ABSTRACT: This essay attempts to show that science, in common with other complex human creations, defies uniform, essentialistic descriptions. Most mainstream schools in the philosophy of science make valuable contributions; each tells us much that is worth knowing about one side or another of this complex subject. But, insofar as they believe that this one side is the only one, or that it is by far the most important aspect of the many-sided human endeavor we call science, they are mistaken. If you want to understand what science is, you can only do so by combining the fruitful ideas of various practitioners and theoreticians into one meaningful whole. This article begins by pointing out the similarities and common origins of rationalist disciplines like metaphysics and science, and the differences between both and authoritarian disciplines like shamanism. It then moves on to describe a few well-known episodes from the history of science. From these episodes, from the reflections of various philosophers, and from my own personal experiences as a scientist, it distills a commonsense, integrated portrayal of science.  In this portrayal, good science does not owe its achievements and its intellectual and aesthetic appeal to a universal scientific method, but to constellation of such distinguishing characteristics as anti-authoritarianism, experimentalism, and rationality.

Rabbi Hillel the Elder, who flourished a generation or so before Christ, was approached once by a man who wished to convert to Judaism. This man made his conversion conditional, however, on being told the essence of Hillel's religion in a few words. We can imagine the acclaimed theologian trying to convince this man to enroll in his academy, offering him perhaps a scholarship of some sort or another, and perhaps even, in view of this prospective convert's life experiences, waiving a few of the credits that would be normally required for ordination. We can imagine Hillel telling his visitor that he himself had been trying to find an answer to this elusive question for many years, by much study and note-taking, and that still, despite his fame, he had a long way to go. Attempts to distill the essence of a complex subject are charming in their simplicity, Hillel might have told him, but they suffer from the obvious problems of brevity and subjectivity.

All this to no avail. Hillel's whimsical student was only willing to listen, as he purportedly put it, for as long he could remain standing on one foot, unaided; not a second longer. At the end, Hillel was equal to the task. He converted him, and handed him a rule: "Do not do unto another that which is hateful to you." This, Hillel told him, is the whole of Judaism, the rest is commentary.

Like Hillel's ephemeral student, I too have over the years tried to decipher the essence of a complex system, albeit in my case this system was science. I put the question to a few scientists of my acquaintance and read what the best known experts had to say. Nevertheless, the picture remained blurry. The scientists' view appeared too simple, while the experts were at great odds with each other, each appearing to place too much emphasis on one aspect of a reality with which I myself was familiar as a working scientist and too little emphasis on all the others.

I was content to let the matter rest, but just then I was asked to teach a class to liberal arts students which went by the name "What is Science?" I began by searching for an article which would provide what I intuitively felt were the distinguishing characteristics of the subject, but failed to find anything that came even close to doing this. Given this futile search, and given moreover the complexity and controversial nature of the topic, I decided to approach it indirectly, by relating to my students a few episodes from the history of science and let the episodes themselves answer the question implicitly.

But, like the would-be convert to Judaism, one student went on demanding an answer. And it was this legitimate demand that I try to explain the nature of science in a class titled "What is Science?" that led me to try to capture the essential characteristics of science.

This essay, which is a slightly modified version of a class handout, relies for the most part on experiments and ideas of others. It ignores fine distinctions and countless technical details. It ignores important philosophical questions; for instance, the question whether the computer screen in front of me can be said to exist anywhere else besides my mind. This essay only attempts to provide a rough sketch of the nature of science, and it does so by deliberately selecting for discussion a few misleadingly simple experiments.

What is different here is the approach:

—Most writings on the subject are abstract, and hence somewhat removed from the reality they seek to illuminate. They begin with generalizations, and only then proceed to illustrate them with a few actual examples from the history of science. I chose instead to begin with specific historical episodes, and only then to try to find what these episodes tell us about the nature of science. By recapitulating the process through which philosophers of science arrive at their theories, this approach underscores the accessibility, commonsense nature, and simplicity of science.

—I shall proceed in two steps, starting with the demarcation line between rational and authoritarian fields of inquiry. Only then will I move on to distinguish the sciences from other rational fields of inquiry, and especially from philosophy. Historically, the separation between rationality and authoritarianism, which preceded the separation between science and philosophy, played a decisive role in the subsequent development of science.

—Perhaps owing to the origins of this article in a freshman setting, I shall try to describe a few distinctive features of science in a language that the uninitiated to either science or philosophy can understand.

—By far the most distinctive and useful feature of this essay is its eclectic nature. It deliberately gives up the idea that the essence of something as complex as science can be described by means of a few words or ideas. Most professional philosophers of science are so caught up with controversies and with attempts to discover the heart of their subject that they tend to forget the obvious: complex systems defy uniform explanations. Judaism has no essence; only a fairly long list of salient attributes can bring us closer to understanding its nature. Hillel's dictum tells us something very important about his religion, but it says nothing about God and it fails to distinguish Judaism from either Christianity or secular humanism. Similarly, Marxism provides some interesting insights into history, but it errs in thinking that it is only these insights which are worth knowing. In this essay I shall take the approach that if you want to understand what Judaism, history, or science are, you can only do so by integrating the fruitful ideas of various practitioners and theoreticians into one meaningful whole.

This approach, incidentally, provides an interesting answer to the question:  Is there such a thing as "the scientific method?"  Overwhelmed by the idiosyncrasies of science and scientists, by the fact that no rule seems to ever apply to every aspect of science, some philosophers reject the very notion of a scientific method, claiming that in science "anything goes."  Others, in contrast, insist that there is an underlying universal method to real science. The truth, as usual, lies somewhere between these two extreme positions.  There is no universal method, but in good science one often finds such distinguishing characteristics as letting nature be the final judge, anti-authoritarianism, rationality, hypotheses, predictions, controlled experiments, observations.  Clearly, any one of these characteristics may be absent, ignored, subverted, distorted, or justifiably absent in any actual historical case.   So these characteristics do not constitute a method in a sense of being able to tell scientists how to conduct their next experiment or build their next grand theory--in the same way that skiing manuals can't tell anyone how to win the World Cup.  But, taken together, they do provide useful guidelines and ground rules and they do tell scientists what they should and should not do.  They tell the would-be great scientist such things as:  Listen respectfully an open-mindedly to others, but do not treat anyone's opinion (including your own) as infallible: today's Nobel Prize winner could well be tomorrow's fool.  Cherish your intuition, but be willing to subject it to logical and experimental tests.   Love your hypothesis, if you have to, try to confirm it by all means, but do not forget that in this business we call science, its the Devil's Advocates who have the final say!     

To sum up. The following eclectic review of the scientific method forgoes any attempt at a comprehensive, unified description of science. It lacks therefore the elegance and coherence of the more one-sided approaches upon which it is based. Still, the combination of a few scientific experiments and the insights of a few philosophers of science might allow us to come up with a fairly coherent, commonsense view of the scientific method.

Philosophy and science share a few characteristics which are not commonly found in earlier disciplines like theology and astrology. The birth of philosophy therefore constitutes an important landmark in the subsequent development of western science. To illustrate this point, one can begin by saying a few words about the metaphysics of Thales of Miletus, who flourished in the early Sixth century B.C.

Thales was one of the so-called Seven Wise Men of Greece. According to one story, he was once taunted about his poverty, the implication being that philosophy is useless. To prove that philosophers are poor out of choice (they don't want to waste their lives in the futile pursuit of riches) and not because they can't make money, Thales is reputed to have made a great deal of money by applying his scientific knowledge to speculation in the olive market. But let me go back to the main point. Thales said that everything is water, by which he probably meant that water is the original substance out of which all others are formed. We know now that he was wrong; everything is made of elements like hydrogen and iron, all of which are in turn made of the same kinds of atomic and subatomic particles. But still, his proposal is remarkable in many ways, and it justifiably marks the birth of western philosophy.

Thales of Miletus, 624-547(?) B.C.

To begin with, it was most likely based on observations. The Aegean Sea on whose shores Thales lived was made of water; burning wood gives off steam; water comes out of a squeezed fig. As a first approximation then, the proposition that everything is water is not as silly as it sounds.

Note also that Thales, unlike, for example, a Delphian priestess, does not appeal to authority. A better known anecdote of the conflict between authoritarianism and rationality is provided by the scientists who refused to observe Jupiter's newly-discovered moons through Galileo's telescope. Such moons, they felt, could not exist because Aristotle did not say they existed. Moreover, if these moons existed the number of planets would not be seven. Seven, in turn, is the only proper number for something as important as the number of planets. Have not our heads seven holes? And did not God rest from the task of creating the world on the seventh day?

In contrast to this, Thales's dictum suggests a partial throwing away of the shackles of authority. Somehow, one gets the feeling that Thales would have been willing to consider alternatives to his proposal that everything is water. From its remote beginnings, philosophy and science are marked by skepticism towards conventions and dogmas, and by tolerance and open-mindedness towards new and unfamiliar ideas.

Thales seems to appeal only to the observed world and the intellect. We see here evidence of a new confidence in the power of one's mind, an indication of incipient individualism that would have been unthinkable to the ancient Hebrews or 19th century Inuits. I shall defer to my judgment, both philosophers and scientists at their best say, not to that of my parents, elders, or tribe. Traditionalists, on the other hand, often perceive this assertive individualism as subversive and irreverent.

Another aspect of Thales's adage is the use of creative imagination and intellectual courage—it takes some creativity and daring to come up with such a radically new proposal.

Unlike most of his contemporaries, Thales is curious about a problem, or a puzzle. In his case the puzzle is: What is everything? He is not content to accept traditional solutions. Instead, he limits his search to rational explanations. He concludes with an account which seems to satisfy our own criteria of rationality, and which puts the world around us in a new light.

Certainly, then, philosophy is not shamanism. But, we feel, something is still missing here. Philosophy is rational and open-minded enough, but it isn't science. How then do you make a scientist out of a philosopher?

To begin answering this question, let us leap from ancient Greece to late Renaissance Italy and the work of Galileo's student Evangelista Torricelli. Again, he begins with a puzzle: Why can a suction pump only raise well water to approximately 34 feet? Most people take such observations for granted; like philosophers, scientists are too curious to simply accept them. Like detectives, both scientists and philosophers want to understand causes of observed phenomena. They reject, however, mythical explanations of the type that ascribes leglessness in snakes to a divine curse; they limit the search to rational explanations.

Both philosophers and scientists imbibe the ideals of an intellectual tradition. We know little about Thales' predecessors, but we can be fairly certain that Thales and the philosophers that followed him were well-versed in the traditions, methods, and proposed solutions of their discipline. Similarly, Torricelli became aware of the existence of a problem and the method by which it could be tackled by studying the works of his predecessors and contemporaries. Thus, others before Torricelli, including his teacher Galileo, tried to answer this question, came up with guesses, or hypotheses, tested them, and found out that they were inadequate. We remember Torricelli because he was imaginative and lucky enough to hit upon the correct explanation, and because he was creative and handy enough to devise ways of testing it.

Torricelli hypothesized that water moves up the pump because the air which envelopes the earth exerts pressure on the water's surface. Normally, this atmospheric pressure is balanced everywhere, so water stays where it is. But suction creates vacuum. That is, it virtually eliminates air and air pressure below the pump's piston. Now the atmospheric pressure on the well water's surface is no longer equalized by pressure inside the pump, and the water rises to the point where atmospheric pressure equals the pressure exerted by a column of water.

Torricelli and others then proceeded to test his hypothesis in two ways. First, he reasoned, if his hypothesis of atmospheric pressure is correct, then the same setup will raise the liquid mercury—whose specific gravity is 14 times that of water (e.g., a glassful of mercury will be roughly 14 times heavier than a glassful of water)—to 34/14 feet, or roughly 30 inches. This is exactly what happened. Second, in a mountainous area the atmospheric air column is shorter, hence lighter. Therefore, the mercury column at a mountain's top should be shorter than at its bottom. Indeed, for the particular mountain which was first tested, the column was shorter by 3 inches.

There is certainly something more here than Thales's philosophizing, but what is it? I think this boils down to a few additional elements.

—Above all, it is the leap from speculations to empirical tests that marks the scientist from the philosopher. Thales, as far as we know, stopped at the water's edge. The typical Greek philosopher believed in the superiority of the intellect over the senses and would have been inclined to stop with the hypothesis of atmospheric pressure. For a scientist, this rational explanation merely marks an important starting point. The final arbiter of the truth is not a Platonically divine and introspective mind, but observation and experiment. Unlike philosophy, science entails more than casual observations and hard thinking. It entails, rather, a critical encounter between valid rules of thinking (logic) and observations.

—Science at its best can make reasonably accurate predictions, e.g., that at sea level a column of mercury will rise to 30 inches. Such predictions, if confirmed by empirical tests, increase our confidence in the veracity of the hypotheses from which they were derived. Accurate predictions might also possess an interesting psychological quality. Without Torricelli's hypothesis, there would have been no reason on earth to expect that mercury will rise to 30 inches at sea level or to 27 inches at 4,800 feet above sea level. Because such observations are so unexpected, they lead us to eagerly embrace the insight from which they were deduced.

—The questions of religion and philosophy, which include fundamental questions about the existence of God and the trustworthiness of our senses, are certainly more profound and interesting than day-to-day science and its preoccupation with such seemingly trivial topics as the height of water in a suction pump. But, some philosophers of science argue, religion provides comforting fairy tales, philosophy provides endless controversies, while science gives us successive approximations of the truth. Even though, with few exceptions, each practitioner makes a miniscule contribution to this never-ending Tower of Babel we call science, the scientific enterprise—taken as a whole—does tell us a great deal about reality. Not, it is true, as much as religion and philosophy attempt to tell, but much, much more than they succeed in telling. Science does not give us as unified a worldview as that provided by theology and metaphysics, but what science does tell is far more likely to be true. Also, despite its seemingly trivial and abstract nature, science gives us considerable power over the natural world and our destiny.

—Failed predictions play a critical role in the progress of science. To begin with, under ideal conditions, scientific hypotheses are falsifiable (refutable). If, for example, mercury rose to 34 feet, Torricelli most likely would have rejected his hypothesis of atmospheric pressure. In view of its importance, let me present this aspect of science a bit more formally. Let us call:

Then the argument assumes the form:

"If H is true then P is true; P is not true; therefore, H is not true;" which is a perfectly valid argument.

One philosophical school argues that this refutability-in-principle is the most distinguishing characteristic of science. You can say or do little to shake off the convictions of true believers in Islam, Judaism, Marxism, or astrology. But in science, these philosophers say, you need only common sense and experiments to discard hypotheses. Indeed, in simple cases like Torricelli's, conjectures are likely to be rejected in the face of failed predictions.

But the real world is not that simple: precious little in it can be shown false on logical grounds alone. Thus, a failed prediction rarely forces us to reject the original hypothesis. In Torricelli's case, we might surmise, for instance, that our ruler was grossly inaccurate, and that mercury has in fact risen to the predicted height. Or we might say that mercury is not as dense as we thought it was. So, in the majority of cases the role failed predictions play in the progress of science is a bit more complex than the immediate refutation of hypotheses.

By 1943, for example, it was known that genes of higher animals and plants are carried in chromosomes, and that the only two chemical constituents of chromosomes which could conceivably carry the genetic information were proteins and DNAs. It was also believed, in view of the diversity and functional attributes of the genetic material, that the molecule which served as the genetic material must be complex. At that time, the available biochemical evidence suggested that proteins were exceedingly complex and that DNA was a simple, fairly uniform molecule. Naturally then, most biologists believed that the genetic material was made of protein and not of DNA.

In 1944 this protein hypothesis encountered its first serious setback when it was shown that nonvirulent bacteria could become virulent by exposure to DNA (but not to protein) extracted from dead virulent bacteria. Needless to say, the substance that could accomplish such a transmissible transformation would appear to qualify for the role of the genetic material. So, in the imaginary world of formal logic, this experiment might have constituted a refutation of the protein hypothesis. But real life has more to it than plain logic. In this case, this new experiment could be explained in many ways, of which only two need to be mentioned here. First, it could be supposed that DNA was the genetic material and, therefore, that either the premise that the genetic material must be complex, or the observations about DNA's allegedly simple structure, were incorrect. Alternatively, it could be reasonably argued that protein was the genetic material, that DNA was indeed too simple, and that the new observations were either inaccurate or irrelevant. At any rate, while the first significant failed prediction of the protein hypothesis could not settle the issue, it did point to an anomaly. It raised the DNA hypothesis to the status of a respected rival of the protein hypothesis. By suggesting that the protein hypothesis might be mistaken, it spurred and guided future research along lines that might have not been pursued in its absence. Historically then, this failed prediction could not, and did not, serve to decide the issue right there and then, but it made a critically important contribution to its final resolution through the observations it made and, perhaps even more important, through serving as a potent catalyst for further research and discovery.

At the heat of the moment, the scientists involved in such historical crossroads, and in countless other controversies of lesser significance, are too preoccupied with their reputation, egos, and careers to give the scientific side of such debates their undivided attention. There is no question, however, that such debates play positive roles in the evolution of science. They lead the protagonists to subject their own and their rivals' hypotheses to more rigorous tests. Since both camps play the game by the same fundamental rules, and since most people are not too good at being judges of their own cause, the mutual criticisms engendered by a debate, and the scientific convention that reasonable objections be answered, lead the protagonists to improve and clarify their procedures, and thereby come closer to a resolution of the debate. Science is committed in principle to trial and error methodology, rationality, open-mindedness, and learning from one's mistakes. But, needless to say, this is the claim of the ideal which can perhaps be approximated, but which can only rarely be reached. Scientific controversies thus serve the highly constructive role of enabling scientists to come closer to this ideal. Such controversies thereby make it possible for science to advance far more rapidly than it would have if harmony always prevailed. We may note here in passing the resemblance between all this and other interdependent entities like national economies or ecosystems, in which an individual is "led by an invisible hand to promote an end which was no part of his intention."

To distill a few more facets of the scientific enterprise, let me briefly mention Ignaz Semmelweis and his lonely struggle against childbed (puerperal) fever. The setting is a mid-19th century Austrian hospital. At Semmelweis's maternity ward women's mortality rate is 8.8%. At the adjacent ward it is 2.3%. That is, in his ward almost four times as many women suffer an agonizing death shortly after childbirth.

Like Torricelli, Semmelweis's quest begins with a puzzle. For him the puzzle is: What is the cause of this striking difference between the two wards? In this instance the problem for this uncommonly compassionate physician is particularly urgent, because one is talking here about the seemingly needless suffering of women all over the western world. And the cause for this butchery? Most likely, Semmelweis and his colleagues. Probably, then, Semmelweis was driven to find an explanation not only by curiosity, by his fundamental humanity and an overwhelming sense of guild.

Like Torricelli, Semmelweis's approach is a bit like that of Sherlock Holmes at his detective best. But unlike Torricelli and Holmes, he experienced many false starts. And, as we shall see, he hits upon the correct explanation through a serendipitous event.

First, he applies the force of logic to the widely accepted explanation that childbed fever is attributable to epidemic influences. This is unlikely, in view of the following: 1. Epidemics should afflict the two wards equally. 2. They should afflict the entire city, yet while the fever was raging in his ward, hardly a case occurred in the city of Vienna (in which his hospital was located). 3. The death rate among women giving birth before reaching the hospital (street births) was lower than it was in his ward. After similarly eliminating diet, overcrowding, and rough handling of patients as possible causes, Semmelweis examined a psychological explanation. The priest bearing the last sacrament to a dying woman was in the habit of passing through Semmelweis's ward, preceded by an attendant ringing a bell. In contrast, in the other ward, the priest had direct access to a dying woman's room. Could this explain the differential mortality rates of the two wards? This hypothesis could not be disposed of with logic alone, and was tested by persuading the priest to reach dying women unobtrusively in both wards. Yet, mortality rate in the ward remained unchanged.

A simple test likewise led Semmelweis to discard the hypothesis that the difference could be attributed to the fact that mothers in his ward delivered on their backs while in the other ward they delivered on their sides.

In early 1847, a chance occurrence gave him the decisive clue. A colleague of his received a puncture wound from a scalpel while performing an autopsy. This colleague died from this minor cut, showing identical symptoms to those usually observed in patients afflicted with childbed fever. This led Semmelweis to hypothesize that childbed fever is caused by infection: during routine examination of women in labor, doctors infected them with cadaveric matter from autopsies they had performed earlier.

Now, this hypothesis, like the others, could be easily tested. If the cause for the disease is infection through cadaveric matter, then washing hands in chlorinated lime before making an examination might eliminate the disease. This last hypothesis was confirmed: disinfection of hands by all attending physicians brought mortality rates down to 1.27% during the year 1848—a rate comparable to the 1.33% observed in the other ward during that year.

Moreover, this hypothesis seemed to explain well all other observations: 1. Mortality rates in the other division were lower because patients there were attended by midwives (midwives did not perform autopsies). 2. Likewise, lower mortality rates following street births would be traceable to the fact that women who delivered on way to the hospital were rarely examined after admission.

A critical feature of scientific experiments, conclusions, and theories—in fact, everything scientific—is their tentative, uncertain, nature. Experiments can prove fairly conclusively that some hypotheses are mistaken (e.g., that the priest did not cause childbed fever), but they can never provide a conclusive proof for anything (e.g., that childbed fever is caused by contact with cadaveric material). The reason for this grave shortcoming of the scientific enterprise is directly traceable to logic. The argument: "If proposition H is true then proposition P is true; P is not true; therefore H is not true," is valid, and therefore a failure to confirm a statement strongly suggests that it is not true.

But the argument: "If H is true then P is true; P is true; therefore H is true," is a fallacy (logicians call it the fallacy of affirming the consequent). To see this, let us say:

H = Thoreau was a Canadian.

P = Thoreau was a North American.

Now, it's certainly true that if Thoreau was a Canadian he was a North American. It is also true that he was a North American. But H here is decidedly not true—Thoreau was an American, not a Canadian.

So, many circumstances might add support to a scientific proposition: the unexpected nature of some of its predictions, its power to explain all the known facts, the fact that it has been subjected to many tests and passed them all with flying colors, and so on. But all this never amounts to a conclusive proof. There is always a margin of error in science. Absolute truth has not been given to man.

Let me illustrate this inherent uncertainty with two examples. Semmelweis provided strong evidence for the hypothesis that childbed fever was caused by infection with cadaveric material, but this, as the following episode suggests, was merely a closer approximation to the truth than hitherto available, and certainly not the whole truth. One day he dissected cadavers, properly disinfected his hands, and proceeded to examine a festering cervical cancer of a woman in labor. He then examined 12 other women without disinfecting his hands. Eleven of these women later died of childbed fever. This led him to modify his earlier hypothesis. The cause of infection, he now concluded, was not only cadaveric matter, as he thought, but all putrid matter, regardless of whether this matter was derived from dead or living bodies.

Consider, as another example, the two curious reversals of opinion which took place in the scientific community concerning the transformation of one chemical element into another. Ancient and medieval alchemists believed they could strike it rich by finding a stone or a substance that could transform cheap metals like lead into gold. But because they had failed and because their successors adopted the new atomic theory (which "proved" that such a transformation was physically impossible), the alchemists' belief in the philosopher's stone came into disrepute. Thus one orthodoxy gave way to another and alchemy passed away. But the physical impossibility of one age often becomes the everyday occurrence of another, and twentieth century atomic scientists have learned to transform some distinct chemical elements into others.

Ironically then, science—perhaps our most satisfactory tool of exploring the world—is an imperfect tool. At best, the combination of logic and the trial-and-error methodology it wields gives us successively closer approximations of a reality which in some important sense remains elusive. Semmelweis's hypothesis of priestly involvement was wide of the mark. That of cadaveric matter was better, but still deficient. Putrid matter was still better. The germ theory of disease—which explains why putrid matter is infectious—was better. Today our conceptions are even more inclusive and more firmly based on theories and observations. Without a doubt, however, and as a matter of logical necessity, they are still deficient and tentative.

Another important lesson from Semmelweis's story has to do with history and psychology. It is not only science itself which is inherently fallible. The scientific community is made up of fallible human beings too. Reason demanded that Semmelweis's conclusions be immediately accepted all over the western world, thereby saving innumerable lives. Both logic, observation, and morality were on his side, and what more can one ask?

Unfortunately, a great deal more. We are all afflicted with selfishness, greed, and other unwholesome motives and desires, and scientists are like the rest of us. Nor are scientists immune from such psychological failings as conformity, closed-mindedness and conceptual conservatism. They too sometimes dismiss new ideas out of hand without even considering them, and they too have fixed conceptions which they find difficult to change. It seemed absurd to Semmelweis's colleagues that a step as simple as washing hands with chlorinated lime could solve such a serious problem. Besides, who was this upstart Hungarian anyway? And, if proven true, what would the implications of this discovery be to their career? And what if some women needlessly died, aren’t we all going to die at the end anyway?

The geneticist Mendel was an obscure monk working in isolation from the scientific establishment of his day. Even now, it takes a great deal of intellectual effort to understand his writings. He failed his university examination because of an alleged "lack of insight." So his astounding experiments had to be independently rediscovered 34 years after their publication by three scientists who then found out that theirs was a twice-told tale. A few more decades went by, Soviet scientists who upheld Mendel's views suffered persecution and martyrdom, before his theories became accepted all over the world. Needless to say, even the most impressive credentials and lineage do not always suffice to overthrow a strongly held world view, as shown, for instance, in the case of nineteenth century evolutionary debate. So Semmelweis died insane, Mendel anonymous, and Darwin and Huxley amidst a controversy which—because of its wider-ranging implications—is still raging.

So, historically, scientific progress is a turbulent affair. Sociology and psychology, cynical power plays, follies, suppression of opponents, hierarchies, closed mindedness, and conceptual conservatism, play an important role. Beyond all these human failings, what is often needed is a change of Gestalt—a different way of looking at reality. When most of us examine for the first time frame C of the figure below, we interpret it as either the picture appearing in frame A or in frame B, and it is only with some effort that we are able to move from one image to another when examining frame C. Similarly, it takes a psychological adjustment to switch from a creationist to an evolutionary view of the world.

Fig. 1: An illustration of a Gestalt shift (A=hag; B=young lady; C=young lady and the hag)

Cognitive psychology similarly points to the great difficulty of replacing one belief with another. In one telling experiment, some twenty scientists at two major research universities in South-Eastern Michigan were led to believe that balls are 50% larger than they are. They were then asked to determine the volume of two actual balls (thorough liquid transfer). Under these—surprisingly trying—circumstances, many of the scientists adjusted their observations (by about one-third) to fit their preconceived notions and all but one clung to a belief which their eyes had just told them was absurd.

Undoubtedly, other human failings contribute to this turbulence too. For instance, although we ought to examine the issue at hand, not the class origins or previous record of the individual who proposes it, we often find it difficult to separate propositions from their proponents. Sometimes scientists simply fail to understand a new theory. In Mendel's time, for example, biologists were unaccustomed to think in the quantitative terms which Mendel's experiments demanded. Even decades later, it took a mathematician like Hardy to correct an elementary misconception to which some eminent biologists fell prey. Another difficulty stems from our tendency to interpret an intellectual error as a character deficiency; we therefore tend to resist recognition of our errors far longer than a perfectly rational machine might have. Admission of error entails other costs. Not every scientist possesses the intellectual courage of men like Copernicus and Descartes—men capable of privately embracing a theory which put them at odds with prevailing systems of belief. It goes without saying that life has always been easier for believers in conventional wisdom. It is even more difficult to take a public stand in favor of positions which are inimical to one's interests. The majority does not ordinarily partake in the heroism of figures like Aristarchus, Spinoza, Bruno, or Galileo. It is a matter of everyday observation that an expert's public stand and inner beliefs on issues that affect his pecuniary and other short-term interests are not as objective and dispassionate as those of disinterested but equally well-informed observers. Our tendency to conform to established systems of belief also contributes to our resistance to new ideas. We find it difficult to break away from established figures of authority which, in this case, are older scientists embracing a more traditional way of thinking. In more general terms, we can probably say that we are not as rational and idealistic as we sometimes believe ourselves to be.

All this raises yet another question: How many Mendels, Semmelweises, and Darwins never made an impact on science and the scientific record (click here for a more detailed study of this question)? This is not an easily approachable issue, but one incident should suffice to suggest that this question casts a shadow of doubt on many textbook attempts to assign a particular scientific discovery to a single individual. Copernicus, it has often been believed, discovered the heliocentric theory. Yet a few extant ancient writers briefly mention an astronomer by the name of Aristarchus who—some 1,750 years before Copernicus—concluded that the earth revolved around the sun. Now, if it weren't for these brief excerpts, we would have never known of Aristarchus's existence. (Note also that in this case the scientific community remained closed-minded for millennia!)

The choice between conflicting hypotheses (e.g., priestly involvement vs. cadaveric matter) is not always as simple as the foregoing might suggest. Often, as we have seen, the main difficulty is not closed-mindedness in face of conclusive evidence, but the compatibility of the available evidence with two or more radically different hypotheses. Given what was known about the structure of proteins and nucleic acids in 1943, the available experimental evidence, and prevailing speculations on the subject, it made at least as much sense to suppose that the genetic material was made of protein molecules as it was to suppose that it was made of nucleic acids such as DNA. Two striking landmarks in the shift to the nucleic acid hypothesis were two crucial tests whose outcome simultaneously supported the DNA hypothesis and discredited its protein competitor. The first has already been mentioned: DNA, but not protein, induced bacterial transformation. The second, carried out in 1952, showed that bacteriophages reproduce and take over the bacterial cell machinery by injecting their DNA, and not their protein, into a bacterial cell.

This brings us to another important aspect of the history of science. At times, an experiment might constitute a crucial test between two particular hypotheses or theories. Let me demonstrate this aspect of scientific progress through a less technical example. Consider the two conflicting hypotheses that the earth is round and that it is flat. As the figure below suggests, assuming that light travels in a straight path, the way a ship appears to an observer from the shore constitutes a crucial test between these two hypotheses. If the earth is flat, all the ship's parts should disappear from view simultaneously. If it is not, lower parts should disappear first, with the masthead disappearing last.

An illustration of a crucial test

There is clearly more to science than the testing of scientific hypotheses. Naturalists have been observing, describing, and cataloging the living world for millennia, crossing oceans and jungles in chase of some rare butterfly they wanted to capture and describe. Some anthropologists spent years with tribesmen in a remote jungle or a desolate, mosquito-infested arctic tundra because they wished to learn about Inuit culture, not necessarily to test any hypothesis. Many geographers do not seek to explain new terrains, but to describe and map them. It follows that some scientific experiments and activities are not conducted to test preconceived notions but to discover, observe, and classify facts. This is especially true of sciences and fields of investigations which are in a less advanced stage. Galileo directed his telescope to the heavens to find out what they are like; only later did these observations lead him to accept the heliocentric theory. Leeuwenhoek used his new tool (the microscope) to observe anything he could put under the lens—drops of water, semen, or blood—and described it all with few or any preconceived notions. So, in part, the scientific enterprise involves descriptions, observations, and classifications which do not aim at testing particular hypotheses or theories.

Finally, it is important to distinguish between hypotheses, observations, and descriptions, on one hand, and theories on the other. Unlike the former, theories unify many scientific observations, hypotheses, experiments, and the like, into one large, interconnected framework. The ancient remains of a hominid constitute an observation. Their reconstruction, dating, and speculation that they belonged to a Neanderthal man a hypothesis. The similar sequences of the building blocks that make the same functional protein in humans and chimpanzees constitute an observation. One can then go on to catalog and classify molecular affinities of other proteins and nucleic acids in these two species, and to hypothesize that the chimpanzee is man's closest living relative. The assertion that increasingly smaller skeletons of horse-like creatures are found in deeper and deeper geological strata of the earth relies on a large set of observations, descriptions, and classifications. The proposition that all these ancient species constitute the line of descent which gave rise to the modern horse constitutes a hypothesis. But the unification of all these and numerous other observations, classifications, and hypotheses into one comprehensive framework constitutes the modern theory of evolution.

For the most part, scientific theories are similar to the observations, classifications, experiments, hypotheses, and explanations that contribute to their formation. Theories are, for instance, inherently uncertain and they seek to provide rational explanations. They do, however, provide a more coherent picture of a far greater portion of the universe than their contributing parts and they give us greater mastery over nature.

But this greater explanatory power, coherence, and mastery are purchased at a very heavy price. Theories are far more complex than their contributing parts, and their assimilation requires a far greater intellectual effort, dedication, and time. Many scientific experiments and observations can be readily understood by laymen, but their full theoretical import—the way they fit within a larger and more meaningful framework—can only be fully appreciated by specialists. Theories therefore contribute to the esoteric nature and inaccessibility of science. As a rule, theories evolve far more gradually than each of their contributing parts.

Given these characteristics, it is not surprising that the acceptance of a new theory, and the displacement of one theory by another, are often a drawn-out, stormy affair. Indeed, history shows that the process of replacing an old theory with a new one might at times continue long after the new theory acquired a greater explanatory power than the old. Though the reasons for this persistence are not entirely clear, they are probably similar to the ones discussed earlier in relation to new hypotheses, observations, and classifications. The greater complexity of theories, the greater commitment of many individual scientists to them, and the greater difficulty in shifting from one worldview to another, typically make the switch from one theory to another even more wrenching than a switch involving a theory's contributing part.

Consider, for instance, the plight of an individual brought up to believe in a literal interpretation of the biblical creation story. She is asked to replace a purposeful God with the purposeless forces postulated by the modern theories of the origin of life and of evolution. She is asked to replace a simple, commonsense view of the universe with counterintuitive theories that can only be mastered with a great deal of study and intellectual effort. She is asked to embrace theories which contradict one of her most deeply held belief systems. She is asked to replace the comfort of a certain explanation with the anxieties associated with admittedly tentative and uncertain theories.

As we have seen, most scientific transformations are associated with excruciating birth pangs. This undeniable historical observation leads some people to assign to these pangs a pivotal role in science. Thus, some observers talk sometimes as if science and political history are fundamentally alike. They allege that, in both, the system which prevails at any given moment is not necessarily better than its pre-revolutionary predecessors. Others go even farther and argue that science is fundamentally irrational; the choice between one theory and another, they seem to say, is a matter of subjective preference.

Several reasons lead me to reject this radical relativism. First, it does not come even close to the intuitive beliefs that most scientists have of their subject. Second, it places too much emphasis on one side of a many-sided reality, and too little emphasis on all other aspects. Third, it ignores the fact that science, in the final analysis, attempts to explain nature. Why then should one theory be superior to its predecessor—in the operational sense of increased mastery over nature—if neither one objectively possesses greater explanatory power than its rival? Fourth, there is an air of unreality to this relativistic view. Would not a subjectivist requiring genetic counseling, or wishing to construct a bomb with the highest attainable yield-to-weight ratio, seek the advice of the newest theory's practitioners? Would he even dream of carrying his subjectivism to its logical conclusion by seeking the advice of practitioners of older "paradigms" or "Voodoos"? Similarly, can the connections between science and such social ills as nuclear war, overpopulation, and the greenhouse effect be denied? Fifth, it could also be persuasively argued that most established scientific theories possess greater coherence and explanatory power than their predecessors. Granted, nineteenth century atomic theory was inferior to medieval chemistry in its insistence that one element cannot be transformed into another. Granted, evolutionary biology suffers from grave shortcomings and it might very well be displaced one day by a radically different theory. But cannot both theories lay claim to greater rationality than their predecessors? Is it really reasonable to believe that Mendel's explanation of his results, or the chromosome theory of inheritance to which his experiments contributed, owed their widespread acceptance to indoctrination of school children? Does anyone really believe that if naive but intelligent observers were presented with all the facts and asked to choose between just about any scientific theory and its earlier competitor, that the majority would choose the latter? (Here, incidentally, is a testable prediction of subjectivism). Consider, as just one example, the following fairly typical episode from the history of science.  In 1818, the French Academy of Sciences conducted an annual competition, which was won by Augustin Fresnel.  Some members of the panel--accomplished scientists in their own right--"were skeptical if not downright hostile" to Fresnel's submission.   One member, the famous mathematician Poisson, "observed that Fresnel's integrals entrained a paradox . . . an obstacle and a hole would thus be indistinguishable by diffracted light. They challenged Fresnel to test this alarmingly implausible consequence. He performed the experiment--with perfect success." There are many such occurrences in the history of science, and they do capture some of its uniqueness.   Has anything like that episode ever occurred in religion, in art, in psychology even?  Sixth, and most important, subjectivists seem to forget the critical differences between science and its creators; the scientific method and its practitioners; process and product. Let me illustrate this last point with an analogy.

If the stepwise construction of the Tower of Babel were allowed to continue, the builders might have never achieved their goal of reaching the heavens. Being human, they might have been cantankerous and not wholly rational. Some of them might have been certifiably mad. But all this would have detracted little from the reality and rationality of their collective undertaking. If we wished to judge their undertaking, we would need to focus our attention on the unfinished tower itself. We would also need to ask whether, despite the many false starts, the quarrelsome creators, and their critics, the construction program is rational and popular enough to make successive additions to the tower's height possible.

We ought to apply the same criteria to the scientific tower our species has been creating for the past 2,500 years or so. This analogy reminds us that, although human beings cannot unfailingly adhere to a rational course of action, a community of individuals dedicated to an ideal of rational conduct might manage, over the long term, to resolve most problems in conformity with this ideal. Similarly, although human beings cannot create perfect structures, they might be able to create reasonably good ones. The claim that, despite their humble origins and countless imperfections, science and the scientific method must be counted among our species' most rational creations, cannot therefore be summarily dismissed. Nor can we dismiss the notion that many of the political problems we face are traceable in part to the double standard we apply to science: we are happy enough to use the powerful tools it creates but we shy away from the rational approach that made their creation possible.

Source:  Moti Nissani.  Permission for the free use and publication of this material is hereby granted.

Back to Atoms & Stars, Main Page