Three Dimensions Of Belief

Ted Everett

(draft 11/08)

1. Three lousy essays.

Here is an ordinary essay question on an ethics exam:

Do you believe that the death penalty for murder is ever justified? Why, or why not?

And here are three little essays in response:

A. Throughout history, most philosophers and other thinkers have supported the death penalty, but today the majority of experts are opposed. It seems to me that the thinkers of today are usually right when they disagree with thinkers of the past, for example about slavery, but not always. In light of these facts, I would put the overall probability of the death penalty for murder being justified at about twenty or thirty percent.

B. Murder is a supremely wicked act, and people who do wicked things ought to be punished in proportion to their crimes. Punishment means suffering, so murderers ought to suffer as much as possible. Dead people do not suffer; therefore, the death penalty for murder is not justified. Instead, murderers ought to be tortured. Ideally, they should be kept alive and tortured forever.

C. I am a Christian pacifist. My family and friends are all pacifists, and I have lived in a pacifist community all of my life. We have a great reverence for all life, inspired by the example of our savior Jesus Christ and many Christian and non-Christian saints. Therefore, I oppose the death penalty for murder or any other crime, in accordance with my personal commitment to life, love, and universal peace.

None of these is a very good essay, but I think that a philosophy professor would be bound to give preference to essay B. It seems unreasonable, perhaps, and hasty, but it at least passes a certain threshold of acceptability that the other two essays do not, in that it addresses the question in a direct and substantive way. It deals, however poorly, with some actual arguments in favor of the death penalty, and expresses a conclusion based on the right general kind of reasoning. It is bad philosophy, but it is at least philosophy. By contrast, the other two efforts do not seem to meet the minimum standards for an essay in philosophy, however brief. Essay A completely fails to address the substance of the death penalty issue. It offers nothing but the kind of indirect inductive argument that one might make about an unread message in a box. The third essay, C, is even worse, in a way, in that it gives no argument at all, direct or indirect, for its thesis, merely a personal statement of commitment to it. Both seem entirely to miss the point of writing an essay in philosophy.

This is not to say that the two rejected essays having nothing in their favor. Indeed, they both present answers to the question, "Do you believe that the death penalty for murder is ever justified?", that would be acceptable in many other contexts. Essay A represents a fairly common, reasonable attitude toward things of which the author admits to knowing very little. He may be quite right, at least given how little he actually knows about the subject, to evaluate the issue only with reference to the opinions of other people who know more. Indeed, this probabilistic reasoning is the way that most of us account for issues where we have no expertise at all, such as the question of life on other planets. If we had just asked the author casually, not on an exam, what he thought or knew about the death penalty, what he says would have been a perfectly acceptable response. Essay C also makes a kind of statement that is generally accepted, even respected and admired, outside of philosophy (or, more broadly, intellectual debate). We like people who take principled stands on things, and we like people who state their commitments frankly and consistently. A pacifist ought to oppose the death penalty, and ought to say so forthrightly. Nevertheless, there is something plainly wrong with saying so on a philosophy exam, at least by way of complete response.

2. Three rules governing belief.

What is going on, here? What accounts for the different virtues and vices of these three kinds of statement of belief? I think that there are three quite different, and sometimes conflicting, principles that we all subscribe to when it comes to our beliefs. I will call them the Principles of Rationality, Autonomy, and Integrity:

The Principle of Rationality (PR): Believe the truth. If this is not directly knowable, believe whatever is most likely to be true, given the total evidence you have available. More precisely, you should believe with greater confidence whatever is more likely to be true, given your pool of available evidence, and you should adjust your confidence accordingly whenever new evidence is added. Failure to do so makes you irrational.

The Principle of Autonomy (PA): Think for yourself. Do not depend on others for your own beliefs. You must assume responsibility for your own conclusions, based on the objective evidence and arguments available, not just ad hominem considerations. Failure to do so makes you dependent and unproductive.

The Principle of Integrity (PI): Stand up for your beliefs. It is important to make your beliefs consistent with your most basic commitments, and to make your actions consistent with these beliefs. If you believe something is right, do not equivocate or back down when you are called upon to state or act on this belief. Failure to do so makes you cowardly and hypocritical.

These rules are ordinarily consistent with each other. There is no reason in principle why one cannot simultaneously think for oneself, believe what is likely to be true, and act accordingly. Abraham Lincoln believed autonomously, primarily on substantive grounds, that slavery was wicked (PA); this was the most rational conclusion for him to have drawn, given all the evidence available to Lincoln concerning that institution (PR); and he spoke and acted according to that belief as well as he could have, given the many constraints on his actions as President, both from circumstances and from his other basic commitments (PI).

The three principles can, however, and often do, conflict in practice. PR tends to conflict with PA whenever a reasonable person has less knowledge than the relevant experts. I do not do my own particle physics, for example, but defer to physicists for my beliefs about the subatomic world, such as they are. I let my doctor tell me whether I am having a heart attack or just acid reflux. I take testimony from my family, friends, and colleagues about all sorts of matters great and small, relying wholly on my confidence in them for many of my beliefs, rather than on my own observations and inferences. In a general way, my desire to be rational in my beliefs restricts my efforts to think things out for myself, because I know that other people have a better direct grasp of the truth of many things than I do.

PR conflicts with PI whenever a rational person has sufficient doubt about something that they nevertheless believe. For example, I believe with some modest degree of confidence that Shakespeare's works were largely written by one Edward de Vere, the Seventeenth Earl of Oxford. That is, I take it as fairly probable, given the moderate amount of reading I have done about the subject, that de Vere was heavily involved in these projects, and that the man William Shakespeare was to some extent a front. But this is not a belief that I stand up for in any way; in fact, I am rather ashamed of it, given the way that my colleagues have responded on the few occasions that I have had the nerve to bring it up. I do not belong to the Oxford Society, or subscribe to any relevant journals or blogs, or go to any conferences, or even spend much time pursuing the matter. Indeed, I do not seem to have any stake in this merely probabilistic belief at all, and I do not feel there is anything that it requires me to do.

PI also conflicts with PA whenever one's belief, however substantively and autonomously derived, still fails to meet some threshold of commitment. Especially when we are working in a creative or hypothetical context, as in science or philosophy, or, often enough, just in casual conversation, we will throw ideas and arguments out with little or no sense of personal assurance or ownership. Perhaps this article contradicts other things that I have written in the past – so what? It will still be a good or bad paper on its merits, independently of who wrote it, and I will still have done a good job as a philosopher if it is a good paper, independently of anything else that I believe or have done. The proper question in evaluating me as a professional philosopher is not whether I am sincere, or whether I act according to what I say in this or any other paper, but whether I am productive of significant ideas. Even in ordinary, casual discussion, the important question is not always whether we are being sincere or showing moral worth in any way, but sometimes only whether what we say is interesting or potentially useful.

It seems that there is a mixture of moral and practical concerns that properly govern our beliefs, on top of purely epistemic ones. PR is the one purely epistemic principle. If all you want is true belief, then you have no need either to think for yourself or to act in any particular way, except as those two principles assist in forming further true beliefs. PA is not about maximizing the likelihood of one's having the correct belief, but seemingly rather about coming up with a certain kind of intellectual product, namely novel or interesting propositions. And PI seems to be a strictly moral principle, governing some but not all relations of action to belief. Thus, in the exam essays above, A is unacceptable despite its evident rationality, and C is unacceptable despite its evident integrity, while B is at least minimally acceptable despite its evident frivolity, because the main purpose of writing an essay in school is to develop autonomous ideas based directly on substantive evidence, so that the student can learn to produce useful thoughts as an adult.

3. Three dimensions of belief.

What is belief, then, that it should be pulled in these three different directions? Shall we say that belief is one thing, simultaneously governed by three competing principles? Shall we say that the word belief is ambiguous, and that each principle governs its own kind of belief? I want to say that belief is not one unified thing and not three distinct things, but that there are three aspects or dimensions of belief, rather than three kinds, in that the concept retains an ultimate unity despite its tendency to split into three types. It is essential to the total human practice of belief, especially in modern societies, that tensions exist among the three dimensions. I will call them perception, opinion, and conviction.

Perception is belief as mental representation. It is the epistemic model of the world that one forms in the first place through sensations, which are connected by memory and reasoning into something like a total picture or theory. It includes indirect sensory perception through instruments like glasses and telescopes, and testimonial perception, including reading and all other forms of gathering symbolic information. Belief as perception tends to be conditional and probabilistic, and lends itself to constant adjustment through the stream of new experience.

Opinion is belief as discursive contribution. In many situations where decisions need to be made socially, it is important that multiple perspectives be independently represented, to be evaluated as partial or alternative ideas of the way things are or of what ought to be done. There are both practical and intellectual areas where discussion is required, sometimes disciplined, systematic discussion as in democratic politics, the law, philosophy, and science. To have an opinion is to be prepared to engage in such discussion. Belief as opinion can be either hedged or categorical, fixed or fluid, depending on the type of discussion in question.

Conviction is belief as presupposition for action. One's convictions are those perceptions or opinions that one has fixed through judgment as not subject to revision or question, at least for the time being. In judging or forming a conviction, one commits oneself to the truth of a belief, and accepts responsibility for its intellectual and moral consequences. A conviction thus acts as both a (partial) cause of decisive action, and as a (partial) explanation for decisive action. Convictions are by their nature categorical rather than conditional or probabilistic.

My three dimensional system contrasts with the two-level typology of belief that has been developing lately through the work of Ronald de Sousa, Daniel Dennett, and Keith Frankish. These writers distinguish between "belief" and "opinion", using Dennett's terms, which seem to have been widely accepted. "Belief" is held to be the lower form, corresponding to what I call perception, and "opinion" the high form, corresponding to what I am calling conviction. It is a mistake to call the lower form "belief", I think, because all (two or) three things are commonly called belief. It is similarly misleading to call the higher form "opinion", because our ordinary concept of opinion is of something that often requires little commitment, and that need not result in any action other than speech. I say that there are three rather than two aspects of belief, then, because the essential functions of perception, opinion, and conviction are three, clearly separable things. But I want to call these three dimensions rather than three types because they are not wholly distinct categories, but overlap and interact in systematic ways. Some beliefs seem to fall into only one category, but most ordinary beliefs have all three dimensions. There is no saying in such cases what kind of belief it is; rather, the same belief can be looked at in each of the three ways.

Here is an example. I believe that it is good for philosophy to be taught in colleges. This is a perception of mine, in that I have seen students improve their thinking and become better people through the study of philosophy, and also in that I have read a good deal of material touting the value of philosophy, and of liberal education in general. It is part of the way I see the world that people's lives are enhanced by the study of philosophy – though a revisable part, in principle. At the same time, this belief is one of my opinions. It is something that I carry into arguments about curriculum changes at my college. It is a point that I am prepared to argue for when I am confronted by people who question the value of philosophy or liberal education, or wish to impose a more doctrinal, practical, or "consciousness raising" approach to education. I accept that other people will decide over the long run what is the best way to educate young adults, and that mine is only one of many perspectives on the topic. Yet it is also one of my convictions that a philosophical education is a good thing. It gets me to go to work in the morning (okay, afternoon), when I don't feel like it, and could as easily take the day off. It motivates me to engage in arguments beyond the level of brainstorming – in other words, to fight – about the future of my college. It impels me to spend more time dealing thoughtfully with students and colleagues, to put up with more stress, embarrassment, and (very occasionally) hostility than I would do if I were not convinced that this sort of work is important.

I do not ordinarily separate out these three aspects of my belief that studying philosophy is good for students. I'd just say that it is one of my beliefs. Yet there are many cases where a belief serves only one or two of these three functions. For example, it is neither an opinion nor a conviction of mine that the population of China is between 450 million and 3.2 billion. I have never articulated that exact idea at all until just now. Nevertheless, I would say that it expresses one of my (implicit) perceptions, in that it is derived directly from my model of the world. I don't hold the opinion that the population of China falls in that range, because there isn't any debate about the topic, as far as I know. It is just a matter of fact, not, as we say, a matter of opinion (though I imagine that there are demographers somewhere who do have opinions about the Chinese population, I doubt that this particular belief is one of them). And I would not call it a conviction of mine that China has such-and-such a population, though I still think that it is probably true. I just don't have any decisions to make that might be sensitive to such a claim. If some urgent dispute arose about the Chinese population, this all could change – but so would the nature of my belief.

Similarly, there are perceptions and opinions that I have that are not convictions. I mentioned above my guess that Shakespeare's plays were written not by William Shakespeare, but by the 17th Earl of Oxford. This is a probabilistic perception, with a subjective degree of confidence that I would place somewhere around fifty percent. As far as I can rationally determine, the actor Shakespeare might have written the plays himself, or it might have been de Vere, or it might have been Marlowe or one of the other candidates, or, perhaps most likely, some kind of flexible collaboration. But I find de Vere a very plausible main author of the plays, and I think the case for him as author has been overlooked, often for bad reasons, by hidebound and contemptuous academic scholars. So, it has become a point that I like to argue for. It is thus my opinion that de Vere is the Shakespearean author, even though I would not be terribly surprised to be found wrong. It is certainly not my conviction, however – I am not a convinced "Oxfordian", for I am not convinced of anything about the authorship question. It is a mere opinion.

Here is how I think the three dimensions of belief correspond to the three rules of belief defined above: PR regulates perception; PA regulates opinion, and PI regulates conviction.

PR regulates perception in that the goal of perception is an adequately complete and reliable model of the world and its workings. This requires that we gather large amounts information through the senses, including through testimonial and other indirect sources, and work this information rationally into the most likely-to-be-true available theory of the world (in fact, a system of probabilized alternative theories), maximally sensitive to rational combination with all new experience. The Bayesian account of evidence absorption through adjustments in subjective conditional probabilities seems to be the right account of rationality for belief as perception.

PA regulates opinion in that the purpose of having opinions is to provide diverse hypotheses for that “continuous and fearless sifting and winnowing by which alone the truth can be found” (Wisconsin). This means that people must be creative in their thinking as individuals, in order to be rational as groups. Only people who think for themselves can fully participate in this social process, thus it is necessary for people to be taught to think for themselves, even when this flies in the face of rational probability, as between essays A and B above. The virtue of belief as opinion is insight, not accuracy, and insight can only come from largely independent points of view.

PI regulates conviction in that the purpose of having convictions is to stop thinking and act in a determined and consistent way. A person with integrity is someone whose beliefs on certain matters of importance are held fixed, so that the person can be counted on by others as a reliable agent, and can count on himself to execute long range plans without potentially destructive levels of doubt. The virtue of belief as conviction is neither insight nor accuracy, but stability, which sometimes requires the suspension of one’s ordinary rational and creative faculties.

Here is how I think the three principles and three dimensions of belief all interact ideally. Imagine sending a group of small robots to explore a distant planet, so distant that there is no effective method of Earthly control over the robots once they depart, and no expectation of receiving back anything more than an eventual final report of their discoveries. How would we design these robots? One possibility is to create a system of interlinked, partly but not entirely autonomous explorers, held together by some means of forming a collaborative picture of this foreign world. Each robot comes with a set of sensory devices; each can independently create a partial map and partial analysis of the large planetary surface. Each is also equipped with devices for communication, so that a total theory can be built out of these parts. Simple enough, if all robots are exactly the same in design, and the parts of the total picture can just be added together without conflict. This produces no “disputes”, but it leaves open the possibility that something important in sensation or analysis will be missed by each, hence by all, of the robots. A more sophisticated system, and one with a greater chance of finding out the truth about this planet, would send robots of various different designs, with somewhat different sensory devices and analytical software, each capable in its own way of downloading and combining information from the other models. It would be important for each robot to “respect” the “opinions” of its fellows, but not to the extent that its own tentative analysis is easily overwhelmed by an emerging consensus. Thus each robot should be programmed to be “rational” in treating the reports of other robots, but not too rational, in case its own minority “opinion” should be the one that works best in the long run. Thus, each robot ought at least to store its own tentative analysis, along with combining and compromising into a group picture. If the planet is a tough one to understand, a good variety of hypothetical approaches should be maintained, at least until the same overall theory finds ready acceptance among all the robots as individuals.

4. Three central problems with belief among humans.

There are three central problems with human understanding of the world we live in. I do not see any principled way to solve them. Here they are.

Problem 1. Our individual perceptions are limited by human size, lifespan, and resources. There is too much in the world for any one person to know. Therefore, intellectual life must somehow be social. Our private perceptions must give way to public opinions in the realms of controversy.

Problem 2. Our individual opinions are easily overwhelmed by the testimony of others, either through the others’ greater proven reliability (especially when we are children), through their greater acknowledged expertise, or simply through their greater numbers. A perfectly rational believer is far too likely to conform to the beliefs of those around him. Therefore, we must be taught to uphold our own beliefs as convictions.

Problem 3. Our individual convictions are liable to be false, and to involve us in violent error in the world. To the extent that we all stick to our guns, we resist reasonable compromise and consensus. Yet without such personal convictions, we risk falling into static schools of thought, even potentially absurd ones, from which the individual has no rational means of escape.

The first problem necessarily subordinates our own beliefs to those of others, in the interests of learning from others in the short run, and of producing more true beliefs for everybody in the long run. How do we handle this subordination? I think that the natural thing, something terribly hard for us all to resist, is for consensus to develop around the views of the majority of experts or elders on any topic. This is why religion, not philosophy or science, is the normal framework for belief throughout the history of the world. In order for philosophy or science to develop at all, some sufficient number of individuals must become what I have elsewhere called epistemic altruists, that is, people who believe in their own view of things despite its being more rational for them to accept the testimony of their peers and elders. This has happened only a few times in history, and only once has it produced a seemingly permanent institution of progressive thought, namely in modern Western Europe and its colonies. Even here, though, and even amoung the people most proud of their principled independence of mind, there is constant pressure to regress back into static and intolerant “paradigms” or schools of thought. Galileo was famously dragged back by his contemporaries into acknowledging the primacy of Church doctrine over independent science. Darwin was terrified of what would happen to his reputation among his Victorian contemporaries that he held his theory in a drawer for decades until forced by looming competition to release it. Einstein began campaigning against radical quantum mechanics almost as soon as his own radical relativity had been accepted. In psychology, one domineering paradigm succeeds another every thirty years or so, with fairly brief eclectic periods in between. Most recently, the Western social sciences seem to have evolved a consensus on a rigid ethnic and gender egalitarianism in response to moral pressure from contemporary movements for civil rights, and even the physical sciences seem to have frozen up somwhat because of the involvement of so many scientists with environmentalism. Thus it is a danger to the careers of scientists to doubt in public the hypothesis that men and women have identical intellectual capacities, or the hypothesis that global warming is a man-made crisis. Most scientists who poke their heads up on these issues have been quickly brought to heel; the few who maintain a principled, public independence on these issues have been ostracized to the extent that only genuine cranks are likely these days to oppose to dominant positions. From within the relevant consensus movements, such suppressions of dissent seem like rational demands for individuals to respect the institution of science itself. And they are rational demands – in my view, this is the main problem. Only a tolerance for irrational dissent allows us to progress beyond the local perspectives of any given age. But too much tolerance promotes irrationality and anarchy in general, where “independent thinkers” like L. Ron Hubbard (or, for that matter, Charles Manson) would be accepted as the equals of Watson and Crick. What is needed is some sort of balance in the individual and in the intellectual society. None of the three basic problems of belief can be solved decisively. We must be as rational as possible for current needs, consistently with maximizing forward-looking creativity and fostering the kind of principled independence that resists our natural descent into the gravitation of consensus theories.