One hundred and thirty years of being experts

Charles Francis Himes taught natural science at Dickinson College for thirty-one years, took a doctorate at Giessen, photographed the total solar eclipse of 1869 for the federal government, and lectured the citizens of Carlisle on the telephone and the phonograph while both were still novelties. He was, in other words, exactly the sort of person the courts of the 1890s kept calling to the stand: a university scientist with a laboratory, a European degree, and a habit of explaining new technology to audiences who had never seen it work or knew how it did. On the evening of February 10, 1893, he stood up at the Franklin Institute in Philadelphia and told a room full of engineers what he thought of the position that role (expert witness) had put him in. The lecture ran in the Institute’s Journal that June under the title “The Scientific Expert in Forensic1 Procedure.”

When I read it, I was expecting a period piece; what I found was a diagnosis I could hand to a laboratory director tomorrow.

Himes opens by collating the abuse foisted on the courts2 by so-called expert witnesses. He pulls from Taylor on Evidence, from jury charges, from judges speaking off the bench in the freer air of a lecture hall, and the accumulated complaints do the work:

Perhaps the testimony which least deserves credit with a jury is that of skilled witnesses. It is often surprising to see with what facility, and to what an extent their views can be made to correspond with the wishes or the interests of the parties who call them.

They do not, indeed, wilfully misrepresent what they think, but their judgment becomes so warped by regarding the subject in one point of view, that, even when conscientiously disposed they are incapable of expressing a candid opinion.

Even men of the highest character and integrity are apt to be prejudiced in favor of the party by whom they are employed.

Read those three passages without the nineteenth century diction and you have the standard contemporary account of contextual bias. Neither lying nor incompetence are evident but rather a judgment warped by regarding the subject from one point of view, in people of the highest character, who would be outraged3 if you asked them whether they were biased because they sincerely believe they are not. Himes’s sources had the mechanism right in 1893 and were writing about it as settled knowledge, not as a discovery.

He also records the joke, which I had assumed was a twentieth century invention:

May it please the court there are three kinds of liars: the common liar, the damned liar and the scientific expert.4

And he records the line that I have not been able to stop thinking about since I read it. It comes from a judge lecturing on medical expertise, and it describes how a lawyer holds two attitudes toward the same person:

The lawyer discriminates between the practitioner and the expert, embodied though they be in one and the same person; at the bedside he reposes a practically unlimited confidence in the man as a physician and in his science; as an expert he regards him as a tool whose cutting edge5 he will employ, or dull, as exigency may require. [emphasis added]

Himes calls that last sentence the real ground of the expert’s dissatisfaction, and he is right: Everything else in the complaint literature is downstream of it. It’s why scientists whine about being in court: “They twisted my words!” Yes, dummy, that’s their job—to make a better argument than the other attorney.6

What makes the lecture worth your time is that Himes refuses the character explanation. The legal profession, then as now, treated the problem as a moral defect in scientists.7 Himes, however, locates it in structure: The expert, he argues, was dropped into a system whose categories had already hardened around four functions, witness, jury, judge, advocate, each so sharply differentiated that a witness who estimated land value at five thousand dollars before the damage and one thousand after could not be permitted to say the damage was four thousand, because subtraction belonged to the jury. Into that system arrives a person who is legally a witness and practically something else entirely:

He is legally a witness, an ordinary witness, but practically with extraordinary functions8, and loaded with extraordinary responsibilities, and one might add, frequently loaded with extraordinary, and even absurd, expectations.

He is subpoenaed9 like a witness, sworn-in like a witness, paid unlike a witness, helps counsel prepare the case, drafts the cross-examination of his opposite number, and then takes the stand and is treated as though none of that had happened. Himes’s conclusion is that the witness cannot be blamed for a position they did not create. “He is being shaped and fashioned rather than finally fixed.” The system, he says, has feebler powers of adaptation than the science being thrust into it; so, the science must bend and flex to the law’s requirements.

That is the sentence that has aged best, and not in a way that flatters us. Himes assumed the shaping was underway and would finish. He was writing about a thirty year old problem in 1863 and expected resolution within his lifetime. We are a hundred and thirty-three years past the lecture, and the scientific expert is still classified as a witness with no procedural category of his own. Daubert did not create one; it gave the trial judge a screening obligation and left the underlying taxonomy exactly where Himes found it. Rule 706 lets a court appoint its own expert, which is precisely the reform Himes endorsed by quoting Judge Redford, and which is used so rarely that its disuse has its own literature. The one structural change he did not anticipate is the one that mattered most.

Where the lecture stops being modern is instructive. Himes’s central defense of the expert is that bias mostly precedes the call rather than following it. His question, and it is a good one, is how many experts are absent from a case because the party who consulted them did not want what they found. In his account the expert reaches a decision and is then hired by the side that shares it, so what looks like purchased opinion is really the market sorting itself. That defense survives partially; the file drawer of unwanted consultations is real and nobody has ever measured it.

But look at what he assumes to make the argument work:

In his laboratory, in conducting his investigations, the scientific expert may keep himself free from bias.

That is a load-bearing premise, and it is the one thing in the lecture that is unambiguously false. The laboratory is not a bias-free zone from which conclusions emerge clean and are only afterward exposed to partisan pressure. It is where the domain-irrelevant information arrives, where the case narrative shapes what counts as sufficient, where the reference sample is examined before the questioned one. Himes could not have known this, but the structure of his defense depends on it, and the defense collapses without it.

The same optimism runs through his showpiece case. Four experts are asked whether arsenic could still be detected on a stove after three months of use. Three say no, one says yes, the stove is scraped, the arsenic is found, the conviction follows. Himes calls this evidence supplied “with inerrant certainty” by “infallible methods.” The result was almost certainly correct but the epistemology is pre-modern. There is no error rate anywhere in it, no question of what the test would have shown on a stove that had never held arsenic, no acknowledgment that three of his four scientific experts were wrong about the very question their expertise was called upon to settle. He tells the story as a triumph and never notices that his own example has a seventy-five percent error rate in it.

Now the change he did not see coming. Himes’s expert was a private man of science, hired by a party, paid by private agreement, whose bias question was always framed as a question about a particular case. Ours mostly is not. The modern forensic scientist is a public employee in a laboratory that is organizationally part of, or funded by, one of the two parties to every criminal case it touches. The alignment is not produced by the need; it exists before the evidence arrives and persists after the case closes, across every case, for the length of a career. Himes worried about a fee agreed upon between the expert and counsel. We built something more durable than a fee, gave it a budget line, and then spent a hundred years litigating the bias of individuals inside it.

So what is the same after all this time?10 Well, let’s see. The complaint, almost word for word, the mechanism, correctly described before anyone had a name for it, and the structural diagnosis, which was right and which nobody acted on. What is different is that Himes thought he was watching a system in the middle of adapting, and we know now that it was not adapting. It was settling.11

He closed by telling his audience that the courts are powerless to exclude this evidence even if they wanted to, that its status is unsatisfactory at many points, and that it demands the most serious consideration of everyone with an interest in what happens in our courts. I cannot improve on that as a statement of the current problem, which is the least comfortable thing I can say about a lecture given 130 years ago.

Share