THE CONFUSION BETWEEN SCIENCE AND TECHNOLOGY IN THE STANDARD PHILOSOPHIES OF SCIENCE

The distinction between pure and applied science seems too trivial to draw, since applied science, as the name implies, aims at practical ends, whereas pure science does not. There is an overlap, to be sure, which is known as fundamental research and which is pure science in the short run but applied in the long run; that is to say, fundamental research is the search for certain laws of nature with an eye to using these laws.

My own concern with all this comes from my studies in the theory of confirmation. Contrary to most, if not all, writers in the field, I hold that confirmation plays no significant role in science, pure or applied.¹ The contrary impression seems to me to stem from the fact that both in­vention and the implementation of novelties - from applied science or from invention - require confirmation. The standards of confirmation are legal, and they are set by patent offices in the case of invention and by institutions in charge of public safety and of commercial practices in the case of implementation.

I shall return to the distinction between applied science and invention later on and also touch on the other points of technology. Here let me say that the literature in the philosophy of science is oblivious of all this, usually confusing pure science with applied science and both with tech­nology. The inductivist philosophers of science insist that pure science concerns probable beliefs and that technology concerns decisions, both relating to the same probable hypotheses. They confuse belief and decision. Instrumentalist philosophers do not believe in the probability of pure science but in its usefulness: not only do they equate pure with applied science and applied science with technology, but they even consider all three as identical. Both leading groups of philosophers of science appeal to the fact that science and technology are so very enormously successful. To go to the root of their confusion² let us examine views on success. We shall later be able to show how different scientific success is from practical success.

i

Living in a robust progressive society, we are surrounded by the tendency to evaluate people by the measure of their success, more specifically by the measure of their success in attaining high economic and social stand­ing, which is usually known - rather inadequately, but let this ride - as material success. Obviously, success philosophy is rather vulgar, and material success philosophy particularly vulgar and even stupid. I have no intention of speaking against success, for without some success there is no action; nor do I wish to speak against material success.

But to value success is not to say that success is all that matters, or that material success is all that we value, with the exception of saints, perhaps. Many works of fiction comprise plausible counter-examples to the identification of success with material success, usually by describing convincingly heroes who are wretched though materially successful (e.g., Bergman’s Wild Strawberries).

He even brought upon himself material ruin so as to put pressure upon himself to wrok hard at his inventions in order to make money which he then quickly disposed of again, and so on, until he was too old to con­tinue.

This example shows how complex and unconvincing may be the case of any person’s personal success. And so let us ignore personal success from now on, material or otherwise. When we speak of success imper­sonally, we speak of the successful execution of a task regardless of its personal significance to the one who carries it out. Can we evaluate a person by the measure of the success of his performance? Can we measure the man by the measure of his achievement? Obviously, this, too, is rather vulgar. The latest biography of Edison, for instance, by Matthew Joseph­son, has been written partly, if not mainly, in order to break away from the vulgar success philosophy which permeated previous biographies, thus making them unpalatable to the more sophisticated.³ At the very least, we all must agree, if we do measure people by their success, then we should take into account the fact that success or its absence depends part­ly on luck, good or bad; we must adjust the evaluation of the person, then, by considering the factors which were beyond his control yet affected the outcome of his activities. Judging people by their luck is poor judg­ment. How they can stand up against bad luck and how they can make use of their good luck contribute to their personal makeup much more than how lucky they were.

ii

My colleagues in the field of the philosophy of science love to speak of the greatness of science as being identical with the success, the great achievements, of science. Consequently, when we speak of the adventure of science, we are hypocritical: we all admire Captain Scott, since though he failed he was very brave; but Joseph Priestley was seldom accorded similar consideration. Not Priestley, but his opponent, Lavoisier, is ad­mired - and not for his sense of adventure either, but for his alleged success.⁴ When you stop to think about it, you may find that success is rather bewildering and calling for much explanation and re-examina­tion.

Somehow the assuredness and faith some of us display in the success of science seem rather smug. This impression is usually dismissed by the claim that the faith in science is amply justified. But when the faith in science is justified, science may become much less of an adventure and, hence, not much of an achievement. The justification of science makes it stupid rather than prudent not to apply it. Moreover, the justification is a principle of induction, and the principle of induction is usually justified by our successful reliance on science in practical affairs. This is circular, but never mind; the principle of induction may indeed be a method for success; it may be the golden goose which lays golden eggs regularly, or a computer which makes its followers rich. If so, then we may compliment the scientists and their followers no more than we may compliment the owner of a golden goose or of a computer. If there exists an inductive algorism which can be fed into a computer and which is most successful, then science must be least exciting and devoid of all adventure. This ideal of science, especially of applied science, as based on a precise algorism, can be found already in Laplace’s Philosophical Essay on Probabilities, where he expresses his hope to see even judges on the bench replaced by computers.⁵ Perhaps the judges of his day were so inhuman that replacing them by machines would have been progress from the negative to the zero degree of humanity; still, the very thought is chilling.

Those, however, who think that viewing science as based on an alogrism debases science usually stress that in science we need both luck and intui­tion. Both words, ‘luck’ and ‘intuition,’ are indicative of our ignorance, perhaps of our essential ignorance. For, obviously, if this ignorance were temporary, science based on algorism would be attainable. Though they are both mysterious, luck and intuition seem to be opposite poles: whereas we admire those who contribute to scientific progress when we view them as inspired, we somewhat deprecate them when we say they were lucky. Understandably, Pasteur and Edison alike felt deprecated when their success was viewed as due mainly to luck. They both repeated Lagrange’s mystical formula about luck coming only to the prepared mind. One might push the argument and say that even having a prepared mind depends on luck. Indeed, both Oersted and Einstein humbly viewed their own inspiration and talent as mere luck. And yet pushing the argu­ment so far is scholasticism: if a typical scientist were lucky enough to have one of two wishes granted, these being for talent and for material success, he would doubtlessly grasp his luck and wish for talent, as King Solomon is alleged to have done, hoping to have the chance to use it.

Hence, it is neither talent nor luck but, rather, adventure, namely, the bold use of talent and the grasping of luck when it comes one’s way, that we wish. It is all too easy not to use either talent or luck by the mere lack of a sufficient degree of courage and perseverance. As Edison said so aptly, “It has been so in all my inventions. The first step is an intuition - and it comes with a burst; then, difficulties arise.... I have the right principle, and am on the right track, but time, hard work, and some good luck are necessary too.”⁶ And again, “The trouble with other inventors is that they try a few things and quit. I never quit until I get what I want.”⁷ Of course, Edison was very lucky to be able to work so hard, so boldly, so cleverly: most people can do nothing to alter their predicament.

in

Query: can we admire science as the product of the bold use of the in­vestigator’s imagination and of his good fortunes, and yet justify our trust in science? If we justify our trust by a criterion or a formula, the criterion or formula may turn into a Laplacian algorism which renders science mechanical, but if we have no criterion or formula, we may have no justification. Thus, in his Logical Foundations of Probability, Carnap tries to differ from Laplace but fails. He indorses the view - which he attributes to Einstein and Popper, but which might be more justly attri­buted to Brewster and Whewell, if not to Galileo and Kant - of the imagination as an essential ingredient in science.⁸ And yet he provides the Laplacian formula for deciding which hypothesis to believe in and to act upon, and his formula provides the Laplacian algorism of projecting the past into the immediate future and thus generating mechanically the best possible hypothesis.⁹

One cannot argue cogently against the creation of any working al­gorism, that is to say, against any systematization. Descartes system­atized geometry and enabled people to prove theorems mechanically rather than cleverly; solid-state physics is now systematizing one of the most known fields of clever invention, namely, metallurgy; and yet no one is any the worse for these systematizations. When Abraham Wald systematized some decision-procedures, his ideas were so powerful that they were kept for a time as war secrets.¹⁰ Any partial systematization, such as Descartes’s or Wald’s, merely covers some ground and sends the adventurer further afield in search of new frontiers. But total systematiza­tion excludes all adventure. In any field, any alogrism is welcome; yet, were an algorism universal in that field, creative thinking in that field would be redundant altogether, once and for all. If the field in question is the generation of ideas in general, algorism in it is the end of creative thinking in all science and technology.

But it is possible to invent a criterion or formula for justifying our faith in and reliance on a scientific theory without taking away the spirit of adventure from science. A formula for trusting science which is not a means of generating mechanically the theory to be trusted is possible, though Carnap, at least, has tried to produce it and failed. Such a formula was invented by William Whewell¹¹ and reinvented with improvement and modification by Karl Popper;¹² my dissent from his philosophy is rooted in my rejection of the formula, but I must acknowledge its supe­riority. Let me present it first and discuss the question whether it is a mea­sure of success later.

It runs as follows. A belief in a theory is justified to the degree to which it was corroborated by experience; that degree depends on the explanatory power of the theory in question, of the degree of testability of that theory, and of the degree of severity of the test which it has thus far passed success­fully.

Let us not examine this formula in detail, so as to avoid a rather aca­demic exercise. Let us first try to get the spirit, the general feel and ap­proach behind it, and see how acceptable that is. Up to a point, I think, it is, but only up to a point. To begin with, let us consider the positive general characteristics of this formula. The formula takes full account of the challenge of research. Thus, it is in full accord with the above quota­tions from Edison, as well as with the following one: “I would construct a theory and work on its lines until I found it untenable, then it would be discarded and another theory evolved. This was the only possible way for me to work out the problem.”¹⁸

Both Whewell and Popper stress the need for problems, for inspired solutions, for the usefulness of criticism in the development of new solu­tions, and for the success of theories which stood up to criticism and thus proved their mettle. The chief difference between Whewell and Popper, it is well known, is that Whewell believed in the finality of such success whereas Popper believes in its tentativity. Also, one need hardly say, in this division no one sides with Whewell today, even though most philos­ophers of science accept neither view. The majority indorse a view which is nearer to Whewell’s than to Popper’s in that they replace Whewell’s finality not with Popper’s tentativity but with probability. This notion is either meaningless or else it is a theory of probability in the classical sense, and thus it yields a Laplacian algorism.¹⁴

iv

So now we have a formula - the Whewell-Popper formula - which perhaps justifies our faith in science, and if so, without doing injustice to intuition and luck. Let me now show you that to judge the greatness of science by its success, when success is measured by the Whewell-Popper formula, does not accord closely with the admiration of science as an active adventure, which I am advocating.

Imagine an Einstein, racking his brain, developing his most inspired ideas, sifting them, elaborating them, deducing from them both explana­tions and tests; he then retires to his den, full of nervous anticipation. The world notices his work, considers it, concedes that it is possibly admirable, and sits patiently and waits. Then comes an Eddington, translates the test into action by designing all sorts of instruments, by mobilizing funds, by organizing workshops to prepare the instruments, by organizing an expedition, by supervising and participating in the experiment, by calculat­ing the experimental results and comparing them to Einstein’s results as calculated from Einstein’s theory. All the while Einstein is supposed to be sitting back and waiting, and all the time we are supposed neither to ad­mire nor to dismiss him; and then Eddington may give the green light, and we all burst in admiration for Einstein; or he does not, and we do not.

In such a case we want the scientist to intuit not only a brilliant idea but also certain truths. In such a case the scientist is more of a fortune­teller than I would like to be the case. It was said that Faraday had a nose for the truth; Edison was called the “wizard of Menlo Park”; maybe all this is true. If so, the truth is not very palatable, at least for those who find it less enjoyable to admire Faraday or Edison because he is a wizard than to admire them as intellectual adventurers.

Have we not gotten into an impasse? That we want success is self- evident: to say that we want success is to say no more than to say that we want to achieve something. If we did not want anything or if we had no expectation of achieving it, we would not act at all. And if we do not want the success of applied science to be a matter of wizarding or uncanny insight, we must have an algorism. So, it seems, if we do not want an algorism, we do want some wizarding! Considering this impasse, we may be more patient with my colleagues who view science as a success-algorism. Michael Polanyi¹⁵ is surprised that most philosophers of science deny that applied science or technology contains an intuitive element when this claim is a standard criterion of all patent offices: a machine than can be created by anybody according to a publicly known alogrism will not be granted a patent; the claim for a patent must involve (by legal standards!) a claim for originality. Polanyi exaggerates the wissdom of the legal patent­ing apparatus (consider the Selden case, and see Edison’s ‘My 40 Years of Litigations’), yet by and large he is right.¹⁶ Still, we may well understand the philosophers whom he criticizes: reluctant to acknowledge the wizar­dry of the inventor, they decide that invention has an algorism to it, and thus they imply - perhaps unwittingly - that inventors need not be original. They equally imply that scientists need not be original, even though we are not supposed to grant Ph.D.’s for dissertations which we do not judge original - this also by some legal standards. This does not disturb the in­ductive philosophers who insist on the idea of science as unoriginal.

The idea that science is no magic and hence it is based on some al­gorism is by no means new, by no means peripheral to the traditions of science, and by no means a merely unintended aspect of inductive phi­losophy. Its chief corollary from our viewpoint is no news either; it has been stressed again and again that the measure of a scientist’s success is, provided he is a proper scientist, the measure of work he has invested in science and no more. This calls for equal admiration for every hard-work­ing contributor - all contributions will be thankfully received - as was stressed by Bacon, by Boyle, and, more recently, by Malinowski as well as A. P. Usher, leading historian of technology.¹⁷ Bacon also said that inventors should be honored by statues made of different material from gold to wood and in different sizes, depending on the greatness of their in­ventions. Similarly, the fact that we do not admire in equal measure every hard-working inventor and every anthropologist who has conducted field­work is well known. Usher dismisses this inequality as cults of personality and historical fiction; young anthropologists are told that every fieldwork is of equal importance, and young chemists and engineers are told similar stories. Yet most young engineers and anthropologists think that all this is a mere scientific myth. Professors reiterate the myth of equality of all contributions, and students reiterate the myth of great adventures. Who is right? What is the real criterion of success?

The instrumentalist philosophers who identity - with Duhem - pure science and applied science have the idea of simplicity as the criterion of success.¹⁸ Mach viewed simplicity as success of mental economy, which is very different from material success.¹⁹ Material success, no doubt, is often awarded to very complicated theories,²⁰ and to be honest we need economists to assess such material success. And, of course, we should choose materially successful economists to do this job. Moreover, all economists will agree that the archeology of extinct cultures, for instance, is of hardly any practical value, except as a highbrow means of entertain­ment, perhaps. Most philosophers of science, however, do not indorse this instrumentalist philosophy, preferring to it some inductive philosophy which justifies their claim that science is useful by extrapolating from past to future material success. They say, “when we make a decision, we base it on induction.”

These people just do not understand the essentially adventurous nature of decisions. Neurotic people, it is well known, who do not possess the ability to decide, often say, “I cannot decide because I do not know enough.”²¹ They obviously deceive themselves: when we know enough we have no problem at all, and so we need not decide at all. When we say we must decide which route to take, we imply that we have no knowledge as to which route is the shorter, or the safer, or the better one by any accepted criteria; alternatively, if we have not accepted a criterion, we may be saying that we do not know which criterion is the correct one. So, one thing is quite clear: we just do not decide by induction.

Let us then examine the sort of decisions which pure scientists, applied scientists, and technologists face; the sort of questions they have to try to answer. This will give us, I hope, some insight to the sort of adventure they may face.

V

One reason why I admire Popper’s philosophy of science is just its playing down of success. Roughly, but very seriously, let me put it thus. Every­body admires science for its ability to predict. Meaning, of course, its ability to predict correctly or, more precisely, sufficiently correctly to pass as the ability to predict successfully. (The criteria of success are vague and invite us to confuse science with technology.) Only Popper, to my knowl­edge, has stressed that the ability to predict with some degree of precision is what characterizes a scientific theory - regardless of the correctness of the prediction. Popper had predecessors, to be sure; Galileo, Boyle, Faraday, Whewell, Peirce, and Edison spoke favorably of mistaken pre­dictions in science, and doubtless others did too. But they all insisted that finally it was the successful prediction that counted, that the errors were important only as stepping-stones to imminent success which really characterizes science. It was Popper’s province to discover that this is not necessarily so.

According to Popper’s theory, a scientist needs a good hypothesis. He may work for many years with no success, or with almost no success. No­thing in Planck’s researches after 1900 equals his success in 1900, for in­stance. And so, according to Popper, the adventure of science may lead to no result; there is, in science, such a thing as utter failure. But, accord­ing to Popper, once success was achieved in that a good hypothesis has been found, and once the work has been invested to show the goodness of the hypothesis, then even if the hypothesis leads to predictions which are refuted we may value that hypothesis. The goodness of the hypothesis, according to Popper, may be measured by its explanatory power, as well as by its testability, which is the ability to deduce from it predictions which may be checked by experiment and observation, quite independent­ly of whether the predictions are later found to be true or false. Hence, our appreciation of science need not follow the Whewell-Popper formula, since that formula speaks of theories which have yielded only correct predictions thus far.

To my regret, Popper has expressed recently a view quite different from the one I have just ascribed to him. Referring to discussions with myself, he has said a good hypothesis should also be corroborated before it be refuted.²² He uses the Whewell-Popper formula as a measure of the good­ness of a hypothesis or of our appreciation of it, thus demanding from scientists some measure of wizardry, though he does not go as far in ap­plying the formula as others would: he does suggest that once a theory is refuted we then withdraw our belief in it, but he does not suggest that in such a case we also withdraw our appreciation of it. Nevertheless, I wish to ignore this aspect of Popper’s writings, and if you do not like it, you may view the ideas I attribute to Popper as his ideas in my modification or in my distortion, as you wish.

My best reason for rejecting the use of the Whewell-Popper formual as a measure of success is that I follow Popper’s theory of the process of trial and error in science as something essentially different from trial and error in many other situations, such as more everyday ones. Trial and error is progress by elimination. In small sets of alternatives, elimination by trial and error may insure success. Most everyday situations are like that, and indeed, no one in his senses would view researches in such situa­tions as adventurous. When our television set breaks down, we take it for granted that there is a finite set of possible faults it may have. Usually we test them one by one and finally eliminate the fault; there is no adven­ture in the search. The classical argument against the theory of induction, as you well know, is that there exist infinitely many possible explanations of known facts. Lord Keynes has acknowledged this criticism and has postulated the principle of limited variety which he has found implicit in earlier inductivist writers.²³ Assuming, quite ad hoc, by the way, that only a finite number of hypotheses may explain known facts, he advocates the most unadventurous theory of science.

Consider now the adventurous method of trial and error in everyday life. This is known as looking for a needle in a haystack. The haystack is supposed to be infinite for all practical purposes. It is quite conceivable that one goes on eliminating more and more possibilities about where the needle may be and yet not progress to any appreciable extent toward the correct position of the needle. Looking for a needle in a haystack is ad­venturous; it demands intuition and luck.

According to Popper, the search for a good hypothesis is looking for a needle in a haystack. But, according to his philosophy, once a hypothesis is found which solves the problem at hand, that is, which explains the facts which puzzle us, and once it is found testable, then progress has been achieved. (He even claims ability to prove this,²⁴ but let us not press him on this point.) Even if a good hypothesis is refuted, things will never be the same again. The refutation of a previously unrefuted hypothesis is surely a new discovery we partly owe to its inventor. The task after the refutation of the new hypothesis is not the same as before: we now wish to explain, not the same set of facts which the refuted hypothesis has ex­plained, nor the same set of facts plus the refuting facts, but the refuted hypothesis, as a special case and as a first approximation, plus the new fact.²⁵ What Einstein explained is not all the known astronomical facts but Newtonian astronomy plus the facts it failed to explain, just as New­ton explained Kepler’s theory and the deviations from it. This is a funda­mental methodological point. The only way to evade it is to claim that pure empirical data exist; as you may know, such claims are becoming increasingly difficult to maintain.²⁶

The metaphysical theory corresponding to this fundamental method­ological point, of explaining not bare facts but previous theories and their refutations, is the theory of the approximation to the truth by levels of explanations. This theory may be false; it does, however, justify Popper’s rejection of the claim that scientists must intuit the truth, and his claim that the intuition of a possibly true explanation is enough of a step for­ward for the time being. If so, then though finding a good solution is like finding a needle in a haystack, the eliminating of that solution as false is not like finding that the needle is not in a small portion of a haystack. Pictorially, scientific theories are not explorations of small portions of a given haystack; they are sieves through which the whole haystack is passed and whose meshes are used to build sieves of ever finer mesh. In this picture of science as progress through conjecture and refutations, there is no room for confirmation or corroboration.

VI

So much for pure science. What I wish to show next is that all this does not apply to applied science or to technological invention. In applied science there are two standard kinds of problems, deducibility and ap­plicability. Given a theory and given a problem, the applied scientist asks himself, can I solve the problem while using that theory, and is my solu­tion true? The first kind of question is essentially mathematical, and this explains why the terms ‘applied science’ and ‘applied mathematics’ are so often used interchangeably by people who will never use the words ‘science’ and ‘mathematics’ interchangeably. Mathematics, pure or ap­plied, has its own kind of adventure of which I shall not speak beyond saying that it is the only kind of adventure an applied mathematician or an applied scientist may encounter. One may be even more specific here and add that in applied science, unlike pure science, the problem of de­ducibility is to find initial conditions which, together with given theories, yield conditions specified by practical considerations. This is, indeed, how Popper characterizes technology at large,²⁷ which is a rather narrow view of technology.

The second question, namely, is the solution to a given problem which he has deduced true, hardly allows itself of intellectual adventure, though designing the test may be, and executing it may also involve physical ad­venture. That the testing of results of applied science does not belong to applied science is rather obvious: even the applied scientists in industry, not to say in the universities, wear white collars, not white lab coats or blue overalls. They cannot and need not perform tests. That they cannot is a point of great significance, which I wish to label as ‘Hatfield’s law.’ It is this: there is always a gap between applied science and the implementa­tion of its conclusions, to be filled by invention.²⁸ The law is trivial in the sense that applied science does not issue programs for computers which implement its results; it is less trivial if we understand that applied science does not issue programs even for skilled workers without gaps to be filled by clever inventors.

To take a very simple instance²⁹ of the difference between invention and applied science, let us take a case involving both. The inventor Edi­son wished to replace gas street lighting with electrical lighting. The crux of the difficulty which disappointed other inventors was practical: the amount of copper needed to conduct electricity along the streets of a city would be too large to warrant investment. It then occurred to Edison that this obstacle could be overcome if high tension and high resistance were employed. This was a technological idea, which could be correct or incorrect: Edison needed an applied physicist to tell him that, and his applied physicist, Upton, applied Ohm’s law to the problem and con­cluded that by using a tension of 100 volts the quantity of copper needed could be cut into one one-hundredth of the originally calculated quantity required. The conclusion surprised even Edison, though it followed logically from a law he knew very well. Upton’s job was finished on paper. Applied science is, thus, providing answers to given questions when these are implicit in a given theory. Posing the question and seeing the possible technological significance of the answer to it was one of Edison’s inven­tions on the road to electric lighting.

Edison’s opponents did not think the obstacle - the need for immense quantities of copper - could be overcome except by connecting the lights in series. Applying Ohm’s law to this solution, they found it impracticable and showed that the light given by the many electric lamps needed to light a city would be too small to be of any use. Doubtlessly they were quite right in their applied science but rather wanting in imaginative in­vention. This shows how different are applied science and invention. That the two can overlap was shown by my example of the collaboration of Edison with Upton; that they are distinct is symbolized by the distinctness of these two individuals, their training, and dispositions.

One need not blame Edison’s opponents, or any other people, for lack of imagination; for by doing so one may underrate the power of the imaginative inventor. Not only was Edison’s idea so very new, it also required many other innovations to adjust to his new scheme. For in­stance, whereas dynamos in his days gave out constant currents, his idea required the invention of a usable dynamo with constant tension. It is conceivable that his idea of high-tension currents would not work if he could not invent such a dynamo. How much more understandable, then, it is that his opponents were thinking with existing dynamos in mind. And this is not all. When all means for high-current lighting were found, there was no filament which could stand the conditions imposed by Edison well enough to be practicable. For all he knew, there may not have been such a filament. He himself conceded that much at the stage when he worked on platinum filaments which could not be produced commercially prior to the commercial development of machines to trans­mute base metals into platinum. It is incredible how lucky Edison was and how much he dared gamble on his good fortune.

Invention depends on finding facts for which we have no clues, at least a sufficiently wide absence of clues to make the haystacks in which they lie practically infinite. Otherwise, a team of applied scientists would find it, to be sure. And so the failure of a technical inventor is as final as the fail­ure of a failed wildcat prospector.

Not seeing the wildcat character of invention, not seeing it as the miracle of finding a needle in an infinite haystack, often leads to misun­derstanding and to underestimation of inventors and their sense of ad­venture. The most definitive and thorough history of photography is probably that of the Gernsheims.³⁰ Yet their view of Daguerre’s inven­tion is incredibly naive; they claim³¹ that Herschel was much greater than Daguerre, since after merely having heard about Daguerre’s success Herschel repeated and improved in a few days the results that initially took years to develop. The truth is that there was no reason at all for anyone to think of the possibility of latent images to be developed. The problem of photography for decades had been how to fix an image which is visible at once (perhaps while turning it from negative to positive). No one thought of latent images to be developed, and everyone knew of visible images which soon disappeared. Thomas Young applied this knowledge of temporary images to photographing infrared interference - the first piece of photography of any value, and a typical case of applied science which is in no way an invention, as was Daguerre’s thought of the theory of latent images to be first developed and then fixed. Daguerre had the thought under most peculiar circumstances, and the thought appeared surrealistic to him at the time and years after.³² Once his idea proved successful, hosts of improvements were possible within a relatively short period.

It is because invention is a theoretical rather than a practical activity - though to a practical end, of course - that inventors may illustrate their inventions while using most crude machines; but it is because their claims for inventions must be corroborated that they must use some machines, unlike theoretical scientists and even applied scientists who may never bother about practice.

We see, then, that though corroboration is needed in technological in­vention, it is of no use in science, pure or applied. The question is whether corroboration has anything to do with belief, and I think I have shown that in technological invention corroboration has something to do with the nature of the problems at hand, which is always, how one can do successfully something specified; whereas in science, pure or applied, belief is not a matter of scientific method either. As Popper has stressed, in pure science we may try to refute the most corroborated view, and in applied science truth is of little importance. In applied science, that is, the question of the truth of the theories to be applied hardly even enters, though the question of the applicability of results from it its crucial and is answered by simple tests. Thus, we do not believe, but we still apply, Newtonian mechanics, though not to systems involving high velocities.

VII

Technology involves various factors, and I have thus far concentrated on only two of them, namely, applied science and invention. The im­plementation of the results both of an invention and its maintenance are important fields which I cannot discuss here beyond saying that in the field of implementation corroboration is most important, and the degree of corroboration required prior to implementation is socially and legally determined. It is doubtless that here the requirement for corroboration is a matter of public safety and that though no amount of corroboration insures success, the legal requirement for corroboration does eliminate some dangerous technical innovations prior to their implementation in the market. But if we demand much corroboration, there will be no im­plementation.

To show how complex the problem of implementation is, and how difficult, I should mention the classic trial of General Billy Mitchell, who thought he had corroborated his theory of the importance of air force bombers and the consequent diminishing importance of naval warfare. His sincerity is beyond question and is the main reason for the fame of his court-martial. Another reason for his fame is rather vulgar, namely, the truth of his prediction of an unannounced Japanese air attack. But, of course, he was mistaken in most of his views except about the power of air raids, and implementing all his proposals might have been a disaster.³³ Success philosophy is what makes the various books (as well as the movie) about him so very unsatisfactory. Without success philosophy the story of Billy Mitchell becomes much more interesting and enlightening, es­pecially in relation to the requirement of the complacent - such as his superiors - for more corroboration before implementing any novelty. But I shall not go further into it. I wish only to say that it exemplifies one point of significance from our present viewpoint.

Those who believe that empirical evidence renders a theory credible view the implementation of that theory not as an adventure but, on the contrary, as the reduction of risk and thus success. In truth, the imple­mentation of a novelty is risky even when no risks seem possible. Testing is performed so as to eliminate some risks - some ‘bugs,’ to use Edison’s jargon - and some risky innovations. But sometimes the prevention of one risky implementation entails the taking of other risks, as Mitchell’s story shows. This argument has been used in a recent U.S. Senate subcom­mittee investigation into the disaster of the atomic submarine ‘Thresher,’ which sailed in spite of known and eliminable risks. Whether the argu­ment that too much testing and insisting on safety may lead to other risks was used properly or in order to shield comrades in arms I do not know, but such questions can be discussed rationally, and will be, perhaps, by future historians of technology.

As I have said before, success is not something to be proud of, but a puzzle to be explained. The discoveries of modern biology and modern physics about the hostile world in which we live should make our very survival a great puzzle. But, on the contrary, philosophers of science re­fuse even to see the obvious fact that luck and intuition are essential to success. Yet I must say this in their favor: When any phenomenon takes place regularly, we wish to explain it and thus systematize it. Discovery and invention look regular enough nowadays, and so they seem to require a systematization. That there is a flaw here I have tried to argue, but I cannot put my finger on it. The dichotomy between algorisms versus lucky intuition seems a fundamental fault in our way of thinking. Obviously, mixing algorism with intuition and luck will not destroy the dichotomy. One way out is to view systematization dynamically: some past successes which looked lucky can be explained by later systematization. This eases the pressure of the difficulty, but no more. Take any unsystematic but regular process: to use Chomsky’s work,³⁴ take a child’s intuitive learn­ing to speak. Not only do we not know how to explain it, it seems as fun­damentally mystifying as Faraday’s or Edison’s uncanny ability to dis­cover and invent to order.

To conclude, the confusions in the field are rooted in a difficulty which is shared by all of us. Our dichotomy between algorisms and lucky in­spirations is very unsatisfactory. Philosophy managed to progress in various directions in spite of this fundamental obstacle. But this obstacle may be a most serious impediment for the progress of the philosophy of technology. So, this field may be at a dead end for the time being, or, if it will get going, may lead to spectacular results which are bound to revolu­tionize all philosophy and much of our way of thinking in many depart­ments.