Duhem and Kuhn
An alternative framing of the relativist-anti-relativist debate often revolves around the notions of empirical equivalence and underdetermination. In its simplest gloss, according to the relativist, a rational decision regarding the competition between two theories is radically underdetermined because these theories are empirically equivalent or can be made so with enough effort and ingenuity in terms of auxiliary patching. Pierre Duhem's To Save the Phenomena and The Aim and Structure of Physical Theory have been massively influential in this regard. In Aim and Structure Duhem hammers home the possibility of auxiliary patching in blocking allegiance to a naive falsificationism. In Save the Phenomena he highlights the continuity, from Plato to the Renaissance, of the importance of empirical adequacy, destroying with massive historical detail the modernist myth of the "scientific night of the Middle Ages," that empirical inductive science began with Copernicus, Galileo, and Kepler. By showing the progressive continuity of the evolution of astronomy and physics from the ancient Greeks, through the Middle Ages, to the Renaissance, Duhem makes his famous plea that the aim of science should acknowledge the wisdom of Plato, Geminus, Ptolemy, Proclus, Posidonius, Simplicius, Maimonides, Aquinas, Bonaventura, John of Jandun, Lefevre d'Etaples, Osiander, and of course Bellarmine and Pope Urban VIII -- and all who "protested without letup against the realism of thinkers like Adrastus of Aphrodisias and Theon of Smyrna, the Arab physicists, the Italian Averroists and Ptolemaists, Copernicus and Rheticus themselves."(1) According to Duhem, the former were "logical and prudent" while the latter were guilty of an "impenitent realism" and succumbed to "folly and delusions."
The realism-anti-realism debate is much too involved to address in this thesis. Rather, my concern with Duhem will be to show:
1. By oversimplifying the debate between the supposed realists and anti-realists, Duhem misses the sustained concern for consistency in astronomical theory by those that he caricatures as naive realists. This point will be important to the development of claims I will be making in chapter 5 regarding the contextual force during the Renaissance of the emerging constraints of systematicity and parameter determination.
2. In stating the classic argument against realism, Duhem set the stage for post-modern relativists to misuse this argument by blanket application to all attempts to support the traditional epistemological project.
3. Although Duhem does not himself attempt to undermine the traditional epistemological project by applying the argument against realism to all cases of rational acceptance, he leaves us with only a vague and unarticulated "good sense" as a method for accepting one theory rather than another, and/or as a method for, in the language of this thesis, continuing to pursue and patch an adjudicatory trail or abandon it.
Although at the very end of Save the Phenomena, Duhem claims that what Copernicus, Kepler, and Galileo were actually doing was showing the importance of conceptual consistency and unity as pragmatic aims of scientific methodology, there is neither mention nor sympathetic attachment of these aims with the Arab thinkers when attacking their realism. Not until the very end of Duhem's historical story do we find, " . . . the truth which, little by little, they (the Renaissance realists) were introducing into science was that one form of dynamics, by means of a single set of mathematical formulae, must represent the movement of the stars, the oscillations of the sea, the fall of heavy bodies."(2) As such, we are left with a very disjointed portrayal of history. As Duhem acknowledged, given the epistemological stance he wished to defend, it is ironic that it "received its greatest setback precisely at the time when astronomy and physics were making new and rapid progress."(3) There is no explanation by Duhem as to why unity and consistency would suddenly become so important. Instead we find that from Plato to Kepler there were two competing views on the aim of science: the good guys who understood that empirical adequacy and "saving the appearances" should be the principal constraint on the "intellectual freedom of the astronomer," and the bad guys who, in advocating a natural philosophy, meddled with this freedom by forcing the astronomer to "reconcile . . . hypotheses with the teachings of that philosophy."(4) Thus, we are left with only irony and hand-waving false consciousness explanations like, well, the "greatest artists are not necessarily best at philosophizing about their art,"(5) as to why realists like Copernicus, Galileo, and Kepler would pop onto the scene and be so successful.
As Lloyd, Gardner, and Ragep(6) have shown, Duhem's bifurcated portrayal of history ignores the strong evidence that his anti-realist heroes made and were guided by many referential commitments. Lloyd, especially, shows in detail how Duhem's instrumentalists needed realist commitments to attack what Duhem considers naive realism. In other words, before one can decide what theoretical postulates are purely fictive calculation devices, one must have a realist theory as to what referential commitments are permissible given the nature of evidence.
For instance, in discussing Proclus, Duhem's star witness, Lloyd shows that Proclus' concern with withholding referential commitment to epicycles, eccentrics, and equants in the Ptolemaic system had to do with consistency with the realist commitment to the continuity of the spheres and the dynamical commitment to, in Proclus' words, the "common axiom of the physicists', namely that every simple movement is either round the centre of the universe or to or from that centre."(7) (Recall the problem of the location of Saturn's equant point discussed in item 5 of Ptolemaic problems in the previous chapter.) Also according to Lloyd, unlike Duhem, Proclus was aware of problems with both the realist and the instrumentalist positions, that although Proclus "argues against the realist way of taking these hypotheses on realist assumptions, he argues against the instrumentalist way of taking them also on realist assumptions."(8) In other words, the instrumentalist must be wary of unwittingly being a naive realist! He must be aware that not only are referential postulates necessary, but also that one must have good reasons for the referential postulates used in separating out what are purely calculation devices.
On this point, consider Ptolemy's acknowledged struggle in making his geocentric model consistent with the realist commitment that there should be no gaps between the spheres. In his Planetary Hypotheses, Ptolemy notes that after computing lunar and solar distances, and then nesting the epicycle radii of Mercury and Venus, such that the greatest distance of the moon will equal the least distance of Mercury and the greatest distance of Mercury will equal the least distance of Venus, "there is a discrepancy (gap) between (Venus and the sun) . . . which we cannot account for."(9) According to Ptolemy, an adjustment must be made by increasing the distance to the moon slightly, such that "this distance to the Sun will be somewhat diminished and it will then correspond to the greatest distance of Venus."(10) Thus, Ptolemy is forced to make the moon's orbit more eccentric causing a significant clash with observation (see item 2 of Ptolemaic problems, previous chapter).
As Lloyd further shows, Duhem's instrumentalist reading of Ptolemy cannot be right given the "straightforward realist account offered in the second book of the Planetary Hypotheses."(11) Nor does Duhem's instrumentalist reading of Geminus and Simplicius work either. All told we find numerous referential commitments made by these historical figures: that the earth is at rest in its natural place, that the movement of the heavenly bodies are circular and orderly, and the sphericity of the heavens as a whole, that there are no gaps between the celestial spheres, that the stars and spheres are made of a special kind of fire (Proclus), that the nature of the heavenly region is eternal, unchanging, homogenous, and transparent. Not to mention, of course, the obvious ampliatively derived realistic conclusions and agreement reached by most ancient astronomers on relative distances and sizes of celestial objects: that the sun is at a greater distance from the earth than the moon, that the planets are at a greater distance than the moon, that the sun is larger than the moon, and that the planets are not always the same distance from the earth.
The real historical continuity that Duhem misses, emphasized by Lloyd and Ragep, is a concern for and debate over consistency. According to Lloyd, in discussing Proclus for instance, the concern was for "a consistent physical account."(12) And according to Ragep,
"If Duhem had simply drawn the obvious conclusion from his own research that the major Greek philosophers and astronomers were committed in varying degrees to the proposition that the principles of mathematical astronomy must come from both mathematics and physics, he would have been more sensitive to the fact that the Arab thinkers . . . were not so much naive realists as scientists interested in reconciling the inconsistencies in astronomical theory that they had inherited from the Greeks. . . (that) Arab astronomers reached a rather simple conclusion -- the mathematical models had to be consistent with the physical principles."(13)
Ragep is speaking primarily of the tension that exists in Duhem's Le Systeme du monde,(14) where Duhem has become aware ("no doubt sadly," according to Ragep) of the authenticity of Ptolemy's Planetary Hypotheses. Even in To Save the Phenomena we can see Duhem struggling to sustain his purely instrumentalist interpretation of his heroes. Even here Duhem acknowledges an "apparent contradiction" within the positions of the holders of the instrumentalist aim, in that they were committed to numerous realistic-physical conceptions. In commenting on the apparent inconsistency in Peucer's geographical commitment to the earth as a globe and its real size compared to the tallest mountains, Duhem asks, ". . . how can he without risk of inconsistency, drop the requirement of truth when dealing with astronomical hypotheses?"(15) Duhem's response to this is to invoke the then accepted celestial-sublunary distinction, and to cite, with apparent time-indexed approval of the rationality of this move, the religious, metaphysical, and physical reasons for this distinction. It was permissible to be a realist about sublunary matters, but not celestial matters. But this appeal fails as well. Duhem's instrumentalists also made referential commitments regarding some features of the celestial realm. Francesco Giuntini, a developer of astronomical tables from Ptolemy's astronomy, argued that some combinations of eccentrics and epicycles must be considered more reasonable than homocentrics, because "the planets do not have constant apparent diameter, (so) . . . they are not always at the same distance from the earth. . . .(and) it must be admitted that some celestial revolutions do not have the earth as their center." According to Giuntini, "this is not a mere assumption."(16)
This illustrates the same point cited previously: to be an instrumentalist one must first make some sort of realistic commitment, given what one counts as reasons and evidence, to separate what items in one's theory will be counted as mere calculation devices.
Thus, the real issue, the continuity that Duhem missed, was conceptual consistency: How can Ptolemaic astronomy be understood as a complete system, a cosmology with a physical dynamics and a metaphysical view of the world. Viewed this way, we see that the marriage between Aristotelian physics and Ptolemaic astronomy was not an easy marriage, and that apparent instrumentalist stances were only a symptom of this tension. Accordingly, where Duhem sees imprudent metaphysical constraints being invoked by the Arabs inhibiting progress in astronomy, we should see instead the constraint of conceptual consistency operating. Many of the Arabs may have been wrong by the standards of our modern physical knowledge as to what dynamics astronomical models should be consistent with, but only an historical chauvinism should keep us from understanding the validity of their concern.
Less noted but equally influential is Duhem's technical argument against realism that occurs in To Save the Phenomena. Throughout his exegesis of the history of astronomy, Duhem pummels the realist with the simple logical point that referential commitment to the concepts of a particular hypothesis is justified if and only if one is able to prove "that no other set of hypotheses could possibly be conjured up that would do as well at saving the phenomena."(17) Such a demonstration can, of course, never be given. So realism is thus a quixotic quest. According to Duhem,
"Never will our understanding lay hold of truth in so exact a manner that it may not grasp it still more exactly, and it will do so indefinitely. . . . Reason is a possibility ever susceptible of new development, . . . . no matter how numerous and exact the confirmations by experience, they can never transform a hypothesis into certain truth, for this would require, in addition, demonstration of the proposition that these same experiential facts would flagrantly contradict any other hypotheses that might be conceived."(18)
And elsewhere, Duhem wrote,
"To prove that an astronomical hypothesis conforms to the nature of
things, it is necessary to prove not only that the hypothesis is sufficient
to save the phenomena, but also that these same phenomena could not be
saved if the hypothesis were abandoned or modified. . . .(and) If the hypotheses
of Copernicus succeed in saving all the known appearances, the conclusion
will be that these hypotheses may be true, not that they are certainly
true. To make the latter conclusion valid it would be necessary first to
prove that no other combination of hypotheses could be devised which permitted
the appearances to be saved equally well; and this demonstration has never
In other words, given any hypothesis, even if that hypothesis is far better supported by the evidence than an extant rival, there remain an infinite number of potential hypotheses that could save the evidential context equally well. It is not possible to separate the reasonable from the conceivable, if reasonable must mean establishing decisive conditions for truth. However, as Laudan has argued, truth and rational acceptance and/or pursuit between extant rivals need not covary, although historically many readers of Duhem have concluded that they must and haven fallen prey to relativism.(20) Although even well-supported criteria of ampliative inference may be powerless in terms of supporting referential commitment, because given any rationally selected theory we must admit that it is only one member of a larger class of possibly true theories, given many evidential contexts such criteria can potentially allow us to accept and separate the reasonable from extant alternatives. In other words, once one recognizes that traditional problems of confirmation and consequent extremist claims of empirical equivalence and underdetermination are parasitic upon the logical shortcomings of the traditional realist project, one can recognize that it is an unreasonable demand to place a burden on scientific methodology of separating the reasonable from the conceivable in terms of acceptance. According to Laudan, "The fact that a theory is deductively underdetermined (relative to certain evidence) does not warrant the claim that it is ampliatively underdetermined (relative to the same evidence)."(21)
Much skepticism and relativism are based upon failing to appreciate this point. A case can be made against blanket claims of empirical equivalence and strong underdetermination, and for rational acceptance, albeit not against potential empirical equivalence and in-principle underdetermination, and for truth. An epistemological wedge can be driven between transparently underdetermined and unwarranted truth claims founded upon bankrupt correspondence theories of truth and even, Laudan claims, nebulous notions of approximate truth,(22) and relativistic stances motivated by the failure of such theories and notions.
Thus, however strong Duhem's technical argument against realism may be, it should not automatically indict the traditional epistemological project of complete failure. Only if one views the matter through foundationalist spectacles need we despair over the purely logical truth that, in the language of this thesis, our hypertextual adjudicatory trails can be potentially patched "come what may," (Quine) or that we will never encounter a type of evidence that will allow us to be "decisive" (Feyerabend) or will leave us "speechless" (Rorty) at any node along an adjudicatory trail.
As noted in the previous chapter, Rheticus, Tycho, Kepler, and Galileo recognized that the Ptolemaic and Copernican systems were not empirically equivalent with respect to Mars. Even if they were de facto empirically equivalent with respect to naked-eye resolution, it does not follow that a future technological advance and/or change in strength and confidence in auxiliaries could not resolve the apparent empirical standoff. Relativists and the defenders of Legend that they criticize are often guilty of the same mistake, although they commit this mistake in an opposite fashion. Both fail to time-index the adjudicatory trails and the relative strength and weaknesses of particular auxiliary nodes at particular times. Defenders of Legend view history through the spectacles of well-supported modern auxiliaries and see the decisions of their historical heroes as virtually inevitable. Relativists note momentary standoffs between adjudicatory trails, the messy and unclear nature of formative networks, and ignore that there is constant debate involving good reasons for individual as well as group pursuit, with eventual ampliative consensus reached concerning the rationality of the particular auxiliary nodes.
To maintain that the reaching of an eventual consensus is not rational, relativists are committed to the belief that any future advance could be matched by a good auxiliary patch of the adjudicatory trail now out of favor. But even relativists are willing to admit that there are "bad" auxiliary patches. If there are always potential good patches distinguishable from bad patches, this implicitly commits us to a process that would be used to make this distinction. Once parallax was measured for Mars, no one argued that it could be accounted for by a swelling of the epicycle at the exact moment a telescope was in use. Once Galileo claimed to have observed a full phase of Venus, as telescopes became better calibrated and confidence in this new instrument increased, while confidence increased that the moon and planets were physical places that reflected sunlight, and other aspects of Aristotelian cosmology became increasingly vulnerable, the reemission theory fell out of favor and it became very difficult to save Ptolemaic astronomy with respect to Venus. Hence, the relativist's position unravels because one needs a methodology to distinguish between good and bad patches.
Duhem, contrary to many whom he influenced, recognized this. According to Duhem, if one scientist decides to continue to patch a theory in the light of experimental contradiction, and another to boldly abandon it, neither scientist has the right to accuse the other of "illogicality." However, "Pure logic is not the only rule for our judgements." We do not require foundationalist logical certainty to make rational decisions. According to Duhem, in the light of experimental recalcitrance there are times that we can recognize the foolish "haste" of abandoning a "vast and harmoniously constructed theory" rather than simply making technical adjustments and searching for auxiliary patches. On the other hand, there are times when it is clearly "unreasonable" for a scientist to continue to seek auxiliary patches for a theoretical structure that resembles "the worm-eaten columns of a building tottering in every part."(23)
But how does one know when it is "proper" to patch or reject a theoretical structure? At this point Duhem opens the door to a robust relativism. Because "experimental contradiction always bears as a whole on the entire group constituting a theory without any possibility of designating (deductively) which proposition in this group should be rejected," it is left to "our sagacity the burden of guessing" what to do based upon "'reasons which reason does not know' and which speak to the ample 'mind of finesse'. . . what is appropriately called good sense." Furthermore, it is often the nature of "genius" that allows the right move at the right time.(24) Other than prohibiting self-contradiction, logic only demands a unity based on "an intuition we are powerless to justify, but which it is impossible for us to be blind to," for "only on this condition will theory tend towards its ideal form, namely, that of a natural classification."(25)
Given what Duhem leaves us with, the indecision that an objective observer confronted in the first decade of the 17th-century regarding the choice between geostasis and heliostasis did not last forever, because "The day arrives when good sense comes out so clearly in favor of one of the two sides that the other gives up the struggle even though pure logic would not forbid its continuation." The most we can say in terms of articulating the conditions of a nonfoundationalist rationality is that the intuitive sagacity that allows us to reach such "fortunate reform" is enhanced by encouraging scientists to be "impartial and faithful judge(s)," for "nothing contributes more to entangle good sense and to disturb its insight than passions and interests."(26) In short, Duhem astutely diagnosed a problem without offering a robust ampliative solution, leaving the impression that there was none.
Given Duhem's great influence, it is not surprising that we can find modern commentators concluding that Copernicus, Galileo, and Kepler "sleepwalked" into a revolution (Koestler), or that it was their "leap-of-faith irrationality" (Feyerabend) that was the key to success, or that they made only a psychological "gestalt switch" to see how to "live in a different world" (Kuhn). For Duhem has left us only with a story about "great artists" who were not very good "at philosophizing about their art."
Thus, for a defense of the rationality of scientific change we seem
to be left with three positions. First, although science may appear illogical,
we know how to prepare the mind of scientists to enhance the likelihood
that they will make the right decisions via intuitions (Duhem and to a
large extent Kitcher). Second, we can see science as virtually an algorithmic
process of simply comparing theories with experience applying fixed rules
(Legend). Third, we can see science as a very complex process of the gradual
ampliative strengthening and weakening of hypertextual adjudicatory trails
of reasoning with the criteria used to judge strength and weakness themselves
gradually emerging and learned via connection to the networks. The first
two choices are easy targets of relativism. The third position is defended
in this thesis. It is my contention that from a nonfoundationalist perspective
it can be shown to be coherent -- it avoids circularity and infinite regress
-- and historically well-supported.
According to Kuhn, an historically grounded post-positivist philosophy of science shows that scientific assessment and change are a "mixture of objective and subjective factors."(27) Although it has become fashionable for both critics and supporters alike to emphasize the implications of the latter in Kuhn's work, let us begin by giving Kuhn his due by discussing what he at least sees as the objective and rational basis of the scientific enterprise. In this way we can not only better understand Kuhn's own "difficulties in understanding" both his critics and supporters and the "surprise and chagrin"(28) over how his The Structure of Scientific Revolutions (hereafter SSR) has been read, but also properly understand the historical context of Kuhn's work and consequent reaction, as well as the shortcomings of Kuhn's analysis from the point of view of my thesis.
Even viewed sympathetically there is an obvious tension in SSR between what Kuhn has to say about "gestalts," "the switch of gestalt," "faith," "conversion experiences," and "viewing different worlds," on the one hand, and the role of "anomalies," the recognition of "crisis," "progress," and "problem solving" on the other hand. It is worth noting that the infamous reference to a "switch of gestalt" does not occur in SSR until midway through the book, and it is introduced at first with caution, as unoriginal, and the insight of others in terms of a possible parallel with scientific change. Citing the work of Butterfield and Hanson, Kuhn says that although the "parallel (with paradigm shifts) can be misleading," it is nevertheless, "because it is today so familiar, . . . a useful elementary prototype."(29) On the very next page Kuhn discusses the frequency of scientists facing "admittedly fundamental anomal(ies)," and on the pages that follow Kuhn talks of the "suggestive" nature of the analogy with switches in visual gestalt and current psychological research, but cautions that psychological experiments and concepts cannot be more than suggestive. According to Kuhn, in referring to the work of Bruner and Postman (the card experiment), and others, psychological experiments
". . . do display characteristics of perception that could be central to scientific development, but they do not demonstrate that the careful and controlled observation exercised by the research scientist at all partakes of those characteristics."(30)
For Kuhn, it is clear that there is an objective world that will inevitably penetrate our conceptual webs, given "prolonged exposure"(31) to nature and the careful and controlled activity of scientists. In a nutshell, for Kuhn the locus of objectivity in the scientific enterprise is to be found in the following:
To experiment, to confront nature, to perceive anything meaningful at all, to overcome James's "bloomin' buzzin' confusion"(32) we must have a framework, a "conceptual web," a "disciplinary matrix," a "paradigm." Once a paradigm is in place, nature will be articulated in earnest via the overlay of this rational grid. In the language of this thesis, hypertextual adjudicatory trails will be created, probed, and followed; puzzles will be confronted and patched with new auxiliaries; features, relationships, aspects, and pictures of nature will be articulated that were always there in one sense. Like potential music contained in the electromagnetic waves broadcast from a radio station, newly discovered relationships are objective, independent features of nature, but they require a receiver before they become manifest to human observers.
For Kuhn we should distinguish between "that something is" and "what something is."(33) Successful paradigms are required to "pull out" connections and will do so given the backing of a scientific community. However, given enough time, because no paradigm is complete, because no paradigm will ever match reality in a one to one correspondence of concepts and details, it is inevitable that the world will overwhelm our rational grids. The very articulation of nature from a perspective will eventually reveal static and dissonance that cannot be assimilated. In Kuhn's language, "the very nature of normal research ensures that novelty shall not be suppressed for very long. . .. (for) normal science repeatedly goes astray," it is a "uniquely powerful technique for producing surprises," and it is a pursuit that although "not directed to novelties and tending at first to suppress them . . . nevertheless (is) . . . so effective in causing them to arise."(34) Unlike disciplines that lack a careful and controlled interface with nature, because normal science will "explore nature to an esoteric depth and detail otherwise unimaginable. . . . that exploration will ultimately isolate severe trouble spots."(35)
According to Kuhn, although it would be a mistake to consider that the features articulated by a paradigm represent truth in the traditional sense -- "there is . . . no theory-independent way to reconstruct phrases like 'really there'"(36) -- nevertheless "scientists (cannot) choose any theory they like so long as they agree in their choice and thereafter enforce it. . . . (because) nature cannot be forced into an arbitrary set of conceptual boxes."(37) An objective world resists our paradigm mapping, even though it is ambiguous to the point of allowing for different (but not arbitrary) successful maps and articulations at different times. This may be relativism, Kuhn says, but it is not "mere relativism" because it is "fundamentally evolutionary": scientific change is "unidirectional and irreversible. One scientific theory is not as good as another for doing what scientists normally do."(38) In other words, Kuhn claims his philosophy of science is no more mere relativism than natural selection is mere randomness.
It is here that we will begin our criticism. Kuhn's use of an analogy with natural selection theory to explain how he is an anti-realist but not a mere relativist is very revealing. To Kuhn's credit he remains consistent on this feature of his philosophy of science in his major writings.(39) However, it is in the defensive Postscript to the second edition of SSR that Kuhn's position most clearly unravels in his attempt to defend an objectivist position with the connection between scientific change and evolution. According to Kuhn,
". . . scientific development is, like biological, a unidirectional and irreversible process. Later scientific theories are better than earlier ones for solving puzzles in the often quite different environments to which they are applied. That is not a relativist position, and it displays the sense in which I am a convinced believer in scientific progress. . . . I do not doubt, for example, that Newton's mechanics improves on Aristotle's and that Einstein's improves on Newton's as instruments for puzzle-solving. But I can see in their succession no coherent direction of ontological development."(40)
But what does Kuhn mean by "quite different environments"? If a point to point correspondence with natural selection theory is maintained, this would mean that the natural world literally changes under our feet so to speak and we must adapt with a new paradigm, just as life forms must adapt to a changing local environment or become extinct. Read this way there is then "no coherent direction of ontological development," because there is not one world, one environment giving the same feedback. Instead, we must deal with "different worlds."
Although Kuhn tells us that an expanded theory of rationality must make some sense of this locution,(41) Kuhn is obviously uncomfortable with interpreting the reference to different worlds ontologically. He tells us that he wishes to explore the "possibility of avoiding this strange locution," asserts that scientists are not literally geographically transplanted into different environments because the world stays the same during revolutions, and acknowledges that the more traditional view that only interpretations of fixed, objective observations change is, although "somehow askew . . . . neither all wrong nor a mere mistake."(42) In these contexts Kuhn does not want to say that the natural world independent of human constructs changes as does the biological environment forcing species diversification, adaptation, and extinction. For this would be to make an ontological commitment about a theory-independent world -- a "really there" ontological pluralism. Kuhn does not want to say that at one time the objective world was really Aristotelian, but that the world changed such that Newtonianism solved more puzzles.
So what does Kuhn mean by different environments? In natural selection there is only local adaptation. Animals do not inevitably improve towards some ideal set of physical characteristics. Eyesight and flight in one environment are advantageous, in another disadvantageous, i.e., blind moles and fish that no longer need eyesight, and the ostrich and many other examples of birds that have evolved from ancestors that could fly. Even "complexity is only a broad trend, not a grand highway toward life's primary goal."(43) Likewise, according to Kuhn, in science we must account for both science's existence and success by substituting "evolution-from-what-we-do-know for evolution-toward-what-we-wish-to-know."(44) In natural selection theory it is appropriate to speak of local progress only; there is global progress only in the sense of more examples of local progress, more examples of adaptation and diversity. But no particular adaptation is better or optimal. Should we push the analogy with science this far? Evidently, the answer for Kuhn is yes.
Although Kuhn cautions us that the analogy can "be pushed too far," when it is combined with what Kuhn says about progress and problem solving we see that the analogy is "very nearly perfect."(45) Since we must not think of problem solving success in terms of scope and cumulativity -- historical counter examples of problem solving loss make this untenable -- we can think only of inevitable problem solving success in terms of "an increase in articulation and specialization."(46) Although Kuhn will go kicking and screaming, we have fallen back into Feyerabend's notion of progress. Over time perspectives proliferate and competition between them may force greater articulation within each, but descriptively and normatively, diversity is all we see.(47)
For Kuhn, "a sort of progress . . . inevitably characterizes the scientific enterprise."(48) Other than the total list of problems solved over time, we do not have better articulations and specializations in any cumulative sense. For "it is only during periods of normal science that progress seems both obvious and assured," and to some extent progress "lies in the eye of the beholder," because the victors in a scientific revolution tend to rewrite history and "are particularly blind to . . . losses as well as gains in scientific revolutions."(49) Thus, it would seem that we can say only that a modern theory is better for a modern environment analogously to the trivial fact that driving a car and knowing how to tie a tie are better in a big city environment of today than would be knowing different types of snow and hunting tactics, skills more appropriate for early European homo sapiens. Apparently for Kuhn, other than an expansion of the sheer number of specializations and the depth of articulations there is no growth aspect to problem-solving; we just solve different self-induced problems that may have a historical unidirectionality, but are also the result of massive historical contingency. Thus, for this reason the ability to solve problems cannot be "an unequivocal (argumentative) basis for paradigm choice."(50)
We are, of course, left with this conclusion because for Kuhn the acceptance or pursuit of a new scientific theory is never simply a matter of applying objective, a-historical rules. The application of the normally agreed upon but inherently vague standards of greater problem solving ability, accuracy, fruitfulness, and simplicity cannot be decisive or absolutely compelling. It is here that we reach the heart of the matter and the diagnosis of false dilemma in need of a post-foundationalist therapy. For Kuhn, because rule-based decisions on evidence can never be absolutely decisive or compelling, the "conversion experience that I have likened to a gestalt switch remains, therefore, at the heart of the revolutionary process. Good reasons . . . (only) provide . . . a climate in which (conversion) is more likely to occur."(51) In short, either we have an absolutely compelling process that "resembles logical or mathematical proof"(52) or we must resort to the suggestive perspective of psychology.
Thus, according to Kuhn,
"What is vague . . . about my position is the actual criteria . . . to be applied when deciding whether a particular failure in puzzle-solving is or is not to be attributed to fundamental theory and thus become an occasion for deep concern. That decision is, however, identical in kind with the decision whether or not the result of a particular test actually falsifies a particular theory."(53)
"There is no neutral algorithm for theory-choice, no systematic decision procedure which, properly applied, must lead each individual in the group to the same decision."(54)
"What I am denying then is neither the existence of good reasons nor
that these reasons are of the sort usually described. I am, however, insisting
that such reasons constitute values to be used in making choices rather
than rules of choice."(55)
Seen in its own historical context, after decades of an algorithmic progress myth of logical positivism and empiricism (Kitcher's Legend), Kuhn's analysis of scientific change in SSR as an historically contingent, messy process was liberating. He viewed his work as the preliminary stage of an expanded understanding of the rationality of science. From hindsight, we can even find Kuhn warning present day sociologists not to make too much of psychological and sociological variables:
"To understand why science develops as it does, one need not unravel the details of biography and personality that lead each individual to a particular choice, though that topic has vast fascinations."(56)
However, for Kuhn acceptance or pursuit of a new perspective is not the result primarily of evidence but of "comfort," a sense of global satisfaction that connects and gives meaning to all the parts of a hypertextual adjudicatory trail, to use the language of this thesis. Whether one talks of the decision to accept or pursue a new theory, to pursue an auxiliary patch for a theory in the light of anomalies, or that a theory has indeed been falsified by evidence, these decisions can never be "straightforward . . . or comfortable."(57) Decades removed from the intimidation of logical positivism and the backlash it produced in Kuhn and others, we can now see the mistake in giving this much prominence to psychology: As fallibilists, scientists should never be completely comfortable with any ampliative justification -- all adjudicatory trails are nonterminating -- and this lack of complete comfort need not be made so much of as to engender a slide to socio-cognitive relativism. My argument is that even if we accept the problematic notion of a paradigm and its importance in defining a "horizon" of perception, paradigmatic hegemony does not entirely obscure the rational scaffolding of debate between rival paradigms. In fact, it is the very messiness that Kuhn wishes to draw our attention to that insures that holistic hegemony and incommensurability are never complete.(58)
For Kuhn the influence of a paradigm is so overpowering that it redefines what science is in terms of not only substance but methods, problems, and standards.(59) An individual scientist must grasp the whole before he or she can understand the parts. Any debate within the scientific community over the parts is indecisive and a result of miscommunication until there is a switch of gestalt and an understanding of the whole. Scientists will "inevitably talk through each other," because "in learning a paradigm the scientist acquires theory, methods, and standards together, usually in an inextricable mixture."(60) For instance, in the case of the Copernican revolution Galileo and Kepler compiled "impressive evidence for the earth's status as a moving planet," but a switch to Copernicanism was incomplete until there were adjustments to the whole field of thought.(61) In the case of the Copernican appeal to harmony, considering the entirety of the evidence drawn from it, it "is nothing if not impressive." But without adjustments to the whole field of thought such an appeal "may well be nothing" in the sense that it was merely an aesthetic, subjective consideration that appealed to an "irrational subgroup" of neoplatonic mathematicians.(62)
My argument, brought to closure in Chapter 5, is that even if the rules are different, scientists are capable of recognizing this difference and entering into debate over the merits of various constraints on choice. Furthermore, I will argue that what is typically referred to as harmony and aesthetics by commentators of the Copernican episode can be couched in terms of parameter determination and seen as an emerging, albeit debatable, constraint. Put simply and preliminarily, Copernicans were arguing that parameter determination ought to be a constraint on theory choice and that it was an important node on an adjudicatory trail that supported heliostasis. Initially, non-Copernicans can be seen as either arguing that it was not important because of their non-realist goals for astronomy or that it was important but did not necessarily support heliostasis. Eventually, a substantial number of influential non-Copernicans were persuaded that parameter determination was an important constraint on theory choice, so important in fact that it drove the construction and pursuit of geostatic transforms (geoheliocentric systems) of Copernican planetary models.(63) Kuhn, because of his holism and his write-off of parameter determination as subjective, sees early heliostatic pursuit and commitment to its future promise as "only . . . made on faith" due to "only personal and inarticulate aesthetic considerations."(64) My argument will be that although parameter determination as a constraint was not decisive, it was and ought to have been an independent feature relevant to theory choice, one capable of being judged itself by ampliative evidence.
Scientists during the Copernican episode were capable of understanding and debating the merits of a particular node on an adjudicatory trail independent of their theoretical allegiances. As we will see, the nature of this debate was not simply a matter of "talking through each other," and it had a reasonable chance of ampliative closure. Finally, I will argue that too much has been made of the fact that parameter determination as a constraint emerged out of particular cultural biases. That parameter determination as a constraint emerged out of a cultural environment of mystical pythagoreanism and neoplatonism need not automatically indict this constraint as subjective and merely aesthetic. Percepts may be blind without some conceptual orientation, but they are quite capable of having evidential content even if the original conceptual orientation is later abandoned in favor of a new one. It is not unusual in the history of science to see that an important feature of modern methodology and/or theory had a suspect origin by modern standards. The concept of natural law evolved out of the ancient Greek notion of cosmos and the "fate" to which both humankind and the gods were subject. Linnaeus believed that he was establishing a system of organization that revealed God's created order. But by designating species as basic units, providing principles for their uniform definition and naming, and arranging them in a wider taxonomic system, he severed the natural order from interpretation in terms of human preference and convenience -- a crucial, preparatory conceptual shift for Darwin's development of natural selection theory.(65)
Similarly, Kepler was surely biased in his pursuit of heliostasis. Given all the confusing choices that lay before him in the late 16th-century, it would have been impossible for him to objectively pursue each one, and his mystical neoplatonism (the sun as the "material domicile of God") was a major factor in his pursuit of heliostasis rather than a geostatic transform of Copernican models. However, he would not have pursued just any theory regardless of its evidential features. Heliostasis had to work. It also had to have been selected by Kepler because of certain constraints on choice that Kepler felt were important -- and later one did not have to be a neoplatonist to appreciate the accuracy of elliptical orbits and the Rudolphine Tables As Kuhn even admits,
"Ultimately his version of Copernicus' proposal would almost certainly have converted all astronomers to Copernicanism, particularly after 1627 when Kepler issued the Rudolphine Tables, derived from his new theory and clearly superior to all the astronomical tables in use before."(66)
The insight that relativists invariably miss is that being biased allows one to notice certain evidential features of the world, and our biases do not always overwhelm these features.
Notes for Chapter 3
1. Duhem, 1969, p. 67.
2. Duhem, 1969, pp. 116-117.
3. Ibid., p. 61.
4. Ibid., p. 28.
5. Ibid., p. 61.
6. Lloyd, 1978; Gardner, 1983; Ragep, 1990.
7. Lloyd, 1978, note 39, pp. 208-209.
8. Ibid., p. 211.
9. Goldstein, 1967, p. 7.
11. Lloyd, 1978, p. 217.
12. Ibid., note 52, p. 211.
13. Ragep, 1990, pp. 209-210.
14. 1985, Ariew translation.
15. Duhem, 1969, p. 77.
16. Ibid., p. 85, emphasis added.
17. Ibid., p. 110. This point allows Duhem as a Catholic to argue for the "wisdom of Bellarmine" and the excellent logic of the Pope against the "impenitent realism" of Galileo (pp. 107-113)
18. Ibid., pp. 57-58, p. 111. Emphasis added.
19. From "Essai sur la notion de theorie physique," quoted from Rosen, 1971, p. 33, note 111.
20. Laudan, 1984, 1990b, 1991, and forthcoming. Laudan's 1984 shows how a reticulation model (evidence, methodology, and cognitive values) of scientific rationality can be used to defend rational acceptance and simultaneously reveal the shortcomings of realism. His 1990b attacks Quine and is probably one of the best sources for showing that the mere conceivability of auxiliary patches and future hypotheses need not impair decisions of acceptance between extant alternatives. His 1991 and forthcoming further articulate a comparativist methodology.
As near as I can ascertain, the historian Edward Rosen was the first (1939) to protest against the questionable dilemma invoked by Duhem. In his introduction to Copernicus's Commentariolus, he notes in passing that
". . . it should be indicated that Duhem's view is not without alternative. We are not limited to the choice offered by Duhem between realism and fictionalism: any proposition or hypothesis is either the Ultimate Truth or mere fiction. We may properly accept a hypothesis as the best statement at the moment and be ready to revise or to reject it when fresh empirical data require a modification of it, or a rival and superior hypothesis emerges to replace it." (Rosen, 1971, p. 33; 1st edition 1939.)
21. Laudan, 1990b, p. 291. Although there is considerable difference in Laudan's position and that of Kuhn's on the viability of the traditional epistemological project and our ability to delineate well-supported ampliative criteria that offer objective constraints on theory choice, it is worth noting that in the Structure of Scientific Revolutions Kuhn makes this same point concerning the problems produced by realism. According to Kuhn,
"If we can learn to substitute evolution-from-what-we-do-know for evolution-toward-what-we-wish-to-know, a number of vexing problems may vanish in the process. Somewhere in this maze, for example, must lie the problem of induction." (1962, p. 170)
22. According to Laudan,
"Virtually all the proponents of epistemic realism take it as unproblematic that if a theory is approximately true, it deductively follows that the theory is a relatively successful predictor and explainer of observable phenomena. Unfortunately, few of the writers of whom I am aware have defined what it means for a statement or theory to be "approximately true". . . . This reservation is more than perfunctory. Indeed, on the best-known account of what it means for a theory to be approximately true, it does not follow that an approximately true theory will be explanatorily successful." (1984, p. 118.)
23. Duhem, 1954, p. 217.
24. Ibid., pp. 211, 216-217.
25. Ibid., p. 220.
26. Ibid., p. 218.
27. Kuhn, 1977, p. 325. In this context, although Kuhn is referring to action on the part of "individual" scientists, it is clear that the philosophy of science he develops commits him to the subjective factors being present at the social level as well. "If subjective factors are required to account for the decisions that initially divide a profession, they may still be present later when the profession agrees." (p. 329) Kuhn's problem, as shown below, is that his philosophy of science leaves us (and Kuhn himself) in a quandary as to exactly how much power the subjective factors have. He does not deny that epistemic factors are present and move the scientific community to some extent. But he leaves the possibility open -- and his unintended socio-psychological relativists supporters clearly read him this way -- that the subjective factors always overwhelm the objective factors. My thesis does not deny that subjective factors are always present. However, it not only sees this as often a positive development in terms of supporting the discovery of objective factors -- subjective factors often help us focus on certain objective factors -- it denies also that subjective factors always overwhelm the objective factors. I believe Lakatos had the best handle on this; see the development below in Chapter 4.
28. 1970, p. 263 and note 3.
29. Kuhn, 1962, p. 85. Emphasis added.
30. Ibid., pp. 110, 112. Kuhn's emphasis.
31. Ibid., p. 111.
32. Ibid., p. 112.
33. 1962, p. 55.
34. Ibid., pp. 5-6, 52, 64.
35. 1970, p. 247.
36. 1970, Postscript, p. 206. To complete the above analogy with electromagnetic waves and music: we will never be in touch directly with the electromagnetic waves, but only with the music. But the latter is to some extent at least a human artifact; it is not "really there" independent of the receivers we use to articulate its presence.
37. 1970, p. 263.
38. Ibid., p. 264; 1970, Postscript, pp. 205-206.
39. The analogy appears first in 1962, pp. 145, 170-172. Also see 1970, p. 264, and 1970, Postscript, pp. 205-206.
40. 1970, p. 206, emphasis added.
41. 1962, p. 120. According to Kuhn, "we must learn to make sense of statements that at least resemble . . . though the world does not change with a change in paradigm, the scientist afterward works in a different world."
42. 1962, pp. 117, 110, 120. According to Kuhn, "In the absence of a developed alternative, I find it impossible to relinquish entirely that (traditional) viewpoint." (p. 125) And, "Whatever he may then see, the scientist after a revolution is still looking at the same world." (p. 128)
43. Gould, 1993b, p. 18.
44. 1962, p. 170.
45. Ibid., p. 171.
47. See Introduction, p.2 and note 4; Chapter 1, p. 39.
48. 1962, p. 169, emphasis added.
49. Ibid., pp. 162, 166.
50. Ibid., p. 168.
51. 1970, p. 204.
52. Ibid., p. 199.
53. 1970, p. 248. Emphasis added.
54. 1970, Postscript, p. 200.
55. 1970, p. 262
56. 1970, Postscript, p. 200.
57. In other words, one cannot even understand the rational arguments -- accept T because E, accept E because M, accept M because of another trail of T's and E's, etc. -- until one grasps something about the whole. See Kuhn's discussion of disagreeing scientists as members of different language communities, of the need for persuasion, translation, and conversion, and how this process is never "straightforward, or comfortable." 1970, Postscript, pp. 201-203.
58. See Figure I-1 again, especially the last paragraph.
59. 1962, p. 102.
60. Ibid., p. 108.
61. 1957, p. 229.
62. Ibid., p. 180; 1962, p. 155.
63. Furthermore, it achieved sufficient independence to be used to criticize heliostasis! Or at least versions of heliostasis. One can see Kepler pursuing a better version of heliostasis, not only because of predictive inaccuracies generated by Copernicus's epicyclet system, but also because of the complex number of circles used by Copernicus and the concomitant parameter flexibility this entailed. Thus, contrary to Kuhn's view on the hegemonic power of a paradigm, a constraint, rule, or methodology that originates within a paradigm, or has a certain attachment originally to a theory, may achieve a sufficient amount of independence to later be used to criticize that very theory.
64. 1962, p. 157.
65. Gould, 1993b. According to Gould,
"Linnaeus's definition fractured the conceit of a human-centered system with basic entities defined in terms of our needs and uses. Linnaeus proclaimed that species are the natural entities that God placed on earth at the Creation. They are his, not ours -- and they exist as they are, independent of our whims. . . . (Thereafter) species are real whether created by God or constructed by natural selection -- and Darwin's conceptual shift . . . required little revision in Linnaean methods." pp. 18-19, emphasis added.
66. 1957, p. 219. Note what a strange statement this is for Kuhn. Conversion based upon predictive accuracy alone? Without a complete conceptual shift to a new whole which incorporated physics, cosmology, and religion? Throughout his work he never quite resolves the question of the exact mix of subjective and objective features in scientific change and assessment.