HEAD

Chapter 4
Demarcation and the philosophy of science

I remarked earlier that there has been an outbreak of new `sciences' in the past century, many of which share few characteristics with the old natural sciences, and that this is in part attributable to scientism. We now need to consider in more depth the extent to which it is justifiable to criticize this development. We need to address whether there is a clear enough definition or understanding of what natural science is to justify distinguishing it from non-science.

4.1 The History and Philosophy of Science discipline

With roots extending back to the early nineteenth century, but blossoming into an identifiable independent discipline in the mid twentieth, a whole field of academic study has grown up, known as the History and Philosophy of Science . Its practitioners study the particular questions surrounding how science works, what scientists have done and said, and what the influence of science is upon the wider culture. The field encompasses historians who specialize in scientific topics or personalities, philosophers who try to understand and explain the logic, methods and epistemology of science, and sociologists who study the social dynamics of science and scientists. To give an impression of the relative size of the field, the largest academic Department of History and Philosophy of Science (HPS) in the United Kingdom, at Cambridge University , has 5 professors, while the Department of Physics (the Cavendish) at Cambridge⁵⁰ has 37. So the professional discipline of HPS, which at Cambridge is taught as part of Natural Sciences, is rather modest in size. Many universities don't have departments but have HPS scholars whose professional homes are in other more traditional departments. Actually, a lot of the published analysis of science, particularly what is written for the general public, is authored not by HPS specialists but by working scientists themselves, and by journalists. This is a situation not unlike, for example, political history, where the books and articles are written by three main groups: historians, the politicians themselves, and journalists. These roles are not mutually exclusive, of course. Politicians can sometimes be historians and journalists, and vice versa; but its is worth being aware of the distinctive emphases of the various perspectives. I write obviously as a working scientist, but in the present context with the hope that what I write is accessible and interesting to a broad audience. I also intend to account fairly for what the academic HPS specialists have discovered, although I don't pretend to their professional focus. I have had the unalloyed pleasure of personal acquaintance with some distinguished historians and sociologists of science, and the more ambivalent experience of being the object of study of others. My first move is now to turn the tables on these folks for a moment and draw some conclusions from simple observations of what they do.

Recall that I have insisted on using the word science not as it was classically derived, referring to any systematic knowledge, but as referring to the study of nature : natural science, what was commonly called natural philosophy prior to the nineteenth century. You might have lingering doubts about the appropriateness of this insistence. And there might be experts in what I have called non-scientific fields who are still smarting at the designation, who feel in it an implied slight. If so, then I want to draw authority from the History and Philosophy of science for my position. Not so much in what HPS scholars say is science or non-science, but what they actually demonstrate by their actions is the primary subject matter of their field. If one takes any sample of the writings of a spectrum of HPS academics or of their journals, it becomes immediately obvious that what they mean by science is in fact natural science. Here and there one finds a few studies of borderline fields like psychology, for example, for which considerable hesitation is justified in calling it a natural science; there is also a certain amount of reflexive study of HPS concepts and development itself; but by far the overwhelming bulk of what HPS specialists study is what the common person would call science: physics, chemistry, biology, geology, astronomy, physiology, and so on. I notice that from time to time philosophy of science scholars refer, often with some detectable awkwardness, to `social science' . They might, apparently somewhat reticently, grant to social science the title. But even if that lip service is rendered, it is almost vanishingly rare that they would actually study the field of sociology as scientific knowledge, or instance sociological discoveries as examples of science at work. So this is my first conclusion from observation of the HPS field: it supports by its actions the predominant usage of `science' to mean natural science, and not just knowledge of any systematic type.

4.2 Scientism, sociology and socialism

This state of affairs would undoubtedly surprise the figures of the early nineteenth century who offered analyses of scientific methods that they sought to apply not only to nature but to society. In doing so they addressed the philosophy of science, but also helped to found the discipline of sociology , and through their more indirect influence were instrumental in the birth of various political movements and theories, such as Marxism. Of these figures, Auguste Comte (1798-1857), the founder of what became known as Positivism , is undoubtedly the most notable.

In a study of some of these sociological figures by Nobel-prize-winning economist F. A. Hayek, The Counter-Revolution of Science (1952), the English word scientism acquired (from French antecedents) the derogatory implication that is widespread today. At the beginning of the first part, entitled "Scientism and the study of society", Hayek delineates scientism thus

During the first half of the nineteenth century a new attitude made its appearance. The term science came more and more to be confined to the physical and biological disciplines which at the same time began to claim for themselves a special rigorousness and certainty which distinguished them from all others. Their success was such that they soon came to exercise an extraordinary fascination on those working in other fields, who rapidly began to imitate their teaching and vocabulary. Thus the tyranny commenced which the methods and technique of the Sciences in the narrow sense of the term have ever since exercised over the other subjects. These became increasingly concerned to vindicate their equal status by showing that their methods were the same as those of their brilliantly successful sisters rather than by adapting their methods more and more to their own particular problems. And, although in the hundred and twenty years or so, during which this ambition to imitate Science in its methods rather than its spirit has now dominated social studies, it has contributed scarcely anything to our understanding of social phenomena, not only does it continue to confuse and discredit the work of the social disciplines, but demands for further attempts in this direction are still presented to us as the latest revolutionary innovations which, if adopted, will secure rapid undreamed of progress.⁵¹

I repeat. This is a definition and assessment of scientism in sociology not by me or by a scientist but by an economist, a world leading practitioner of one of the most distinguished social disciplines. Hayek's work provides a valuable overview and insight into the scientisms of the early nineteenth century from which, however, we shall have space to mine only a few nuggets.⁵²

Auguste Comte's apprenticeship, his early thinking and writing, was under the tutelage of Henri Saint-Simon . Saint-Simon, an adventurer and failed entrepreneur who at the age of 38 turned to intellectual pursuits, was himself the architect and advocate of a scientistic approach to the analysis of society that was to regard "our social relations as physiological phenomena". His intellectual ambitions were adorned, from the beginning, with bizarre religious overtones. In the "Cult of Newton" he proposed, apparently in all seriousness, that a council consisting of three mathematicians, three physicists, three chemists, three physiologists, three authors, three painters and three musicians, elected by the whole of mankind, should become the representatives of God on earth.⁵³ Saint-Simon's ability to attract to his coterie dynamic young thinkers, particularly from the École Polytechnique, opposite which he lived, provided him with an intellectual and literary depth that he lacked in himself. That symbiosis was handicapped by many acolytes' eventual estrangement because of Saint-Simon's unsupportive stance toward liberty. Nevertheless, after Comte, a nineteen-year-old, expelled polytechnicien, became his secretary in 1817, much of subsequent Saint-Simonian doctrine bears the strong stamp of Comte; and Comte's most influential System of Positive Policy (1824) first appeared as part of Saint-Simon's Catéchism des Industriels.

Comte's positive system, in this first incarnation, looks forward to the disappearance of the liberty of conscience from society. Just as such liberty has no place in astronomy, physics, chemistry, etc., it ought not to do so in society. Its disappearance will happen when politics has been elevated to the rank of a natural science: the new science of Social Physics . This expectation is in accordance with Comte's most famous dictum, the Law of Three Stages. "Each branch of knowledge is necessarily obliged to pass through three different theoretical states: the Theological or fictitious state; the Metaphysical or abstract state; lastly the Scientific or positive state", which of course is what constitutes true knowledge.

Comte's classic work began in 1826 as a series of lectures and extended to six volumes appearing between 1830 and 1842 as the Cours de philosophie positive . His task is defined by the Law of Three Stages. Physics, chemistry and biology have passed through the preliminary stages and thus become positive knowledge. Comte is the one who is going to conduct this passage for the knowledge of society. Without delving in detail into what this transition was supposed to require, let the following characterization and critique from Hayek's analysis of Comte's scientism stand as a pithy summary of the program.

... since the habits of thought which man had acquired in interpreting the actions of his own kind had long held up the study of external nature, and the latter had only made real progress in proportion as it got rid of this human habit, the way to progress in the study of man must be the same: we must cease to consider man anthropomorphically and must treat him as if we knew about him as little as we know about external nature. Although Comte does not say so in so many words, he comes very near doing so, and therefore one cannot help wondering how he could have failed to see the paradoxical nature of this conclusion.⁵⁴

The sociological part of the Cours expanded to a size exceeding the sum of the other three volumes which Comte devoted to the natural sciences. The most crucial part of the sociological analysis, according to Hayek is

... the attempt to prove the basic contention, which Comte, as a young man of twenty-six, had expressed in a letter to a friend when he promised to show that "there were laws governing the development of the human race as definite as those determining the fall of a stone." History was to be made a science, and the essence of all science is that it should be capable of prediction. The dynamic part of sociology was therefore to become a philosophy of history, as it is commonly but somewhat misleadingly called, or a theory of history as it would be more correctly described . The idea which was to inspire so much of the thought of the second half of the nineteenth century, was to write "abstract history," "history without the names of men or even people." The new science was to provide a theoretical scheme, an abstract order in which the major changes of human civilization must necessarily follow from each other.

As Hayek, writing during the Second World War, was increasingly aware, this program's eventual chilling product would be the rationale for the twentieth century's most despotic totalitarian regimes. It makes little difference whether the subsequent development set out to be principled, left-wing, academic, and intellectual, or whether it was employed by manipulative right-wing demagogues to add verisimilitude to their populist mythologies. The scientistic dehumanization of history and the portrayal of society as as subject to inevitable laws of dynamics became the suitably malleable base material from which could be fashioned political rationalizations bearing the spurious honorific of `scientific'. From today's perspective of a post-Soviet and postmodern world, we can be thankful for the much reduced persuasiveness of this scientistic sociology. But while scientisms are still alive and well, as they undoubtedly are, it is salutary to recall the consequences, however indirect, of some of the earliest identifiable examples.

I have pointed earlier to the religious character of scientism, its status as a comprehensive world-view, based on prior metaphysical commitments, and as a guiding framework for understanding all of existence and motivating moral decisions. A major objection of secularist advocates to this characterization rests upon the absence in scientism of the clerical hierarchic authority and public rituals that characterize most theistic religions. Whether such features are universal in religion is debatable. And whether scientism today possesses them is also open to interpretation when one thoughtfully considers the ways by which scientific orthodoxy is adduced and enforced, or the ways that academic ceremonies, for example, mimic those of organized religion. But regardless of the current state of scientism, a remarkable feature of the early nineteenth century scientisms was the attempt to embody them in explicit new religions, complete with all the trappings of traditional faiths.

The last of Henri Saint-Simon's works published in his lifetime was entitled New Christianity and appealed to a divine core of truth focused on "the improvement of the moral and physical existence of the poorest class". It reflected his view, articulated a few days before he died, that "The Catholic System was in contradiction with the system of sciences and of modern industry; and therefore, its fall was inevitable. It took place, and this fall is the signal for a new belief which is going to fill with its enthusiasm the void which criticism has left in the souls of men".⁵⁵ His followers, led by Barthélemy-Prosper Enfantin, set out to found the organized religion of this new belief. Their efforts began with a short-lived journal, moved into what might be considered a `house church' hosted by Hippolyte Carnot (brother of Sadi, the discoverer of the Carnot cycle), and progressed to public lectures on the Doctrine de Saint-Simon, and thence to a "Family over which Enfantin and Bazard presided as the two Supreme Fathers - new popes with a college of apostles ... "⁵⁶, services, public confession of sins, itinerant preachers, and the founding of local centers throughout the country. It ended in a move to a monastic community complete with menial labor and vows of celibacy.

Auguste Comte quickly disassociated himself in the 1820s with the Saint-Simonian religion. But in his own later writings, the Systeme de politique positive (1851-54), and the Catéchism positiviste (1852) he is concerned to move beyond intellectual analysis to "moral regeneration" which is emotional, imaginative, and subjective. To do so led him to found his own new Religion of Humanity complete with four-fold hierarchic priesthood, and a high-priest, who was of course Comte himself. Its public worship was to be practiced through commemorative acts in celebration of our beloved dead. Comte's obsessive prescriptive detail, which led many even of his supporters to allege mental instability, governs, for example, the number and length of daily private prayers, his nine personal sacraments, eighty-one annual festivals, the saints, the icons to be used in Positive churches, and the stipulation that they should all face towards the source of their enlightenment: Paris.

The Religion of Humanity gained a significant following even outside France. Although in England the main London church never had more than 137 contributors, and there were never more than six or seven English congregations outside London, it did appeal to some influential figures. John Stuart Mill for example, whose System of Logic (1843) brought Comte's sociology to the attention of the English public, gave support which, though qualified by his distaste towards the cultic details he regarded as "ridiculous", included financial support. And novelist George Eliot and her partner Henry Lewes attended Positivist church services, gave financially, and personally promoted Positivism.⁵⁷

4.3 Scientific method and demarcation

For much of the twentieth century philosophers generally, and philosophers of science particularly, sought mightily for methodological descriptions or definitions of science. These efforts can be thought of as being of two types. The more ambitious is to try to identify and explain the methods that science uses to obtain its knowledge. One such method is Induction . Francis Bacon's analysis of Induction is a large part of his claim to be the father of modern science. But in today's terminology, his ideas were an early example of the philosophy of science. Indeed, since Bacon practiced little or no science himself, it is better to consider him the father of the philosophy of science, than of science itself. Subsequently Induction has been examined logically and historically, in the hopes of being able to articulate a more or less well defined procedure, or set of procedures, that science uses, or should use, to ensure that it discovers reliable descriptions of the world. Nineteenth century philosopher John Stuart Mill, whose `Mill's Canons' are still considered an important encapsulation of inductive methods, considered Induction the foundation of all knowledge, including the knowledge of logic itself. His view was that the rules of logic are themselves discovered by Induction. In the twentieth century, the logical strengths and weaknesses of Induction, its presuppositions, and the ways that it has in fact been used by scientists have been extensively studied. Other logical and epistemological schemes have also been proposed and subjected to similar critical evaluation, as we shall shortly discuss.

The second type of definition of science is less ambitious. It takes as its objective not a comprehensive description of science's methods, but more modestly a criterion or set of criteria that distinguish science from non-science. This would be a solution to what is called the problem of demarcation : demarcating the boundaries between science and non-science. Obviously if one were successful in the more ambitious program of comprehensively describing science's methods, that would provide a solution to the problem of demarcation. Yet one can imagine a solution to demarcation that does not incorporate a solution of the more ambitious methodological challenge.

To give away the punch line, the current opinion in philosophical circles is that both of these programs has failed. Before discussing more fully this apparent failure however, I want to point out that it gives rise to a paradox. It is this. Despite having concluded that there is no satisfactory working definition of what science is, the History and Philosophy of Science has not collapsed and vanished as an academic field. It is tempting to describe colloquially the apparent failure of demarcation by concluding "HPS does not know what it is talking about". The conclusion I actually draw from the paradox that HPS still seems to be alive and well is different. It is that in fact HPS does know what it talking about, but that it is unable to say what it is talking about. In other words, there actually are some intuitive ways by which science is identified, as evidenced by the pretty clear boundaries of the topics that HPS does actually study, but that, at least thus far, these intuitions have not been sufficiently clear-cut or logically explicable to survive the rigors of HPS criticism.

What brought a new perspective to the philosophy of science in the late nineteenth century is generally thought to be the modern understanding of logic. George Boole is best known today as the progenitor of Boolean Algebra, the symbolic treatment of logic, in his An Investigation of the Laws of Thought on Which are Founded the Mathematical Theories of Logic and Probabilities (1854). Boolean Algebra is an embodiment of the rules of logic, abstracted (from the syllogism, for example) into an algebra, in which statements about sets are represented by symbols and are subject to combining operations such as Or, And, Not, etc. Actually, in Boole's own analysis, the rules of this algebra are not quite in their modern form, which was developed by others in the succeeding 30 years. The rules are somewhat comparable to, but subtly different from, the high-school algebra of numbers we all know. Boole's legacy was not widely known except to mathematicians and logicians until the 1930s, when it was realized (notably by Claude Shannon ) that these laws are precisely what is needed to analyze the operation of combinations of switches and relays. Today, because of the importance in electronics of digital circuits, which fundamentally consist of electronic switches, no engineering undergraduate education can be considered complete without at least a passing acquaintance with Boolean Algebra. But, to return to our topic, in the late nineteenth century the development of symbolic formalization was carried further, and with even higher ambitions, by Friedrich Frege . He sought to demonstrate that arithmetic is a branch of logic in his Basic Laws of Arithmetic (1893). This path was pushed to what was hoped to be a logical conclusion by Bertrand Russell and Alfred North Whitehead , in their Principia Mathematica (1910-13), whose aim was to derive the whole of mathematics from well-defined axioms and symbolic logic.

The undoubted clarification of logic itself through these symbolic mathematical developments, and their apparent⁵⁸ success in establishing a firm logical foundation for mathematics, was of great significance for the philosophy of science. It removed much of the mystery from classical logic, for anyone who cared to master the mathematical methods; and it offered a rigorous representational model, an example or paradigm, which it appeared might be extensible to the whole of science.⁵⁹

Logical Positivism was so called mostly because of its ambition to fill this role. It asserts that there are only two types of statement that are meaningful. First, Analytic statements, including logic and mathematics, are true, in a sense by definition. Second, Empirical statements must be verified empirically; they include, and are best represented by, the statements of natural science. Any other types of statement, especially statements that cannot in principle be experimentally verified, are nonsense - not just false, but meaningless. Thus Logical Positivism not only reinforced the ambitions of the old Positivism to turn all of knowledge into positive science, but it also proffered a definition of what would qualify as science, in effect a demarcation, and ruled out definitively the cognitive value of any undertaking that failed to conform to its standard. If Logical Positivism were right, then every discipline had to be reinvented in its terms. As Rudolph Carnap, the best known of the Vienna Circle that founded Logical Positivism, put it "To pursue philosophy can only be to clarify the concepts and sentences of science by logical analysis. The instrument for this is the new logic."⁶⁰ Passing over the elementary, but none the less telling, critique that Logical Positivism is itself neither Analytic nor Empirical, and therefore by its own lights is nonsense, its most serious difficulty is perhaps the whole notion of verification. It rapidly became clear that what counts as verification could not be adequately articulated, and that in any case, acknowledged theories of natural science frequently did not seem to qualify.

Falsificationism (sometimes called Deductivism) is associated primarily with Karl Popper, whose Logik der Forschung (1934) followed closely on the heels of the logical positivists, although its English translation, Logic of Scientific Discovery⁶¹, did not appear till 1959. Popper's position is critical of the positivists, but still seeks a description of science's methods in terms of logic. Popper's approach is based on the elementary observation that a universal proposition such as "all swans are white" can not logically be proved, no matter how many white swans we see, but can be logically disproved by the observation of a single black swan. He concluded that verification of universal propositions through processes of induction and confirmation is not what science does. Instead, the way science works is by systematic attempts to falsify supposed universal laws. When a law successfully survives many such attempts, when it passes the most stringent of potentially falsifying tests, it is regarded as having thereby gained strong corroboration . Although Popper does not advocate a thorough-going scientism like the logical positivists that dismisses unfalsifiable statements as ipso facto meaningless; and he thus avoids their worst excesses of self-contradiction; he does promote falsifiability as a demarcation criterion, as the test of whether a statement is or is not scientific. We shall see in a moment that replacing verification with falsification, while it appeals to a clearer deductive logic, has some serious problems of its own.

Logical empiricism was the term Carl Hempel (1905-97) preferred to describe his position, though he was part of the positivist Vienna circle, and unlike Popper accepted Induction as a process of verification. He was predominantly again a logician, and an advocate of the analysis and identification of science with logic. He is most famous for his Raven Paradox⁶², which runs approximately like this.

The statement "All ravens are black" is logically equivalent to "All non-black things are non-ravens". The observation of a black raven is inductive evidence in support of "All ravens are black", and hence of its equivalent, "All non-black things are non-ravens". The observation of a non-black thing that is a non-raven, for example a white shoe, is inductive evidence in support of "All non-black things are non-ravens", and hence of its equivalent "All ravens are black". But it appears unreasonable to regard the observation of a white shoe as evidence that all ravens are black.

This paradox has proven extremely fruitful in the analyses of Induction that it has provoked. It would be inappropriate here to try to list or evaluate all the different types of resolution that have been proposed. However, they all have in common that they invoke something more than the operations of pure logic on the isolated case. Either they quantify the degree of support by appealing to presumptions about the wider context, or else, seemingly more drastically, they abandon Boolean two-value logic, and appeal to other more generalized `logics'. Let's illustrate with a couple of examples of the appeal to external information. One can argue (in a way that can be formulated quantitatively using Bayes' theorem in probability) that the weight of support given by the white shoe is far less than that of the black raven because we know that ravens are a tiny fraction of all the things there are, whereas the number of non-black things is not. Or one can argue that it depends on one's knowledge of the external structure of the universe whether in fact a black raven observation supports the universal proposition all ravens are black. Suppose we happen to know that we are standing in one of two aviaries of which one aviary possesses just one raven (and it is black) out of ten thousand birds, while the other aviary has one hundred ravens, just one of which is white, out of ten thousand birds, and we see a black raven among the thirty other birds in our field of view. The best guess is then not that "all ravens are black" in our little aviary universe, but actually that, because we see a raven at all, we are probably in the one-hundred-raven aviary, in which not all are black. The observation of a black raven in this situation is evidence against the proposition that all ravens are black. Although this example may seem contrived, it illustrates that the evidentiary value of an observation, even whether it is for or against a universal law, depends upon the structure of the `universe', and our prior knowledge of it. The reason why we don't consider the white shoe evidence that all ravens are black is that we intuitively incorporate into our thinking our background knowledge of the universe, which is not present in the purely logical analysis of the statements.

Hempel's objective was to acknowledge and solve a serious challenge to describing Induction, conceived as a purely logical process, as the foundation of science's method. The predominant view today is that what he succeeded in doing was to clarify the seriousness of the challenge to such an extent that the whole logical positivist/empiricist program became untenable.

Under-determination is a label that gathers together several critiques of simplistic views of the relationship between theory and empirical data. At the beginning of the twentieth century, Pierre Duhem had made the important point "... the physicist can never subject an isolated hypothesis to experimental test, but only a whole group of hypotheses; when the experiment is in disagreement with his predictions, what he learns is that at least one of the hypotheses constituting this group is unacceptable and ought to be modified; but the experiment does not designate which one should be changed."⁶³ There is no such thing as a simple empirical test of a theory; "hypotheses are tested in bundles". A theory is always tested in conjunction with lots of other auxiliary hypotheses which are necessary to render the results of the theory into experimental form. Acts of observation are not independent of theory, they are `theory laden', at the very least because even the techniques of observation rely upon theory. Therefore there is no such thing as the unambiguous experimental falsification of a theoretical hypothesis. When a result is observed which contradicts deductions from a hypothesis, either the hypothesis is false, or something in the rest of science used to deduce the falsified result is erroneous, or the observation itself is mistaken. Duhem believed that this meant there could be no such thing as a Crucial Experiment , because theories are under-determined by the experimental data. We have to choose not between two competing hypotheses but among an infinite number of possible hypotheses. Duhem illustrates this with reference to the experiment of Foucault on the speed of light in water that helped to distinguish between the wave and particle theories of light, saying "Light may be a swarm of projectiles, or it may be a vibratory motion whose waves are propagated in a medium; is it forbidden to be anything else at all?" He believed not.

In his celebrated critique entitled The Two Dogmas of Empiricism⁶⁴, W. V. Quine, in 1951, argued that neither Logical Positivism's distinction between analytic and empirical statements, nor the supposition that individual empirical statements can be reduced to immediate experience were supportable. Instead "... our statements about the external world face the tribunal of sense experience not individually but only as a corporate body." This echoes Duhem's assertion, but extends it to all of knowledge, including logic itself: "The totality of our so-called knowledge or beliefs, from the most casual matters of geography and history to the profoundest laws of atomic physics or even of pure mathematics and logic, is a man-made fabric which impinges on experience only along the edges. Or, to change the figure, total science is like a field of force whose boundary conditions are experience. A conflict with experience at the periphery occasions readjustments in the interior of the field."

Sociology of Science in the form of Thomas Kuhn's seminal The Structure of Scientific Revolutions⁶⁵, which first appeared in 1962, gave the tottering edifice of logicisms the final shove that brought their ambitions to ruins. However, these particular philosophies of science could hardly be considered more than collateral damage from the bombshell he dropped on the whole enterprise of systematizing science's methods. Kuhn's perspective was based on an observational sociology of science. He treated theorizing about logical structures of scientific warrant as a fruitless form of armchair philosophy, and set out instead to document and analyze what scientists actually do. While both the details and the conclusions of his work have been extensively criticized, many of the scientific traits he identified rang true in the experience of scientists, and he gained such credibility that no subsequent HPS work could ignore it. [Scientists themselves can ignore it, and many have done so, just as they have in large numbers ignored the whole of the History and Philosophy of Science discipline; but that is another tale.] Kuhn observed that, in contradiction to Popper, scientists don't in general try to falsify their theories; they defend them. What's more, theories aren't immediately considered defunct if a single example of a falsifying observation arises. Instead, they are shored up by adjustment of auxiliary hypotheses .

This much was not really revolutionary, but Kuhn's more major point concerned those rare but exciting moments of history when a transformation occurs of our conceptions of the universe, or more modestly at least, of a scientific discipline. Such revolutions replace, for Kuhn, the picture of slow, steady, or stumbling, self-correction, which is the more traditional image of science. He regards `normal' science as being the quiescent period of gradual self-consistent development of the reigning paradigm. It leads eventually to an accumulation of explanatory strains arising from the accommodation of perplexing experimental results or theoretical dissatisfactions. Those strains continue to grow, but the paradigm is not overthrown until it can be replaced wholesale by a new one that relieves the strain by a more satisfying explanation of the results that puzzled the prior one. And the process begins over. The formation of paradigms, and the replacement of one paradigm by the other is, by contrast with normal science, a discontinuous and logically inexplicable revolution. In the end, science seems to be little more than opinion, expert opinion granted, but still just opinion. There is, in Kuhn's words "no standard higher than the consent of the relevant community": a situation that has been colorfully characterized as scientific mob rule.

So much so, that Paul Feyerabend seemed to many to represent an extrapolation of the same intellectual path when he argued in Against Method⁶⁶ (1975) that there is no scientific method; that science is, and should be, anarchic; and that theories are successful because of the no-holds-barred rhetorical tactics of their advocates, not because of any particular correspondence with reality, or conformity with accepted epistemological schemes. Feyerabend is completely explicit that his stand is motivated to a large degree by a rejection of what he dubs the chauvinism of science, "... let me repeat that for me the chauvinism of science is a much greater problem ... Scientists are not content with running their own playpens in accordance with what they regard as the rules of scientific method, they want to universalize these rules, they want them to become part of society at large ...". It is a charge of scientism.

Later sociological studies of science have gone to extreme in developing this theme, arguing that there is nothing special about science, that its knowledge is no more certain than any other type of knowledge, that its knowledge is culturally determined, and that purely cultural analyses of science (e.g. feminist, or third-world perspectives) might lead to sciences with different technical content. We'll explore these postmodern viewpoints later, but I remark here that scientists (and I mean natural scientists) generally regard the notion that there might be a feminist chemistry (for example), having different technical content from our current chemistry, as utterly ludicrous.

Research Programs is the terminology advocated in place of paradigms by Imre Lakatos, Feyerabend's contemporary and sparring partner in the metaphorical bout of rationalist (Lakatos in the blue corner) vs irrationalist (Feyerabend in the red corner). A Research Program "... is a series of successive theories with their associated auxiliary assumptions, instrumental procedures, etc." What distinguishes the worthy from the unworthy Research Programs is whether or not they are `progressive' . All right, but what is progressive? How is that to be measured or assessed? Several plausible measures of progressiveness are offered by the advocates of this viewpoint. Independent testability arising from the ability to test auxiliary hypotheses independently, "in different bundles"; Unification, meaning the successful application of a small family of problem-solving strategies to a broad class of cases; and Fecundity, designating the property of theories that open up new and profitable lines of investigation.⁶⁷ The vagueness of these `progressiveness' measures is not just my brevity here. It is the outward evidence that the attempt to define science in terms of mathematically precise logic has been abandoned.

4.4 Demarcation disputes

One might think these increasingly abstruse debates about the status and identity of science an area of philosophical specialization with little practical significance. However, these matters of demarcation have been brought very much into the American public eye in recent years by the role they play in battles about high-school biology. The 1925 show trial of John Scopes for teaching evolution in Dayton, Tennessee, is surrounded by sufficient contemporary and subsequent commentary - and drama - to fill several bookshelves. We shall not expand it further here and will postpone discussion of evolution itself to a later section. But it is appropriate to comment on the influence of the ongoing legal skirmishes upon demarcation.

Although the guilty Scopes verdict was overturned on the absurd technicality that the $100 fine exceeded the judge's authority, anti-evolution law remained on the statutes of Tennessee and several other US states. Text-books worked cautiously around these laws until in 1963 a Pennsylvania court ruled that Bible reading and the Lord's Prayer were not appropriate for public schools but that neutrality between "those who believe in no religion [and] those who do believe" must be observed. Then in 1968 the Arkansas Supreme Court struck down its anti-evolution statute, and was followed in this action by Mississippi two years later.⁶⁸ Thereafter repeated attempts by activists to introduce creationism into the curriculum have repeatedly been overturned by the courts based on the opinion that it advanced particular religious preferences and entangled the state with religion, in contravention of the First Amendment clause "Congress shall make no law respecting an establishment of religion, or prohibiting the free exercise thereof;". The logic by which it follows from this clause that religion is to be effectively banished from public school curricula is hard to discern, but that is not our present topic. The strategy adopted by `Creation Science' activists increasingly, in the face of these reverses, was to portray creationism as science and to argue that, as such, it should be taught alongside evolution. A statute worded along these lines in Arkansas was struck down in 1982, after testimony from a host of expert witnesses. Judge Overton's decision was widely reprinted, even in the journal of the American Association for the Advancement of Science (AAAS), Science. Louisiana's similar statute arrived by a tortuous legal route at the US Supreme Court on 10 December 1987. Its defenders argued that it was not religious but scientific. Seven of the nine justices were unconvinced. Thereafter, the Intelligent Design (ID) movement⁶⁹ went to even further efforts to ensure that their ideas were free from religious taint. A number of local school-board initiatives to include it in the curriculum, or to mandate disclaimers weakening evolution, gained some temporary advantage, leading to the celebrated case in Dover, Pennsylvania, 2005. Judge Jones ruling on that case, identifies ID as "a religious alternative masquerading as a scientific theory" and castigates the school board for precipitating a pointless trial.

The key observation for us here, is that these high-profile legal decisions hinge on the question of whether certain opinions and teachings are or are not science. That this has become the deciding question is a remarkable sign of the dominance of scientism in our culture. Scientism leads to acrimonious arguments about whether opinions are or are not science because the scientistic ethos gives special status to science that it does not give to non-scientific disciplines. The result is that the demarcation of what is or is not science becomes not merely an academic philosophical discussion, but a vital legal matter that decides practical questions of deep importance and emotional significance in the minds of most of the American public.

It is impossible that even academic discussion should escape the influence of this practical significance. In the first place, the various legal verdicts have undergone strong scrutiny by commentators sympathetic to the losing side (the creationists). There is plenty of scope for the criticism that the verdicts are based on an over-simplified understanding of what constitutes science. It is in the nature of legal argument to require sufficient simplification and crispness as to justify the decision. In the process many of the subtleties are inevitably lost. The critics are naturally led to emphasize the difficulty of demarcation,⁷⁰ because the broader and the fuzzier are science's boundaries, the easier it is to argue that creationism is inside them. One might have expected the opposite influence upon those who support the verdicts: a preference to regard the demarcation as easier or more definite. However, there does not seem to have been a significant countervailing initiative among philosophers of science to clarify demarcation; although one can perhaps detect a trend away from the more extreme relativism of the sociological critiques of science. Scientists themselves are generally of the opinion that they know science without difficulty when they see it.

4.5 A more modest characterization

Since the program to define science by uncovering its logical methods has thus far failed; since every attempt to identify a process by which one could even establish what is science and what is not science is judged to have been found wanting; since, in fact, it has become a celebrated cause that the problem of demarcation is unsolved; and since this question has become a high-profile legal matter: I am in dangerous waters. I am asserting that there are two identifiable characteristics of science, reproducibility and Clarity. Am I therefore claiming to have solved the demarcation problem?

No, I am not claiming to have solved the demarcation problem. What I am observing is that, despite the difficulties that undoubtedly exist in specific demarcation, there are in fact identifiable characteristics of science. These characteristics don't provide algorithms either for the practice or the identification of science, but they are nevertheless truly part of science. They are not exhaustive. They are not all the characteristics of science; many others might be identified to differing degrees in different branches of science; but reproducibility and Clarity are vitally important ones. Moreover, and this is the most important point, these relatively non-controversial characteristics on their own constitute a substantial restriction of the scope of science. They don't seek to be definitive enough to resolve all disputes about vexed questions of identifying and demarcating science. They certainly don't resolve all the logical or methodological puzzles about how science actually works. But they are important qualitative elements that are, in actual fact, part of the way that science operates. In short, I am not setting out to provide a comprehensive solution of demarcation, but I am claiming to be able to identify some characteristics of what any solution must look like. Modest answers to parts of problems are sometimes what one must settle for.

Another factor that I want to plead in mitigation of my apparent hubris is that the difficulty of demarcation is substantially amplified by scientism . What I mean is this. First, philosophically, demarcation between science and non-science in the context of scientism, is equivalent to the demarcation between sense and nonsense, rationality and irrationality, knowledge and superstition. If one accepts scientism, the belief that science is all the knowledge there is, then that demarcation of meaningfulness is what one is trying to solve. Many philosophical arguments are about that demarcation. The demarcation of meaningfulness, of all knowledge, is beyond my interests here. I have a much narrower intent which is to identify (some of) those principles which characterize natural science - regarded as a subset, not the totality, of all valid rational knowledge. I think that this problem is soluble, at least in part, and that one should not discount the identifiable characteristics of natural science just because of failures of a wider program. Second, politically, since scientism has embroiled the problem of demarcation in high-profile legal questions that raise emotions on both sides, the difficulty of demarcation is made significantly greater. People see that practical decisions, about which they care very much, hinge on questions of scientific demarcation. No wonder there is great sensitivity about this issue from all quarters. But my whole aim here is to repudiate the scientism that leads to the enhancement of these difficulties. If, as I am saying, science is not all the knowledge there is, then the weight that demarcation has to bear is reduced to a scope that is both more manageable and less sensitive.

My ultimate defence is that I am content if necessary to regard the characteristics of repeatability and Clarity as partial definitions of what I mean by science. In doing so, I might be ruling out some studies that claim to be, or perhaps even widely are regarded as, science. If that happens, then I say, so be it. This is not simply a semantic stratagem on my part. The characteristics I have identified are possessed by those sciences that are responsible for the high epistemological prestige and compelling explanations of the natural world that we attribute to science. It is those natural sciences which are being implied almost always when science's authority is being cited. So if I have by definition ruled out some disciplines that have a reasonable claim to being scientific in the sense of systematic, I have ruled out only those that are irrelevant to the most significant debates. I recognize that this will sound chauvinistic coming from a physical scientist. But I insist once more that any perception of chauvinism in this position arises from the self-same scientistic viewpoint I am at pains to deny. I find it utterly inconsistent, for example, that toward the end of his splendid introductory book What is this thing called science?, philosopher of science, Alan Chalmers criticizes philosophers who emphasize experiment, observing that their consequent failure to provide a "universal account of science" ... "could conceivably be avoided by identifying science with experimental science, but this would hardly serve to appease those who wish to call themselves political scientists or Christian scientists, for example."⁷¹ Yet Chalmer's whole book has been about natural science. I didn't notice a single one of his many instructive illustrative exemplars of science being about sociological, political, or religious topics. If political scientists are offended, so what? Why should these folks have to be appeased? Why should one suddenly regard exclusion of political science as a disqualifying fault in an identification of science's topics, characteristics, or methods?

Finally, and this seeks immediately to soften any harshness of my previous impatient sentences, I need to be crystal clear that I have no intention to discount or disparage academic disciplines that I regard as not being science. That political science is not a science in the way I mean it, does not change its scholarly or practical value. I do not subscribe to scientism. I believe there is deep meaning, truth, relevance, and insight in non-scientific studies pursued with intelligence and rigor. But their merits have to be really their own, not the reflected glow of a terminological anachronism.

In summary, then, the discipline of History and Philosophy of Science does not have simple answers to the questions, what is or is not science? or what methods does science use? Science cannot be, or at least has not so far been, formalized into a purely logical structure. But HPS, like science itself, nevertheless appears to have intuition about what science is, which is definite enough to know what it is talking about. Natural science is what HPS studies. Although strict demarcation is fraught with peril, a peril greatly enhanced in recent debates by a scientism that artificially inflates the stakes, there are identifiable characteristics of science. Attempts to turn other disciplines, especially social disciplines, into explicit positive science, after the manner of the natural sciences, have a long history - of failure.