs, we could alter our beliefs accordingly. But this is a bad choice, even were it possible. For the alterations proposed are massive, since the range of beliefs dependent upon our senses is extensive. Moreover, that we are attributing beliefs that defy common sense to a distinguished thinker is some evidence that we hardly yet comprehend the reasoning. The Principle of Charity tells us to search for believeable, even if not true, assumptions which would make the defiance of common sense more plausible, and which assumptions are plausibly attributed to the philosopher. So the ascription is responsive to two factors: textually accurate and plausible--we cast doubt on our own attributions, if we ascribe to Descartes beliefs that are wild and lack credibility.
At the next stage then we need to uncover assumptions to ascribe to Descartes that would move us to construe possibilities of error that we think way-out as genuine possibilities to doubt?
We began this earlier (14) by looking for the connection between Descartes sceptical doubts and his basic project.
15. One line of reasoning is that if you can know that there is a blue screen in front of you by looking, and someone else does not, then there should be some difference in your evidence or grounds to explain the difference. This does not have much basis in Descartes' text except that he begins by observing that there actually have been times when he has gone wrong when using his senses, so how can he trust them in this particular case, when, presumably, he has no better evidence? The scope of the reasoning depends upon whether there are actual cases [only past or present?], where someone in a similar position was mistaken (about the same judgment).
The reasoning assumes that the agent (Descartes) can only have knowledge in one case, but not the other due to some internal difference--something in his evidence that is better now than then. Should we accept this assumption--call it the 'like cases alike' assumption:
If A can not know that p in epistemic position V, then B can not know p (or some proposition akin to p) unless his epistemic position V' is better than A's [mainly, where B has evidence to discriminate his case from A's]
Imagine that in one case he really sees an oasis and in the other it is a mere illusion. Why cannot we say that there is a difference: in one case he is right and the other wrong, even if it is not a difference that he can discriminate. This is an externalist resolution, since the betterness of his position in one case, but not the other, is not available to Descartes. So, from his point of view, it remains accidental that he is correct compared to the other person. [But might the externalist go further to insist that the very possibility of Descartes' thinking of an oasis in one case (the successful) one requires the presence of the oasis, which is lacking in the other?]
16. The second line of reasoning has more of a textual basis in Descartes, though I shall illustrate it with a markedly non-Cartesian example. Descartes' project is to search for absolute knowledge--beliefs that are so firmly established as true that we can now (absolutely) know will never be overturned. How is such a search to be conducted--what criterion should we adopt so that when a belief satisfies that criterion we have absolute knowledge? Descartes recognizes that this is not an ordinary search: "the task now in hand does not involve action but merely the acquisition of knowledge." (p.15).
Whenever we undertake any inquiry, we are interested in truth: Is the proposition we are inquiring about true or not? But we have many other--practical--interests competing for our limited time and energies. So, you are inquiring as to what new bicycle you should purchase. Your aim is find out what truly is the best for you--that optimally meets all your aims: affordable, works well on the terrain you mainly ride on, light enough, proper gears and brakes, etc. You investigate these matters--reading up in various magazines, and even entering information explicitly about different locals according to your criteria. If you observe your own behavior with this kind of purchase, you will find, I expect, that you would stop investigating well before you have exhausted all the relevant information. There are bicycles in your price range that you have not examined at all, despite reasonable ratings. They were not carried by local stores, your test rides have been limited, and you have not spoken much to those in the know.
Still, I assume that your investigation was intense, and yet you decide before it is exhausted. What is left out is not squeezing the handbrakes one more time, after numerous previous ones, to check that they feel right. The possibility that the previous checks were just luck is a way-out, not a genuine, possibility. Here I am concerned only with genuine possibilities as that the gears do not shift as well on three of the top 6 bikes as the others. The point is simply that you have other things to do with your life--you have practical interests (e.g. to spend time with friends, to complete a work assignment, to paint your bedroom) regularly tugging on you, and the level of satisfaction of those practical interests diminishes as you invest more time and effort to satisfy your epistemic: to determine which is truly the best bike for you.
This second way to understand Descartes' extraordinary criterion of absolute certainty has more textual evidence, and it takes off from the earlier observation that you will devote more time to an inquiry as the truth matters to you more. Imagine the town mayor places a large glass jar on the desk filled with tiny colored balls. You win a prize if you are the first to guess correctly the number of balls: this is the analogue of Descartes' goal of absolute knowledge or truth. Let's represent the relation between the prizes and the efforts you are willing to place in determining the number of balls in a chart:
Prize Time Spent to Determine Number
$10 15 minutes
100 60 minutes
1000 2 hours
10000 5 hours
100,000 12 hours
1,000,000 24 hours
highest ?
The exact amount of the prizes and the corresponding time spent are irrelevant. What matters is simply that as the prize increases, you will spend more time and effort in determining the number of balls. In each case, what matters is your being right--that is your goal. But as the prize increases your willingness to invest more in being right increases. That is, you put aside more of your practical interests, e.g. to spend time with friends, when the prize goes from, $1000 to $10,000.
Now imagine that you are offered a prize as high as you want (“highest”). At this point, virtually nothing matters to you as much as being right. Then, correspondingly, you would put aside virtually all of your practical interests to discern the correct number. So as the prizes increase the force of your practical interests against your interest in just being correct, decreases. Correspondingly, your standard for being correct, increases. As you put more time in, you are eliminating more possibilities of error and raising your standard for belief. When the prize is unlimited, you are willing to invest most all of your time and energies in being correct--your standard approaches that of Descartes' absolute certainty, where you will not enter a determination unless there is (virtually) no possibility of error.
17. It is not evident that there is any inquiry that is pure, as Descartes requires. We understand the extrapolation to an infinite payoff, but does that extrapolation really correspond to any real demand on us as inquirers? Science is a paradigm of inquiry, and yet its research programs are very practically bounded. Scientists seek funding, and they must promise to have completed work by particular dates. Imagine a scientist investigating whether ravens are all black. Any raven not yet examined is a potential counter-example, and so examining it will confirm the hypothesis, if it is black. Yet, does it make sense for our scientist to continue to investigate into the far-future for the diminishing, but genuine, possibility that the next ravens will not be black? Research takes time and resources. If scientists did so act, little would be accomplished, and we would never realize such simple assertions as that 'Yes, all ravens are black'. So the question of extrapolating to the absence of any practical interests is required by the model, but questionable for its realism.
18. Nevertheless, the distinction between epistemic and practical interests (or reasons) is an important one, especially for one topic that motivated Descartes: The ethics of belief. What are the rights and wrongs of believing? Descartes proposed that one should not believe anything for which one could not be absolutely certain. Aside from that criterion being extremely strong, Descartes' formulation leaves it open how one should determine whether a belief is absolutely certain. On what basis should one seek to justify one's beliefs? We have noted, as Descartes would agree, that since belief aims at truth, one's reasons for believing should be reasons of truth, rather than practical or prudential reasons. One should believe that one is tall only if one has evidence that one is tall, not merely because one wants to believe it.
Section 2: The Ethics of Belief
1.The most famous debate over the question 'what ought one to believe?'--what are the rights and wrongs of believing--occurred in an exchange between W.K. Clifford and William James. Clifford argued that if a shipbuilder allowed a ship to pass inspection, and subsequently, the ship is destroyed due to a defect he neglected, then he is responsible for the death of the persons on board. The blame attaches to the shipbuilder's beliefs--he convinced himself that the ship was safe, under a desire to cut costs, and it is that belief that guided his action in approving the boat to sail. Clifford dramatically concluded:
It is wrong always, everywhere, and for anyone, to believe anything upon insufficient evidence.
Let us call the suggested view 'evidentialism'.
2. I say 'suggested' because as it stands it is unsatisfactory. One problem is that someone might be forced or tricked into holding a belief, and then, unlike the shipbuilder, Clifford would not blame him. So an emended version to handle this problem would be:
It is wrong always, everywhere, and for anyone, to believe anything upon what he believes to be insufficient evidence.
For a doctrine that we want to examine fairly, we should subtract off its rhetorical or dramatic devices, and try to express it at its most general and minimal (or simplest). By its most general, we get the clearest picture of what is being proposed--the key properties. Clifford has already handled this problem for us, since he does not offer us a formulation restricted to this shipbuilder:
It is wrong for the shipbuilder to have believed the ship was safe on insufficient evidence.
To stop with this conclusion would be misleading--suggesting that the wrongness here is restricted to the shipbuilder, when the point of the example is that it is the ignoring of evidence, which anyone can do (anywhere) that is crucial. So Clifford gives us instead:
It is wrong always, everywhere, and for anyone, to believe anything upon insufficient evidence.
But, for the sake of analysis, this formulation should be shorn of its embellishments. So let's take as our final (so far) version of evidentialism:
It is wrong to believe anything upon what one believes to be insufficient evidence.
3. James' response is complex. He begins by acknowledging a range of cases in which he and Clifford agree--in which James thinks evidentialism works:
Does it not seem preposterous on the very face of it to talk of our opinions being modifiable at will?...Can we, by just willing it, believe that Abraham Lincoln's existence is a myth, and that the portraits of him in McClure's Magazine are all of some one else?...We can say any of these things, but we are absolutely impotent to believe them; and of just such things is the whole fabric of the truths that we do believe in made up...(James 1948: 90)
But he thinks that there are many other cases for which it would not be wrong to believe without evidence. Specifically, he thinks this is true for many religious (and ethical) propositions like 'People are basically good' or 'God answers prayers'. For persons for whom these propositions are significant--believing in them will affect how they live--there is the possibility of belief without evidence. The two other conditions he adds for the possibility of belief without evidence are that the matter is (1) Not decidable by the evidence:
Our passional nature...must decide an option between propositions, whenever it is a genuine option that cannot by its nature be decided on intellectual ground;(James 1951: 95, his emphasis)
and
(2) 'Forced'. James observes that the choice of loving me or hating me is not forced because "your option is avoidable." However,
if I say, 'Either accept this truth or go without it', I put on you a forced option, for there is no standing place outside of the alternative. (James 1951: 89)
If you do not "accept this truth" then you thereby "go without it", and vice versa.
Given these further, and crucial restrictions (that the matter is essentially undecidable on intellectual grounds and that it is forced), James argues for the right to choose to believe that is meant to be a choice in the interests of the truth. But it is a choice that can be made rightly without evidence and so contrary to Clifford and evidentialism.
4. Earlier, James disparagingly refers to a famous argument of Pascal's for believing in God. Pascal argued that either God exists or not and either you believe in him or not. Since believing in him, when he exists, assures infinite rewards, while the other options offer nothing comparable or far less, you should believe in God. Pascal's reasoning can be represented, roughly, as follows:
God
Exists Does not Exist

Believe in God infinite blessing -(a little)

Not Believe infinite damnation 0
in God

Since the combination of God's existing and your believing in him yields infinitely more blessing than the your not believing when he exists, and since the time wasted if you believe, when he does not exist, compared to your not believing when he does not exist, is comparatively negligible, believing in God seems the right choice. Noticeably, it is the right choice regardless of how probable you think it is that God exists, so long as you do not regard it as impossible or close to it.
James disparages not the argument, but trying to have people believe for practical (or prudential) reasons (because it would be to their benefit), rather than epistemic reasons (supportive of God’s existence/blessing). He writes,
We feel that a faith in masses and holy water adopted wilfully after such a mechanical calculation would lack the inner soul of faith’s reality; and if we were ourselves in the plac of the Deity, we should probably take particular pleasure in cutting off believers of this pattern from their infinite reward.
So now James turns to offering an epistemic ground for coming to a belief, though a ground of choice, not evidence. He observes that
...There are two ways of looking at our duty in the matter of opinion...We must know the truth; and we must avoid error--these are our first and great commandments as would-be knowers...
If I am searching for an opinion on whether Bill is sociable in order to decide whether I should invite him to a party, and if my only concern is to avoid error, then I do best by never forming an opinion. I cannot go wrong. Clearly, this will not do. I must sometimes form an opinion, come to hold a belief, even if by doing so I risk error. If I seek the truth, I must be willing to risk error. As James suggests to adopt the policy of just avoiding error would be like adopting the policy that whenever possible, you do not put yourself at risk of rejection or embarrassment. If you adopt that policy, you'll never put yourself forward to meet someone, since you risk rejection. But we know from the friends and romances in our lives that the risk is often worth it, even if we suffer the occasional embarrassment (of rejection). So the right attitude in forming an opinion or a belief is a combination of caution and risk. Clifford, he thinks, is guilty of advising a caution-only policy. (Why is the demand for evidence a cautious policy?)
So James finally concludes that since intellectual grounds leave the religious question wide open, you cannot be faulted for answering it in the affirmative--believing. In so doing, you will provide yourself with the opportunity for discovering whether religious belief is true or not. You need the experience to decide. So James writes,
There are...cases where a fact cannot come at all unless a preliminary faith exists in its coming. And where faith in a fact can help create the fact, that would be an insane logic which should say that faith running ahead of scientific evidence is the 'lowest kind of immorality' (James 1951: 104-105)
And so he rejects the evidentialist rule:
a rule of thinking which would absolutely prevent me from acknowledging certain kinds of truth if those kinds of truth were really there, would be an irrational rule. (James 1974: 107)
5. James' argument as we have it so far seems to be this:
1. Assume that the religious hypothesis is significant for you.
2. There are two options: You can believe it or you can disbelieve it.
3. The evidence and intellectual grounds do not favor either of these options.
So, 4. it cannot be irrational for you to take either option, and you must take one.
5. Only by believing the religious hypothesis do you have the opportunity to discover if it is true.
6. So, it would not be irrational for you to 'will to believe' the religious hypothesis, and it has the potential advantage of providing the opportunity to discover whether it is true.
Some comments:
1. Missing assumption. Since each step in an argument is supposed to validly follow from previous ones, there is a problem in the step from 2 to 3. Even if the evidence and intellectual grounds do not favor either option, couldn't there be other grounds on which the choice is irrational? Couldn't it be irrational because harmful or unethical? When a step in an inference blatantly fails, it is a good idea to look for a hidden or missing assumption–-what must be added to the premises to make it credible that the writer would believe the conclusion. A plausible suggestion in this case is:
3.' The only kind of irrationality that is relevant is irrationality due to opposition to the evidence or intellectual grounds.
2. Incompatibility (inconsistency). The conclusion that it is acceptable (or permissible) for you to 'will to believe' without evidence is incompatible with Clifford's view (evidentialism) only if no evidence is insufficient evidence. For if evidence is simply inappropriate--as it might for whether it is permissible to decide to eat a hot fudge sundae--then the matter of sufficiency does not arise. So if no evidence is insufficient evidence, the incompatibility must be located among the following:
Clifford: (a) It is wrong to believe without sufficient evidence.
(b) It is sometimes wrong to believe without sufficient evidence.
James: (c) It is never wrong to believe without sufficient evidence.
(d) It is sometimes not wrong to believe without sufficient evidence.
(b) cannot be Clifford's position because of what he explicitly states ('always, and everywhere...'), and because James accepts it too. (Recall his remark about Lincoln's existence as a myth.). So (a) is really
(a) It is always wrong to believe without sufficient evidence.
So if Clifford holds (a), it will be incompatible with (c)and
(d), as well as
(e) It is often (mostly) not wrong to believe...
But we should focus first on the minimal incompatibility--the narrowest disagreement. Clearly, this will be (d), and not just because James explicitly implies rejection of (c). (c) and (a) cannot both be true (i.e. they are incompatible). But they are minimally incompatible (contradictory)--they cannot both be false either.
3. Shared Presuppositions. Even though James and Clifford sharply disagree, there is an important claim that they both accept. Both arguments presuppose that it is possible to will or to choose one's belief Clifford holds that it is wrong to believe without sufficient evidence, and he blames the shipbuilder for so believing. But this only makes sense if the shipbuilder believes on the basis of a choice. So I could state Clifford's view as "It is wrong to choose to believe (or choose to continue to believe).." The presupposition is more evident for James: You cannot will to believe unless willing to believe is a genuine possibility.
Finding shared presuppositions identify points of agreement between opposed view. These provide a common ground to advance discussion between the opponents. But it also opens up the possibility that the opposed positions are not exhaustive–an alternative position may be available, which rejects the shared presuppostion. Perhaps, as proposed below, one can hold that belief is not a matter of choice at all, and so dissent from both positions?
4. Forced Choice and the Law of the Excluded Middle.
Premise 2 is not generally true:
There are two options: You can believe it or you can disbelieve it.
In general, there are three options, you can believe, you can disbelieve (e.g., atheism), or you can hold no belief on the matter at all (e.g., agnosticism). You can 'suspend judgment', as you presumably do on the question 'are the number of blades of grass in central park even?' James attempts to rule out this option on practical grounds, so that the religious hypothesis is a forced option:
We cannot escape the issue by remaining skeptical and waiting for more light, because, although we do avoid error in that way if religion be untrue, we lost the good, if it be true, just as certainly as if we positively choose to disbelieve. (James 1951: 106)
If we fail to believe, and not just if we disbelieve, we do not act as believers. In not acting as believers, in not taking part in religious practices or rituals, we effectively lose any truth that the religion has to offer.
If the agnostic is no better positioned than the atheist to grasp religious truth, if it is true, then there is a (practically) forced choice. But neither attitude (belief/disbelief) can be favored or disfavored by the evidence. But we must take one. Since neither option is more rational than the other, James's resolution is that it is open to us to do either. (Much more on the question of negation and forced choice in the next section.)
To now draw the argument to a close, recall the conclusion stated above
6. So, it would not be irrational for you to 'will to believe' the religious hypothesis, and it has the potential advantage of providing the opportunity to discover whether it is true.
From (6), the further conclusion is meant to follow, now addressed by an individual to himself
7. So, since religious belief is significant for me and I care to know the truth about it, and the question is forced and not answerable by evidence, I ought to will to believe.
8. So, I do--I will to believe.
(7) is what I ought to do, and it is supposed to imply (8), which is not about what I ought to do, but what I will do. Evidently, there is an assumption between (7) and (8):
7.' If I ought to do A, if it is my duty, then I can do A. (Can it be my duty to save a drowning child, if I am unable to swim?)
What kind of association is there between "ought" and "can"? Is it that if I ought to do something, I'll do it, and so I can? No. Clearly, there are times when I do not do what I ought. Here’s a counterexample: I ought to mow the lawn today, but I do not. So there is a direct connection between ought and can: if I ought to do something then it must (at least) be possible for me to do it, even if I actually do not.
Why is this? How do we know? The thought seems to be along these lines: If I ought to do something e.g. drive on the right, then if I do not do it e.g. I drive on the left, I am blameworthy (especially if I cause an accident). But it does not make sense to blame anyone if he or she could not have done otherwise. So you can do--it must be possible for you to do--what you ought. Since the connection is between the concepts of 'ought' and 'can', it is a conceptual connection, implying a conceptual possibility.
6. James is arguing that, in the circumstances proposed, one ought to will to believe, and, given the conceptual connection just noted, it is appropriate to ask 'can' one? The motive to answer this question is that in many cases it does not seem possible to do what James, like Clifford, presupposes, and James himself provides an example. In section II, James writes
Does it not seem preposterous on the very face of it to talk of our opinions being modifiable at will?...Can we, by just willing it, believe that Abraham Lincoln's existence is a myth, and that the portraits of him in McClure's Magazine are all of some one else?...We can say any of these things, but we are absolutely impotent to believe them; and of just such things is the whole fabric of the truths that we do believe in made up...(James 1948: 90)
As he observes, and we all concur, you cannot just will yourself to believe this. Of course, this case does not meet James' other conditions--that it is significant for you, a forced choice, and evidentially indeterminate. Do these conditions make the difference?
Well, why cannot you just will to believe that Lincoln's existence is a myth? Pascal also makes a similar observation in just the setting where James wishes to distance himself from Pascal. Pascal imagines someone moved by his wager-argument, but worries as to how to accomplish what is advised (to believe that God exists):
‘I confess, I admit it, but is there really no way of seeing what the cards are?’–‘Yes. Scripture and the rest,etc.’–‘Yes, but my hands are tied and my lips are sealed; I am being forced to wager and I am not free; I am being held fast and I am so made that I cannot believe. What do you want me to do then?’
The natural explanation in James’ case for why you cannot bring yourself to believe that Lincoln’s existence is a myth is that since it is evident to you that he was a real person, you have overwhelming grounds that it is false that his existence is a myth, and so you cannot believe it (true). So the barrier here--the 'cannot'--is a conceptual barrier, not a mere inability. The barrier is due to the concept of belief's aim of truth, and the recognition of a violation of it. So in order for James argument to be consistent--for his admission that in many cases you cannot believe without evidence and for his recommendation to believe (in other cases) without evidence--these conditions must make a difference on the count of truth.
But do they? For my choosing to believe something, e.g. that there is an all-good God, does not have anything to do with its being true. My choosing to believe it is not an epistemic reason for it. [How does this explanation apply to Pascal’s response to the willing, but unable, potential believer?]
James might say that the issue is not establishing that it is true, but placing oneself in a position to discover the truth. Does this help? Perhaps I can never discover the truth about whether Jane will like me unless I build up confidence to talk to her. But the confidence requires that I believe that she will like me. (Can I ever succeed, if the belief requires that I spend time close to her and I will not spend time close to her unless I have the belief?) Can I just decide to believe this, and at the same time, recognize that my decision is the basis for my belief? (Try it, in an analogous case, for yourself.)
Would not my real approach be to try to convince myself that she will like me, while trying to hide from myself that this is what I am attempting? So I might focus on all the people who like me, and not attend to all the people who have ignored or scoffed me. Why do I have to go through this roundabout route except that I cannot simultaneously take myself to believe something and that my reasons for believing it are not reasons that make it true? They are not epistemic reasons. Recall James' discussion of Pascal: Pascal responded to those who accepted his 'wager', not by telling them to just will to believe that there is a god–he quotes, recall, someone who wants to do just that, but cannot.
Instead, Pascal recommends following those who engage in religious practices, and thereby, as a psychological matters, we expect that he belief will come. Here we have a roundabout way to induce belief–specifically, for it to work, one must forget why one undertook to follow the religious practitioners. Why must it be roundabout (ones that I attempt to do, while trying to hide from myself that I am attempting it)?
Belief’s aim of truth is satisfied only if the belief is true. Although the best way to realize that aim is typically through investigation or inquiry, the difference between believing correctly, which is simply a matter of the belief being true, and properly believing, which is a matter of applying appropriate methods, remains. Even if James was correct that the only way to know whether religious claims are true is through experience as a believer, this would not meet the aim of belief itself. The analogy with trust breaks down at this point. One is trusting, or extends oneself to another, in the hope of reciprocation, perhaps even in order to inspire it. The trust may be well founded, but one can have a trusting attitude while well aware of one's limited chances of success. In contrast, belief's claim to truth applies in the very adoption of the attitude, it does not wait upon a future payoff.
The argument here assumed that since belief’s aim is truth to properly believe a proposition, one must have [epistemic] reasons establishing that it is true. As noted earlier, the ‘reasons’ are doing dual work–they are reflecting that there is a genuine or reliable process (or connection in nature) involved in the belief’s formation and its being true, but they are also providing the person with a justification or defense of his or her belief. Is the second really necessary? Does it confuse properly or rightly holding a belief, which may be just a matter of correct connections, regardless of the person’s awareness of them, with being able to show or justify that one properly or rightly holds a belief?
Section 3: Scope, Negation, and Belief
1. Previously, in regard to James’ forced choice premise, I wrote:
Premise 2 is not generally true:
There are two options: You can believe it or you can disbelieve it.
In general, there are three options, you can believe, you can disbelieve (e.g., atheism), or you can hold no belief on the matter at all (e.g., agnosticism). You can 'suspend judgment', as you presumably do on the question 'are the number of blades of grass in central park even?'
In this section we want to understand better why there are three options, and why it is easy to overlook the third option. The problem is to read ‘believing not’ or ‘believing false’ or ‘disbelieving’ as simply ‘not believing’ (having no belief-attitude on the matter):
John believes that Marcia loves him.
However, Marcia does not give him a present on his birthday, and so John starts to doubt whether Marcia loves him. Should we say:
John does not [no longer] believes that Marcia loves him.
Or
John believes that Marcia does not love him [i.e. John believes that it is false that Marcia loves him]
The first is much weaker and more appropriate: John starts to have doubts, and so withdraws from all-out believing that Marcia loves him. But he does not conclude that she definitely does not love him, unless he is neurotically insecure. Maybe she was just terribly pre-occupied, and forgot his birthday.
In the following dialogue, the question is can B's position consistent?
A: Abortion is impermissible.
B: I don't believe that.
C: Abortion is permissible.
B: I don't believe that either.
When I have posed this question, many students answer 'no'. B either believes that abortion is permissible or that abortion is impermissible. So for B to deny agreement with both A and C is inconsistent.

The answer cannot be right. For a hint at what’s bothering me: Isn't there a difference between students who raise their hand to answer 'no', students who raise their hand to answer 'yes', and students who do not raise their hand at all?
Do you believe
(1) There are an odd number of atoms in the Washington monument.
No? Ok, do you believe
(2) There are an even number of atoms in the Washington Monument.?
But out of (1) and (2) we are able to construct a dialogue which parallels the opening one with you as B*:
A*: There are an odd number of atoms in the Washington Monument.
B*: I don't believe that.
C*: There are an even number of atoms in the Washington Monument.
B*: I don't believe that either.
Not only can B*'s position be consistent, but it mirrors each of your own:
I do not believe that the Washington Monument contains an even number of atoms and I do not believe that it contains an odd number of atoms.
And it is false for each of you that
Either I believe that the Washington Monument contains an even number of atoms or I believe that it contains an odd number of atoms.
(What's the logical relation between these last two reports on what you believe?)
The consistency of B* is due to a lack of evidence one way or the other. B* just holds no belief on the matter. Besides belief and disbelief, there is the possibility of no belief. We drew this connection earlier in the discussion of James. He had to argue that the choice between belief in god and athiesm was forced because he recognized that logically they are not forced. Logically, there has to be the third option of non-belief or agnosticism.
2. An analogous third option applies to relations. Think of the relation of symmetry. Some relations are symmetrical: if Bill is the sibling of Mary, then Mary is the sibling of Bill. (aRb==>bRa). Other relations are asymmetrical (or anti-symmetrical): If Colleen is taller than Oscar, then Oscar is not taller than Colleen (aRb==> not(bRa)). But there is another possibility: non-symmetry. If Romeo loves Juliet, Juliet may or may not love Romeo. (Not (aRb==>bRa)). Asymmetry permits a definite implication or conclusion (that b is not the brother of a); non-symmetry does not permit a definite conclusion (b may or may not be the brother of a) With asymmetry a relation is denied, with non-symmetry an implication is denied. (Other applications: In jury trials, is there a difference between finding a defendant innocent and finding him not guilty? Is there a difference between showing that someone's argument fails and showing that the opposite conclusion holds?)
3. We can express the formal relation as follows:
(I) X believes that not p. e.g. John believes that the school is not closed on Tuesday
(II) X does not believe that p e.g., John does not believe that the school is closed on Tuesday.
(III) X believes that p. e.g., John believes that the school is closed on Tuesday.
(I) implies (II), but not conversely. (I) and (III) are contraries, both cannot be true, but they could both be false. The true negation--contradictory--of (III) is (II). Denying that one holds a belief is just that: a denial to hold a belief. It is not to affirm that one holds a different--specifically, opposite--belief.
One reason for the assimilation of (I) and (II) is a misapplication of the law of excluded middle. From the fact that either p or -p it does not follow that either one believes p or one believes -p. Contrast:
(3) For every proposition p, either p or -p.
and
(4) For every proposition p and for every person X, either X believes p or X believes -p.
(3) is truly the law of excluded middle and what follows from it is not (4), but
(5) For every proposition p and for every person X, either X believes p or it is not the case that X believes p.
We can present these points pictorially. Let "X believes that p" be "Bp". Then we have four possibilities:
(I) Bp
(II) B-p
(III) -Bp
(IV) -B-p
We may now invoke the traditional square of opposition:
(I) Bp [------ contraries ---------] (II) B-p
: : : :
: : : :
implies contradicts implies
: : : :
: : : :
: :
(IV) -B-p [-----sub-contraries-------] (III) -Bp
The crucial representation is that (III) does not imply (II) (and (IV) does not imply (I)). If A doesn't believe that the coffee beans are in the cabinet, he need not believe that they aren't there. He may simply not be sure.
By parallel reasoning, (III) and (IV) are kinds of contraries ("sub-contraries"), not contradictory. Both cannot be false. For if both were false, it would have to be that their denials are true. But the denial of (III) is (I) and the denial of (IV) is (II), and we just observed that (I) and (II) cannot both be true.
The pictorial representation is handy for presenting other logical relations. Traditionally, it is how the logical relations between statements of the following forms are represented:
All As are Bs
No As are Bs
Some As are Bs
Some As are not Bs.
(The "implies" relation requires that the "A" term not be empty.)
4. Other applications of the "square" are closer to our use of it to capture distinctions of scope. When belief is withheld, negation has wide scope, unlike when belief is denied:
(6) Jim does not believe that Mary likes him.
Read (as written) with negation having wide scope, belief is denied. Read (as heard) negation has small scope, belief is affirmed, and (6) is construed as
(7) Jim believes that Mary does not like him.
One other application is worth noting since it concerns central modal terms--necessity, possibility (and related notions). The following argument begins with a conditional concerning permissible drinking of alcohol. Is it valid?
If a person drinks alcohol, he or she must be over 21.
Bill Clinton drinks alcohol.
So, Bill Clinton must be over 21.
The argument looks valid--if the premises are all true, the conclusion is, so we may assume, true. It also seems to be sound--each premise true. However, unfortunately, it is not necessary that Bill Clinton be over 21, any more than for any of us. There was a time when Bill Clinton, like the rest of us, was under 21.
You are tempted to protest that this is being literal-minded. What the conclusion means is that given that Clinton drinks, he must be over 21. It does not say that he simply (for all times) must be over 21. But regardless of what is intended, that is not what is said, and the failure (to properly express what one meant) is an important logical one, not a nitpicking error.
The problem here is the scope of 'must' in the opening premise. What it is really modifying is the relation of the 'if' statement to the 'then' statement, when, as it is written, it is just attached to the 'then' statement. The difference is between
(8) MUST (if a person drinks alcohol, he or she is over 21.)
(9) If a person drinks alcohol, he or she MUST be over 21.
With wide scope (8), it is true, but not in small scope as in (9).
In exposing this scope fallacy, we have found an orientation or bias in it that may help to explain why the fallacy is made. The orientation or bias is to misread (ignore) wide-scope uses of negation (or modal operators) for small-scope ones. So returning to the earlier examples
(10) John believes that the school is not closed on Tuesday
(11) John does not believe that the school is closed on Tuesday.
The tendency is to (mis)understand (11) as (10), rather than the reverse. Why?
Think of a natural context where you would want to assert either (10) or (11). A typical one is where John's beliefs are in question, and, it may be added someone has asserted
(12) John believes that the school is closed on Tuesday.
You disagree and so you assert (11). Why is it heard as if you had asserted (10)? Because that is probably what you (the speaker) meant. For if you asserted (11), you probably know John and know what he believes about the matter of the school closing. If so, you would sooner assert (mean) what (10) says, since it is more informative than (11). (Recall: 10 implies 11, not conversely.)
5. The basic form of reasoning just suggested can be formulated systematically, and the resulting theory explains a distinction between what is logically or semantically implied, and what is conversationally or pragmatically implicated. The theory also explains why it is easy to confuse the two.
The theory, due to the late H. Paul Grice, was originally motivated to respond to doubts about the adequacy of logic to describe natural language logic and inferences. For example, in logic
A & B
is equivalent to
B & A.
(Informally, 2 statements are logically equivalent if they have the same meaning. Formally, each implies the other.) However, the following two statements seem different in meaning:
(13) Jane got married and she became pregnant.
(14) Jane became pregnant and she got married.
Is this a counter-example to understanding 'and' as the logical '&'? Grice argued that it is not. The ever so natural mistake is to treat each statement as asserted. But once asserted we impose conversational expectations, without so recognizing it. The imposition adds content that the speaker probably means to communicate. But it is not part of the statement itself.
In conversation, hearers expect speakers to be cooperative. Grice referred to the Cooperative Principle (CP), as the basic expectation or rule governing conversation. Hearers expect speakers to speak--to make conversational contributions--that serve conversational purposes. Under the CP, more exactly, hearers expect speakers (and speakers know that hearers so expect) to obey various maxims: to speak truthfully, informatively, relevantly, and in an orderly, especially brief, manner.
The last maxim is most germane to the case at hand, as well as for understanding why implicatures are such pervasive devices of communication. Brevity is important for comprehension and because our time is precious. Consequently, we want to say as much as we can with as few words as possible. Speakers leave much implicit or tacit, if they can expect hearers to recover the speaker's meaning without them. In the case at hand(13-14), it is simple to adopt a general rule that if events occur in temporal order to so report them. So (13) and (14) would both be heard as implicating with 'and' 'and then'. Still, what they both state involves only 'and', so that the two are logically equivalent. But they can be used to communicate much different thoughts.
The most fundamental test for distinguishing genuine implications from implicatures is cancellability. Does (13) (logically or semantically) imply that that the marriage came first or only conversationally implicate it. Well, if it implies it, then to assert
(13') Jane got married and she became pregnant, but, actually, maybe she became pregnant first.
would be contradictory, which it is not. Since the inference is cancellable, it is an implicature, not an implication. By contrast,
(15) John knows that Lucy dislikes him
implies that Lucy dislikes him, as the following is contradictory:
(15') John knows that Lucy dislikes him,
but Lucy does not dislike him.