I should like to know [said Mr Chichely] how a coroner is to judge of evidence if he has not had a legal train-ing?” “In my opinion,” said Lydgate, “legal training only makes a man more incompetent in questions that re-quire knowledge of another kind... No man can judge what is good evidence on any particular subject unless he knows that subject well. People talk about evidence as if it could be weighed in scales by a blind Justice. [But a] lawyer is no better than an old woman at a post-mor-tem examination... You might as well say that scanning verse will teach you to scan the potato crops.¹

Mulling over the invitation² to write 6,000 or so words on “different legal and scientific modes of thinking,” I felt like a panicked exami-nation candidate who’s just realized that the question he has twenty minutes to answer would take at least a Ph.D. dissertation to handle adequately. There are so many issues entangled here that I hardly knew which loose end to tug first.

Eventually, I decided to begin with two simple observations: that the work of a scientist is very different from the work of an attorney or a judge; and that, when scientists give expert testimony or advise a court, a regulative body, etc., communication can be difficult and very imperfect. Lawyers, judges, and juries don’t always understand scientific evidence very well; and scientists don’t always convey clearly to lawyers, judges, and jurors what’s solidly-established and uncontroversial, what’s very likely true but still contested by a few, what’s still conjectural, and what’s as yet beyond the reach even of plausible scientific conjecture. In consequence, the law sometimes asks more of science than science can give, and sometimes gets less from science than the best that science could give.

Is this simply because the kind of person who becomes an at-torney or a judge will likely have a very different temperament and very different talents from the kind of person who becomes an astrophysicist or an anthropologist, and will certainly have had very different education and training? No: while true enough, this is hardly explanation enough. We need to understand why the two enterprises require such different temperaments and talents and such different kinds of specialized apprenticeship. Well, might those supposed “different legal and scientific modes of thinking” provide an explanation? No; this still doesn’t go deep enough. True, no one would mistake a scientific article for a legal brief or a judicial ruling; but there’s a deeper reason for this difference: the different goals of science and the law.

The aspiration of science, I take it, is to find out how the world is; the aspiration of a legal system, to decide what to do, in these or those circumstances, to make our human world, in some mea-sure, more just and more livable. If we want to understand why the law experiences such difficulties in handling scientific evidence, we need to go beyond those supposed “different modes of thinking” to consider these very different purposes, the very different constraints under which scientists and legal players pursue those purposes, and the very different cultures of the two enterprises.

The core business of the sciences is inquiry, investigation, finding explanations of natural, or in the case of the social sciences, social phenomena, events, etc. To be sure, scientists do many other things as well, and sometimes instead: attend seminars on “grant-writing,” apply for grants, read others’ grant applications, design apparatus, computer programs, etc., write up their results, referee others’ pa-pers, make slides and posters to present at conferences, and per-haps apply for patents, attend to their start-up companies, and so on. And, to be sure, if a scientist feels passionately about some issue, he may find himself morphing into an advocate for action to combat global warming, to get Intelligent Design Theory out of the public-school classroom, or to bring clean water to the tribe he is studying; or, if he discovers a taste for this kind of thing, may find himself be-coming a professional expert witness, biotech executive, scientific administrator, dean, provost, or university president.

Still, the essential business of science, as of history, detective work, investigative journalism, legal scholarship, etc., is inquiry: i.e., an effort to find answers to some question or questions. And what this requires is that you identify a phenomenon in need of explana-tion, come up with a potentially explanatory conjecture, figure out the consequences of the conjecture, look at the evidence you have and any further evidence you can find a way to lay hands on, and as-sess where all this evidence points and what conclusions might be drawn, with what degree of confidence.

It’s often assumed that there’s a distinctively scientific way of inquiring, the “scientific method,” and that it’s this method that ex-plains the remarkable successes of the sciences. In Daubert,³ Justice Blackmun took this for granted; by now it is even enshrined in Black’s Law Dictionary.⁴ Scientists themselves, put on the spot to say something about how they do what they do, often gesture at some half-understood philosophical conception of this supposed method⁵ -often, unfortunately,⁶ like Justice Blackmun’s conception, some-thing vaguely Popperian in tenor. But there is no method used by all scientists and only by scientists. On the one hand, there are the familiar, inescapable kinds of inference involved in all serious em-pirical inquiry. But these are not used only by scientists. On the other hand, there are the myriad specialized tools, techniques, and pro-cedures devised by generation after generation of scientists to get more evidence of the kind they need and a subtler sense of where it points: instruments of observation, mathematical and statistical techniques, experimental, epidemiological, and other protocols, etc., and the internal social arrangements that have grown up in scientific communities to enable the sharing of results, encourage creativity, discourage dishonesty, and so on. But these scientific “helps”⁷ to inquiry, which are constantly evolving and often local to a specific field, are not used by all scientists.

This is not to deny that scientists make inferences -of course they do: about what might explain a puzzling event or phenomenon, about what the consequences would be if this or that conjec-ture were true, about the degree to which a conjecture is warranted by the evidence. But such inferences aren’t unique to the sciences -everyone trying to answer an empirical question makes them; and doing science isn’t simply a matter of inference. For one thing, it requires not only reasoning, but also appropriate connections with the world; ultimately, all scientific theories rest on experience, the evidence of the senses. For another, science can’t confine itself to a language fixed in advance, but will require the development of new vocabulary or the adaptation of older vocabulary in a special-ized language more closely approximating newly-identified kinds of stuff, thing, event, etc. -as witness, for example, the hundred-year history of the development of the concept of DNA.⁸ That’s why, as Dr Lydgate realized a long time ago, scientific work requires a grasp of content, not just of form.

Perhaps all this sounds radical; and from the perspective of twen-tieth-century proponents of inductivist, deductivist, probabilistic, Bayesian, game-theoretical, etc., models of the Scientific Method, it is radical. But it would have been entirely familiar to Thomas Hux-ley, according to whom the “man of science... simply uses with scru-pulous exactness, the method which we all, habitually and at every minute, use carelessly”;⁹ to Albert Einstein, who once observed that “the whole of science is nothing more than a refinement of common sense”;¹⁰ to Percy Bridgman, who commented that the supposed scientific method is “something talked about by people... on the outside,”¹¹ when the crucial thing is that a scientist “do his utmost with his mind, no holds barred”;¹² to James B. Conant, who wrote that “what the scientist does is simply to carry over, into another frame of reference, habits that go back to the caveman”;¹³ or to Gus-tav Bergmann, who described science as the “long arm” of common sense.¹⁴

I’m tempted to say, with Paul Feyerabend, that in science, “the only principle that does not inhibit progress is: anything goes”.¹⁵ This is not, however, to suggest, as he did, that there’s nothing more to sci-ence than power, politics, rhetoric, and negotiation; but only to rec-ognize the creative, open-ended, improvisational character of scien-tific work. And neither is it to suggest, as he did, that science isn’t a rational enterprise; but only to acknowledge that the reason the sciences have been able to make their remarkable discoveries isn’t that they have a unique method of inquiry, but that they have ampli-fied and refined the familiar procedures of the most ordinary every-day inquiry with tools and techniques that enable them to get more, more exact, and more reliable evidence, and better assess where it points. As the saying goes, nothing succeeds like success: each step forward enables scientists to correct past missteps, and perhaps to glimpse new ways to go about tackling the next problem.

If this is right, there is no peculiarly scientific method a scientist is obliged to employ. And neither, obviously, is there anything in the sciences analogous to legal rules about burdens and standards of proof or the admissibility of evidence.¹⁶ There are, however, other kinds of constraint on scientific work. The most important, internal to the enterprise, is that -like a historian or a legal scholar or any other empirical inquirer- a scientist has an obligation to respect the evidence, the facts. But there are also external constraints, such as rules about the ethical treatment of human or animal subjects; and professional constraints, such as the need to find funding for your work, the risk that you won’t be able to publish anywhere re-spectable if your work is too far from mainstream ideas, or the fear that you will be written off as a crank if your ideas are too radically at odds with what is taken to be already well-established.

Again, scientific inquiry isn’t bound by anything like the statutes of limitations or rules limiting the introduction of new evidence to which legal proceedings are subject; it takes the time it takes. Depending on the difficulty of the problem in the existing state of knowledge, the availability of resources, the abilities, backgrounds, and motivation of the people involved, and how hospitable the en-vironment in which the research is conducted happens to be, prog-ress may be painfully slow or exhilaratingly fast, steady or uneven. There are, however, external time constraints, such as the need to report results in a timely way to whatever body funds your work, or to produce something publishable this year to get tenure, a raise, or a promotion; not to mention the need to publish quickly to ensure that it’s you, and not your rival, who gets credit for a discovery.

In any scientific community, probably, some will be more ready to try a new conjecture when the existing hypothesis most of their col-leagues are content to work with encounters difficulties, and others will be more disposed to keep trying to adapt the old idea. Still, con-sensus will gradually form. Ideally, we would find agreement among scientists in a field when, and only when, the evidence is sufficient to indicate that it’s probably safe to rely on this conjecture, but likely to be a waste of time pursuing work on that rival idea. But there’s absolutely no guarantee that consensus on a scientific question will always faithfully track the state of the evidence. The work is hard; and scientists are only human. Sometimes they cut corners; some-times they deceive themselves into thinking that the evidence points more decisively to a commercially-desired or politically convenient conclusion, or to the conclusion they have long defended, than it re-ally does; and they are almost always subject to pressures both from the larger world and from within their profession. There’s a consid-erable gap between science as it would ideally be conducted, and science as it actually is conducted in the real world.

Ideally, again, given their core task, the culture of the sciences would accord with the values Robert Merton described long ago,¹⁷ with respect for evidence and willingness to share results taking center-stage. But scientists don’t work in a vacuum; they are part of, and depend on, a larger society. In practice, the core values of science can be eroded by the competing priorities -political, com-mercial, legal, and institutional- of the bodies that support scien-tific work. Governments, pharmaceutical companies, etc., that fund scientific research almost always have other concerns besides dis-covering the truth, and sometimes have an interest in playing down, disguising, or suppressing unpalatable facts.¹⁸ And recent changes in the management of universities have created a raft of perverse incentives -especially, the insistent demands that everybody be “productive” all the time- which encourage haste and careless-ness, and damage the scientific ethos. The result? -an explosion of journals, a torrent of publications, salami publishing, misleading multiple attributions of authorship, an increasingly overburdened and corrupt peer-review system, and an enormously time- and en-ergy-consuming preoccupation with “writing grants,” and the vast bureaucracy that accompanies it.

Over time, the physical and intellectual helps to scientific inquiry -the instruments, techniques of measurement and calculation, etc.- have grown steadily more sophisticated, more discriminating, more powerful. But the social helps, the institutional arrangements to encourage creativity, honesty, evidence-sharing, and the like, have not. Instead, as science has grown larger, more expensive, potentially more profitable, and politically more consequential, as scientific publishing has become big business, and as university administrators have come increasingly to rely on badly flawed surrogate mea-sures of the quality of the work produced, such social helps as the pre-publication peer-review system¹⁹ are under severe strain. The titles of recent articles -“Who’s Afraid of Peer Review?”²⁰ “The Cor-ruption of Peer Review is Harming Scientific Credibility,”²¹ “Peer Re-view Fraud,”²² “Scientific Regress”²³- hint at a disturbing truth: the integrity of the scientific enterprise is presently under real threat.²⁴

It is sometimes said that the purpose of a trial, like the purpose of scientific work or historical research, is to discover the truth.²⁵ But a common-law trial²⁶ isn’t remotely like a scientific experiment or a historical investigation; indeed, no one involved in such a trial is trying to discover who committed the crime, who or what caused the injury, etc. When judges, attorneys, or jurors try to figure out the truth of factual questions, no doubt they go about it in much the same way as the rest of us; but this isn’t what they’re doing at trial. The task of the attorneys is to make the best case they can for their side; which may require them to turn the peculiar constraints imposed by our evidence law to their side’s advantage. The task of the judge presiding over a trial is to ensure that the legal rules are followed and the law appropriately applied; which may require him to exclude relevant, but inadmissible, evidence. And the task of the jury (or of judges acting as triers of fact in a bench trial)²⁷ is to de-termine whether the party with the burden of proof has presented admissible evidence sufficient to establish guilt, liability, etc., to the degree required by the standard of proof; which may oblige them to ignore relevant evidence that has been ruled inadmissible.

I would say, rather, that the purpose of a trial is to arrive at a de-termination of guilt, responsibility, punishment, liability, or what-ever, in a legally-correct way. This, however, is not to suggest that factual truth is just irrelevant to our legal process. For the hope is that arriving at conclusions in a legally-correct way will mean that, by and large and on the whole and in the long run, justice is served; and part of what this means is that, if all those involved do their jobs adequately, often enough the truth about what actually happened will come out. Insofar as there is an epistemological rationale for our adversarial legal system, it would be this: provided that at-torneys are effective in seeking out evidence favorable to their side, discovering the weaknesses of evidence favoring the other side, and exposing those weaknesses in cross-examination, that judges rule correctly on the (in)admissibility of evidence, and that juries under-stand their duty and perform it adequately, this admittedly-oblique procedure is a tolerably-efficient way -given the inevitable limita-tions of time and resources- of arriving, often enough, at factually-correct verdicts.²⁸

There is a distinctive kind of reasoning used by attorneys and judges -the process we describe by the phrase “analogize and dis-tinguish”. But this isn’t a distinctively legal way of discovering truths about the world, but something quite different. Like all truths about human societies, truths about what the law is in some place at some time are made true by things people do; but unlike a legal scholar, whose task is to discover such truths, a judge interpreting statutes and precedents may be contributing to making something legally true. Sometimes the application of a legal provision to a specific case is a matter of deduction from a statute or a rule, or from a statute or a rule plus legal principles of statutory interpretation;²⁹ and when it is, what’s involved is, simply, deduction, not some peculiarly le-gal kind of reasoning. But where this doesn’t suffice, a judge may argue that relevant similarities between this case and these other previous cases and relevant differences between this case and those other previous cases mean that that the legally-most-reasonable upshot is this rather than that; and the attorneys for each side will try to persuade the judge that this case is analogous to these previ-ous cases where the upshot was the one their side seeks, but dis-tinguishable from those other previous cases going the other way.

In short, the peculiarly legal reasoning in which judges engage is aimed at determining how to interpret and extrapolate a law or a rule in the light of the precedents; while the peculiarly legal reasoning in which the parties to a case engage is aimed at persuading the judge to stress these precedents, these analogies, these distinctions, rather than those.

Perhaps this too sounds radical; and from the perspective of those who suppose that judges and attorneys are finding out the real meaning of the law or statute concerned it is radical. But it would have been entirely familiar to Oliver Wendell Holmes, who wrote that the appearance that a judge is deducing the legally-correct result is often just the “evening dress” the proposed interpre-tation puts on “to appear respectable”;³⁰ or to Benjamin Cardozo, who wrote of the “creative element” in a judge’s work, the frequent forks in the road, and the incompleteness of the signposts.³¹ You only have to look at some of those tangled plurality rulings from the Su-preme Court -Allegheny v. ACLU³², say, or Williams v. Illinois³³- to realize how implausible it is to suppose that, when judges disagree about the interpretation of some legal provision, at least one of them must have made a logical mistake and got the wrong answer. Rather, they disagree about how to apply a law in the circumstances of this case because they give different weights to various historical, legal, moral, political, economic, etc., considerations.

Granted, scientists often argue analogically,³⁴ and even, some-times, go through a process that might be described in terms of anal-ogizing and distinguishing. Molecular biologists working in the wake of Marshall Nirenberg and Johan Mattaei’s successful solution of the first “word” of the code by which RNA builds proteins,³⁵ for example, were presumably thinking that their word would be like Nirenberg and Mattaei’s in certain respects, but unlike it in others. True, too, legal reasoning by analogy³⁶ involves spotting a pattern common to a class of previous cases and the case at hand; which is in some re-spects quite similar to the first, abductive step of a scientific argu-ment by analogy, which involves spotting a pattern common to these and those natural or social things, phenomena, or events. But there’s a crucial difference: the upshot of a scientific analogy is a factual conjecture, a potential explanation that then needs to be checked against what is taken to be already known, explored for explanatory power, and tested (directly or indirectly) against the world; the up-shot of a legal analogy, by contrast, is an interpretation or applica-tion of a rule or a law -which may be better or may be worse, and may become established in the legal system or may not, but which just isn’t susceptible to this kind of testing against the facts.

Unlike the work of scientists, the work of attorneys and judges is formally constrained in many ways: by rules of procedure, timeli-ness, etc., and by the rules about burdens and standards of proof, admissibility, and discovery to which the legal scrutiny of evidence must conform. Moreover, in the law promptness and finality are, rightly, valued: we want legal decisions to be reached without inter-minable delay, and (normally) to stand once the appeals process has been exhausted. In this, the law is very different from the sciences: while scientists would no doubt prefer that they find the answers to the questions they’re struggling with sooner rather than later, the idea that they might simply decide to accept whatever answer looks best now, regardless of how inadequate present evidence may be, is just bizarre; and while, again, they would no doubt prefer that the currently-accepted theory stand firm, the idea that they might simply decree that it will stand firm no matter what new evidence comes in is no less so.

To be sure, the sciences aren’t wholly open and improvisational, and our legal system isn’t wholly rigid and formalized. The sciences have their rigidities: protocols for going about this kind of experi-ment or that kind of study, the routinized “write-by-numbers” style of almost all scientific articles today, and so on. And our legal sys-tem is constantly adapting to changing circumstances: developing, for example, a raft of ways of handling mass tort claims,³⁷ diversion programs in sentencing,³⁸ a whole body of internet law,³⁹ alterna-tive forms of dispute resolution,⁴⁰ and so forth. And it also gradually adapts to scientific developments -as, for example, in response to new DNA identification techniques and databases, some jurisdic-tions issued “John Doe” warrants or tolled the statute of limitations on certain crimes.⁴¹

Still, the culture of our legal system is very different from the culture of the sciences: strongly adversarial, focused on advocacy rather than on inquiry; shaped by a thicket of rules and formalities; and encouraging, and in some circumstances even mandating, the sequestering of information. True, there’s sometimes competition in the sciences (between rival individuals, or rival teams); but this isn’t, as in our legal system, built-in, structurally-required. True, there’s sometimes advocacy in the sciences (for one approach to a prob-lem, or one proposed explanation, over another); but this isn’t, as in the law, built-in, structurally-required -indeed, it is something to be regretted. And true, again, results in the sciences are sometimes withheld (to prevent a rival’s beating you to a discovery, to disguise a weakness in your approach, etc.); but this isn’t, as with the law’s protection of proprietary information, something required -but, again, something to be regretted.

Of course, just as there’s a considerable gap between science as it would ideally be conducted and science as it actually is conducted, there’s also a considerable gap between our legal system as theoreti-cally conceived and justified, and our legal system as it actually func-tions in practice. After all, like scientists and the rest of us, lawyers, judges, and jurors are only human: not every cross-examination is equally well-informed and searching,⁴² not every judicial interpreta-tion of the law is equally reasonable, not every verdict based strictly or appropriately on the evidence presented. Moreover, our legal sys-tem is seriously overburdened; and as fewer and fewer cases go to trial and more and more go to arbitration or are concluded with a plea bargain or a settlement, the assumptions on which the epis-temological rationale for adversarialism depends become less and less realistic.

Scientific inquiry, I have argued, is continuous with everyday em-pirical inquiry, but enormously refined and amplified by specialized tools and techniques. So one reason it’s hard for those outside the relevant field fully to understand scientific evidence is, simply, that they are unfamiliar with those tools and techniques and with the technical vocabulary that grows up in every serious scientific spe-cialty -vocabulary that can be at best superficially understood by someone who lacks real knowledge of the subject-matter. Yes, some-times it’s possible even for someone without specific knowledge of the field to spot methodological flaws: an epidemiological study has no controls, say, or wasn’t conducted double-blind, or relied on self-reported disorders; but some flaws are very hard for an outsider -whether a judge, an attorney, a juror, or for that matter a scientist from a different field- to detect.

Moreover, the adversarial culture of our legal system not only en-courages lawyers and judges to assume that, on any scientific question, there must be two sides, but also tends to draw in, as expert witnesses, those scientists who are ready to settle on a conclusion when most people in their field take the view that they must wait for further evidence (and who, to make matters worse, are apt to grow more dogmatically certain as they testify over and over to the same effect). And though, on many scientific questions -including, very often, the scientific questions at the heart of toxic-tort litigation- the most reasonable position is “As yet, we just don’t know,” neither party is likely to want to hire as their expert a scientist who would candidly say just that. At the same time, the procedures of the law can come as quite a culture-shock to legally-naïve scientists, who may find their treatment under cross-examination disrespectful, the legal preoccupation with conflicts of interest baffling,⁴³ and being “Dauberted out” by a judge who rules that their work isn’t really sci-entific downright insulting.

And it’s not surprising if, faced with the need to reach some con-clusion when the relevant science is as yet unsettled -“to decide,” as Learned Hand put it long ago, “where doctors disagree”⁴⁴- law-yers and judges often look for something like procedural principles to appraise scientific testimony, in effect falling back on their skills in scanning verse to scan the potato crops. This, I suspect, partly explains Daubert’s preoccupation with “methodology.”⁴⁵ It also sug-gests why statistical significance, relative risk > 2,⁴⁶ faithfulness to the “Bradford-Hill criteria,”⁴⁷ and the like are sometimes given a com-pletely artificial legal importance; why the pseudo-rigorous ACE-V fingerprint-identification “methodology” passes legal muster;⁴⁸ why adversarial testing under cross-examination in court is sometimes confused with real-world empirical testing;⁴⁹ and why the scien-tific peer-review process, developed for very different purposes, is sometimes taken as a straightforward indication of evidentiary reliability.⁵⁰

Well, as I warned you, this is only the twenty-minute version of that hypothetical Ph.D. dissertation; and now I’m just about out of words. I hope, however, at least to have shown that, if we are to un-derstand why the interactions of science and the legal system can be so problematic, we need to do much more than gesture vaguely to-wards supposed “different legal and scientific modes of thinking”- to think through the consequences of the different purposes, con-straints, and cultures of science and of the law.⁵¹