Fallacies in the Name of Science
from Stephen Cory, A Beginner’s Guide to the
Scientific Method, 1994, pp. 89-107.

The faith of most people in the credibility of science is nearly unshakable. When we read in the newspaper or see on television that there is "new scientific evidence" for X or that "scientists have discovered" Y, our tendency is to assume that the evidence for X or Y is impeccable. Certainly, the material we have covered in the previous chapters suggests that careful scientific investigation is perhaps the most powerful tool we have for getting at the truth of things. But the methods used by the scientist to investigate the natural world can be abused.

In. this chapter, we will examine a number of fallacies committed in attempting to employ the methods introduced in the previous three chapters.  In logic, a fallacy is a mistake in reasoning. Thus, if from the two statements

(1) Morris is a mammal
(2) Dolphins are mammals

I conclude that

(3) Morris is a dolphin

then I am guilty of a fallacy -- a mistake in reasoning. My conclusion (3) does not follow from (1) and (2), even if (1) and (2) are true. Similarly, a fallacy in applying the methods of science occurs when one draws a conclusion one is not logically entitled to draw, given the evidence available.

We must keep in mind here the difference between an instance of fallacious scientific reasoning on the one hand, and a mistaken scientific belief on the other. Many ideas in the history of science have turned out to be mistaken, but the mistake they involve is not a product of fallacious reasoning. Prior to the mid-18th century, for example, scientists believed in the existence of something called phlogiston, also known as the "fiery substance." Phlogiston, it was thought, was responsible for a number of observable reactions in matter; among other things, it was thought to be the stuff released rapidly into the atmosphere during combustion and slowly as metals decay. Now, as it turns out, there is no such thing as phlogiston; the scientists of the time were mistaken. However, the theory of phlogiston reactions was well supported by a large body of experimental evidence-indeed, the best evidence available at the time. Among other things, the formulas by which metals were produced from ores derived from phlogiston theory. Subsequent experimentation revealed a better explanation for reactions accounted for by phlogiston theory, one involving a new chemical element later to be identified as oxygen. The point here is that both the work that established and that ultimately overturned phlogiston theory involved correct applications of the experimental methods we have been discussing.

By contrast, a fallacy occurs when the methods of science are illicitly applied. Proper applications of scientific method may, as in the case of phlogiston theory, lead to inaccurate results. But they are inaccurate results arrived at by honest investigation. Fallacious applications of the methods of science lead only to a false impression that something has been established with great care and rigor. Indeed, many of the fallacies we shall consider involve ways of lending the appearance of scientific evidence where, in fact, there is little or none.

One well-known fallacy in informal logic is called argumentum ad hominen-attacking the person rather than his or her argument. If, for example, I argue that every student ought to know something about science and, therefore, ought to read this book, you might reply that I receive a royalty from the sale of the book. If the point of your reply is to mount an objection to my argument, you are guilty of an ad hominem fallacy. Even though what you say is true, the point you make is not relevant to the argument I have given. By pointing out that I stand to profit if students buy this book, you attack my motives for arguing as I have, but you have not shown that my argument is somehow defective.

At the risk of committing an ad hominem fallacy, let me propose a generalization. Most, though certainly not all, of the fallacies we will discuss are typically committed by people on the fringes of science, not by mainstream scientists. I By "people on the fringes of science," I mean people who engage in fallacious scientific reasoning for either or both of two reasons.

First, people commit fallacies because they have little knowledge of what rigorous scientific inquiry involves but nonetheless believe they are capable of undertaking such inquiry. Second, fallacies are sometimes committed by people who may well know a great deal about science but who are trying to create the impression that there is some real measure of scientific evidence for something when, in fact, there is very little. Thus, errors of the sort we will discuss are committed sometimes' inadvertently, but sometimes intentionally. Later in this chapter, we will have more to say about the distinction between mainstream science and fringe or pseudoscience, but for now let's begin looking at various types of scientific fallacy.

Our discussion of scientific fallacies will center around three important questions that need to be asked in any scientific investigation:

1. What precisely are the facts of the case?
2. If an explanation is required, have we carefully considered all plausible rival explanations prior to proposing a novel explanation?
3. If a novel explanation is required, can we devise an effective test for its correctness?

Associated with each of these questions are a number of fallacies. The first set involves mistakes in making initial observations of apparently puzzling phenomena. The second involves a failure to consider rival hypotheses.

The final set involves mistakes in testing explanations or claims of extraordinary abilities.

Anecdotal evidence. In Chapter 2, we discussed the dangers of basing generalizations on anecdotal evidence; such generalizations are all too frequently misleading because they are founded on memorable but atypical cases. I've noticed that when I need to do some last-minute preparation for an upcoming lecture, a student invariably knocks on my office door. What is it with students? Do they just have an instinctive sense for the wrong time to come to see me? I'll leave it to you to comment on the fallacy in my observations.

Omitting facts. One way to make something appear mysterious is to ignore certain facts in describing the phenomenon, facts that suggest that the phenomenon in question may not be all that mysterious. In Chapter 2, we mentioned an apparently puzzling phenomenon, crop circles. Large, symmetrical geometric figures, circular and otherwise, have mysteriously appeared in wheat and corn fields in southern England. What we failed to mention is that near almost every crop circle, and in some cases even running through the circles, are what are called "tram lines." Tram lines are the indentations made by tractors as they travel through the crop fields. One of the most puzzling things about crop circles is said to be the fact that there is no sign of human intrusion. There are no footprints or bent plants leading to the circles; thus, it seems unlikely that the circles are hoaxes. It is conceivable, however, that a person could simply walk in the tram lines to the point where the circle was to be constructed without leaving any sign of intrusion. Thus, accounts of the crop circles retain much of their sense of mystery only when certain facts about tram lines are conveniently omitted.

Another example of creating a sense of mystery by omitting crucial facts involves the strange happenings said to occur in the Bermuda triangle, an expanse of several thousand square miles off the coast of southern Florida.

Hundreds of boats and planes have mysteriously disappeared in the area over the years. Books about the mysterious happenings in the Bermuda triangle typically describe in great detail cases in which it is clearly documented that a boat or plane, known to be traveling in the vicinity of the Bermuda triangle, has disappeared, never to be heard from again. Yet two interesting facts are conspicuously missing from most of these reports. In many of the instances described, wreckage is subsequentlyy found, suggesting an accident, not a mysterious disappearance. Moreover, in just about any large expanse of ocean near a populated area like the coast of Florida, there will be a number of disappearances caused by accidents, storms, inexperienced sailors and pilots, and the like. Only when these facts are omitted do the disappearances in the Bermuda triangle take on an air of great mystery.2

Distorting the facts. Another way to create a sense of mystery is to subtly change the content of a factual description. For example, much research has been done in recent years on "near-death experiences." Some researchers claim that people who have been near death, typically during a medical emergency, but who have been revived, have reported a remarkable experience. Here is an account of that experience from one of the best-known books on the subject, Life After Life, by Raymond Moody:

A man. . . begins to hear an uncomfortable noise, a loud ringing or buzzing, and at the same time feels himself moving very rapidly through a long dark tunnel. After this he suddenly finds himself outside of his own physical body, but still in the immediate physical environment, and he sees his body from a distance, as though he is a spectator. . . . After a while, he collects himself and becomes more accustomed to his odd condition. . . . Soon other things begin to happen. He glimpses the spirits of relatives and friends who have already died, and a loving warm spirit of a kind he has never encountered before-a being of light-appears before him. . . . At some point he finds himself approaching some sort of barrier or border, apparendy representing the limit between earthly life and the next life. Yet he finds that he must go back to earth, that the time for death has not yet come.

Now, if this precise experience were reported by many people, we would have quite a remarkable thing on our hands. In fact, the description provided in this passage is based on the reports of hundreds of people. But no two reports are precisely the same. The description we have just read combines elements from many varied experiences. Moreover, no single element in this description occurs in all reports, and no single subject has given precisely this description. Although Moody quite openly admits all of this, many people who argue that near-death experiences provide evidence of life after death accept this artificial account as an accurate description of strange experiences people report when near death. The fact they are liable to report any of a number of things, that reports are frequently at odds with one another, and that many people who have been near death report no such experience, all suggest that there may be a more n explanation for the things people report when near death. At any rate the appearance of a great mystery here is exacerbated by the subtle fabrication of an experience that, strictly speaking, no one has ever had.

Fallacious argument by elimination. Suppose we know that either A or B or C must happen, and we subsequently discover that B or C will happen. Logically we can conclude that A will happen. This pattern of reasoning is sometimes called argument by elimination; it involves establishing one alternative, A, by eliminating the possibility of all others. A fallacious argument by elimination occurs when possibilities other than A, B, or C are ignored in the process of arguing for one of the explicit alternatives. Imagine that I want to establish a particular explanation. I first list possible explanations and then proceed to show that none of the rivals can be correct. Have I established my favored explanation? The answer is no, for two reasons. First, there may be other possible explanations I have failed to consider. Second, even if I succeed in ruling out all the rival candidates I can think of, the failure of these rival explanations entitles me to conclude that the phenomenon in question needs explaining, not that my favored explanation is correct.

A common strategy in ESP research is to claim that an explanation involving some sort of extrasensory mechanism can be established by showing that experimental subjects can achieve results that would be highly likely by chance or luck alone. Thus, for example, a study might claim a particular experimental subject has the gift of mental telepathy (the ability to read the mind of another) because he or she is able to guess the playing card an experimenter is thinking about more frequently than would suggest. Implicit in this claim is a fallacious argument by elimination. That the subject is telepathic follows only if we assume there are two possibilities -- the subject did it either by telepathy or by sheer luck -- and can effectively rule out luck or chance under tightly controlled mental conditions. Yet this assumption is flawed. First, there may be other possible explanations. Maybe an invisible imp peeks at the cards and whispers the right answer in the subject's ear. As wild as this "explanation" seems, it would appear to be as well supported by the experimen1 come as is the telepathy hypothesis.4 Second, even in the absence explanations, the outcome of this experiment does not confirm th that the subject has telepathy. The only conclusion we are warranted drawing, based on the results of this experiment, is that something interesting, something we do not fully understand, is going on. What we are conspicuously not entitled to conclude is that we have evidence for any particular explanation.

Fallacious inference to a causal link. In Chapter 3, we considered the kind of evidence required to establish a causal link. People all too often draw conclusions about causal links based on attractive, even plausible, but unfortunately insufficient bits of evidence. In most cases, the inference seems plausible only because rival explanations for the suspected causal link are overlooked. Conclusions about a causal link between A and B are often drawn on the basis of a number of specific kinds of evidence, none of which, taken alone, is sufficient to support a claim of causal connectedness.

The most prominent of these are:

1. a correlation between A and B
2. a concomitant variation between A and B
3. the fact that A precedes B

Let's consider an example or two of each and the plausible rival explanation each fails to take into consideration.

1. A correlation between A and B. In Chapter 3, we noted that a correlation involves a comparison of two characteristics in a population. From the simple fact that A is correlated with B in Cs, it does not follow that there is a causal link between A and B. Yet people frequently make the illicit inference to a causal connection based on nothing more than a correlation.

Imagine we were to read the results of a study that purported to show a link between a person's astrological sign and his or her profession. Reading further, we discover that the birth dates of a large group of lawyers were examined, and it was discovered that more were born under the sign of Leo than under any other sign. Clearly, there is a positive correlation between being a lawyer and being a Leo. Now, this may suggest that there is a causal link between the two factors. However, there seem to be at least two plausible explanations for the data--explanations that do not involve any sort of causal link between profession and astrological sign.

The first is that the correlation is just a coincidence. If we look at a number of groups by profession, we may now and then find one where there is a significantly greater number of people born under a particular sign, particularly if we restrict our investigation to groups that are none too large. Imagine we were to do a study of plumbers and astrological sign. If we restrict our sample to one or two dozen subjects, chances are quite high we will not find an even distribution under all signs. What we will find is some entirely expectable "clumping." Some signs will have more subjects than others. From here it is but a short step to a claim about a remarkable correlation between being born under a few astrological signs and the profession of plumbing!

The fact that our study cites only one profession and one correlation suggests another possible explanation. It may be that the researchers who undertook the study have presented us with only one small part of their overall data, the part that appears to confirm the possibility of a causal link. Or it may be that, convinced of the truth of astrology, they have inadvertently pruned away just enough data-say, by excluding certain subjects-to lend support to the idea of a correlation.

The explanation for a correlation need not be coincidence nor even fudging, inadvertent or otherwise. Frequently, correlations are explained by some third factor that suggests a possible indirect link between the correlated factors. Suppose, for example, that we discover from careful observation of a number of classes that students who sit near the front of the classroom tend to achieve higher grades than do students who sit near the rear. It may be that this is a coincidence. At any rate, it hardly seems likely that I can improve my grade simply by moving to the front of the classroom. What seems a more likely explanation is that students who want to do well are enthusiastic and want to sit "where the action is "-namely, near the front of the classroom. Thus, it may be that some additional motivational factor accounts for the correlation between the two factors in question.

2. A concomitant variation between A and B. Concomitant variation occurs when a variation in one factor, A, is accompanied by a variation in another factor, B. It is quite tempting to conclude that there must be some connection between A and B if changes in the level of one are invariably accompanied by changes in the level of the other. The problem with such a conclusion is that an enormous number of entirely unrelated things tend to vary in very regular sorts of ways. Over the past ten years, there has been a dramatic increase in popularity of country-western music. At the same time, there has been a corresponding increase in the cost of a loaf of bread. What is the explanation here? A genuinely baffling causal link? Some overlooked third factor? The most likely explanation is that we have managed to pick two completely unrelated trends that happen to be going in the same direction at the same time.

3. The fact that A occurs prior to B. We have all had experiences like this before: Just as you think of someone, the phone rings, and it is the person you were thinking about. Recently, a repairperson fixed my furnace. A few days later, I noticed that the clock on the thermostat that controls the furnace was not working. It seems natural to conclude that something the repairperson did caused the clock to stop. In such cases, the fact that one event precedes another is probably best explained as nothing more than a coincidence. What would be required to discount the possibility of coincidence, in the case thermostat clock, would be some sort of explanation linking the activities of the repairperson and the subsequent behavior of the thermostat.

In summary, many kinds of claims are taken to suggest a causal link.

Among them are (1) claims about a correlation; (2) claims about a concomitant variation; and (3) the claim that one thing happened just before another. Although no such claim should be dismissed out of hand, none should be assumed to establish a causal link. The first step in trying to decide whether there may be a causal link is to consider possible rival explanations. As our examples suggest, likely candidates are coincidence, fudging of data-inadvertent and otherwise-and possible third factors.

Exploiting analogies and similarities. In attempting to explain something puzzling, it is sometimes useful to consider something similar whose explanation is well understood. Thus, for example, in the late 19th century, physicists hypothesized about the existence of what was then called the luminiferous ether, the medium in which light waves are propagated. They arrived at this notion by thinking of certain similarities between light and sound. Both appear to be wave phenomena, and sound waves are propagated in a medium, our atmosphere, much as the waves created by dropping a pebble in a pond are propagated out of the surrounding water. Thus, physicists reasoned, there must exist a medium for the transmission of light waves as well, a luminiferous ether. Subsequent experimentation demonstrated that there is no such stuff, and physicists went on to consider other possible explanations for the propagation of light waves. Interestingly enough, physicists next thought about light in terms of another well-understood phenomenon, electromagnetic fields.

This example illustrates the way in which thinking about a puzzle in terms of something similar but better understood can lead to possible explanations. But it also illustrates the need for independent testing of the explanation arrived at in this way. Analogies and similarities are fallaciously exploited when the fact that an explanation works in one case is given as evidence for the correctness of a similar explanation in another case. At the very most, a well-chosen similarity can guide us to a possible explanation; it should not be thought to provide evidence that the explanation is correct. Only careful testing can provide such evidence.

Consider one explanation often proposed by astrologers. Grant, for the moment, that there may be something to astrology and that, indeed, the position of the stars and planets at the time of birth can influence personality or even choice of profession. What is the explanation? How is it that the stars and planets influence our lives? Astrologers are likely to give something like the following explanation:

Thus, the stars and planets affect our lives much as the moon influences the tides. Of course, there is no claim here that the relation between stars and lives is precisely the same as that between the moon and the tides, or the sun and radio transmissions. What we have, then, is the barest suggestion that an explanation may be possible for astrological effects and that it may somehow be similar to whatever it is that explains the relation between moon and tides, sun and radio transmissions. What we do not have is any of the details of what that explanation might be. Nonetheless, by appealing to something that is understood and suggesting that the explanation for something else must be similar, our astrologer has managed to create the impression that something like an explanation has been given.

Proposing unfalsifiable explanations. To test an explanation, we begin by devising a set of experimental conditions under which we predict that something will occur if the explanation is correct. If the predicted result fails to occur, we conclude that the explanation is probably wrong. \VJ1at this means is that, to be subject to scientific testing, an explanation must, in principle, be falsifiable. Don't confuse falsifiability with falsehood. Correct explanations, as well as incorrect ones, are, in principle, falsifiable; all this means is that they can be tested in the way we have described. By contrast, an unfalsifiable explanation would be one whose falsity could not be detected by any conceivable test. It may seem that an unfalsifiable explanation is simply true, but this is not so. An explanation that is, in principle, unfalsifiable is not a scientific explanation at all. Precisely why this should be so can be illustrated by means of an example or two.

A group of people calling themselves "special creationists" claim that there is "scientific evidence" that the universe was created by God. Some believe creation occurred only a few thousand years ago, while others believe it may have occurred billions of years in the past. Both groups, however, claim that the processes by which God created the world are "special," in the sense that they no longer operate in the natural world; the laws of nature by which God created the world are different from those we currently observe. Well, this is all very interesting. But what prediction about the world could we make, provided this claim is true? If the processes by which God created so quickly and completely are no longer in existence, then we should not expect to find evidence of their continuing operation. And for precisely the same reason, we should expect to find no evidence against the theory of special creation. It would seem that the creationists' explanation is consistent with everything that is happening or could conceivably happen, and thus could not possibly be falsified.

But this means that the creationist account of how things began is not an explanation at all! To explain something is to try to make clear how or why it and not something else happened. A proposed explanation that is consistent with what happened and anything else that could have happened instead, explains nothing. I cashed a large check yesterday and today. discover that it bounced. Looking over my check register, I discover a glaring error in addition; I had much less money in my checking account than I thought. My miscalculation, then, explains why my check bounced. Had I not miscalculated, I would not have written a bad check. Thus, my miscalculation explains why the check I wrote bounced. Imagine instead I gave this as the explanation for my bad check: "It must have been fate. What happens, happens." But what if my check had not bounced? Once again, fate, I say, is the real culprit. Now, it may be that fate determines what we do and do not do. But insofar as the notion of fate is consistent with everything that happens, it cannot be invoked to explain why a particular thing and not something else happened. Maybe fate determined I would bounce a check, maybe not. But by invoking the notion of fate, I do not thereby explain why my check bounced as opposed to not bouncing.

We can say something similar about the creationist's account of the origin of things. Perhaps God created all things and did so in a very short time using special processes no longer in operation. But by venturing this scenario, the creationist has not explained why things are as they are and not some other way; the creationist's scenario is consistent with anything that could conceivably happen. Though the creationist's account is interesting, it is not a scientific account of things. Does this mean the creationist is wrong? No. What it does mean, however, is that special creationism does not constitute a scientific explanation.

So, if we find that an apparent explanation cannot be falsified, we have uncovered a powerful reason to reject it as an instance of genuine scientific explanation. As a rule of thumb, it is always a good idea to ask of any proposed explanation, "Under what conditions would we be willing to set aside the explanation on the grounds that it is false?" If no such conditions can be imagined, we are dealing with something that is at best fascinating speculation, perhaps even an article of faith, but not a genuine scientific explanation.

Many conspiracy theories seem attractive and plausible largely because they are impervious to falsification. Imagine, for example, that I claim to understand why gasoline prices continue to rise at a much greater rate than the cost of living. There is, sorry to say, a plot, a conspiracy, among the major oil companies to ensure that just enough gasoline is refined to keep demand slightly ahead of supply. Might I be wrong, you ask? After all, there have been many congressional investigations of the oil industry, and none has yet turned up evidence for such a plot. Well, what doyou expect, I reply. The one thing we can be sure of in a conspiracy of this magnitude is that the conspirators are going to do everything necessary to cover their tracks, even if this requires buying the services of a few congressmen. Note here how I have attempted to turn the lack of any evidence against my theory into evidence that it is so. Thus, far from viewing its inability to be falsified as evidence that my theory is not scientific, I take this to be evidence that it must be correct.

A common tactic of conspiracy theorists is to attempt to vindicate their theories by reference to the very facts that have occasioned them. You asked whether my theory about the oil companies could be shown to be false. But you didn't ask for my evidence that it is true. That I am on the right track, I might contend, is shown by the fact that if there were such a conspiracy, we would expect gas prices to rise at an artificially high rate. And isn't this just what we find? The problem with my reply here, of course, is that I am using the very facts which have prompted me to give my conspiratorial explanation in an attempt to vindicate it. My thinking here is going in a circle. I explain P by reference to T and then claim that P constitutes independent evidence that T is so.

Much of the plausibility of many conspiracy theories stems from the fact hat they seem to provide a simple and elegant explanation of a number of apparently unrelated but puzzling facts. So, for example, I might go a bit further and point out that it is because of the oil company conspiracy that we see not only the artificial rise in the price of oil but also that lobbyists represent the entire oil industry, not individual oil companies, in the halls of Congress. Moreover, it explains why a few very influential congressmen accept large political donations from the oil industry and even why it is that we see so few independent gas stations today -- gas stations not owned by the major oil companies. Now, a whole series of rather interesting facts are explained by a single conspiracy. Yet in bringing in these additional facts, I am only showing that my theory can be extended to explain a lot. I have yet to provide any evidence that it is true. Though it no doubt sounds intriguing (who among us does not enjoy a good conspiracy?), my theory has yet to be supported by a single independent test.

No doubt there are conspiracies and conspirators, but their existence cannot be proven simply by spinning stories that would, if true, account for a myriad of interesting facts. One antidote to fallacious conspiracy theories involves considering the possibility of a discrete explanation for each of the facts the theory purportedly explains. It may be, for example, that the reason that congressmen accept large donations from the oil industry has little to do with the actual explanation for the demise of many independent gas stations.

In addition to conspiracy theories, we should be wary of any attempt to indicate an explanation by treating known facts as though they were predictive consequences of the explanation. If I know X, I cannot "predict" X as a means of defending a particular explanation for X. One evening not too long ago, I passed a person I had never seen before just prior to entering my unlocked office to pick up some tests that needed to be graded. But the tests were missing! My initial hunch was that the stranger took the tests from my office. Now, it may be that my hunch is right. But suppose some one were to doubt the correctness of my explanation. I do not provide independent evidence for my explanation by again citing the facts that have prompted my explanation -- namely, that I observed the stranger near my office, that the office was unlocked, and that the papers were missing.

Claims of extraordinary abilities and events are often made in such a way that they are unfalsifiable. An astrologer claims to be able to tell us something of what awaits us in the future based on our astrological chart. "Expect to move to the east coast within the next six months," says our astrologer. "However, always remember, the stars impel, they don't compel." Thus, the prediction is vindicated if we move east but also if we do not, for we can always choose to change the possible future laid out in our chart.

Equally unfalsifiable, though for a different reason, would be this prediction: "You'll be making a significant trip within the next six months." No doubt this is right. I, for one, plan on making an important trip to the bank to deposit a large sum of money to cover my overdraft, just as soon as I receive my paycheck. Here our astrologer has guaranteed success not by explaining in advance a possible failed prediction, but by making a prediction that is so vague that it will "fit" an enormous variety of likely events.

Illicit ad hoc rescues. Explanations and claimed extraordinary abilities need not be dismissed simply because, in a given test, they appear to be false. As we said in Chapter 2, it is always worthwhile to consider various auxiliary assumptions made in conjunction with a test. It may be that failure is due to a questionable assumption we have made in setting up the experiment. In the event that we fail to get the result we expected, we might want to modify our test and try again. But this sort of holding maneuver can only take us so far. If numerous modifications yield no different results, there is a point at which we must admit that our initial expectations were wrong. To persist in defending our expectations after it is clear they are probably wrong is to engage in what is called an illicit ad hoc rescue.

Suppose someone has advanced an explanation, but subsequent tests fail to confirm it. The effect of an ad hoc rescue is to suggest a new explanation for the failure to confirm the original explanation. Nothing is wrong with such a maneuver provided the new, ad hoc explanation can itself be independently tested. In fact, such maneuvers are part and parcel of the way in which science is done. Such a move is illicit, however, if it is advanced only to "save" the original explanation by proposing something that, if true, would account for the failure to confirm the original explanation.

The discovery of the planet Neptune provides a good example of the sort of legitimate ad hoc rescue that occurs in scientific research. In the early 1800s, six of the seven known planets in our solar system seemed to obey laws set forth by Kepler and Newton. But the outermost planet, Uranus, moved on a course considerably different from that predicted by these laws. Now, one possibility was that the laws in question were a special case, only capable of explaining the motions of some of the planets. Another possibility, however, suggested a way for the laws in question to retain their generality. In the mid-1800s, astronomers speculated that the peculiar movement of Uranus could be explained in a way consistent with Newton and Kepler if another planet were to exist out beyond the orbit of Uranus. (The gravitational attraction of the newly postulated planet would account for the problems.) Now, at this point in the story, we must regard the proposed new planet as part of an ad hoc rescue. If there is such a planet, the laws in question retain their generality. Fortunately, astronomers were able to predict just where the new planet should be in order to exert the postulated gravitational influence on Uranus, and shortly thereafter Neptune was discovered precisely where predicted. The ad hoc rescue thus turned out to be justified.

By way of contrast, consider the following. Imagine that a psychic has agreed to be tested and further agrees that he or she can perform under the experimental conditions we have set up. Unfortunately, the psychic fails. Nevertheless, claims our psychic, this does not show that he or she cannot do the things in question. For psychic abilities are subject to something called the "shyness effect": Psychic abilities ebb and flow and frequently seem to ebb just when we want them to flow. It is almost, adds our psychic, as though they don't want to be tested. It would seem that the psychic's appeal to the shyness effect is not calculated to help us rethink our experiment, particularly if there is no independent way of testing for its presence or absence. Rather, it is nothing more than an attempt to make sure that, no matter how carefully we design our experimental test, no conceivable result need be taken as repudiating the psychic's claimed ability. By contrast with the legitimate ad hoc rescue that ended in the discovery of Neptune, our psychic's maneuver seems clearly to constitute an illicit ad hoc rescue. It would appear to be untestable, and its only redeeming feature is that, if true, it would "save" our psychic in the face of his or her failure to perform under controlled conditions.

All of the ways we have considered in which scientific inquiry can go astray suggest something of a problem. How do we determine whether a result, advanced in the name of science, is genuine or bogus? Our discussion of fallacious uses of the methods of science suggests one crucial difference between genuine science on the one hand, and pseudoscience -- fake science -- on the other: Genuine science involves the rigorous testing of new ideas, employing the methods introduced in Chapters 2 through 4; pseudoscientific ideas will frequently be backed by evidence that is the product of one or more of the fallacies discussed in this chapter. Though adherence to the methods of science is at the heart of the distinction between genuine science and pseudoscience, there are a number of other important differences between the two, as well as number of mistaken ideas about what the distinction involves.

Science cannot be distinguished from pseudoscience on the basis' of the quality of the results each produces. In science, at any rate, ideas earn their respectability not because they are right, but because they have been tested in the right sort of way. Many of the examples we have considered here and in preceding chapters serve to confirm this. At one point in the history of Western thought, the best-informed scientific view was that the earth is at the center of the universe. Though this view was ultimately shown to be wrong, it nonetheless constituted the best science of the time. Though Ptolemy and his followers were mistaken, their view of the cosmos provided a coherent, testable explanation for a wide variety of phenomena. Our discussion earlier in this chapter of the luminiferous ether provides another striking example of genuine, though ultimately mistaken, science.

The distinction between science and pseudoscience cannot be drawn along lines of scientific discipline. We cannot say, for example, that astronomy is a science whereas astrology is not, that psychology is but psychic research isn't. This is not to say that astronomy or psychology does not deserve to be called a science. But the notion of a science, or scientific discipline, is much too broad for our purposes. My dictionary defines astronomy as "the science which treats of the heavenly bodies-stars, planets, satellites and comets," and I suppose this is as good a definition as any other. But within this broad discipline, we sometimes do encounter instances of pseudoscience as well as of genuine science.

For example, in the 1950s, a self-proclaimed astronomer and archeologist, Immanuel Velikovsky, hypothesized that the planet Venus was created out of an enormous volcanic eruption on Jupiter. Velikovsky speculated that as the newly formed planet hurled toward the sun, it passed by the earth, causing several cataclysmic events, and eventually settled down to become the second planet in our solar system. Yet careful examination of Velikovsky's work has shown that this sort of cosmic Ping-Pong is quite impossible and that Velikovsky either ignored or was unaware of certain physical constraints that his hypothesis violated. One of Velilovsky' s most glaring mistakes involves a well-known law of motion: If one body exerts a force on a second body, then the second exerts a force that is equal in strength and opposite in direction. An explosion of sufficient magnitude to allow an object the size of Venus to overcome the gravitational attraction of Jupiter would simultaneously send Jupiter off in the opposite direction, despite Jupiter's great mass. Yet in Velikovsky's theory, the orbit ofJupiter remains unaffected by this most cataclysmic of events. Here, then, we have an example of pseudoscience, yet one that we can certainly classify under the broad heading of astronomy.

Similarly, early in this century, British psychologist Sir Cyril Burt claimed to have decisive evidence that heredity, not environment, plays the dominant role in determining intelligence. Yet as it turned out, much of Burt's work was based on fictional or distorted data. Burt apparently invented experimental subjects and altered test results to conform to his expectations in the process of trying to make his findings appear to be scientific.

The distinction between science and pseudoscience has nothing to do with the distinction between "hard" and "soft" sciences. The sciences that study human behavior-sociology, anthropology, psychology, political science, to name a few-are sometimes characterized as "soft," as opposed to the "hard" physical and biological sciences. Although the soft and hard sciences differ in a number of respects, none of the differences is sufficient to support the complaint, occasionally leveled against the soft disciplines, that they are pseudosciences. The hard sciences, in their attempts at describing and understanding nature, do not have to deal with the complexities posed by the human ability to choose what to do. It is sometimes said that only the hard sciences are "exact," and this is generally taken to mean that predictions about human behavior cannot hope to be as precise as, say, predictions about what will happen to a gas under a specific set of conditions. Moreover, it is difficult to think of a single "soft" scientific theory that is as broad in scope as the theories of modern physics and chemistry.

The law of gravity describes the behavior of all gravitating objects; it is hard to imagine a similar law describing a single aspect of the behavior of people, societies, economic systems, or political institutions.

Yet despite their obvious differences, the hard and soft sciences are all properly called sciences. All aim at explaining phenomena of the natural world, be it the behavior of matter or the behavior of human beings. And both hard and soft sciences adhere to the methods we have discussed in Chapters 2 through 4 in advancing and testing their "hows" and "whys." Many philosophers argue that the social sciences will never produce the kinds of grand, unifying theories characteristic of the physical and biological sciences; it may be that the "soft" sciences will have to be satisfied with discrete bits of explanatory material, each of which is suited to a limited aspect of human behavior. But insofar as research in the social and behavioral sciences conforms to the more general methods of good scientific research, we have no reason to doubt their qualifications as disciplines capable of delivering genuine scientific results.

Genuine science tends to be self-correcting; pseudoscience is not. We have examined a number of instances in which the results of scientific inquiry have been overturned.; in most cases, mistaken ideas have been rejected on the basis of further scientific inquiry. It is estimated that there are currently about 40,000 active scientific journals worldwide. The function of such a journal is to report new findings and results and, equally important, to provide a forum for the critics of current research. Interestingly enough, most criticism of potentially pseudoscientific research comes from within mainstream science as well. Recent criticism of the work of the special creationists, for example, has been leveled by mainstream anthropologists, zoologists, biologists, and evolutionary theorists. Although there are a few journals devoted to creationist science, it is rare to find a single article by a noted creationist critical of the work of other creationists.

As a scientific discipline develops, it will gradually produce a maturing body of explanatory or theoretical findings; pseudoscience produces very little theory. One major aim of science, as we discussed in Chapter 1, is to "make sense" of nature by providing better and better and, often, more and more encompassing bodies of explanatory material. Think, for example, of all that is known about the mechanisms involved in the transmission of genetic information from one generation to the next, by contrast with what was known 150 years ago when the science of genetics was born. Gregor Mendel, the first great figure in the field of genetic research, began by speculating about "genetic factors" that might be responsible for observable characteristics in some simple varieties of plants. Today, modern geneticists speak of the subtle and complex methods by which DNA is transmitted in any organism.

By contrast, pseudoscientific research almost always produces spectacular claims for extraordinary abilities and events, but little else. Moreover, the nature of the claims produced varies little over time. As it turns out, ESP research began only a little later than did genetic research. Yet today we find little more than an enormous body of controversial evidence that a few people have psychic ability and almost no theoretical understanding of how ESP might work. In many pseudoscientific endeavors, what little explanatory material emerges is likely to be based on vague analogies and similarities drawn from some well-understood area of science. Thus, for example, a book on ESP published in the 1930s was entitled ESP: Mental Radio. An interesting idea, but hardly a reliable explanation.

The findings, theoretical and otherwise, of genuine science are always open to revision; rarely do pseudoscientific claims change much over time. It is hard to imagine a thriving scientific discipline today in which much of what was believed 100 or even 50 years ago has not been supplanted by a more accurate picture of things. Fifty years ago, particle physics provide us with a picture of the world in which the most fundamental particles were the electron, proton, and neutron. A few stray experimental results conflicted with this picture, but few physicists questioned its rough fit with reality. Today, physics provides a more comprehensive picture in which protons and neutrons are composites built out of even more fundamental particles, quarks. The landscape of particle physics has change dramatically in a brief period of time.

By way of contrast, it is interesting to look at the work of modern astrologers. If you were to have a competent astrologer draw a detailed horoscope, his or her work would be based on classic astrological texts written nearly 2000 years ago. Pseudoscientists often claim the long history of their ideas as evidence for their correctness. Thus, an astrologer might boast that his or her techniques are derived from the discoveries of ancient Babylonian and Egyptian astronomers. But from a scientific point of view, any idea that undergoes little or no revision over such a long period of time is probably not the product of careful scientific investigation.

Genuine science embraces skepticism; pseudoscience tends to view skepticism as a sign of narrow-mindedness. The first reaction of a competent scientist, when faced with something new and unusual, is to try to explain the phenomenon by fitting it into what is already known. Many people who engage in pseudoscience see this as the worst sort of skepticism; the fact that one's initial reaction is to try to rob something of its mystery is taken as a sign that one is unwilling to entertain new ideas. It is perhaps this attitude toward scientific skepticism more than anything else that contributes to the tendency in pseudoscience to accept claims in the absence of solid scientific evidence.

The question of whether a piece of "scientific" research is genuine or bogus is not always easy to answer. Though the points of contrast we have highlighted here can provide us with some initial sense of when we are in the presence of pseudoscience, we should not wield them dogmatically. If people purport to have "scientific evidence" for something, we should not dismiss their work simply because, for example, they refuse to countenance serious criticism, complain that their critics lack an open mind, or proclaim the longevity of their ideas. However, such moves should be taken as a sign that something may be seriously amiss. The fundamental difference between genuine and bogus science is really a difference in method. The results of genuine scientific inquiry are the product of open, honest applications of the methods we have discussed in previous chapters. Pseudoscientific results, by contrast, are produced with little regard for these methods.

A person claims to have "scientific evidence" for X. Are we confronted with genuine science or pseudoscience? To answer this question, there is no substitute for taking a careful, critical look at the methods employed in establishing X.

The Limits of Scientific Explanation

In Chapter 1, we said that one major goal of science is to further our understanding of how and why things happen as they do. In Chapters 2 through 5, we have taken a close look at the method by which science attempts to accomplish this goal and at some of the ways in which the methods of science can be abused. One further issue, however, deserves some brief discussion before we conclude: Are there "hows" and "whys" that science cannot help us to answer? Are there things, that is, that science is powerless to explain?

With respect to questions about processes occurring in the natural world, it is hard to imagine a limit to the potential of science to explain. This does not mean that science, given enough time and effort, will provide us with an understanding of all natural processes. What it does mean is that there appears to be no limit to the questions-questions about the natural world-that science, properly carried out, cannot profitably address. And if it turns out that there are such limits, we will discover them only by approaching them scientifically and discovering just how far this approach can carry us.

But there other "hows" and other "whys"-"hows" and "whys" that take us beyond the interests normally associated with scientific inquiry. These are the great questions of metaphysics, questions that have vexed philosophers and ordinary people alike for as long as people have thought. They are questions you have probably wondered about in some idle moment: Why is there anything at all, rather than nothing? Is there some benevolent, creative force responsible for the natural world? Is there, in other words, a God? Why are we here? Do our lives have some ultimate meaning, some cosmic purpose?

Deep metaphysical questions like these, I suspect, will not be settled by scientific inquiry. This is not to suggest that science is somehow deficient. The methods of science are not designed to answer questions of this sort. Science aims to explain processes occurring within the natural world; these deep metaphysical questions raise issues about the nature of the natural world itself. They are not concerned with mechanisms, causes, laws, the very stuff of scientific explanation. They involve, rather, an attempt to understand the purposes behind the sum total of the natural world. If scientific questions are, by definition, questions about how and why things happen in the natural world, then metaphysical question are, by definition, not scientific questions. Even if science were somehow, someday to provide us with an utterly complete explanatory picture of all processes in nature-a theory of everything-science would still leave our deep metaphysical questions untouched. Why this particular set of explanations and not another? What is their cosmic significance? Their purpose? Who is their author?

Not long before his death, Sir Peter Medawar, Nobel laureate in medicine, made the following observation:

Catastrophe apart, I believe it to be science's greatest glory that there is no limit upon the power of science to answer questions of the kind science can answer.

Metaphysical questions aside, there would seem to be no limit to the ability of science to explain as long as we restrict science to the kind of question, as Medawar says, science can answer.

Here is a brief summary of the fallacies we have discussed and of the telltale signs of a pseudoscience.

Anecdotal evidence: basing a general claim on a few anecdotal reports.
Omitting facts: creating an air of mystery by leaving out facts that might account for the mystery.
Distorting the facts: altering the facts to create the impression that something is mysterious

Fallacious argument by elimination: arguing for a given explanation byattempting to show that rival explanations are wrong.

Fallacious inference to a causal link: inferring a causal link on the basis of a correlation, concomitant variation, or the fact that the suspected cause occurred before its effect. Possible rival explanations are coincidence, fudging of data, and third factors.

Exploiting analogies and similarities: treating explanations for well-understood phenomena as though
they were evidence for a similar explanation for something not so well understood.

Proposing un falsifiable claims: (1) advancing a claim, explanatory or otherwise, that is consistent with everything that could happen; (2) working with predictions that cannot be falsified; (3) explaining away all conceivable experimental results that might suggest that a claim is false; or (4) treating initiating facts as confirming facts.

Illicit ad hoc rescues: advancing auxiliary assumptions that cannot be independently verified as a means of saving an explanation or extraordinary claim.

1. Pseudoscientific claims often involve fallacious scientific reasoning of the sort exemplified by the preceding fallacies.
2. Pseudoscience can occur within the bounds of legitimate scientific disciplines.
3. Pseudoscience tends not to be self-correcting.
4. Pseudoscience produces very little explanatory theory.
5. Rarely do pseudoscientific claims change much over time.
6. Pseudoscientists tend to view skepticism as a sign of narrowmindedness.