What Is This Thing Called Science Part 6

The central text for Lakatos's methodology is his 1970 text, "Falsification and the Methodology of Scientific Research Programmes". Most of the other key papers have been collected in Won all and Currie (1978a and 1978b). Also important is Lakatos (1968), The Problem of Inductive Logic, and (1971), "Replies to Critics". A fascinating account of Lakatos's application of his ideas to mathematics is his Proofs and Refutations (1976b). Howson (1976) contains historical case studies designed to support Lakatos's position. Another such study is Lakatos and Zahar (1975). Cohen, Feyerabend and Wartofsky (1976) is a collection of essays in memory of Lakatos. Feyerabend (1976) is an important critique of Lakatos's methodology. The notion of a novel prediction is discussed by Musgrave (1974b), Worrall (1985), Worrall (1989a) and Mayo (1996). A useful overview of Lakatos's work is B. Larvor (1998), Lakatos: An Introduction.

CHAPTER 10.

Feyerabend's anarchistic theory of science.

The story so far.

We seem to be having trouble with our search for the characterisation of science that will serve to pick out what distinguishes it from other kinds of knowledge. We started with the idea, adopted by the positivists who were so influential earlier in the century that science is special because it is derived from the facts, but this attempt floundered because facts are not sufficiently straigt.i.tiorward for this view to be sustained, since they are "theory-dependent" and fallible, and because no clear account of how theories can be "derived" from the facts could be found. Falsificationism did not fare much better, mainly because in any realistic situation in science it is not possible to locate the cause of a faulty prediction, so a clear sense of how theories can be falsified becomes almost as elusive as a clear sense of how they can be confirmed. Both Kuhn and Lakatos tried to solve the problem by focusing attention on the theoretical framework in which scientists work. However, Kuhn, for his part, stressed the extent to which workers in rival paradigms "live in different worlds" to such a degree that he left himself with inadequate resources for elucidating a sense in which a change from one paradigm to another in the course of a scientific revolution is a step forward. Lakatos tried to avoid that trap, but, apart from problems concerning the reality of the methodological decisions he freely invoked in his answer, he ended up with a criterion for characterising science that was so lax that few intellectual pursuits could be ruled out. One philosopher of science who was not surprised by, and who attempted to draw out what he saw to be the full implications of, these failures was Paul Feyerabend, whose controversial but nevertheless influential "anarchistic" account of science is described and a.s.sessed in this chapter.

Feyerabend's case against method.

Paul Feyerabend, an Austrian who was based in Berkeley, California, for most of his academic career, but who also spent time interacting with (and antagonising) Popper and Lakatos in London, published a book in 1975 with the t.i.tle Against Method: Outline of an Anarchistic Theory of Knowledge. In it he challenged all of the attempts to give an account of scientific method that would serve to capture its special status by arguing that there is no such method and, indeed, that science does not possess features that render it necessarily superior to other forms of knowledge. If there is a single, unchanging principle of scientific method, Feyerabend came to profess, it is the principle "anything goes". There are pa.s.sages in Feyerabend's writings, both early and late, that can be drawn on to severely qualify the extreme anarchistic account of science that is contained in the bulk of Against Method. However, it will be most instructive for our purpose to stick to the unqualified, anarchistic theory of science to see what we can learn from it. In any case, it is the extreme form of Feyerabend's position that has made its mark in the literature and which philosophers of science have, not without difficulty, attempted to counter.

Feyerabend's main line of argument attempts to undermine characterisations of method and progress in science offered by philosophers by challenging them on their own ground in the following way. He takes examples of scientific change which his opponents (including the vast majority of philosophers) consider to be cla.s.sic instances of scientific progress and shows that, as a matter of historical fact, those, changes did not conform to the theories of science proposed by those philosophers. (Feyerabend does not have to himself agree that the episodes in question were progressive for his argument to go through.) The main example appealed to by Feyerabend involves the advances in physics and astronomy made by Galileo. Feyerabend's point is that if an account of method and progress in science cannot even make sense of Galileo's innovations, then it is not much of an account of science. In this outline of Feyerabend's position I will stick largely to the Galileo example, mainly because it is sufficient to ill.u.s.trate Feyerabend's position, but also because the example is readily understood without requiring resort to recondite technicalities.

A number of Feyerabend's points will be familiar because I have already drawn on them for various purposes earlier in this book.

Quotations invoked in chapter 1 of this book ill.u.s.trate the positivist or inductivist view that Galileo's innovations can be explained in terms of the extent to Which he, took the observable facts seriously and built his theories to fit them. The following pa.s.sage from Galileo's Dialogue Concerning the Two Chief World Systems (1967), cited by Feyerabend (1975, pp. 100-101), indicates that Galileo thought otherwise.

You wonder that there are so few followers of the Pythagorean opinion [that the earth moves] while I am astonished that there have been any up to this day who have embraced and followed it. Nor can I ever sufficiently admire the outstanding ac.u.men of those who have taken hold of this opinion and accepted it as true: they have, through sheer force of intellect done such violence to their own senses as to prefer what reason told them over that which sensible experience plainly showed them to the contrary For the arguments against the whirling of the earth we have already examined are very plausible, as we have seen: and the fact that the Ptolemaics and the Aristotelians and all their disciples took them to be conclusive is indeed a strong argument of their effectiveness. But the experiences which overtly contradict the annual movement are indeed so much greater in their apparent force that, I repeat, there is no limit to my astonishment when I reflect that Aristarchus and Copernicus were able to make reason so conquer sense that, in defiance of the latter, the former became mistress of their belief.

Far from accepting the facts considered to be borne out by the senses by his contemporaries, it was necessary for Galileo (1967, p. 328) to conquer sense by reason and even to replace the senses by "a superior and better sense" , namely the telescope. Let us consider two instances where Galileo needed to "conquer" the evidence of the senses - his rejection of the claim that the earth is stationary and his rejection of the claim that the apparent sizes of Venus and Mars do not change appreciably during the course of the year.

If a stone is dropped from the top of a tower it falls to the base of the tower. This, and other experiences like it, can be taken as evidence that the earth is stationary. For if the earth moves, spinning on its axis, say, (the whirling of the earth referred to by Galileo in the pa.s.sage cited) then should it not move frOM beneath the stone during its fall, with the result that the stone should fall some distance from the base of the tower? Did Galileo reject this argument by appealing to the facts? That is certainly not how Galileo did it in the Dialogue, as Feyerabend pointed out. Galileo (1967, p. 125 ff) achieved the desired result by "picking the brains" of the reader. He argued as follows. The speed of a ball set rolling down a frictionless slope will increase, because it is "falling" towards the centre of the earth to some degree. Conversely, the speed of a ball rolled up a frictionless slope will decrease because it is rising away from the centre of the earth. Having persuaded the reader to accept this as obvious, he or she is now asked what will happen to the speed of the ball if the slope is perfectly horizontal. It would seem that the answer is that the speed will neither increase nor decrease since the ball will be neither rising nor falling. The horizontal motion of the ball persists and remains constant. Although this falls short of Newton's law of inertia, it is an example of a uniform motion that persists without a cause, and it is sufficient for Galileo to counter a range of arguments against the spinning earth. Galileo draws the implication that the horizontal motion of the stone falling from the tower, which it shares with the tower as the earth spins, remains unchanged. That is why it stays with the tower, striking the ground at its foot. So the tower argument does not establish that the earth is stationary in the way many had supposed. To the extent that Galileo's case was successful it did not involve appealing to the results of observation and experiment, at his own admit ance, (I point out here that frictionless slopes were eve harder to obtain in Galileo's time than they are now, and that measuring the speed of a ball at various locations on the slope lay beyond what was feasible at the time.) We saw in chapter 1 that the apparent sizes of Venus and Mars were important insofar as the Copernican theory predicted that they should change appreciably, a prediction not borne out by naked-eye observations. The problem is resolved once the telescopic rather than the naked-eye data is accepted. But how was the prefyrence for the telescopic data to be defended? Feyerabend's rendering of the situation and Galileo's response to it run as follows. Accepting what the telescope revealed in the astronomical context was by no means straightforward. Galileo did not have an adequate or detailed theory of the telescope, so he could not defend the telescopic data by appeal to one. It is true that in a terrestrial context there were trial and error methods of vindicating telescopic sightings. For instance, the reading of an inscription on a distant building, indiscernible to the naked eye, could be checked by going close to the building, and the identification of the cargo of a distant ship could be vindicated once the ship arrived in port. But the vindication of terrestrial use could not be straightforwardly emptOyed to justify astronomical use of the telescope. Terrestrial use of the telescope * is aided by a range of visual cues absent in the astronomical case. Genuine images can be distinguished from many artifacts of the telescope because we are familiar with the kinds of things being inspected. So, for instance, if the telescope reveals the mast of a distant ship to be wavy, red on one side and blue on the other and accompanied by black specks hovering above it, the distortions, colours and specks can be dismissed as artifacts. However, when looking into the heavens, we are in unfamiliar territory and lack clear guidance as to what is really there as opposed to an artifact. What is more, comparison with familiar objects to help judge size, and the use of parallax and overlap to help judge what is far and what is near, is a luxury not in general available in astronomy and it is certainly not the case that Galileo could check telescopic sightings of planets by moving closer to them to check with the naked eye. There was even direct evidence that the telescopic data was erratic insofar as it magnified the moon to a different degree than it magnified the planets and stars.

According to Feyerabend (1975, p. 141), these difficulties were such that recourse to argument would have been inadequate for the task of convincing those opponents who wished to deny both the Copernican theory and the telescopic data relating to the heavens. Consequently, Galileo needed to, and did, resort to propaganda and trickery.

On the other hand, there are some telescopic phenomena which are plainly Copernican. Galileo introduces these phenomena as independent evidence for Copernicus while the situation is rather that one refuted view - Copernicanism - has a certain similarity to phenomena emerging from ari,other refuted view -The idea that telescopic phenomena are faiihful images of the sky. Galileo prevails because of his style and his clever techniques of persuasion, because he writes in Italian rather than in Latin, and because he appeals to people who are temperamentally opposed to the old ideas and the standards of learning connected with them.

It should be clear that if Feyerabend's construal of Galileo's methodology is correct and typical of science, then standard positivist, inductivist and falsificationist accounts of science have serious problems accommodating it. It can be accommodated into Lakatos's methodology according to Feyerabend, but only because that methodology is so lax that it can accommodate almost anything. Feyerabend teased Lakatos by welcoming him as a "fellow anarchist", albeit one "in disguise", playfully dedicating Against Method to Lakatos "friend, and fellow anarchist". The way in which Feyerabend construes the two frameworks, the Aristotelian stationary earth framework backed up by naked-eye data and the Copernican, moving earth theory supported by telescopic data, as mutually exclusive circles of thought, as it were, is reminiscent of Kuhn's portrayal of paradigms as mutually exclusive ways of seeing the world. Indeed, the two philosophers both independently coined the word "incommensurable" to describe the relationship between two theories or paradigms that cannot be logically compared for lack of theory-neutral facts to exploit in the comparison. Kuhn avoided Feyerabend's anarchistic conclusions essentially by appealing to social consensus to restore law and order. Feyerabend (1970) rejected Kuhn's appeal to the social consensus of the scientific community, partly because he did not think Kuhn distinguished between legitimate and illegitimate ways (for example by killing all opponents) of achieving consensus, and also because he did not think the appeal to consensus was capable of distinguishing between science and other activities such as theology and organised crime.

Given the failure of attempts to capture the special features of scientific knowledge that render it superior to other forms, which failure Feyerabend considered himself to have established, he drew the conclusion that the high status attributed to science in our society, and the superiority it is presumed to have not only over Marxism, say, but over such things as black magic and voodoo, are not justified. According to Feyerabend, the high regard for science is a dangerous dogma, playing a repressive role similar to that which he portrays Christianity as having played in the seventeenth century having in mind such things as Galileo's struggles with the Church.

Feyerabend's advocacy of freedom.

Feyerabend's theory of science is situated in an ethical framework which places a high value on individual freedom, involving an att.i.tude that Feyerabend described as the 'humanitarian att.i.tude". According to that att.i.tude, individual humans should be free and possess liberty in something like the sense the nineteenth-century philosopher John Stuart Mill (1975) defended in his essay "On Liberty". Feyerabend (1975, p. 20) declared himself in favour of "the attempt to increase liberty, to lead a full and rewarding life" and supports Mill in advocating "the cultivation of individuality which alone produces, or can produce, well-developed human beings". From this humanitarian point of view, Feyerabend supports his anarchistic account of science on the grounds that it increases the freedom of scientists by removing them from methodological constraints and, more generally, leaves individuals the freedom to choose between science and other forms of knowledge.

From Feyerabend's point if view, the inst.i.tutionalisation of science in our society is inconsistent with the humanitarian att.i.tude. In schools, for example, science is taught as a matter of course. "Thus, while an American can now choose the religion he likes, he is still not permitted to demand that his children learn magic rather than science at school. There is a separation between state and Church, there is no separation between state and science" (1975, p. 299). What we need to do in the light of this, wrote Feyerabend (1975, p. 307) , is to "free society from the strangling hold of an ideologically petrified science just as our ancestors freed us from the strangling _hold of the One True Religion!". In Feyerabend's image of a frees ' 'society, science will not se given preference over other forms of knowledge or over other traditions. A mature citizen in a free society is "a person who has learned to make up his mind and who has then decided in favour of what he thinks suits him best". Science will be studied as a historical phenomenon "together with other fairy tales such as the myths of 'primitive' societies" so that each individual "has the infoi [nation needed for arriving at a free decision" (1975, p. 308, italics in original). In Feyerabend's ideal society the state is ideologically neutral between ideologies to ensure that individuals maintain freedom of choice and do not have an ideology imposed on them against their will.

The culmination of Feyerabend's case against method, together with his advocacy of a particular brand of freedom for the individual, is his anarchistic theory of knowledge (1975, pp. 284-5, italics in original).

None of the methods which Carnap, Hempel, Nagel [three prominent positivists], Popper or even Lakatos want to use for rationalising scientific changes can be applied, and the one that can be applied, refutation, is greatly reduced in strength. What remains are aesthetic judgments, judgments of taste, metaphysical prejudices, religious desires, in short, what remains are our subjective wishes: science at its most advanced and general returns to the individual a freedom he seems to lose in its more pedestrian parts.

There is no scientific method, then: Scientists should follow their subjective wishes. Anything goes.

Critique of Feyerabend's individualism.

A critique of Feyerabend's understanding of human freedom will act as a useful preliminary to an appraisal of his critique of method. A central problem with Feyerabend's notion of freedom is from the degree to which it is entirely negative, in the sense that freedom is understood as freedom from constraints. Individuals should be free of constraints to the extent that they can follow their subjective wishes and do what they like. This overlooks the positive side of the issue, the extent to which individuals have access to the means to fulfil their wishes. For example, freedom of speech can be, and often is, discussed in tei ins of freedom from constraints, in the form of state suppression, libel laws and the like. So, for example, if students disru '' Pf"a lecture on campus by an academic expressing views sympathetic to Fascism they might well be accused of denying the speaker freedom of speech. They are accused of putting an obstacle in the way of the speaker's natural right. However, freedom of speech can be considered, from the positive point of view, in terms of the resources available to individuals to have their views heard by others. What access does a particular individual have to the media, for example? This point of view puts our example in a different light. The disruption of the lecture could perhaps be justified on the grounds that the speaker was given access to a university lecture hall, microphone, media advertising and so on in a way that those advocating other views were not. The eighteenth-century philosopher David Hume nicely ill.u.s.trated the point I am getting at when he criticised John Locke's idea of the Social Contract Locke had construed the social contract as being freely adopted by members of a democratic society and argued that anyone not wishing to subscribe to the contract was free to emigrate. Hume responded as follows: Can we seriously say, that a poor peasant or artisan has a free choice to leave his country when he knows no foreign language or manners, and lives from day to day, by the small wages which he acquires? We may as well a.s.sert that a man, by remaining in a vessel, freely consents to the domination of the master; though he was carried on board while a:sleep, and must leap into the ocean and perish, the moment he leaves her. Individuals are born into a society that pre-exists them and which, in that sense, possesses characteristics they do not choose and cannot be in a position to choose. The courses of action open to them, and, consequently, the precise senses in which they are free, will be determined by the access that they have in practice to the resources necessary for various courses of action. In science too an individual who wishes to make a contribution to a science will be confronted by the situation as it stands various theories, mathematical techniques, instruments and experimental techniques. The paths of action open to scientists in general will be delimited by that objectively existing situation, while the paths open to a particular scientist will be determined by the subset of the existing resources to which that individual scientist has access. Scientists will be free to follow their "subjective wishes" only insofar as they are free to chose among the restricted range af_nptions open to them. What is more, a prerequisite for an understanding of that situation will be a characterisation of the situation that individuals face, like it or not. Whether it be changes in science or in society generally, the main theoretical work involves understanding the situations confronted by individuals rather than involving some generalised appeal to unconstrained freedom.

It is ironic that Feyerabend, who in his study of science goes to great lengths to deny the existence of theory-neutral facts, in his social theory appeals to the far more ambitious notion of an ideology-neutral State. How on earth would such a State come into existence, how would it function and what would sustain it? In the light of work that has been done in making serious attempts to get to grips with questions about the origin and nature of "the State", Feyerabend's fanciful speculations about a utopia in which all individuals are free to follow their inclinations in an unrestricted way appear childish.

Criticising Feyerabend for setting his views on science in an individualist framework involving a naive notion of freedom is one thing. Getting to grips with the details of the case he makes "against method" in science is another. In the next chapter we will see what can be constructively salvaged from Feyerabend's attack on method.

Further reading.

Feyerabend develops some of the ideas of his Against Method: Outline of an Anarchistic Theory of Knowledge (1975) in Science in a Free Society (1978). Realism, Rationalism and Scientific Method (Feyerabend, 1981a) and Problems of Em piricism (Feyerabend, 1981b) are collections of his articles, a number of which predate his "anarchistic" phase. "Consolations for the Specialist" (1970) and "On the Critique of Scientific Reason" (1976) are his critiques of Kuhn and Lakatos respectively. I have taken issue with Feyerabend's portrayal of Galileo's science in "Galileo's Telescopic Observations of Venus and Mars" (Chalmers, 1985) and "The Galileo that Feyerabend Missed" (Chalmers, 1986).

CHAPTER 11:.

Methodical changes in method.

Against universal method.

We saw in the previous chapter that Feyerabend made a case against the various accounts of scientific method that have been put forward by philosophers as attempts to capture the distinctive feature of scientific knowledge. A key strategy that he employed was to argue for the incompatibility of those accounts and Galileo's advances in physics and astronomy. Elsewhere (in Chalmers, 1985 and 1986) I have taken issue with Feyerabend's historical account of the Galileo episode and some of the details of my disagreement will be introduced and exploited in the next section. Once that history is corrected I believe it to remain the case that the corrected history poses problems for standard accounts of science and the scientific method. That is, I suggest there is a sense in which Feyerabend's case against method can be sustained ,provided we are clear about the notion of method that has been refuted. Feyerabend's case tells against the claim that there is a universal, ahistorical method of science that contains standards that all sciences should live up to if they are to be worthy of the t.i.tle "science". Here the term "universal" is used to indicate that the proposed method is to apply to all sciences or putative sciences - physics, psychology, creation science or whatever - while the term "ahistorical" signals the timeless character of the method. It is to be used to appraise Aristotle's physics as much as Einstein's and Democritus's atomism as much as modern atomic physics. I am happy to join Feyerabend in regarding the idea of a universal and ahistoric method as highly implausible and even absurd. As Feyerabend (1975, p. 295) says, "The idea that science can, and should, be run according to fixed and universal rules is both unrealistic and pernicious", is "detrimental to science, for it neglects the complex physical and historical conditions which influence scientific change" and "makes science less adaptable and more dogmatic". If there is to be a scientific method capable of judging sciences of all kinds, past, present and future, one might well ask what resources philosophers have for arriving at such a potent tool, so potent that it can tell us in advance what are the appropriate standards for judging future science. If we have a conception of science as an open-ended quest to improve our knowledge, then why cannot there be room for us to improve our methods and adapt and refine our standards in the light of what we learn.

I have no problem joining the campaign that Feyerabend launched against method, then, provided method is understood as universal, unchanging method. We have seen that Feyerabend's response to the case against method is to a.s.sume that there is no method, that scientists should follow their own subjective wishes and that anything goes. However, universal method and no method at all do not exhaust the range of possibilities. A middle way would hold that there are methods and standards in science, but that they can vary from science to science and can, within a science, be changed, and changed for the better. Not only does Feyerabend's case not tell against this intermediate view, but his Galileo example can be construed in a way that supports it, as I shall attempt to show in the next section.

I hold that there is a middle way, according to which there are historically contingent methods and standards implicit in successful sciences. A common response from philosophers of science who reject Feyerabend's anarchism and extreme relativism as firmly as I do is that those like myself who seek a middle way are kidding ourselves. John Worrall (1988), for instance, has given clear expression to the general line of argument. If I am to defend a change in scientific method in a way that avoids extreme relativism then I am obliged to show in what way such a change is for the better. But better according to what standards? It would seem that unless there are some superstandards for judging changes in standards then those changes cannot be construed in a non-relativist way. But superstandards takes us back to the universal method that is meant to yield such standards. So, Won-all's argument goes, either we have universal method or relativism. There is no middle way. As at least a preliminary to a rejoinder to this argument it is useful to take an example from science of a change in standards. The next section is devoted to such a change accomplished by Galileo.

Telescopic for naked-eye data: a change in standards.

One of Galileo's Aristotelian opponents (cited in Galileo, 1967, p. 248) referred to the idea that "the senses and experience should be our guide in philosophising" as "the criterion of science itself". A number of commentators on the Aristotelian tradition have noted that it was a key principle within that tradition that knowledge claims should be compatible with the evidence of the senses when they are used with sufficient care under suitable conditions. Ludovico Geymonat (1965, p. 45), a biographer of Galileo, refers to the belief "shared by most scholars at the time (of Galileo's innovationsi" that "only direct vision has the power to grasp actual reality". Maurice Clavelin (1974, p. 384), in a context where he is comparing Galilean and Aristotelian science, observes that "the chief maxim of Peripatetic physics was never to oppose the evidence of the senses", and Stephen Gaukroger (1978, p. 92), in a similar context, writes of "a fundamental and exclusive reliance on sense-perception in Aristotle's works". A teleological defence of this fundamental standard was common. The function of the senses was understood to be to provide us with information about the world. Therefore, although the senses can mislead in abnormal circ.u.mstances, for instance in a mist or when the observer is sick or drunk, it makes no sense to a.s.sume that the senses can be systematically misleading when they are fulfilling the task for which they are intended. Irving Block (1961, p. 9), in an illuminating article on Aris totle's theory of sense perception, characterises Aristotle's view as follows: Nature made everything for a purpose, and the purpose of man is to understand Nature through science. Thus it would have been a contradiction for Nature to have fashioned man and his organs in such a way that all knowledge and science must, from its inception, be false.

Aristotle's views were echoed by Thomas Aquinas many centuries later, as Block (1961, p. 7) reports: Sense perception is always truthful with respect to its proper objects, - for natural powers do not, as a general rule, fail in the activities proper to them, and if they do fail, this is due to some derangement or other. Thus, only in a minority of cases do the senses judge inaccurately of their proper objects, and then only through some organic defect, e.g. when people sick with fever taste sweet things as bitter because their tongues are ill-disposed.

Galileo was faced with a situation in which a reliance on the senses, including naked-eye data was "a criterion of science itself". In order to introduce the telescope, and have telescopic data replace and overrule some naked-eye data, he needed to fly in the face of this criterion. By the time he had done so, he had effected a change in the standards of science. As we have seen, Feyerabend did not believe it was possible for Galileo to make a compelling case and needed to resort to propaganda and trickery. The historical facts tell otherwise.

I have already considered the case that Galileo made for the veracity of his sightings of the moons of Jupiter. Here I will focus on the case that Feyerabend was able to muster for accepting what the telescope revealed of the changing apparent sizes of Venus and Mars. We have already described, in the previous chapter, the urgency of the question and also accepted Feyerabend's account of the difficulties that lay in the way of accepting telescopic observations of the heavens.

Galileo appealed to the phenomenon of irradiation to help discredit naked-eye observations of the planets and as providing grounds for preferring the telescopic observations. Galileo's hypothesis (1967, p. 333) was that the eye "introduces a hindrance of its own" when it views small, bright, distant light sources against a dark background. Because of this, such objects appear "festooned with advent.i.tious and alien rays". Thus, Galileo (1957, p. 46) explained elsewhere, if stars "are viewed by means of unaided vision, they present themselves to us not as of their simple (and, so to speak, their physical) size but as irradiated by a certain fulgor and as fringed with sparkling rays". In the case of the planets irradiation is removed by the telescope.

Since Galileo's hypothesis involves the claim that irradiation arises as a consequence of the brightness, smallness and distance of the source, it can be tested by modifying those factors in a variety of ways which do not involve use of the telescope. A number of ways are explicitly invoked by Galileo (1957, pp. 46-7). The brightness of stars and planets can be reduced by viewing them through a cloud, a black veil, coloured gla.s.s, a tube, a gap between the fingers or a pinhole in a card. In the case of planets the irradiation is removed by these techniques, so that they "show their globes perfectly round and definitely bounded", whereas in the case of stars the irradiation is never completely removed, so that they are "never seen to be bounded by a circular periphery but have rather the aspect of blazes whose rays vibrate about them and scintillate a great deal". As far as the dependence of irradiation on the apparent size of the observed light source is concerned, Galileo's hypothesis is borne out by the fact that the moon and the sun are not subject to irradiation. This aspect of Galileo's hypothesis, as well as the a.s.sociated dependence of irradiation on the distance of the source, can be subject to a direct terrestrial test. A lighted torch can be viewed from near or far and at day or night. When viewed at a distance at night, when it is bright compared with its surroundings, it appears larger than its true size. Accordingly, Galileo (1967, p. 361) remarked that his predecessors, including Tycho and Clavius, should have proceeded with more caution when estimating the size of stars.

I will not believe that they thought that the true disc of a torch was as it appears in profound darkness, rather than as it is when perceived in lighted surroundings: for our lights seen from afar at night look large, but from near at hand their true flames are seen to be small and circ.u.mscribed.

The dependence of irradiation on the brightness of a source relative to its surroundings is further confirmed by the appearance of stars at twilight, which appear much smaller then than at night, and of Venus when observed in broad daylight which appears "so small that it takes sharp eyesight to see it, though in the following night it appears like a great torch". This latter effect provides a rough way of testing for the predicted change in size of Venus which does not involve an appeal to telescopic evidence. The test can be made with the naked eye provided observations are restricted to daytime or twilight. According to Galileo, at least, the changes in size are "quite perceptible to the naked eye", although they can only be observed precisely with the telescope (Drake, 1957, p.131).

By fairly straightforward practical demonstration, then, Galileo was able to show that the naked eye yields inconsistent information when small light sources, bright compared with their surroundings, are viewed in the terrestrial and celestial domain. The phenomenon of irradiation, for which Galileo provided a range of evidence, as well as the more direct demonstration with the lamp, indicate that naked-eye observations of small, bright light sources are unreliable. One implication of this is that naked-eye observations of Venus in daylight are to be preferred to those made at night when Venus is bright compared with its surroundings. The former, unlike the latter, show that the apparent size of Venus varies during the course of the year. All this can be said without any reference to the telescope. When we now note that the telescope removes irradiation when used to observe planets and that, what is more, the variations in apparent size are compatible with the variations observable with the naked eye in daylight, a strong case for the telescopic data begins to emerge.

A final argument for the veracity of the telescopic data on the sizes of Venus and Mars is that they corresponded precisely with the predictions of all of the serious astronomical theories at the time. This conflicts with the way in which Feyerabend, and Galileo himself, presented the situation, implying, as they did, that the data offers support to the Copernican theory over its rivals. The rivals to the Copernican theory were those of Ptolemy and Tycho Brahe. Both of those theories predicted precisely the same variations in size as the Copernican theory did. Variations in distance from earth, leading to predicted changes in apparent size, arise in the Ptolemaic system because the planets move closer then further from the earth as they traverse the epicycles superimposed on the deferents, which later were equidistant from the earth. They occur in Tycho Brahe's system, in which planets other than earth orbit the sun while the sun itself orbits a stationary earth, for the same reason that they occur in the Copernican theory, since the two are geometrically equivalent. Derek J. de S. Price (1969) has shown quite >generally that this must be so once the systems are adjusted to fit the observed angular positions of the planets and the sun. That the apparent sizes of the planets had posed a problem for the major astronomical theories since antiquity is acknowledged by Osiander in his introduction to Copernicus's Revolutions of the Heavenly Spheres.

We have surveyed the way in which Galileo argued for acceptance of some significant telescopic findings, arguments that, I suggest, were compelling, a suggestion borne out by the historical fact that they convinced all of Galileo's serious rivals in a short s.p.a.ce of time. But in establishing his case, Galileo made the first step in what was to be a common trend in science, the replacement of naked-eye data by data acquired by way of instruments, and in doing so violated, and brought about a change in, "the criterion of science itself".

How does his accomplishment of this bear on the case for and against method?

Piecemeal change of theory, method and standards.

How is it that Galileo has managed to change standards by making a rational case in the face of arguments, such as John Worrall's, to the effect that this is impossible? He was able to do so because there was much that was shared between him and his rivals. There was a large overlap in what they aimed for. Among much else, they shared the aim of giving a description of the motions of the heavenly bodies that was borne out by the empirical evidence. After all, Ptolemy's Almagest is full of recordings of planetary positions, and Tycho Brahe is famed for his construction of ma.s.sive quadrants and the like which dramatically increased the accuracy of such recordings. There were low-level observations pointed out by Galileo that his opponents had no sensible option but to accept, such as the observation that a lamp appears larger than it really is from a distance at night, and that Venus looks smaller in the light of day than in the dark of night. Shared observations such as these, against the background of the shared aim, were sufficient for Galileo to be able to convince his opponents, using "clever techniques of persuasion" that involved nothing other than straightforward argument, that in one context at least they should be willing to abandon the "criterion of science itself" and accept some telescopic data rather than their naked-eye counterpart.

At any stage in its development, a science will consist of some specific aims to arrive at knowledge of some specified kind, methods for arriving at those aims together with standards for judging the extent to which they have been met, and specific facts and theories that represent the current state of play as far as the realisation of the aim is concerned. Each individual item in the web of ent.i.ties will be subject to revision in the light of research. We have already discussed ways in which theories and facts are fallible (remember that supercooled liquids refute the claim that liquids cannot flow uphill) and we ill.u.s.trated in the previous section a change in method and standards. The detailed form that the aim of a science takes can change too. Let me give an example.

The experimental work of Robert Boyle is rightly seen as a major contribution to the scientific revolution of the seventeenth century. Two somewhat conflicting aspects of Boyle's work can be discerned that, in a sense, represent the old and the new way of doing science. In his more philosophical writings Boyle advocated the "mechanical philosophy". According to that philosophy, the material world is seen as consisting of pieces of matter. It is taken as obvious that there is just this one kind of matter. Observable-sized objects are made up of arrangements of microscopic corpuscles of matter, and change is to be understood in terms of the rearrangement of corpuscles. The only properties corpuscles of matter have is the specific size, shape and motion that each one possesses, together with the property of impenetrability that serves to distinguish matter from empty s.p.a.ce. The motion of a corpuscle changes when it collides with another, and this mechanism is the source of all activity and change in nature. An explanation of some physical process will involve tracing that process back to the motions, collisions and rearrangements of the corpuscles involved. In giving expression to a version of this view, Boyle was subscribing to the new mechanical world view that was seen as the appropriate alternative to the Aristotelian one. In it, adequate explanations were ultimate explanations. They appealed to the shapes, sizes, motions and collisions of corpuscles, and these notions were themselves not considered to be in need of explanation. The aim of science, then, from this point of view, is ultimate explanations.

As well as advocating the mechanical philosophy, Boyle did experiments, notably his experiments in pneumatics and chemistry. As some of Boyle's own remarks imply, his experimental successes did not yield scientific knowledge of the kind demanded within the mechanical philosophy. Boyle's experiments on the physics of air, especially those with an air pump which enabled him to evacuate most of the air from a gla.s.s chamber, led him to explain a range of phenomena, such as the behaviour of barometers both inside and outside of evacuated chambers, in terms of the weight and elasticity of air. He was even able to suggest a version of the law connecting the pressure and volume of a fixed ma.s.s of gas that bears his name. But his explanations were not scientific explanations from the point of view of the mechanical philosophy because they were not ultimate. Appealing to weight and elasticity was not acceptable until those properties themselves had been explained in terms of corpuscular mechanisms. Needless to say, Boyle was unable to satisfy that demand. Eventually it became appreciated that Boyle's experimental science sought explanations that were both useful and attainable. By contrast, mechanical explanations in the strict sense came to be appreciated as unattainable. In effect, by the end of the seventeenth century the aim for ultimate explanations was given up in physics. That aim came to be seen as utopian, especially when contrasted with the achievements of experimental science.

The general idea, then, is that any part of the web of aims, methods, standards, theories and observational facts that const.i.tute a science at a particular time can be progressively changed, and the remaining part of the web will provide the background against which a case for the change can be made. However, it will certainly not be possible to make a reasoned case for changing everything in the web at once, for then there would be no ground on which to stand to make such a case. So if it were typical of science that rival scientists see everything differently from the point of view of their respective paradigms and live in different worlds to the extent that they share nothing, it would indeed be impossible to capture an objective sense in which science progresses. But there are no situations in science or its history or, for that matter, anywhere else that conform to this caricature. We do not need a universal, ahistorical account of scientific method to give an objective account of progress in science, and, furthermore, an objective account of how method can be changed for the better is possible.

A light-hearted interlude.

I can imagine how John Worrall, and like-minded opponents of relativism and defenders of universal method, would respond to the line I have taken above. They will say of my Galileo example, for instance, that, although it does ill.u.s.trate a change in standards, an appeal to some higher, more general standards is involved. Both Galileo and his rivals demanded that their account of planetary orbits should be borne out by appropriate evidence, for example. Once we have spelt out these general a.s.sumptions, my critics might well argue, then it is those general a.s.sumptions that const.i.tute universal method, and it is precisely those which form the backdrop against which the change brought about by Galileo is to be judged progressive. Without such a backdrop, I hear them say, you cannot argue that the change is progressive.

Let me make a concession. Suppose we do try to formulate some general principles that any proponent of science from Aristotle to Stephen Hawking might be expected to adhere to. Suppose the result is something like "take argument and the available evidence seriously and do not aim for a kind of knowledge or a level of confil 'nation that is beyond the reach of available methods". Let us call it the commonsense version of scientific method. I concede that there is a universal method in the common sense. But let me immediately attempt to remove any feeling of smugness John Worrall and his allies might be enjoying having won this concession from me. Let me first point out that, to the extent that commonsense universal method is correct and adequate, it puts them all, and myself, out of business, because it is hardly the kind of thing that it takes a professional philosopher to formulate, appreciate or defend. More seriously, I point out that once we do press the issue further, and demand that more detail be given, concerning what counts as evidence and confirmation, and precisely what kind of claims can be defended and how, then those details will vary from science to science and from historical context to historical context.

A formulation of commonsense method might not be sufficiently demanding a task to keep philosophers of science in business. However, I do suggest that an appreciation of it is sufficient to resist some contemporary trends in science studies. I have in mind those sociologists of science and postmodernists (let's call them "the levellers" for short) who downplay or deny the special status to be accorded scientific knowledge on the grounds that establishing its credentials necessarily involves the interests of scientists and groups of scientists, such things as financial or social status, professional interests and the like, in much the same way as any other social task does. In response to this I suggest there is a commonsense distinction between, say, the aim to improve knowledge of how chemicals combine and the aim to improve the social standing of professional chemists. I would even go so far as to suggest that if there are academic movements that fly in the face of this commonsense, then those in possession of such sense should demand that those movements be starved of funds. It is interesting to note that traditional philosophers of science have themselves contributed to the manufacturing of a situation that opens a s.p.a.ce for the levellers. It is they who have presumed that a distinction between science and other kinds of knowledge can only be achieved with the aid of some philosophically articulated account of universal method. Consequently, when those attempts fail, in a way that the preceding chapters of this book have shown them to have done, the way seems open for the levellers to move in. Michael Mulkay (1979), one of the most modest of levellers to be sure, provides just one of the many possible examples of an a.n.a.lyst of science who draws the conclusion that a sociological categorisation of science is made necessary by the failure of what he tee ins "the standard view".1 This brings us to the point at which the debate within philosophy of science stood about fifteen years ago. We cannot leave matters here, because during that period there have been two important movements that have developed since then and which warrant attention. One of these movements involves an attempt to develop an account of universal method by adapting a version of probability theory. We investigate it in the next chapter. The second movement has attempted to counter what it sees as the excesses of the theory-dominated accounts of science that have held sway for some time by taking a close look at experiment and what it involves. This approach is discussed in chapter 13.

Further reading.

My case against universal method is made in a little more detail in Science and Its Fabrication, (Chalmers, 1990, chapter 2), while "Galileo's Telescopic Observations of Venus and Mars" (Chalmers, 1985) and "The Galileo that Feyerabend Missed" (Chalmers, 1986) contain a critique and improvement of Feyerabend's Galileo case-study. Laudan (1977) and Laudan (1984) involve an attempt to find a middle way between universal method and anarchism that differs from mine. More details of the case I make with relation to Boyle's work can be found in "The Lack of Excellence of Boyle's Mechanical Philosophy" (Chalmers, 1993) and "Ultimate Explanation in Science" (Chalmers, 1995).

CHAPTER 12:.

The Bayesian approach.

Introduction.

Many of us had sufficient confidence in the prediction of the most recent return of Halley's comet that we booked weekends in the country, far from city lights and well in advance, in order to observe it. Our confidence proved not to be misplaced. Scientists have enough confidence in the reliability of their theories to send_manned s.p.a.cecraft into s.p.a.ce. When things went amiss in one of them, we were impressed, but perhaps not surprised, when the scientists, aided by computers, were able to rapidly calculate how the remaining rocket fuel could be utilised to fire the rocket motor in just the right way to put the craft into an orbit that would return it to earth. These stories suggest that perhaps the extent to which theories are fallible, stressed by the philosophers in our story so far, from Popper to Feyerabend, are misplaced or exaggerated. Can the Popperian claim that the probability of all scientific theories is zero be reconciled with them? It is worth stressing, in this connection, that the theory used by the scientists in both of my stories was Newtonian theory, a theory falsified in a number of ways at the beginning of this century according to the Popperian account (and most others). Surely something has gone seriously wrong.

One group of philosophers who do think that something has gone radically wrong, and whose attempts to put it right have become popular in the last couple of decades, are the Bayesians, so called because they base their views on a _theorem in probability theory proved by the eighteenthcenturymathematician Thomas Bayes. The Bayesians regard it as inappropriate to ascribe zero probability to a well-confirmed theory, and they seek some kind of inductive inference that will yield non-zero probabilities for them in a way that avoids the difficulties of the kind described in chapter 4. For example, they would like to be able to show how and why a high probability can be attributed to Newtonian theory when used to calculate the orbit of Halley's comet or a s.p.a.cecraft. An outline and critical appraisal of their viewpoint is given in this chapter.

Bayes' theorem.

Bayes' theorem is about conditional probabilities, probabilities for propositions that depend nn (and hence are conditional on) the evidence bearing _on _those propositions. For instance, the probabilities ascribed by a punter to each horse in a race will be conditional on the knowledge the punter has of the past form of each of the horses. What is more, those probabilities will be subject to change by the punter in the light of new evidence, when, for example, he finds on arrival at the racetrack that one of the horses is sweating badly and looking decidedly sick. Bayes' theorem is a theorem prescribing how probabilities are to be changed in the light of new evidence.

In the context of science the issue is how to ascribe probabilities to theories or hypotheses in the light of evidence. Let P(h/e) denote the probability of a hypothesis h in the light of evidence e, P(e/h) denote the probability to be ascribed to the evidence e on the a.s.sumption that the hypothesis h is correct, P(h) the probability ascribed to h in the absence of knowledge of e, and P(e) the probability ascribed to e in the absence of any a.s.sumption about the truth of h. Then Bayes' theorem can be written: P(h/e) = P(h).P(e/h)/P(e).

P(h) is referred to as the prior probability , since it is the probability ascribed to the hypothesis prior to consideration of the evidence, e, and P(h/e) is referred to as the posterior REthgbility, the probability after the evidence, e, is taken into account. So the formula tells us how to change the probability of a hypothesis to some new, revised probability in the light of some specified evidence.

The formula indicates that the prior probability, P(h), is to be changed by a scaling factor P(e/h)/P(e) in the light of evidence e. It can readily be seen how this is in keeping with common intuitions. The factor P(e/h) is a measure of how likely e is given h. It will take a maximum value of 1 if e follows from h and a minimum value of zero if the negation of e follows from h. (Probabilities always take values in between 1, representing certainty, and zero, representing impossibility.) The extent to which some evidence supports a hypothesis is proportional to the degree to which the hypothesis predicts the evidence, which seems reasonable enough. The term in the divisor of the scaling factor, P(e), is a measure of how likely the evidence is considered to be when the truth of the hypothesis, h, is not a.s.sumed. So, if some piece of evidence is considered extremely likely whether we a.s.sume a hypothesis or not, the hypothesis is not supported significantly when that evidence is confirmed, whereas if that evidence is considered very unlikely unless the hypothesis is a.s.sumed, then the hypothesis will be highly confirmed if the evidence is confirmed. For instance, if some new theory of gravitation were to predict that heavy objects fall to the ground, it would not be significantly confirmed by the observation of the fall of a stone, since the stone would be expected to fall anyway. On the other hand, if that new theory were to predict some small variation of gravity with temperature, then the theory would be highly confirmed by the discovery of that effect, since it would be considered most unlikely in the absence of the new theory.