There are, Susan Haack says, two opposing schools of thought about science: the New Cynics, as she calls them, who believe science is a shabby social construct of the ruling class, and the Old Deferentialists, who believe science has a uniquely rational method of reaching the truth. Her plan is to steer a middle course between the two. She pursues this plan across a wide field of issues, a much broader spectrum than is found in traditional philosophy of science or sociology of science books—including, for example, expert scientific evidence in law as well as old chestnuts like religion and science. She writes throughout with the verve of her Manifesto of a Passionate Moderate (1998), and with attention to well-chosen and short examples of real science. She shares with us many juicy aperçus from her reading. One will not soon forget her recurrent use, as a diagnostic tool for some of the less sane pronouncements of the irrationalists, of J. L. Austin’s dictum, “There’s the part where he says it, and the part where he takes it back.”
Haack is convincing on many of the issues. On the sociology of science, she effortlessly exposes the conceptual confusions and internal inconsistencies of those modern enemies of science who think knowledge is “socially constructed” in the sense that it is a power play not subject to any norms of honest inquiry or respect for evidence. But she does not agree with the more hard-headed scientists who think sociology is a swamp from which nothing good can emerge. She shows well how the norms of the scientific community, like peer review, intensive training, and the preservation of scientific research in libraries, lead (fallibly, but by and large) to the weeding out of error and the real improvement of theories. The present controversy over refereeing standards, and whether fabulously wealthy scientific journals can expect a high standard of refereeing without paying scientists anything for it, concerns sociological questions, and it matters for the health and credibility of science.
Haack negotiates well the minefield of the differences or otherwise between the natural and the social sciences (well-trodden minefield might be the apt metaphor, except for its suggestion that the mines are no longer there). The requirements of honest and disinterested inquiry are the same in all kinds of science, and for that matter in common sense and in solving crossword puzzles. If some of the evidence in the social sciences concerns human intentions, that is a way in which they differ from the natural sciences, but there is no need to exaggerate the difference that makes. If the social sciences are poor at prediction, meteorology is nothing to write home about either. On law and science, Haack is well-informed and clear-headed on the recent attempts of American courts to deal with problems of expert scientific testimony, where, as Learned Hand put it, we set the jury “to decide, where doctors disagree.” (And we set a scientifically untrained judge to decide what the jury may safely hear.)
A rarely discussed but important issue to which Haack directs attention is the relation of scientific language and what C. S. Peirce called “real generals,” the causally important properties of things with which scientific terminology should align itself. Logicians once preferred the example “All swans are white.” At a certain point, difficulties arose from some black birds discovered by Dutch explorers on the way to the Indies and it was necessary to make a quick switch to “All ravens are black.” But perhaps with a better marketing department they could have kept to the original? It only requires a shift in terminology to dismiss the swimming ravens of New Holland as a new kind of bird of no significance for generalizations about swans. What is wrong with a ploy of that sort? Haack explains the need for scientific classifications to stay attentive to clusters of characteristics that are found to co-occur. Black swans are very like white swans in nearly all respects except color, which is a property found to be of generally low significance in biological classification. Calling them ravens is dishonest. As Haack writes of the ideal of objectivity in science, it is less a matter of technique than of character.
The eirenic or moderate plan of the book is less successful on the more basic issue of what makes science rational. Who are the “extreme” defenders of scientific rationality, and what is wrong with them? For several centuries, the more intransigent friends of science have had a consistent story to tell. It is in terms of probability, conceived as a species of logic. Just as “all ravens are black and this is a raven” makes it logically certain that this is black, so “99 percent of ravens are black and this is a raven” makes it logically highly probable that this is black (in the absence of further relevant evidence). That is why the results of drug trials give users rational confidence in the effects of drugs.
Galileo and Kepler used the language of objective probability about the way evidence supported their theories, and in the last hundred years a number of books have filled out the theory of logical probability—Keynes’s Treatise on Probability (the great work of his youth, before he went on to easier pickings in economics), D. C. Williams’s Ground of Induction, Polya’s Mathematics and Plausible Reasoning, and, just published in 2003, E. T. Jaynes’s posthumous masterpiece, Probability Theory: The Logic of Science. These works emphasize the continuity between scientific and com- monsense reasoning: the legal standard of “proof beyond reasonable doubt” that a jury of ordinary citizens must evaluate is of the same nature as the relation between a scientific theory and observation. That is, it is a purely logical relation between a body of evidence and a conclusion.
Philosophers of science and other commentators of a humanist bent have been very unwilling to notice this tradition. Probability theorists suspect that the reason for this is a simple one—that humanists hate numbers. To alleviate this difficulty, they explained that exact numbers were beside the point, and that “proof beyond reasonable doubt,” for example, is not a precise probability but needs only the imprecise quantity “very high,” which even jurors can understand. Polya explained that the main point of the admittedly somewhat complex Bayes theorem, which has as many as a dozen symbols, is contained in its simple corollary, “A theory is confirmed by its consequences.” (So, if a detective’s theory that the butler did it implies that the knife is behind the sofa, and the knife is found behind the sofa, then the theory is more likely than it was before.) To no avail: however qualitative and wordy the mush was made, humanists were not eating it up.
Haack is in this respect on the humanist side. Keynes, Williams, Polya, and Jaynes are not mentioned, and the two members of the school who are, Rudolf Carnap and Mary Hesse, are criticized for idiosyncrasies they do not share with the mainstream. Whenever she wishes to damn the view that the relation of evidence to conclusion is a matter of logic, Haack uses the phrase “syntactically characterizable logic.” The view that logic should all be syntactically characterizable, that is, done by manipulating uninterpreted symbols the way a computer does, is a narrow one. It was popular in the first half of the twentieth century but not otherwise, and it is no part of the probabilists’ theory of logic. It is especially surprising to see such a narrow view of logic here, given that the classic work on the wider field of logic and the difficulties of saying what exactly counts as logic and what not, is Haack’s own Deviant Logic, Fuzzy Logic (1974, second ed. 1996).
There is one issue on which Haack is anything but moderate. Religion brings out the extremist in her, and she is keen to revive the theory of Victorian books like Draper’s History of the Conflict between Science and Religion that science and religion are incompatible and that science is right. It is not easy to see where the incompatibility of science and religion lies, since on the face of it they are talking about different things. Nor does it become entirely clear what propositions exactly Haack thinks religion asserts and science denies. She does complain, with some reason, about the religious attitude to belief, on the grounds that faith is an immoral attribution of certainty to beliefs for which one knows one does not have full evidence. Nevertheless, it is hard to believe certainty is a central part of most believers’ faith—common experience, surely, is that most religious are less certain of their beliefs than most atheists are of theirs. Haack attacks adequately some soft creationist targets, but most believers are not creationists.
The heart of the conflict, it seems, concerns matters of value. Does science speak on this matter? According to Haack, it does. She quotes approvingly Stephen Hawking’s view, “We are such insignificant creatures on a minor planet of a very average star in the outer suburbs of one of a hundred thousand million galaxies. So it is difficult to believe in a God that would care about us.” There is a “deep tension,” she says “between a scientific picture of a vast and uncaring universe, and the religious idea of a caring and involved God.” One is surprised to see “insignificant” and “uncaring” listed as scientific properties, and wonders which science is devoted to their study. (Geometry, perhaps? Size, vast or minor, seems to be the only property mentioned as relevant to significance.)
If the universe is uncaring, religious believers are even worse, Haack thinks. She quotes indignantly the religious philosopher Richard Swinburne’s statement, “I am fortunate if the natural possibility of my suffering if you choose to hurt me is the vehicle that makes the choice really matter” and adds
the same day I read this I also read an article by a woman who had been raped, sodomized, terrorized, beaten, and left for dead by a gang of thugs. Can you accept that she was benefited by having been the vehicle that make her attackers’ choice really matter?—I can’t.
I can’t either, but I wonder if Haack is entitled to so much indignation, given her view of human significance. The view, namely, that humans are the same kind of things as galaxies, whose destruction is a mere firework. If science shows humans are too insignificant for a God to care about, why all the anger at God’s failure to prevent injury to them?
Haack leaves us with a useful picture of scientific inquiry as like a crossword puzzle. Some clues are as yet lightly pencilled in, a few conflict, some are firm and well-supported by a matrix of others. Solvers have guesses that often agree, but sometimes their differing bodies of prior knowledge pull them towards different hypotheses. It is a model that emphasizes the continuity between scientific and ordinary reasoning. The continuity was confirmed dramatically when researchers on infant cognition rushed to labor wards to film newborns, and on analysis of the tapes found the babies were studying them back—with, indeed, less background knowledge, but with much the same methodology. That methodology, including, for example, the confirmation of theory by consequence, is only intelligible if it is itself independent of the way the world is, that is, it is logic.