Popular Science Monthly/Volume 72/May 1908/Should Psychology Supervise Testimony?

From Wikisource
Jump to: navigation, search


PROFESSOR MÜNSTERBERG, of Harvard University, who, by his combination of scientific eminence, active interest in public questions, and rare literary skill, occupies a unique position among those who discuss large educational and social problems, recently made a plea,[1] which attracted wide attention, for the introduction of the tests of experimental psychology as a means of judging of the value of evidence given in courts of justice. To show that a deplorably large part of the evidence given in courts and elsewhere is untrustworthy for one reason or another would have been a work of supererogation; nothing is better established in the minds of all who have to deal with the subject. But, while everybody is familiar both with the phenomena of lying and loose statement, and with the fantastic tricks played by our treacherous memories, it is probable that most persons are unaware of the trouble that lurks in malobservation; and it is solely with the errors which arise from this source that Professor Münsterberg deals in the argument to which we have referred. He presents an appalling array of divergences in the outcome of ordinary observation of the same things by one hundred members of his class at Harvard; and he draws from his results the double conclusion that human observation is incomparably more fallible than is generally supposed, and that courts of justice can, and therefore ought to, test the value of the testimony of witnesses by subjecting them to the examination of the expert psychologist. If Professor Münsterberg is right in these conclusions, they are certainly deserving of the earnest attention both of psychologists and of lawyers; but, while his argument drew forth wide-spread comment, it does not seem to have been anywhere discussed with any thoroughness. And it is the purpose of this paper to show that neither the sweeping pessimism as to human reliability which would result from accepting the record given by Professor Münsterberg at its face value, nor the practical plea that he urges for the classifying of witnesses by the methods of experimental psychology, is justified upon careful examination of his argument.

To pass in review each one of the tests detailed by Professor Münsterberg, and attempt to measure the degree in which the whole array falls short of establishing the contentions which it is designed to support, would require the quotation of almost the whole of the article, tedious, and would demand a greater amount of space than this magazine could be expected to furnish. The nature of the objections to the article can, however, be made clear without so comprehensive and minute an examination.

Let us, in the first place, take a single one of the tests, that relating to the size of the moon:

My next question did not refer to immediate perception, but to a memory image so vividly at every one's disposal that I assumed a right to substitute it directly for a perception. I asked my men to compare the size of the full moon to that of some object held in the hand at arm's length. I explained the question carefully, and said that they were to describe an object just large enough, when seen at arm's length, to cover the whole full moon.

The answers ranged from a carriage-wheel to a pea; and on this result Professor Münsterberg makes a number of interesting comments. "To the surprise of my readers, perhaps," he says, "it may be added that the only man who was right was the one who compared it to a pea. It is most probable that the results would not have been different if I had asked the question on a moonlight night with the full moon overhead. The substitution of the memory image for the immediate perception can hardly have impaired the correctness of the judgments. If in any court the size of a distant object were to be given by witnesses, and one man declared it as large as a pea and the second as large as a lemon-pie and the third ten feet in diameter, it would hardly be fair to form an objective judgment till the psychologist had found out what kind of a mind was producing that estimate." And elsewhere he refers to the fact that his students do not know "whether the moon is small as a pea or large as a man."

Now, this experiment, damaging as Professor Münsterberg considers it, has, in point of fact, no bearing whatever upon the perceptive powers of his students. If they understood perfectly what his question meant, and shaped their answers accordingly, those answers were the outcome of some kind of conscious mental calculation or experimentation (unless, indeed, we suppose that some of the students, for example the one with the pea, had made the explicit experiment of the obscuration of the moon by an object held at arm's length, and remembered the result). There is absolutely nothing in our direct perception of the moon to give us any idea of the angle it subtends, or of the size of an object which, held at arm's length, would "just cover it." Professor Münsterberg is, indeed, in this dilemma: either he wants to know what the intuitive feeling of a given man is as to the apparent size of the moon when he looks up at it, or he wants to know just what he explicitly declared—what object, held at arm's length, would cover it from the eye (i. e., one eye). If he means the former, the intuitive feeling, each person is the final arbiter of the question—he is either lying or telling the truth; the moon either does or does not seem to and a commentary of even greater length. Such a procedure would be him to be of the size of a plate, a cart-wheel, or what not, and there's an end. His answer can not be challenged on the score of malobservation; one may impeach either his memory or his truthfulness, but not the accuracy of his observation or even of his judgment. In other words, if the man misunderstood the question, and was really answering this other: "How large does the moon instinctively appear to you?" his answer, "as large as a carriage-wheel" or "as large as a one-cent piece," is better than any that either the psychologist or the astronomer can supply for him. If, on the other hand, the student understood the question as it was "carefully explained" by Professor Münsterberg, he absolutely deceived himself if he imagined that his "perception" of the moon's size had anything to do with the answer. We do not, by direct perception, actually estimate the angle subtended by an object, nor do we perform any equivalent operation. For any object within a small distance, we automatically make absolutely complete compensation for the diminishing angle under which it is seen as the distance increases; a plate, a silver dollar or a pea, held at arm's length, looks precisely as large as it does when held at the distance of a foot—we have no consciousness whatsoever that the angle it subtends is only one third as great in the former case as in the latter. The moon subtends an angle of half a degree; a large pea at arm's length does the same; a large pin's head, perhaps, does the same a foot away from the eye. But the keenest observer in the world is no more aware of these things by direct perception than is the most ill-constructed member of Professor Münsterberg's class. Our automatic compensation for diminishing angle becomes very imperfect both at great distances and in unusual circumstances—such as looking down upon the floor of the rotunda of the capitol from the gallery at the top; and for celestial objects, like the sun and the moon, it of course falls infinitely short of requirements. We make some compensation, though immeasurably less than what is required; and the well-known fact that people differ enormously in their feeling of the apparent size of the moon merely shows that the amount of this compensation (which, in any event, has no simple relation to the actual size and actual distance of the moon) is very different with different persons. But, strangely enough. Professor Münsterberg seems to lose sight of all these facts, and actually to regard the question of the apparent size of the moon as identical with the mathematical question of the angle that it subtends—or, what comes to the same thing, the size of an object which, held at arm's length, will "just cover" it. For while he "carefully explained" the question as meaning the latter thing, his comments relate to the former; and, in particular, in the closing remarks of his article, he speaks of his students not knowing whether "the moon is small as a pea or large as a man." By "is," he of course means "seems"; but this is a radically different question from the one specifically stated at the outset. Thus the master himself clearly slipped from one meaning to the other; and we may be quite certain that some of the answers of his students were intended for one interpretation of the question and some for the other. I feel quite sure that no member of Professor Münsterberg's class really thinks, when he looks up at the full moon, that a solid disk of the size of a carriage-wheel held up at arm's length would just suffice to shut it out from his view. He knows very well that that would shut out a considerable part of the whole sky; and the man who gave "a carriage-wheel" as his answer was almost certainly speaking of how large the moon seemed to him and not of the other question.

This example, from its peculiar nature, has required much space for its discussion; and I hasten to add that, in singling it out, I have put Professor Münsterberg's worst foot foremost. In no other of the instances is there involved, as in this one, a fundamental error. Yet a defect which, in this instance, reached the proportions of downright error is in some measure present in nearly the whole of the article. The defect I have reference to is a failure adequately to discriminate between conscious inference or conjecture, on the one hand, and the immediate dictum of sense-perception, on the other. I am perfectly aware—and it is a commonplace not only of books on logic and psychology, but also of the ordinary text-books of law—that no sharp line can be drawn between these two things. In almost every judgment, however immediately it seems to be given by the impressions made on our senses, an element of inference, conscious or unconscious, enters. Yet there is a vast difference between different cases; and, furthermore, a difference which is distinctly recognized by the wayfaring man. Professor Münsterberg begins his article by citing contradictions of testimony as to whether a road was dry or muddy and as to whether a man had a beard or not; but the staple of his article relates to estimates of the number of spots irregularly scattered on a sheet of cardboard, the rapidity with which a pointer moves around a circular dial, the interval of time between two clicks, and the like. Nowhere does he intimate that there is any vital difference between questions like these and questions of the simpler kind with which he starts out. But when a man is asked how many people he sees in a hall or how fast a train is moving, he knows perfectly well that the validity of his answer is of a wholly different nature from that which attaches to his statement as to whether the road in front of him is wet or dry, or whether a man he is looking at has or has not a beard. In the former cases, he is guessing or consciously estimating, and knows he is guessing or consciously estimating; but when he says that the road in front of him is wet or that the man he is looking at has a beard, he is making an assertion in which he places implicit reliance as the direct result of the evidence of his senses. Most of the examples dwelt upon by Professor Münsterberg are simply proofs of the want of skill of his students—and of most persons—in making certain kinds of numerical estimates. This is interesting, and even important; but it has no such range of bearing as Professor Münsterberg seems to impute to it. And unfortunately in the only instances in which the verdict called for turned on an immediate sense-perception, the subject-matter chosen was of such a character as greatly to reduce the significance of the result. To ask whether a bit of blue paper or a bit of gray paper is the darker is to ask a technical question; the plain man simply can not put himself into the proper attitude of mind to dissociate the question of color from that of illumination, and the most natural conclusion from the failure of a number of Professor Münsterberg's students to answer the question correctly is that they had not succeeded in training themselves to that special task. The other case of direct perception is open to a similar objection. Professor Münsterberg "asked the class to describe the sound they would hear and to say from what source it came. The sound which I produced was the tone of a large tuning-fork, which I struck with a little hammer below the desk, invisibly to the students. Among the hundred students whose papers I examined for this record were exactly two who recognized it as a tuning-fork tone. All the other judgments took it for a bell, or an organ-pipe, or a muffled gong, or a brazen instrument, or a horn, or a 'cello string, or a violin, and so on. Or they compared it with as different noises as the growl of a lion, a steam whistle, a fog-horn, a fly-wheel, a human song, and what not." What does this show but that to the habitual thoughts of a great majority of the men the tuning-fork is highly unfamiliar? Otherwise nothing but diabolical perversity could have caused 98 out of 100 of the men to declare that some other instrument had produced the sound. When driven to a guess as "to what some highly unfamiliar object is whose existence is announced to one of our senses, we do the best we can under the circumstances; but this is something quite different from what we habitually do under ordinary circumstances. I know of a teacher who, whenever his entire class did phenomenally badly in answering a particular examination question, inferred not that the class was stupid, but that the question was an unfair one.

While, then, nothing can be more certain than that great clangers lurk in the possibilities of malobservation, I think I have shown that the case of the average man is by no means as bad as an uncritical acceptance of the specific charges against him so formidably presented by Professor Münsterberg would lead one to believe. One of them falls down completely; most of the others relate to matters that are universally recognized to be matters of estimate or conjecture, and yet are used by Professor Münsterberg as though bearing with full force on the question of the reliability of ordinary simple observation; and even the two remaining ones have peculiarities which greatly restrict their significance. Indeed, a moment's reflection would be sufficient to convince one that things are not as bad as they look through the professional glasses of Professor Münsterberg. If our immediate perception of the things around us, or even our judgment of times, distances, velocities, etc., were as desperately deficient as the whole tenor of Professor Münsterberg's article implies, the world could not be carried on as it is. Business transactions take place every day by the hundred million which turn on the unhesitating recognition of a rarely seen face, and the number of cases of mistaken identity is infinitesimal in comparison; we cross in front of trolley-cars and automobiles and bicycles millions of times just near enough to escape being run over, to once that we actually get run over. To translate that fine judgment of time and distance which brings us safely across Broadway into terms of feet and seconds is a task that most of us perform extremely ill; but this is a particular matter affecting the value of testimony of a special kind, and not touching the general reliability of human observation. This latter is itself undoubtedly highly impeachable; but Professor Münsterberg's tests add little, if anything, to the impeachment.

But there is another side to Professor Münsterberg's article. It is designed not only to show how frequently observation is untrustworthy, but also to advocate a method of classification of witnesses by which the trustworthy ones may be separated from the untrustworthy. "The progress of experimental psychology," he says, "makes it an absurd incongruity that the state should devote its fullest energy to the clearing up of all the physical happenings, but should never ask the psychological expert to determine the value of that factor which becomes most influential—the mind of the witness." That an appeal to psychological experts may in certain special cases be necessary or desirable I do not at all wish to deny; but it seems very clear to me, from the evidence of Professor Münsterberg's own paper, that any attempt to introduce psychological tests as a regular part of the machinery of courts in their dealings with witnesses would be utterly futile. It is conceivable that psychological experts who combined the highest scientific attainments with the most consummate common sense, and the greatest precision of reasoning with the utmost practical caution and shrewdness, could, by subjecting a witness to a sufficiently comprehensive examination, arrive at an authoritative determination of the weight that ought to be attached to his account of the facts which he alleges to have come under his observation; but nothing short of this would suffice. The difficulties in the way are many and great; but first and foremost among them comes the distinction between a laboratory experiment and the involuntary or unregulated experience of real life. Certainly accuracy of observation, whatever other elements it turns on, turns very largely on the question of attention or interest; and the state of the attention or interest of the observer is wholly different in the ordinary circumstances of life from what it is when the observer is being tested by the psychologist. It is entirely possible that when fifty little black squares irregularly pasted on a large sheet of white cardboard are exposed for five seconds to the gaze of A and B, A will make a much better guess than B at their number, owing to his ability to concentrate his attention or to make a swift calculation; and yet that if A and B were in a hall with fifty people in it, B would instinctively have a much better idea of the number of people in it than A, owing to a habit of being interested in the scenes of which he is a natural part and in their significance. Again A, fixing his attention on the end of a black pointer moving over the edge of a white dial, may be vastly better able than B to get that ratio of the consciously observed space to the consciously observed time which is the velocity Professor Münsterberg desired his students to measure in one of his experiments; and yet if A and B were walking casually along the street, B might be an incomparably more reliable witness on the question whether an automobile was or was not exceeding the legal speed limit. And this matter of the different distribution of interest in different circumstances is only one of a vast number of elements which go to making the psychologist's test highly precarious. You must catch a man "in his habit as he lives," you must follow him into all sorts of situations under all sorts of circumstances, internal and external, before you can decide what value to attach to his statement as to the facts that come into his ordinary experience of daily life. The man who may be too dull-witted to understand the psychologist's question, too lethargic to make a decent observation of what is put before him by his examiner, or too "rattled" to state correctly the result of that observation, may be a man who, as he goes about his work or chats with his fellows, misses nothing of the ordinary human occurrences that take place around him. And, on the other hand, the man of quick intelligence and keen activity who, upon demand, can bring all his faculties to bear upon a subject on which he is challenged to make a creditable report may, not only in spite of having this temperament, but actually because of it, be the very man who habitually takes extremely imperfect notice of the visible and audible things that are going on around him all the time and that have for him no significance.

Difficulties like these—I do not say insuperable difficulties, but certainly difficulties that offer enormous resistance to the investigator—are inherent in the subject. But over and above these inherent difficulties are those which attach not so much to the investigation as to the investigator. As a practical proposition, Professor Münsterberg's project must contemplate the employment of the psychological expert as an expert, strictly speaking. His report on the capacity of a witness must be taken by court and jury on his authority; it could not be expected of the court, still less of the jury, that it should examine into the soundness of the methods which he had employed. Yet it must be plain from what has been pointed out above that, even assuming—which is a great deal—the possibility of expending the necessary amount of time and pains upon the inquiry, none but the most highly qualified expert could be safely entrusted with it. Professor Münsterberg is a man not only of the highest professional training, but of extraordinary native powers of mind; yet he not only makes a fundamental error in one instance, but in a number of others overlooks elements essential to the true bearing of the facts upon the question in hand. In addition to the points already noted, one other may be mentioned which throws perhaps an even stronger light on the pitfalls that lie on all sides. It is a curious circumstance that in none of the questions put by Professor Münsterberg to his class does he give any room (or at least any encouragement) to the simple answer, "I don't know." How many of the queer guesses he got in response to the question as to what caused the sound he made by striking the hidden tuning-fork would have been choked off by the simple and straightforward plan of telling the students to answer only in case they felt a reasonable assurance, there is no means of telling. And yet in court a truthful witness would do that very thing. If he had heard a sound the character of which he could not identify, he would so state to the court, and not say it was a bell or a church organ or a human song or what not. A man who, upon being asked to make the best guess he can, makes a very bad guess is not necessarily an unreliable witness; he may be the very man who on the witness stand would refuse to testify to things that he doesn't feel sure of, and who, when he does make a statement, may be implicitly believed. And the same remark applies, in some measure, to nearly all the tests in Professor Münsterberg's questionary. If Professor Münsterberg has laid himself open to criticism in so many points, how much less would it be possible to entrust to an every-day psychologist the decision of so delicate a question as that of the degree of reliability of each of the witnesses in a given case?

There is one very striking test, of a different character from any of the others, to which I have made no reference, and which might be pointed to as concrete proof of the correctness of the method in spite of any criticism that may be brought against it in the abstract. This is an experiment in which Professor Münsterberg, having asked his class to describe everything that he was going to do from one signal to another, did certain conspicuous things with his right hand, upon which he ostentatiously fixed his own attention, while at the same time he did a number of other things with his left hand. The result was that 18 out of 100 students were utterly unaware that he was doing anything at all with his left hand. This of itself would not be surprising, it being a phenomenon familiar in all sorts of sleight-of-hand performances; but the striking fact was developed that of these eighteen non-observers fourteen were included among the twenty men (or thereabout)[2] who had judged a dark blue to be lighter than a certain lighter gray. "That coincidence," says Professor Münsterberg, "was of course not chance. In the case of the darkness experiment the mere idea of grayness gave to their suggestible minds the belief that the colorless gray must be darker than any color. They evidently did not judge at all from the optical impression, but entirely from their conception of gray as darkness. The coincidence, therefore, proved clearly how very quickly a little experiment such as this with a piece of blue and gray paper, which can be performed in a few seconds, can pick out for us those minds which are utterly unfit to report whether an action has been performed in their presence or not. "Whatever they expect to see they do see; and if the attention is turned in one direction, they are blind and deaf and idiotic in the other." That the coincidence is not a matter of chance may be admitted as practically certain; and yet there is ample room for disputing the inference which Professor Münsterberg draws from it. He finds in it a triumphant proof of the adequacy of an extremely simple and special little test for a sweeping conclusion as to the general powers of observation of the men subjected to it. But surely there is another possible explanation. There is one thing that both the color test and the sleight-of-hand test have in common; the danger of a wrong answer in either case may lie chiefly in a failure on the part of the student to grasp firmly and clearly the exact and full import of the question. The man that is alert and keenwitted and intent in his attitude toward the test will both know exactly what is meant by the question of the relative brightness of the two colored papers and be on his guard as to the possibility of a trick (for that is what it is) in the attempt to concentrate his attention upon the spectacular doings of the right hand. The man less keyed up to the requirements of the tests will be in danger both of failing to make the requisite discrimination in the question of brightness and of falling into the trap laid for him in the sleight-of-hand performance. Professor Münsterberg may have (but he certainly does not mention it) confirmatory evidence of the conclusion he draws from the coincidence; but on the face of it that coincidence may quite as plausibly be accounted for in the manner I have indicated as it is by the supposition that an inability to determine which of two differently colored paper squares is the darker carries with it a high probability that the observer is "utterly unfit to report whether an action has been performed in his presence or not." The common observation of every-day life is of a radically different character from what is involved in either of these two associated tests; and the man who distinguishes himself in both may, in ordinary life, with his senses and his intellect in their normal state, be a far worse observer and a far worse reporter of the things going on around him than the man who is either too indifferent or too slow-witted to bring his mental and physical faculties to bear adequately upon the artificial test.

It is far from being the purpose of this paper to cast discredit on the methods of experimental psychology. On the contrary, the writer feels that the advance made by that science has been among the most interesting and important of the scientific developments of the past three decades. In Professor Münsterberg's succeeding paper, "The Third Degree," for example, are to be found a number of illustrations of the remarkable results that have been obtained by the methods of experimental psychology. The account of them given by the distinguished professor with that skill and attractiveness of which he is a rare master, while as interesting as a romance, is full of convincing force. The questions there discussed have at once intense theoretical interest and great practical importance; but there is this difference between them and the questions at issue in the paper I have been criticizing: In worming out of a suspect, or a hysterical patient, the secret he is endeavoring to guard, the thing under examination by the expert is highly definite. It is something in the actual contents of the subject's mind or in his emotional susceptibility on certain definite matters. The study of his reaction-times, association-times, etc., has been shown to furnish astonishingly definite information on these specific things. The point made in the present paper is simply that no such case has been made out in the much broader, looser, more varied and more intangible region covered by the question of the trustworthiness of every-day human observation; that the indictment brought against such observation, not so much by the exact letter as by the whole tenor of Professor Münsterberg's article, is not sustained by his instances; and finally that, so far as regards the discrimination of trustworthy from untrustworthy witnesses (truthfulness aside) as to the affairs of ordinary life, the investigation by expert psychologists, in order to be entitled to authority, would have to be vastly more comprehensive and vastly more able than there is any practical possibility of commanding. That there are special classes of cases in which the expert's investigation of a witness would be of value is certain, but the scope of his usefulness must be regarded as severely limited. So far as the ordinary run of things is concerned, the present homely procedure, imperfect as it is, is to be preferred to a system in which, over and above the question of the trustworthiness of witnesses, there would be injected into every case of importance the further and at least equally puzzling question of the trustworthiness of the tests employed by the psychological expert.

  1. McClure's Magazine, July, 1907.
  2. "About one fifth of the men" is Professor Münsterberg's statement.