1. Introduction.

In his posthumous book, Conquest of Abundance, Paul Feyerabend derides relativism and rails against “the champions of incommensurability.” This is not to say, however, that in the final years of his life he turned against the philosophical stance that made him a philosophical hero to some and a pariah to others. For Feyerabend's late thought arises out of the concerns he felt so deeply throughout his career. As for his style, in this last book of his, he is as provocative as he always was.

Feyerabend's starting point is that the world presents to our senses and our minds a greater abundance than we could handle, were it not for the fact that senses and minds are designed to simplify, as indeed we need them to. This need for simplification and abstraction has been carried further by Western science ever since it took root in Greece. All such simplifications are distortions of our experience, but that seems like a small price to pay when we consider how often they are useful. A problem arises, however, when scientists and philosophers take as “real” the impoverished environment that results from abstraction, to the detriment of many aspects of experience that enrich and make our lives worthwhile. Such a move by scientists and philosophers cannot be justified on either scientific or philosophical grounds. Relativism, however, is not a reasonable alternative to realism, for it, too, makes unwarranted assumptions about the rigidity of human culture. For similar reasons (the flexibility of human language), Feyerabend argues, incommensurability is a problem only for philosophers who insist on the rigidity of meanings. Moreover, what appear to be epistemological and scientific investigations surreptitiously make moral assumptions. Ethics should thus be a measure of scientific truth. Now, what is a reasonable alternative to realism? An approach similar to Bohr's: experiments, ideas, and models strongly interfere with nature, but they do not reveal how nature is independently of that interference—they only reveal how nature responds to that interference (239).

I will do my best to explain Feyerabend's positions, although I will not be able to do justice to all his clever arguments, nor to his many interesting descriptions of the practice and the history of art, science, and philosophy. I will argue, nonetheless, that this is a significant book, that Feyerabend is right about realism, but that he is wrong about relativism. I will also deal with the problem of reflexivity: Isn't Feyerabend's account of human knowledge an abstraction after all? And therefore, if his arguments against abstraction are taken seriously, does he not undermine his own account? A similar difficulty may be raised about my own (unavoidable) simplification of Feyerabend's rich views.

I should warn the reader that my task of interpretation is made all the more difficult because Conquest of Abundance is a fragmented book. Bert Terpstra, the editor, pieced it together from several incomplete drafts found amongst Feyerabend's papers after his death. As Feyerabend's widow, Grazia Borrini, points out, Mr. Terpstra worked at this task “with the attention and the selflessness of a restorer of an ancient vase.” The book is divided in two parts. The first consists of the introduction and four chapters of the unfinished manuscript. A fifth chapter intended by Feyerabend was never found—and was probably never written. The second part of the book consists of several essays on the manuscript's themes. I will generally concentrate on the unfinished manuscript and draw on the essays from the second part for support.

2. Abstraction: The Good, The Bad, And The Ugly.

“There are trees, dreams, sunrises,” Feyerabend tells us, as there are “thunderstorms, shadows, rivers… wars, flea bites, love affairs… lives… Gods… entire galaxies”(3). It is a blessing, not a drawback, he adds, that only a tiny fraction of this extraordinary abundance affects our minds. “A superconscious organism would not be superwise, it would be paralyzed” (3–4). To support this claim, Feyerabend recalls the case of one of A.R. Luria's patients, whose greater-than-normal awareness of nuances made it impossible for him to recognize human faces. To survive, he blocked off large numbers of details, trying to boil his perceptions down to the essential characteristics of faces. Without simplification we could not manage to deal with dangers and illnesses, nor take advantage of opportunities.

In this Feyerabend is correct: Perception involves, to a large extent, pragmatic simplification. In adopting an evolutionary perspective on conscious experiences, we come to realize, V.S. Johnston tells us, that “… they are… properties of the nervous system that have been selected because they amplify and discriminate between attributes of the world that are biologically important” (1999, 18). These discriminations, just as Feyerabend holds, are central to learning and reasoning. They are also unavoidably “distortions” (Johnston Reference Johnston1999, 18–19). For example, they take extremely small differences in wavelengths, between, say, green and red, and turn them into sharply different experiences of color. As it turns out, green covers the range of wavelengths reflected by chlorophyll molecules when illuminated by white light. In this particular case, perceptual exaggeration provides a happy solution to basic survival problems of “diurnal animals that depend on a chlorophyll-based food chain on a planet that possesses particular atmospheric conditions” (Johnston Reference Johnston1999, 15).

Similar plausible philosophical and scientific considerations can be adduced in favor of Feyerabend's remarks about the “considerable merit” of simplification and abstraction in everyday thinking and in science. But there is a rub. Abstractions, he explains, remove the particulars that distinguish one object from another. Experiments go further by removing “the links that tie every process to its surroundings—they create an artificial and somewhat impoverished environment and explore its peculiarities” (5). Blocking things off from the totality that surrounds us, he admits, has often been very useful. “Interestingly enough,” he says, however, “the remains are called ‘real,’ which means they are regarded as more important than the totality itself” (5). This is the problem that intellectuals create when they reify the results of an otherwise sensible process. The problem is made worse in Western philosophy and science when the totality is described as having two parts: “a hidden… real world and a concealing and disturbing veil around it” (5). In religious contexts, this dichotomy may be combined with (or give rise to) a second dichotomy between Good and Evil.

The most extreme example of the reification of abstraction seems to be that of Parmenides, who argued that only Being exists and that change is an illusion. Since the only possible change of Being is into non-Being, but non-Being does not exist, there can be no change. By analogous reasoning, Being can have no subdivisions, hence Being is continuous. Thus, says Feyerabend, for Parmenides “neither tradition nor experience provides reliable knowledge. This was… the clearest and most radical separation of the domains which later on were called ‘reality’ and ‘appearance’. It was also the first and most concise theory of knowledge” (61).

Feyerabend urges those who find Parmenides’ arguments “primitive and linguistically absurd” to consider “how many modern scientists repeat his general ideas without his rigor and coherence” (61). “Nineteenth-century point mechanics posited a ‘real’ world without colors, smells, etc.” (61–62). And in a relativistic world events are laid out in advance. Here the world, following H. Weyl,

And then Einstein:

“The trouble,” Feyerabend argues, “is that ‘convinced physicists’ are also ‘convinced’ empiricists. But how can experiments which occur in time and are therefore illusions tell us about a reality that is beyond all illusions?” (62). Moving time, and change in general, to the domain of consciousness does not help Weyl, in my opinion. Having the “gaze of my consciousness” crawl upward along “the lifeline of my body” assumes an untenable dualism and is inconsistent with the findings of neuro-psychology. This is not the place, however, to argue that certain theories of time depend on a bad philosophy of mind. Suffice it to say, that Weyl's view seems to make questionable philosophical and scientific assumptions, and thus appears to lack, as Feyerabend claims, Parmenides’ rigor and coherence.

To avoid misunderstandings, I would like to stress two points again. The first is that Feyerabend does not fault scholars because variety disappears when subjected to scholarly analysis: “Anyone who tries to make sense of a puzzling sequence of events… is forced to introduce ideas that are not in the events themselves, but put them in perspective” (12). There is no escape: “understanding a subject means transforming it, lifting it out of a natural habitat and inserting it into a model or a theory or a poetic account of it” (12). What Feyerabend objects to is the commitment to the results of this procedure of abstraction as “reality,” to the exclusion not only of other abstractions (cf. his case for pluralism, Feyerabend Reference Feyerabend1975), but of features of experience that may be important to us for many sorts of reasons. This “elevation into the world of ideas” that was now being called the “real world” ignored such features as irrelevant, unreal, or subjective details. Despite the considerable intellectual and practical merit of the abstract procedures, however, “the details did not therefore cease to exist, just as people don't cease to have a nose when being weighed” (14).

The second point is that Feyerabend's plea for abundance is not the first step towards the kind of relativism that many of his old critics might perhaps expect. There may be many different ways for abstraction to simplify and transform abundance, but, he says, “… one transformation may be better than another in the sense that it permits or even explains what for the other transformation remains an unsolvable puzzle” (12).

Let us return now to the reification of abstraction. Feyerabend considers the main arguments that can be given in favor of this move and he finds them wanting, as we will see in the next section. But he also points out that there are alternative notions of reality that do not lead us to categorize as illusory what most human beings may find eminently sensible. “Early Chinese thinkers [took] variety at face value,” and for Aristotle, “Natural is what happens always or almost always” (7). The Chinese concentration on the particular resulted in many valuable astronomical and historical reports, while the “Aristotelians emphasized the local character of regularities” (7). More importantly, Aristotle's notion—which is intermediate between the extremes of Parmenides’ “objective” reality (in the sense that it was “untouched by human idiosyncrasy”) and the subjective approach of commonsense philosophy—“depended on the nature, on the achievements, and especially on the interests of humans” (190). Aristotle, for example, mocks the Platonic version of the Good in itself: “… the physician does not try to find health in itself, but the health of human beings or perhaps even the health of the individual. For he heals the individual” (69, Aristotle, Ethica Nichomachea, 1096b33ff). A variety of similar considerations lead him to conclude that the abstract Platonic Good is not what we should look for. Feyerabend, in sympathy with Aristotle, suggests that we “regard as real those things which play an important role in the kind of life we prefer” (71). For Aristotle, change was a prominent feature of our lives that he was not prepared to give up. Thus he opposed Plato and found Parmenides absurd.

Before considering a multitude of notions of reality, however, let us see what Feyerabend has to say on the arguments that the philosophical tradition may offer to support the conclusion that the abstract is the “real.”

3. Abstraction As “Reality”: The Arguments.

The conclusion that the view of the world that results from abstraction is “real” could be established in three main ways: experience, reason, and excellence of results. I will take them up in reverse order. I will be able to consider only a small fraction of the many arguments Feyerabend marshals against each of these possible avenues of support.

4. Modern (But Naive) Realism.

Modern scientific realism, as a version of empiricism, seems to build a bridge across the schism created by Parmenides and Plato. Carefully collected experience presumably gives us evidence about which of our (abstract) ideas are best. Previous work by Feyerabend (Reference Feyerabend1975) showed that prominent scientists (Galileo, Einstein) did not act in accordance with empiricist method, and that they could not so act and still advance science. In Conquest Feyerabend provides ingenious and interesting twists to his arguments.

Feyerabend notices a parallel between art and science. Medieval paintings were presumed inferior to the more realistic paintings both of the classical Greek period and of the Renaissance, which presumably followed scientific principles to achieve perspective. The most “accurate” work in this regard was Brunelleschi's “stage set,” in which the viewer would look with one eye through a peephole at the mirror reflection of an image of the Baptisterium in Florence. Brunelleschi rigged his contraption so it would reproduce the visual experience someone would have if standing at a certain angle, distance and height from the Baptisterium itself. In spite of the success of this approach, Leonardo, Raphael and other painters preferred to avoid all that complicated artificiality and to paint so that viewing their works under “normal” conditions would achieve the effects they had in mind, even when “realism” was their goal. Feyerabend analyses the situation in the following manner. “There are established ways of viewing animals, people, mountains, houses—they constitute the reality the artist sets out to explore” (103). Likewise there are familiar ways of viewing artworks. “They determine what people experience when [confronting them]” (103). The task of the artist, then, is to “create a physical structure which, when… ‘projected’ in the customary manner, produces an aspect similar to one of the familiar aspects of the things represented” (103). This often goes beyond perspective, for such an aspect may well include fear, respect, or other emotions (the Holy Virgin may look encouraging or threatening, for example). Traditional ways of ‘projecting’ are “natural projections, the aspects they create natural aspects” (104). But even if the stage set involved is a traditional one, it is still a stage set. “The difference between Brunelleschi and tradition is that while Brunelleschi controlled the set, traditional artists are largely controlled by it” (104).

The suggestion is that the notion of “stage sets” is also useful for thinking about the relationship between science and “reality.” At the complex end of science we have, for example, the CERN experiment that led to the discovery of the W and Z particles in 1981: proton-antiproton collisions, detectors to catch the short-lived particles, computers to select candidate events and display them so as to compare them with predictions from electro-weak theory (specially adapted for the peculiarities of the experiment) and so on. According to Feyerabend, “Neither ‘nature’ nor its theoretical image were faced directly; they were both transformed by complex and sophisticated processes” (103). Here the stage was constructed with great effort and ingenuity, and the projections involved were clearly the result of long and difficult training. But we may find their equivalents even at the more Baconian end of science, in which scientists, finding coherence in their projections, “combine them into a world” (115). And then, disregarding the projecting mechanisms, “which by now have become second nature, they assert the objective existence of this world” (115).

The best cure for this naïve realist compulsion, both in art and science, is to have “alternative ways of creating order gain the upper hand” (113). In art we come to realize that medieval painting, for example, had goals other than imitation. Its purpose was rather to portray “power, authority, and permanence” instead of individuality. In portraits, artists aimed to convey the dignity of the office, not the personal characteristics of the office holder. The existence of mechanisms of projection becomes quite evident also when we consider our interpretations of, say, the sorts of drawings popular in anatomical and other scientific treatises from the 17^th Century on, as well in 20^th Century art. We all have seen the illustrations from Descartes’ L’homme. A human figure is drawn with one eye on top of another on the right side of the face, and with lines projecting out of the eyes to several points on a wall in front of the figure. There is nothing like that in the world, but we read it as representing how vision works (lines of sight, angles, nerve projections inside the head—to the pineal gland—and nerves that go out to muscles inside the arm, which without the appropriate projecting mechanism we might see as tattoos on the arm).

As far as science is concerned, the point is made by realizing that science is not one thing but many, in content, methods, and attitudes about experiments and experience. Scientists like Luria and Delbruck showed little patience for subjects “loaded with weak inferences,” such as the big problems of the universe or the early Earth. Einstein, Newton and Poincare exhibited the opposite inclinations. Feyerabend quotes Einstein's famous remark: “Is it not really strange that human beings are deaf to the strongest arguments while they are always inclined to overestimate measuring accuracies?” (149, Einstein Reference Einstein1971) Such overestimation, however, is “the rule in epidemiology, demography, genetics, spectroscopy, and other subjects” (149). Moreover, “The existence of different trends within a comprehensive venture is not surprising… it would be strange if large groups of passionate and imaginative people, who despise authority and make criticism a guide to research, subscribed to a single point of view” (151–152). In the 19^th Century, the astronomical view rested on action-at-a-distance laws; the atomic view was accepted in some areas of chemistry and rejected in others; the kinetic and mechanical view accommodated atomism well; the phenomenological view did not (152). The situation was not very different in the 20^th Century, in spite of some unifications achieved by statistical thermodynamics, molecular biology and quantum chemistry (154). We need not restrict ourselves to the physical sciences. Behaviorism pointed to successes that seemed inaccessible to other procedures. Like Brunelleschi, behaviorists set up a stage (experimental equipment, etc.), “project their subjects, call the projected aspects facts, and identify them with a stage-independent reality” (120). Nevertheless, “the very presence of a stage and of methods of projection throws doubt on the identification” (120). Think of the very different projections made by ethology, and of its even greater successes. And think also of later approaches, such as neuro-psychology, which makes its object of study what for the behaviorists was a black box. At one time behaviorists were confident of giving an “objective” account of human affairs, but alternative ways of “creating order” gained the upper hand (120). In summary: “… science offers not one story, it offers many; the stories clash and their relation to a story-independent “reality” is… problematic” (27).

5. Incommensurability And Relativism.

In Conquest, Feyerabend does not discuss his own version of incommensurability, i.e., that the use of the concepts of a theory or worldview precludes the use of the concepts of another theory or worldview; thus, presenting Einsteinian facts means suspending principles assumed in the construction of Newtonian facts (1975, 267–285). Nor does he discuss the original and sensible notion of incommensurability, soon after hijacked by the concerns of linguistic philosophy (Munévar Reference Munévar, Preston, Munévar and Lamb2000, 70–74). He concentrates instead on a popular account of Kuhn's version: “… some have asserted that the transition from a comprehensive physical theory to its historical successor involves an act of conversion and that the converted no longer understand the older faith” (267). Against this account, he recalls that the transition from classical physics to quantum physics was one of the most radical transformations in the history of science, and yet, “every stage of the transformation was discussed” (267). In the dramatic clash between Bohr and Einstein, he asks, did they talk past each other? “No. Einstein raised an objection; Bohr was mortified, thought intensely, found an answer, told Einstein, and Einstein accepted the answer. Einstein raised another objection; Bohr was again mortified… ” (267). When we look at the details, “we realize that the conversion philosophy simply does not make sense” (267).

Earlier in Conquest, Feyerabend had taken issue with “the champions of incommensurability” (37), for they assume that languages and worldviews “are closed in the sense that they admit, even constitute, some actions, thoughts, perceptions, while others are not merely excluded but rendered nonexistent” (31). This sounds like an attack on his own view, unless we keep in mind a qualification he has often stated:

Of course, “if discourse is defined as a sequence of clear and distinct propositions… which are constructed according to precise and merciless rules, [it will be] often interrupted by ‘irrational’ events and soon… replaced by a new discourse” (32–33). Furthermore, for this new discourse, “its predecessor is nonsense pure and simple” (33). If, on the other hand, worldviews and cultures “are allowed some leeway, so that they either keep their identity through very drastic changes… or change their content without violating [their] worldview, then we have no reason to assume that our ways of conveying meaning have any limits” (33, also Munévar 1981, 91–105). This is not to deny differences between languages or cultures, but to “ascribe them to accidents” of location or history, “not to clear, unambiguous… cultural essences: potentially every culture is all cultures” (33).

We have seen in Section 3 why Feyerabend thinks that meanings are indeed flexible. And it is from that flexibility and his thoughts on incommensurability that he draws the conclusion that relativism is unacceptable. When speaking of “stages,” “projections,” or “aspects,” Feyerabend has “made things more definite than they are” (114). Different traditions may have had their own ideas about the function of art, for example (and perhaps different stages, etc. too). We go too far however when we begin thinking in terms of “systems” (or, one might say, conceptual schemes) and “inferring a general relativity of artistic efforts” (114). This would be “as sensible as trying to define the shape of a body of water by the shape it assumes when frozen and inferring a radical difference between water, ice, and steam” (114). Relativism assumes that traditions are not only “equally truthful messengers of reality,” but that they are “well defined and clearly separated” (122–123). If this second assumption fails, then relativism will “cease to be acceptable” (123). This failure will also do in realism, but this matter has already been discussed. In any event, the discovery that every culture is potentially all cultures clashes with the assumption and brings it into question.

Now, some traditions may be quite rigid, but “most groups, societies, traditions not only interact, they are built for interaction.” Moreover, the pressures brought about by those interactions can transform their systems of projection in a fundamental way: “Not being well defined is part of their nature” (123), and thus the ambiguities they contain, “when becoming manifest, dissolve all relativistic judgments” (127–128). He states his new position strongly: “Relativizing existence to a single ‘conceptual system’ that is then closed off from the rest and presented in unambiguous detail mutilates real traditions and creates a chimera” (143).

I believe that Feyerabend's account on this matter is mistaken. It may work against some forms of relativism, but surely not against all. In the first place, and against the caricature of the last 2,300 years, a view counts as relativistic if it denies that there exist one single truth or “reality” (e.g., absolute truth). It need not claim that all points of view are equally valid—only that there may be more than one valid point of view. Second, relativism need not “relativize” with respect to a theory, a tradition, or a culture, although it may. A relativization with respect to the cognitive potential of a species’ genome may be sufficient to show that there is no single truth or “reality” (Munévar, Reference Munévar1981, Reference Munévar1998). Third, a biologically-based relativism will unavoidably be flexible and open to change (being guided, after all, by the conceptual machinery of evolutionary theory):

In this, I take culture in general, and science in particular, to be human phenotypic expressions. There are, incidentally, additional circumstances within which this evolutionary relativism accommodates change:

6. Ethics As A Measure Of Scientific Truth.

To many intellectuals, says Feyerabend, “the universality of a principle means that it corresponds to universal features of an observer- and history-independent world.” But, he argues, “such a correspondence is not obvious” (245). According to him, the evidence tells us only that we get quarks, leptons, and space-time frames when we approach the world, or Being, with certain concepts, instruments, and interpretations (which are often the “highly accidental outcome of complex, idiosyncratic, and rather opaque historical developments” (245–246)). What the evidence leaves open, he says, is whether “the response is the way in which Being reacted to the approach, or if it belongs to Being independently of any approach” (246).

The first option is consistent with quantum theory, which “suggests that the familiar spatiotemporal properties of elementary particles are not inherent in them but emerge as the result of special interactions” (246). Yet scientific realists prefer the second option. In this they reflect “a preference for forms of coherence that can be managed without much effort” (247). What agrees with their preference is called “real.” All else is illusion. According to Feyerabend, this notion of reality also contains a normative component: “behavior should center around what is real and must avoid being influenced by illusions” (247). But as we have seen earlier, we have a choice in our notion of reality (and abstraction as reality did not fare too well). We should thus not be misled by the fact that “some phenomena seem to form a coherent whole; if reality were required to produce coherent effects, then shy birds, people who are easily bored, and entities defined by statistical laws would be very unreal” (247). But this intellectual preference has affected our choice of “reality” already, although surreptitiously and without debate. Instead of easy manageability, why not emphasize human freedom? Different values will then lead to different ontologies. But these are ethical values, since ethics is the discipline that guides our choices between forms of life. This is not too far, at least in Feyerabend's reading, from Aristotle's notion that reality depends on what is important to the lives of human beings. In any event, it seems that “ethics, having once been a secret measure of scientific truth, can now become its overt judge” (247).

7. Two Models Of Knowledge.

“We need simplifications,” Feyerabend acknowledges: “we need bodies with restricted motions and brains with restricted modes of perception” (241). But, as we have seen, there are many simplifications, “and they can be changed to remove the elitism which so far has dominated Western civilization” (241). A very important moment in the development of that elitism, according to Feyerabend, was the rejection by Socrates of the Homeric worldview. In particular, Socrates would ask his fellow citizens to tell him what virtue, justice, or knowledge were. When they gave him a list of examples in which the word was appropriately used (e.g., the virtue of a man, the virtue of a woman, of a child, etc.), Socrates sarcastically replied that he had asked for one thing and his interlocutor had given him many. Socrates wanted a definition, a universal; they gave him particulars. Greeks, Feyerabend thinks, thought in terms of lists. Indeed, “the view that giving an account means enumerating instances, not subsuming them under a single term, retained its popularity right into the classical age of Greece” (38). So we have two competing models of knowledge: the list model and the abstraction model. Or, perhaps I should say, we had two models, since apparently the list model was pretty much run over by Western elitism.

I do not believe that it has died, though. In fact, I would wager that it is the way most human beings still think. And there is a good reason for it: That is how the human brain works. Feyerabend's comrade-in-arms, Kuhn, was the first philosopher of science to call our attention to this matter. He argued (1970) that, when scientists practice their trade, they do not apply rules but instead learn to see problems as being like other problems they encountered before, where “being like” is best explained by Wittgenstein's notion of “family resemblance.” Scientists are thus trained on a collection of particularly instructive examples (“exemplars”) that will enable them to develop a grasp of the way their discipline approaches its investigation of the world. The rules Kuhn would have us do without are the analogs of Socrates’ abstract definitions, and thus it was not surprising that his proposal, which expelled from science the sort of decision procedures dear to the hearts of philosophers of science, met with mumblings against Wittgenstenian obscurantism (cf., also Dreyfus Reference Dreyfus1972). But science has come to Kuhn's rescue.

In the last two decades, scientists and philosophers have collaborated in a program to explain the workings of the human brain (in part) as a complex of neural nets. As Paul Churchland explains, a neural net is designed to compute a large number of functions, even functions that we are unable to specify, “so long as we can supply a modestly large set of examples of the desired input/output pairs” (1991, 6). This process, by the way, is called “training up the network.” In artificial networks, the error in the output in the first run is calculated and fed back to the units in the network. This procedure will lead to a readjustment in synaptic weights in the network (this is the “back-propagation” algorithm). After repeating the procedure many times, the network will finally assume “a configuration of weights that does yield the appropriate outputs for all of the inputs in the training set” (1991, 7). We can, for instance, train a network to discriminate sonar echoes of explosive mines from those of submarine rocks, explains Churchland. After it has been trained, the network will be able to identify reliably echoes it has never heard before. It is important to realize that “neural nets typically have no representation of any rules, and they do not achieve their function-computing abilities by following any rules. They simply 'embody’ the desired function, as opposed to calculating it” (1991, 12).

This account sounds very much like Kuhn's explanation of how scientists typically operate. It certainly seems reasonable to consider it a serious model of the typical workings of human neural nets. Contrast it now with the failure of classical artificial intelligence (AI) to explain human thinking in terms of abstract rules. “Greek morality at the time of Plato,” Feyerabend says, “was a morality of instances and examples, not a morality ruled by abstract properties” (259). I would bet that the same could still be said of most fruitful human moralities (as opposed to ethical models invented by philosophers). Churchland has developed this very point in a highly provocative way (Churchland Reference Churchland1996, 143–150; Munévar Reference Munévar1998, 130–147).

Had Feyerabend lived longer, I suspect that he would have investigated Churchland's approach in great detail. He approved highly of Churchland's embedding of his philosophy into that approach, although I do not believe he had quite assimilated it (Feyerabend, Reference Feyerabend and Munévar1991, 517).

8. Concluding Remarks.

In looking back at my simplifications of Feyerabend's subtle and rich writing, I am struck by how many choices I had to make and how many important developments of his thought I have left out. For often what I presented as a point or a premise in an argument was put to multiple uses by Feyerabend. I could have gone in different directions, and others surely have and will. My choices have been dictated by what I took to be significant contributions to important issues in the philosophy of science. I hope I have been able to convey why the quality of those contributions makes Conquest an important book. There is much to ponder in it.

For all its richness, Feyerabend's account cannot escape being a simplification itself. But I do not think this is a vicious reflexivity. For, as we have seen, Feyerabend tell us two things relevant to this issue. The first is that many different accounts can and should be considered. The second is that not all of them will yield results. He does not want his account accepted while all the others are being excluded, quite the contrary. At the same time, he has tried to offer a compelling case for it. If he succeeds, his account ought to be one of those we take most seriously. And I do think that he has succeeded.