Opinion

Plato divides the states of the mind in the image of the divided line in the Republic. The line has four segments: imagination and belief, which together constitute opinion (doxa), and mathematical understanding and intellectual apprehension, which together constitute knowledge (episteme). The division rests on a difference in the objects to which these states are directed rather than on a difference between true and false assertions. Opinion, formed with the help of sensation, has for its object the sensible particulars which come to be and pass away; knowledge, formed by reason without the help of sensation, has for its object the intelligible forms, which always are and have no becoming. Since for Plato the realm of becoming belongs exclusively to opinion, the study of the physical world yields no science but only, as the Timaeus has it, a likely story, which is to say a plausible composition of probable opinions. Aristotle takes a different view on this point, admitting physics among the sciences, and the disagreement is treated more fully under the idea of Knowledge.

The Meno states the difference in another way, in terms of the act of the mind rather than of its object. Right opinion, Socrates there suggests, is no less true than knowledge and, for the purpose of acting well, no less serviceable; the man who has a correct opinion about the road to Larissa guides the traveller as surely as the man who knows it. But right opinion is compared to the statues of Daedalus, which are beautiful enough while they stand yet run away if left untethered. What tethers an opinion and converts it into knowledge is the ability to give an account of why the thing is so. Lacking such an account, opinion remains insecure against contrary persuasion, and it is also unteachable, since a man who cannot explain the rightness of what he asserts cannot bring others to see it. The Theaetetus examines the proposal that knowledge is true opinion accompanied by an account, finds difficulties in every sense that can be given to the word account, and leaves the question open.

The distinction bears on politics as well as on the theory of the mind. In the Republic the men who have ascended out of the cave to the vision of the good are unwilling to return to the management of affairs, yet Plato holds that they must, on the ground that only those who know can frame just laws and correct the opinions by which the many are guided. The multitude is not represented as wicked but as directed toward appearances rather than realities, and toward images of appearances at that. Because opinion in most men is formed less by argument than by habituation to what they have been accustomed to see and hear, Plato proposes in both the Republic and the Laws a regulation of the tales and poems by which the young are shaped. The questions raised by such regulation belong to the chapters on Poetry and Liberty. What belongs here is the ground on which it is proposed, namely the supposition that opinion, unlike knowledge, is produced by causes other than the object under consideration.

Almost every later treatment of opinion takes some position with respect to Plato's. Aristotle retains the opposition of knowledge and opinion while denying that the sensible world falls wholly on the side of opinion, and he assigns to received opinion a place in inquiry which Plato does not allow it. Hume, dividing the objects of inquiry into relations of ideas and matters of fact, appears to separate the two realms as sharply as Plato does, with the difference that for Plato the intelligible order is the reality which changing things image, while for Hume ideas have no existence apart from the mind which derives them from sense. The political consequence, that those who know should rule those who merely opine, is contested by writers who hold that the multitude may judge some questions better than the expert, and the dispute is pursued under Democracy and Aristocracy.

Aristotle retains the opposition between knowledge and opinion, but states it in terms which allow the physical world a place on both sides of the division. The object of science is the essential and the necessary, that which cannot be otherwise and can be demonstrated from first principles; the object of opinion is the accidental and the contingent, that which admits of being otherwise. Since sensible things have aspects both of necessity and of contingency, being intelligible in their essences and variable in their accidents, it is possible to have scientific knowledge of nature as well as probable opinion about it. Physics accordingly stands among the sciences with mathematics and theology, though the particular facts of natural history remain matters of experience and of probable judgment. Aristotle holds further that the same mind cannot know and opine the same object at the same time, though a man who once accepted a proposition of geometry on his teacher's authority may afterward learn the reasons for it and come to know what he formerly opined.

On the standing of received opinion Aristotle diverges from Plato. Where the dialogues treat the opinions of the many as confusions from which inquiry must be extricated, Aristotle treats them as the material with which inquiry begins. The examination of what is generally believed, and of what is believed by the wise, is the business of dialectic, which proceeds by setting received opinions side by side, exposing the inconsistencies among them, drawing distinctions, and arriving at a position which preserves as many of the appearances as can be preserved. Such reasoning does not demonstrate but persuades, and the comparison of demonstration with persuasion, of the syllogism with the enthymeme, runs through the Topics and the Rhetoric. The method rests on the supposition that what has seemed true to many men, or to men of judgment, is evidence of a sort, even where it is not evidence of the kind from which a science can be built.

If opinion has for its object what can be otherwise, then deliberation and practical wisdom (phronesis) belong to the sphere of opinion rather than of science. The prudent man does not deduce his actions from universal rules; he perceives the particulars of the case and judges what is to be done in circumstances which could have been other than they are. Hence, on Aristotle's view, ethics and politics do not admit the exactitude of mathematics, and the student of them must be content with conclusions that hold for the most part. The same consideration enters his account of political decision. Since matters of action are matters about which reasonable men may continue to differ, they are settled by the counting of votes rather than by demonstration, and Aristotle observes that the majority rules not only in democracies but in every constitutional form. He suggests further that the multitude, whose members individually may judge worse than those with special knowledge, may as a body judge as well or better. These questions are pursued in the chapters on Democracy and Aristocracy.

The Aristotelian settlement leaves opinion with a positive office rather than a merely deficient one, and later writers who defend the sufficiency of probability for the conduct of life, among them Aquinas and Locke, work within terms which Aristotle supplied. The connection he draws between opinion and the contingent reappears in Hume, who counts as matters of belief those propositions whose contraries are conceivable, and excludes from opinion whatever is self-evident or demonstrable. What separates them is that Aristotle does not suppose the real existence of things, or the connection of cause with effect, to lie beyond demonstration. The dependence of rhetoric and of political deliberation upon opinion so understood is treated further under Rhetoric and Prudence.

The skepticism of the Apology for Raymond Sebond is not the provisional doubt which is undertaken in order to discover the foundations of certain knowledge. It is a resolute skepticism, which reduces all human judgments to the condition of equally tenable opinions and leaves man no hope of doing better than to adopt opinions on insufficient grounds or else to suspend judgment altogether. No axiom, Montaigne observes, has ever won the consent of all mankind, and no demonstration has ever escaped the necessity of assuming its own premises; unless men beg the question in this manner, they cannot avoid an infinite regress in reasoning. Illusion and hallucination are taken to show the unreliability of the senses, and the errors of judgment and inference to show a like infirmity in the mind. What he holds and believes today, Montaigne writes, he holds with his whole belief; yet the same instruments, in the same condition, have a thousand times embraced what he afterward judged to be false.

Two consequences are drawn from this position, and they pull in different directions. The first is a rule of intellectual conduct: since our judgments are no better founded than those of the men who differ from us, we ought neither to hold our own opinions dogmatically nor to despise the opinions of others. The second is a rule of practical conduct: since we cannot reason our way to better institutions than those we have inherited, custom and established usage are safer guides than the schemes which reformers construct from first principles. The uncertainty of opinion becomes in this way an argument against acting upon opinion, and Montaigne is unfriendly to reformation, not because he judges the existing order just, but because he doubts the judgment which pronounces another order better. Whether a skepticism of this sort supports deference to custom, or supports instead the toleration and free contest of opinions for which Locke and Mill later argue, is a question the tradition has answered in both ways.

Religious faith is exempted from the general uncertainty. Montaigne withdraws from the reach of his doubt whatever man does not arrive at by the unaided efforts of his senses and his reason. We must, he holds, accompany our faith with all the reason we have, yet always with the reservation that it does not depend upon us, and that our arguments and endeavors cannot of themselves attain to a knowledge so supernatural and divine; faith is distinguished from ordinary belief only if it enters into us by an extraordinary infusion. The difference lies not in the object, which may be an article of faith to one man and a matter of opinion to another, but in the cause of the believing. Aquinas draws the line in a comparable place, though he treats faith as a mean between science and opinion rather than as the sole survivor of a general shipwreck, and the comparison is developed in the chapter on Religion. Readers have differed over whether Montaigne's exemption expresses piety, a considered fideism, or a covering for unbelief.

The Essays furnish the modern discussion with its most sustained statement of the skeptical case, and much of what follows consists in answering it or in accepting it. Descartes takes universal doubt as a method rather than a conclusion, and looks for a proposition which the doubt cannot touch. Hume grants a great part of the argument, yet holds that nature settles by custom what reason cannot settle. Against the extreme form of the position, Aristotle and others had already urged that the proposition that nothing is true or false is inconsequential if it is false and self-contradictory if it is true, to which it may be replied that the proposition that everything is a matter of opinion is itself an opinion, and its contrary an opinion also. The bearing of such skepticism upon toleration is considered under Liberty.

Cervantes places at the center of his novel a pair of men whose opinions about the world they ride through are in constant disagreement, and whose disagreement the novel does not resolve by calling one of them right and the other wrong. Don Quixote, formed by his reading, holds the opinion that the field contains giants, that the barber's basin is the enchanted helmet of Mambrino, that a flock of sheep is an army advancing, that a peasant woman on a donkey is the lady Dulcinea under an enchantment. Sancho Panza, formed by the common experience of his village and by a well-developed sense of his own interest, holds the contrary opinions: that what his master sees are windmills, a basin, sheep, and a peasant woman. The question which runs through the novel is not how to decide between these two sets of opinions but how two men of good faith, looking at the same things, can arrive at views so completely at variance.

The older treatments of opinion, beginning with Plato's account of doxa as the state of mind that falls short of knowledge, had assumed that the correction of opinion was a matter of bringing the mind into closer contact with the object. Aristotle, more moderate on the point, had still assumed that opinion could be steadied by long experience and prudent judgment. Cervantes presents the more disturbing case in which long experience and prudent judgment already belong to Sancho, and closer contact with the object is already available to Don Quixote, and the gap remains. What separates the two men is not a lack of evidence but the prior conviction which each brings to the evidence. The knight's prior conviction is shaped by the books of chivalry, the squire's by the habits of village life, and each prior conviction is powerful enough to select from the available appearances the ones which confirm it.

The question this raises connects to the treatments of Memory and Imagination, where the knight's imagination is shown supplying his perceptions; to Knowledge, where the distinction between opinion and knowledge is considered; and to Truth, where the question is whether a single standard of truth is available for settling such disputes. What belongs to the idea of Opinion is the display of a case in which the formation of opinion is traced not to the object but to the observer, and in which the disagreement between two observers is found to be rooted in what each of them already brings with him to the act of observation.

Montaigne, in his essays on custom and on the force of imagination, had already observed how much of what men take for the testimony of their senses is in fact the testimony of their upbringing. Descartes, reading Cervantes or reading the world in which Cervantes wrote, will try to answer the problem by seeking a method for purifying opinion of its prior formations, a method so rigorous that only what remains after all opinion has been suspended can count as knowledge. Hume, moving in the opposite direction, will conclude that what we call our beliefs about the world are themselves habits of the mind, so that Sancho's confidence in the windmill and Don Quixote's confidence in the giant differ not in kind but only in the particular history which has produced them. The novel's questions survive all of these answers.

Descartes proposes to deal with the uncertainty of received opinion by a method rather than by resignation. Since the opinions taken from parents, teachers, and custom may all be false, he undertakes to withhold assent from whatever admits of any doubt, in order to see what, if anything, survives. The Meditations on First Philosophy carries out this exercise. The testimony of the senses is set aside as sometimes deceptive, the existence of bodies as indistinguishable from the imagery of a dream, and even the truths of arithmetic and geometry as possibly the contrivance of a deceiver. What resists the doubt is the doubting itself, for so long as the mind is thinking it cannot suppose itself not to exist. The proposition that I am, I exist, is put forward accordingly not as an inference drawn from premises but as something perceived with a clarity which leaves the mind no liberty to dissent.

Upon this Descartes builds a criterion for separating knowledge from opinion. Whatever is perceived clearly and distinctly is true, and the veracity of God is invoked to guarantee that a faculty so exercised does not mislead. Error, and with it the formation of opinion, arises not from the intellect but from the will, which is unlimited in its range while the intellect is not, and which gives assent to what the understanding has not clearly set before it. Judgment, on this account, is an act in which two faculties concur, and the mind falls into opinion whenever the will outruns what the object itself compels. Descartes, Locke, and Hume differ in the terms of their analyses, but they agree in placing knowledge where the mind is moved to assent by what it perceives, and opinion where it is moved by something other than the object considered. Whether the will is free in giving or withholding assent is a further question, treated in the chapter on Will.

The suspension of judgment is meant to be theoretical and temporary rather than a rule of life. Descartes adopts in the Discourse on the Method a provisional code of morals, resolving to obey the laws and customs of his country and the religion in which he was brought up, and to follow the most moderate opinions commonly received among the most judicious, on the ground that action cannot wait upon demonstration. In this he stands with those who hold that probability suffices for practice though not for science. Once the existence of the thinking self and of God is established, the reconstruction proceeds in the opposite direction, physics being derived from principles as certain as those of geometry, and the probable opinions of the schools being replaced by demonstrations. How far that program was carried out, and whether the science of nature can be demonstrative at all, are matters considered under Science and Mechanics. The uncertainty of opinion is thus made an occasion for radical doubt rather than, as with Montaigne, a reason for deference to what is established.

The attempt to leave opinion behind altogether has been questioned from two sides. Gassendi and those who followed him argued that the cogito establishes less than is claimed for it, and that the clear and distinct ideas on which the system rests are themselves opinions ratified by nothing but their own appearance of evidence. Others have observed that the proof of God, which is to guarantee the faculties, is itself conducted by those faculties. Locke and Hume retain the Cartesian question while abandoning the Cartesian standard, holding that most of what men need to believe can be no better than probable. The relation of certitude to the degrees of assent, and the question whether any first principle is exempt from doubt, are treated more fully in the chapter on Knowledge.

The Essay Concerning Human Understanding is in large part an inquiry into how far human knowledge extends and what is to be done in the much larger territory where it does not reach. Knowledge, in Locke's account, is the mind's perception of the agreement or disagreement of its ideas: intuitive where the perception is immediate, demonstrative where it proceeds through intervening ideas, and sensitive where it concerns the existence of the particular things presented to the senses. Where such perception is wanting, a second faculty takes over. Judgment puts ideas together or separates them when their agreement is not perceived but presumed, and to judgment belong belief, assent, and opinion, which he describes as the receiving of a proposition as true upon arguments that persuade without giving certain knowledge that it is so. As demonstration stands to persuasion, and certainty to probability, so knowing stands to judging.

Locke is less troubled than Plato or Descartes by the extent of the territory thus assigned to opinion, since he holds that judgment was given to man to supply what knowledge cannot furnish, and that the light granted him is sufficient for the purposes for which it was granted. Probability admits of degrees, and its grounds are of two sorts: the conformity of a proposition with our own observation and experience, and the testimony of others, weighed according to their number, their skill, their disinterestedness, and the consistency of their reports. Where the degrees are high, the assent may rise to a confidence bordering upon certainty; where they are low, assent should be withheld or given with reserve. The rule laid down is that a man proportion his assent to the grounds he has, and the corresponding fault is an assent which outruns them. Enthusiasm, which puts the strength of a persuasion in the place of evidence for its truth, is on this account opinion without warrant.

From the nature of opinion Locke draws a political conclusion. If men of equal sincerity and equal care arrive at different opinions in religion, then the magistrate who enforces one of them decides a question by force in a matter to which force does not apply. Belief, moreover, does not follow upon command, since it answers to what appears evident to the believer, so that penalties may produce conformity of profession without producing conviction. The business of laws, he writes in the Letter Concerning Toleration, is not to provide for the truth of opinions but for the safety and security of the commonwealth and for every particular man's goods and person; truth would do well enough if she were left to shift for herself, and gains nothing from any borrowed force that violence can add to her. The argument separates the political from the logical standing of a disputed opinion, and the same separation underlies Mill's later defense of free discussion, considered in the chapter on Liberty.

Locke's account of judgment supplied the eighteenth century with its working vocabulary for the degrees of assent. Hume takes it up and presses it further, holding that even the propositions Locke counted as knowledge of matters of fact rest upon custom rather than upon any perceived agreement of ideas. Kant, separating opinion from belief and belief from knowledge by the sufficiency of their grounds, refines the same scheme. The connection between the fallibility of opinion and the case for toleration, which Locke was among the first to state at length, is carried further by Mill, and belongs also to the chapters on Religion and Liberty.

Hume's treatment begins from a division of the objects of inquiry. All the objects of human reason, he writes, may naturally be divided into relations of ideas and matters of fact. The first, of which the mathematical sciences consist, are discoverable by the operation of thought alone, and their contraries imply contradiction. The second, which include every question concerning the real existence of anything and every question concerning the connection of one thing with another, admit of no such demonstration, since the contrary of a matter of fact is always conceivable. Whatever is not demonstrable falls therefore among matters of belief rather than of knowledge, and belief is the term Hume commonly employs where others use opinion. The criterion resembles Aristotle's, which assigns to opinion whatever can be otherwise. It differs in the consequences drawn, for Hume places the real existence of things and the connection of cause with effect on the side of belief, where Aristotle does not.

The consequence appears when Hume asks upon what the beliefs concerning matters of fact are founded. Reasoning about what has not been observed proceeds through the relation of cause and effect, and the knowledge of that relation is drawn not from any perceived necessity but from the experience of constant conjunction. What is observed is that one sort of event has been followed by another; what is not observed is the tie between them. The expectation that the sequence will continue is therefore the work of custom, which disposes the mind to pass from the impression of the one to the idea of the other; and belief is distinguished from mere conception, not by anything added to the idea, but by the manner and vivacity with which the mind entertains it. Belief so understood lies nearer to sentiment than to reasoning, and, as Hume says of it, it depends not on the will, nor can it be commanded at pleasure.

Hume does not offer this analysis as a reason for distrusting the beliefs by which men live. Custom is the great guide of human life, and without it we should be unable to act at all; nature has provided by instinct what reason could not have supplied, and the probabilities which physics establishes may be entirely sufficient for action even if they fall short of demonstration. The skeptical result bears rather upon the pretensions of systems which claim to have passed from probability to certainty. The rules by which causes are to be judged, the weighing of testimony against the antecedent improbability of what it reports, and the examination of the arguments for religion in the Dialogues Concerning Natural Religion are applications of a single maxim, that a wise man proportions his belief to the evidence. Whether a position of this sort can be consistently maintained, or whether it undermines the reasonings by which it is itself reached, is a question raised in the chapter on Knowledge.

Kant reported that the recollection of Hume interrupted his dogmatic slumber, and the Critique of Pure Reason may be read as an attempt to preserve the necessity of causal judgment while conceding that it cannot be extracted from experience. Later writers who make belief a function of feeling or of interest, among them James, acknowledge Hume as their predecessor while declining his estimate of what such an account costs. The question left standing runs through the remainder of the discussion: if custom rather than reason produces our beliefs, by what standard is one opinion to be preferred to another? The related problem of how the emotions and the imagination enter into the formation of opinion is treated under Emotion and Imagination.

Kant grants Hume that no necessity can be extracted from experience, and denies the conclusion which Hume draws from it. The forms of intuition and the categories of the understanding, which the mind brings to experience rather than takes from it, are what make an objective and necessary judgment about appearances possible; within the field of possible experience there is knowledge and not merely well-founded custom. Beyond that field, where reason attempts to determine the soul, the world as a whole, and God, no knowledge is to be had, and the antinomies are meant to show that reason falls into contradiction with itself when it makes the attempt. But Kant does not conclude that whatever lies beyond knowledge belongs to habit or to inclination. He distinguishes instead among the ways in which the mind may hold a proposition to be true.

The three modes are opinion (Meinen), belief (Glauben), and knowledge (Wissen), distinguished by the sufficiency of the grounds on which assent rests. Opinion is an assent which the man who holds it recognizes as insufficient both subjectively and objectively; it is entertained with the consciousness that the matter may be otherwise, and Kant thinks it out of place in questions of pure reason, where one ought either to know or to abstain. Belief is an assent sufficient for the person who holds it though not sufficient to compel the agreement of others. Knowledge is an assent sufficient in both respects. The scheme is more articulated than Locke's degrees of probability, and, unlike Hume's account, it locates the difference in the standing of the grounds rather than in the liveliness with which the idea is entertained. Kant applies it to separate the pragmatic belief on which a physician acts in an uncertain case from the moral belief which no evidence drawn from experience can either establish or overturn.

The postulates of practical reason are the principal use to which the distinction is put. Freedom, immortality, and the existence of God cannot be demonstrated by theoretical reason, and the exposure of the attempts to demonstrate them is a chief business of the first Critique. Yet the moral law commands unconditionally, and the commitments which the moral life presupposes are, on Kant's argument, rationally warranted although they can never become objects of knowledge. An assent of this kind is not an opinion held provisionally in the absence of evidence, since no evidence could touch it, nor is it a habit generated by custom. Kant calls it a faith of a rational sort and places it between the two. How such faith stands to the faith which theologians attribute to grace, and which Aquinas ranks as a mean between science and opinion, is a question belonging to the chapters on God and Religion.

The tripartite scheme has been criticized from opposite directions. Some hold that an assent subjectively sufficient but objectively insufficient is opinion under another name, and that Kant has renamed rather than removed the difficulty. Others ask why, if commitment beyond the evidence is warranted where the moral life requires it, it should not be warranted wherever a choice cannot be deferred and the evidence does not decide. James presses the second question and extends the permission considerably further than Kant would have allowed. The comparison of demonstration with persuasion, on which the whole distinction rests, is treated under Reasoning, and the obligations which Kant assigns to the moral agent under Duty.

James treats belief as a fact of psychology before treating it as a problem of philosophy. Will and belief, he writes in the Principles of Psychology, are two names for one and the same phenomenon, which is the attribution of reality to an object of thought. Except for those necessary truths which concern only the relations of ideas, the mind in thinking about reality is free to choose among alternative theories, in science no less than in religion. That theory will be most generally believed which, besides offering objects able to account for our sensible experience, offers those which appeal most urgently to our aesthetic, emotional, and active needs; the conceptions called scientific have so far gratified the intellectual interest while leaving the others cold. James does not hold that objective criteria are absent from the choice, but he denies that they mainly determine it, and in this he continues Hume's account of belief as something nearer to sentiment than to demonstration.

The essay on the will to believe considers the case in which the evidence does not decide and the decision cannot be postponed. James calls an option genuine when it is living, in that both alternatives make some appeal to the man who faces it; forced, in that no third course is open; and momentous, in that what is foregone cannot afterward be recovered. Where an option of this sort presents itself, he argues, our passional nature may lawfully decide, since it will decide in any event: to withhold assent for want of proof is itself to choose, and to accept the risk of losing a truth in preference to the risk of embracing an error. The suspension of judgment is thus denied the neutrality which the skeptic claims for it. Critics have replied that the argument permits a man to believe what he wishes, and James confines it to cases where the question cannot be settled on intellectual grounds and the choice cannot be evaded.

The pragmatic account of truth carries the same tendency into the theory of knowledge. An idea is called true, on this view, in virtue of what the believing of it does, so that truth is less a property residing in the idea than something which happens to it as it is verified in the course of experience. On the psychological side, a belief is a habit, disposing a man to act in one way rather than another, to attend to some features of his surroundings and not to others, and to assimilate what is new to what he already holds. Its stability is of use to him and its revision is at times necessary, and much of James's inquiry concerns the conditions under which men form, retain, and give up their beliefs, and when the holding of a belief against the evidence is to be counted a disorder and when a faith which the evidence may yet confirm. The distinction between knowledge and opinion is on this account softened rather than denied, since what passes for knowledge is opinion verified so far and still liable to further test.

James's position gathers up several lines of the discussion. From Hume it takes the derivation of belief from something other than reason; from Kant, the legitimacy of assent where theoretical proof is wanting; from Locke and Mill, the argument that opinions must be left free to contend with one another. What it adds is the proposal that opinion differs from knowledge in the degree of its verification rather than in kind. Whether what works is a criterion of truth or only a description of how belief comes about has been disputed since, and Dewey carries the question further by proposing to set aside both terms of the old distinction in favor of a single standard of warranted assertion. The connection of belief with wish and interest, which James shares with Freud, belongs also to the chapters on Desire and Religion.