Aug 31, 2015
Decisions, Decisions: A Choice Not an Echo
Editor’s note: This is the last installment of Mel Acheson’s epistemological treatise, originally meant to be a presentation at a Thunderbolts conference in 2009.
Leon Rosenfeld, a coworker of Neils Bohr, commented that a model is not “sufficiently defined without the knowledge of its domain of validity.”18 He had in mind the “mathematical idealizations” that “give an adequate representation of the physical phenomena.” But of course idealizations include more than mathematical representations: every representation idealizes, that is, makes (poiein) an idea. A domain of validity encompasses those idealizations, together with the facts that they determine, that can be verified in some way: that we can do something with.
A simplistic illustration of this concept of domain of validity may be taken from the field of architecture. Architects use a geocentric model of the Earth-Sun system to site a house on a lot so that, say, the morning sun will shine into the kitchen window. The geocentric model is true for the relevant facts: the viewpoint of interest is on the lot, the range of interest is confined to the lot, and the goal of the inquiry is the motion of the Sun in relation to the lot. A heliocentric model would be cumbersome and irrelevant: wrong viewpoint, wrong interest, wrong goal.
Rocket scientists use a heliocentric model to send a probe to another planet. In this case, the heliocentric model is true for the relevant facts: the viewpoint of interest is the solar system, the range of interest is confined to the solar system, and the goal of the inquiry is the motion of the probe in relation to the planets.
This is to insist that literally home-building is not rocket science. It’s to insist on the integrity and honor of each endeavor. Within the domains of each field, with respect to the facts of each matter, each theory is valid, each is true. To claim that the heliocentric model is “really true” is to assert one’s provincialism: For the little green men in the spaceship who stop to ask directions across the galaxy, the heliocentric model is as useless and “wrong” as the geocentric one.
A model is valid within that domain bounded by those facts which are meaningfully interpreted by the model. In other words, a model is true where it’s true, it works where it works. The tautology is useful: much of the effort in science is just this business of exploring where and how far a model is true and workable. Popper’s criterion of falsifiability is then understood as the effort to find domains of falsity, populated by anomalies and different conciliations of facts, where a new theory is required, leading, presumably, to progress (that is, to a conceptual revolution). Falsifiability is not a property of a theory that makes it acceptable and justifies acquiescing in verification; it’s an ability of the theorist to be curious about alternatives, an entrepreneurial attitude of fixing what’s not broken.
Verification is irrelevant outside the domain of the model. Stacking up instances of verification arouses confidence, a feeling that often grows into arrogance. However, using the subjective emotion of confidence to judge the objective truth of a model is another absurd consequence of confusing Weyl’s “pair of opposites.” Reliability would be a better criterion. It goes beyond mere self-congratulation over finding instances where one’s pet theory is true. “[R]eliability hinges on a dual process of confirmation and ruling out error…. [S]everity of tests…depends on how one has probed the possible alternative interpretations of results at all levels…. Fully resolving fact and error means probing for error in addition to demonstrative verification….”19 Reliability requires an active search for and testing of alternatives. A model must be verified (in the area where it will be used) in order to be useful. But mere verification gives no guarantee that another model might not be more useful, especially if it encompasses a wider area of data, ideas, and applications.
The acceptability of a model cannot be determined from criteria derived from the model. Criteria outside the model, outside the science, even outside the intellectual realm become pertinent. Goals are particularly important and not always explicit. The data of other fields, the needs of industries, or political pressures may direct the choice of model. The availability of resources and the financing for experiments influence the development of a model. The state of technology will determine what instruments are available. Cultural ideas and values will suggest what is worth thinking and doing. Because the models are composed of interrelated facts and theories—indeed, one may say that the models are those interrelationships—and depend on external criteria, more than one can be true.
In fact, truth becomes a matter of secondary concern. “Thus it is decisions which settle the fate of theories…. We choose the theory which best holds its own in competition with other theories; the one which, by natural selection, proves itself the fittest to survive.”20 Popper writes only of selection pressures from severe testing and rigorous testability. But “outside criteria” certainly add pressures and may dominate the competition.
Furthermore, the competition never ends: The acceptance of a model generates new instruments and new observations, new technologies and new goals. These produce new anomalies, new questions, and new values, which stimulate the development of still newer models. The process is a dynamic one of co-evolution in which environment and species exert selective pressures on each other.
A better analogy is that of a marketplace in which various models compete for market share.
Incommensurable models are apples and broccoli. It would be absurd to judge apples by the qualities of broccoli or to reject broccoli because it wasn’t red and round and shiny. A chef decides to buy one or the other in response to the requirements of the goal—his menu and the tastes of his customers. A model becomes “accepted” when a large number of scientists decides that it can best explain and help them do things with the questions and observations that most interest them and with the technology, values, and goals of their age.
This analogy also casts light on the situation in which the proponents of one model (the Big Bang comes to mind) can gain a monopolistic dominance: if they can control the warehouse and the truck line, they can force the store to offer only broccoli and the chef to take apples off his menu. People naturally tend to preserve their accomplishments, privileges, and influence. But when the proponents of one theory control the funding, publishing, and employment in the entire field, the uncritical expression of that tendency can exclude competition and force on everyone a product with deteriorating quality and increasing cost.
This evolutionary or market view prescinds from treating knowledge in relation to a “knowing subject.” “Objective knowledge consists of such things as linguistically formulated expectations, or theoretical systems…stored more importantly in books, articles, or programs than in minds….” It is concerned more with “the content, structure, function, growth, use, and development of the product known as knowledge.”21 The importance of a model is not its relationship (“approach”) to some hypothetical reality but how people choose to use it in relation to their other doings. “[O]nce created, it takes on a life of its own…. What is crucial about an item of objective knowledge…is its potential for being understood, or being utilized…in some way that has not yet been imagined….”22 Or, as has happened before, and again thinking of the Big Bang, its being discarded.
In this regard, the narrative that you’re reading claims no privilege. The arguments it makes are reflexive: It is as provisional as the narratives it criticizes. It provides particular insights and enables you to do particular things. It also hides other insights and disables you from doing other things. It has truth in it but it is not the truth. It won’t be useful for everyone; it may be replaced with an improved model—or with an altogether different one. Therefore, caveat emptor! Such reflexivity seems paradoxical, like the sentence: “This sentence is false.” But the reflexivity—and the paradox—ends with the decision to “buy” it or to pass it up.
Skeptical Fashion or Rock of Faith?
The emphasis here is on a social context. Although deciding is an individual act, survival or market share is a trait of populations. For science as a social activity, the deciding is an interpersonal affair in which a large number of people adopt the same model and the same goal. In an environment of personal freedom, this involves conversation and persuasion. In an authoritarian environment, it involves intimidation and coercion. People, being people, mix these behaviors without giving them much thought. Depending on incentives and mores, they’ll argue logically and compete fairly for resources but also slander each other and shut out dissenters.
Despite the ideal of a pure search for truth, much of the effort to understand the stars is a marketing rivalry in intellectual fashion. It involves social giving and taking and political pushing and shoving. Observational ability, methodological integrity, and theoretical acumen may not be as important as the editorship of journals, control of funding, and allocation of telescope time. Progress is not assured; the process may degenerate. Peer pressure, especially in a corrupt form of peer review that values ideological conformity over the changing fashions of methodological innovation, easily turns truth into consensus, followed by mob rule. Skepticism comes to be redefined as disbelieving dissent. Novel ideas are judged by orthodox criteria. The question, What else could it be?, is transformed into a terminal exclamation, What else could it be! This is the first step into fideism, which blows the chill wind of dogmatism, ending in the frozen repetitiousness of orthodoxy.23
This is why cognitive liberty, openness to conjectures, extra-disciplinary criticism, and even crackpots are of utmost importance to the functioning and progress of science. They preserve the possibilities. As disconcerting as revolutions can be, the stasis of orthodoxy is worse. Crackpots on the margins—and remember that proponents of incommensurable models will always appear to be crackpots to each other—must be at least tolerated and given some audience. To keep alive the possibility of writing new things in the book of knowledge, the marginalia must be preserved. One might even recommend a policy of loving one’s enemies, keeping them alive for the sake of perpetuating the competition. This is not a passive condition that merely happens; scientists must decide to preserve openness; they must actually do it.
Decision entails risk. Richard Rorty, in Contingency, irony, and solidarity, writes: “Interesting philosophy is rarely an examination of the pros and cons of a thesis. Usually it is, implicitly or explicitly, a contest between an entrenched vocabulary which has become a nuisance and a half-formed new vocabulary which vaguely promises great things.”24 Read “interesting science” in place of “interesting philosophy.” Established models have a track record of successful applications. New ones don’t. But the old ones also have had their failures, which are recorded in their many patches: the patches are what make them “nuisances.” The promise of a model without patches is alluring: but the promise may not play out; the great things may be impossible.
A second risk comes from the “finder’s fallacy.” A model that works, that is true for its domain, is hard to find. It’s the needle in the haystack. There can be more than one needle, but there is much more hay. The finder’s fallacy jumps to the conclusion that there is only one needle, and the finder stops looking after finding the first one. His risk is that he will end up with a bent and rusty needle, having missed a straight shiny one that was nearby. On the other hand, he could spend all his time looking for a better needle and never getting around to sewing what he has to sew.
Risk inflames the fear of uncertainty. The craving to justify knowledge, to found it on a secure bedrock of unquestionable principles, is understandable. If scientific theories could be formulated in a language based only on incorrigible observations, the inductively entailed general laws would be certain, indubitable, absolutely true. Weyl’s opposition would disappear, reality would become one, and knowledge could be had merely by taking a sufficiently careful look at the reality. Knowledge would be risk free.
But as we’ve seen, observations are corrigible, and induction entails creativity instead of indubitability. Sensation provides only sensation. “The claim that there is an external world in addition to the evidence is a claim going beyond the evidence. Hence, claims about such realms are unjustifiable.”25 They must be taken on faith—or chosen with risk. “This approach may produce in one who is unaccustomed to it an uncomfortable feeling of floating, of having no firm foundation. That would be appropriate: for it is floating; it is doing without a foundation.”26The metaphor of construction cannot be extended to the foundations of theories. Cornerstones are merely ideas—assumptions—from which theories ramify. The world of thought has no counterpart to gravity to impart cognitive weight that requires support. If theories are castles, they are neither founded nor “in the air” but “in orbit,” in free-fall.
A person chooses a theory in order to understand a particular selection of data. As with any entrepreneurial venture, it may not be as profitable as the person originally hoped. A lucid view of the venture requires courage to face the fear of its uncertainties. If lucidity and courage weaken, the temptation arises to seek reassurance by excluding other choices. Disabling choice activates the ersatz certainty of belief. But then one has converted one’s science into another religion—the religion of scientism.
In the late 1880s, John Tyndall issued a challenge in his famous Belfast address: “We claim and we shall wrest from religion the entire domain of cosmological theory.” As it turned out, he was only bickering with the church about who would preside over that department of religion. Cosmology became a part of science, but the religious attitude of belief, of presuming that one possesses a final thought, came with it. Cosmology is still largely a theological quest in which true belief is achieved by fervency and the enforcement of conformity, and the risks of choosing among the “makings” of human cognitive artistry are denied. Perhaps the denial and the relapse into the orthodoxy of belief are forgivable: to adopt scientific skepticism is to accept the mortality of fashion in theories and the mutability of truth in our understandings of observations. Few astronomers want to gaze on the demise of the theory they believe in.
Eschew Eschatology: Carpe Diem!
The new data of the space age have already falsified the Big Bang. Instead of deciding to reject the theory, astronomers have decided to patch, adjust, twist, fudge, rework, and reinvent it to explain away the recalcitrant data. But with the data out of the way, the cosmos that the theory describes is mostly unobservable. The theory has gone from a big bang of explanation to a whimper of intellectual bankruptcy: It’s a nuisance. Still, no one can move against the mob of consensus. Dissenters are trampled, as many have been: denied telescope time, refused publication, fired or not hired, slandered, blacklisted.
The way forward will likely be to establish new publications, new facilities, new departments. The old ones will decline in population and power; the new ones will compete, and one or perhaps more will grow, probably into another orthodoxy. The cycle will repeat: only the demise of the human species can put an end to the skirmishing that accompanies scientific model-making. But we are not concerned here with such eschatological matters. As Stephen Toulmin reminds us, “For progress can be made in science only if men apply their intellects critically to the problems which arise in their own times, in the light of the evidence and the ideas which are then open to consideration.”27
Today the challenge is deciding to replace the Big Bang. The nascent field of plasma cosmology is developing under the auspices of the Institute of Electrical and Electronics Engineers (IEEE). Other models may also find niches in which to thrive. At this time of beginnings, the more alternatives there are to choose among, the more reliable will be the one or few that will stock the shelves of tomorrow’s astronomy.
18 Cited in The End of Certainty by Ilya Prigogine, The Free Press, New York, 1997, p. 29.
19 Douglass Allchin, “The Epistemology of Error,” Minnesota Center for the Philosophy of Science. http://www.tc.umn.edu/~allch001/papers/epist’of.pdf , pp. 12-14.
20 Karl Popper, The Logic of Scientific Discovery, Hutchinson Education, 1959. Routledge reprint, 2000, p. 108.
21 W.W. Bartley III, “Alienation Alienated: The Economics of Knowledge versus the Psychology and Sociology of Knowledge,” in Evolutionary Epistemology, Rationality, and the Sociology of Knowledge, Radnitzky and Bartley eds., Open Court, 1987, p. 434.
22 Ibid, p. 435.
23 This is not an argument for “anything goes” novelty. Innovative heat must be moderated by dogmatic chilling if anything is to get done. The dynamic balance between the two is what the political jostling is about. The balance can’t be prescribed, but a thoughtful moderation should prevail over puritanical fidelity to either one.
24 Richard Rorty, Contingency, irony, and solidarity, Cambridge University Press, 1989, p. 9.
25 W.W. Bartley III, “Philosophy of Biology versus Philosophy of Physics,” op. cit., p. 12.
26 W.W. Bartley III, “Theories of Rationality,” op. cit., p. 213.
27 Toulmin, op. cit., p. 110.