Modern science claims to be founded on experiments, and more precisely on reproducible (ideally, on reproduced) experiments. In practice, of course, mutual trust among scientists and the desire to advance the discipline rather than revisit old territory means that many experimental results are never actually checked (sometimes with unfortunate consequences!).

One might assume that, while researchers are themselves perhaps too busy to verify each other's (uncontroversial) discoveries, historians would be eager to do so. This is not the case, and only in part because historians of science are themselves rarely trained as scientists. A more profound reason is the division between "anachronic" and "diachronic" approaches to history, and the historical profession's distrust of the former.

The anachronic perspective is the view from the present, and is almost automatically assumed by working scientists in their occasional considerations of the past. Most postgraduate students in physics, for example, are required as part of their education to take a laboratory course in which famous experiments critical to the development of atomic science are, in some sense, repeated. The student, of course, knows the desired outcome (and is perhaps graded on attaining it); the equipment used is likely to be of modern design and far simpler to operate than that available around 1900. (In recent years there has even been a growing tendency to move the laboratory into virtual reality, and run a simulation with the correct result built in.) This is on the one hand a replication -- look! charge is quantised! look! matter has a wavelength! -- but it is on the other hand very unlike the experience of the first pioneer wandering about in confusion and uncertainty.

Millikan oil-drop experiment lecture-demonstration unit AP-8210A, manufactured by PASCO.

For most professional historians, such an anachronical replication is no replication at all. The attitude of the historian toward the past is supposed to be diachronic; one should view the past from its own perspective, without reference to subsequent developments. One should try to enter into the mind, or at least the mindset, of the scientist who did the experiment originally.

Millikan's original equipment. The validity of this difficult experiment was debated for years.

This is clearly a difficult task. Indeed, the replication of famous scientific experiments is thought by many philosophers to be literally impossible: one may go through the motions of carefully re-creating, say, the Millikan Oil-Drop Experiment, but one cannot re-create Millikan's mind nor his millieu, so what is the point? Better to focus on the broad socio-economic forces which acted upon, and presumably informed, the network of scientists who accepted or rejected Millikan's claimed result!

This pessimistic attitude, typical of professional historians' conviction that the past is a series of one-off contingencies, would seem to render Clio a muse without music, something the public funding-agencies have gradually noted. Why support historians if history has no lessons? In the last few years, particularly in Britain, scholars' desire to survive has begun to outweigh their love of philosophic purity, and history of science is among the fields affected. One may with confidence predict that more and more History-Channel-ready great moments in scientific history will soon be upon us.

Bertini's Galileo and the Doge.

As of today, however, this boom has clearly not arrived. Experimental History of Science (EHS) remains, as it has always been, a mere twig in the History of Science tree. There are a few outstanding practitioners -- Paolo Palmieri and his associates at Pitt's HPS Lab, for example, or Otto Sibum at Uppsala. To call EHS a major movement, however, would be a gross exaggeration.

This is surprising, in a way. Many contemporary historians focus their attention on material objects, and many historians of science view technology as more important than thought in determining the outcome of events. Walking into an old laboratory, looking at the apparatus, noting their cost and place of manufacture, trying to understand how they were actually used and what social messages were encoded in the choice of this device rather than that: all quite standard historical scholarship. So why not try to actually use the apparatus (or a reconstructed version, if need be)? EHS would seem, on the surface, to fit naturally into postmodern historical practice.

In reality, however, even the most earnestly diachronical re-creation of a famous experiment goes against the subjectivism of many workers in the humanities. This can be disguised: Sibum, for example, has written that EHS "explores the historical and epistemological meanings of the experiencing subject in physical investigations of nature -- knowledge embodied in historical actors ... We have taken up the challenge represented by the long-standing divide between epistemology and practice, and sought to break with traditional concepts of disembodied knowledge. At the heart of [our] conception of embodied knowledge lies the unity of experience of the actors involved in productive work." Nevertheless, behind this rhetoric is the unstated assumption that something objectively happened in the laboratory, that it happened whatever interpretation the "actor" may have put on it, and that it, or something like it, will happen today, in a re-created laboratory, as well. If this were not so, the act of re-creation would be pointless.

There is also an anachronic (and therefore socially disfavoured) aspect to much of the most interesting research involving replication of historic experiments. Re-enactors cannot truly forget what they know, nor can they pretend for long not to be interested in the present-day things which actually and naturally interest them. Let us consider a specific example for clarity.

Alexandre Koyré

In the middle decades of the 1900s, Alexandre Koyré attacked the long-established image of Galileo as an empiricist, presenting him instead as a Platonist philosopher who conducted Gedankenexperimenten and then claimed to have done them in reality as well (or, perhaps, who indeed did them in reality, but without allowing actual results to interfere with the pre-conceived ideal). Koyré presented himself as a proper diachronic historian, pointing out that the word "experiment" has acquired many connotations since the Renaissance, and that primary sources mean what they meant when written, not what they have come to mean over time. This applies even to scholarly ethics: to a convinced Platonist, making up data is not necessarily cheating. A brilliant writer, Koyré skillfully depicted Galileo's Rube-Goldberg clocks and "very smooth" inclined planes as transparently obvious fantasies, not real objects in any real laboratory. This rhetorical campaign was highly successful; by the late 1950s, most historians had rejected tradition and adopted Koyré's view.

One who did not was Thomas Settle, (only a graduate student!), who proceeded to build and operate meticulous re-creations of Galileo's instruments. His papers defending the authenticity of Galileo's work [e.g. Science 133, 19 (1961)] not only sent the pendulum (we must use an appropriate analogy) swinging back again, but also marked the beginning -- some would say the main triumph -- of the EHS movement in physics. Galileo's results, it turned out, were reproducible; his bizarre contraptions functioned as advertised. It was Koyré who had been projecting modern prejudices into the past.

At first glance, then, Settle's work appears to be a good example of experimental mainstream (i.e. diachronic) history: actually building the old apparatus, it seems, gets one closer to Galileo's world than Koyré's lifetime of reading about it. However, while historians have (with many reservations, including some from other experimentalists) accepted Settle's view on the specific subject of Galileo, their hearts remain with Koyré, and perhaps justifiably. Problems still remain. What did Settle actually prove? Only that a sufficiently dedicated and ingenious experimenter could use Galileo's equipment to obtain results similar to those Galileo had published. How Galileo himself obtained those results -- the way he himself claimed, or by the power of thought -- is unknowable.

One obvious reason for supposing that the experiments took place in the real world -- or, just as much, one reason for supposing they did not -- is Galileo's inclusion in his private notes, but not in his published dialogues, of a counterintuitive and counter-theoretical observation. He is discussing the famous problem of dropping two bodies from a leaning tower, perhaps Pisa's, and Aristotle's opinion that heavy weights fall faster than lighter ones. They do fall faster, of course, but Galileo realised that the difference is because of air resistance and would disappear in a vacuum, his celebrated equivalence principle. But he adds something else: " [F]or it is true that wood at the beginning of its motion is carried more speedily than lead; but a little later the motion of lead is so accelerated that it leaves the wood behind, and, if they are released from a high tower, the lead gets ahead of it by a large distance: and I have often put this to the test." Heavy objects moving in air must catch up with lighter ones before hitting the ground ahead of them!

To a modern reader, this seems absurd: how can air resistance do that? Settle, however, repeated the experiment, and photographed the effect. Probably he did not suppose that he was witnessing a breakdown of the equivalence principle; nevertheless, he ceased to act diachronically and began to act anachronically: he found an acceptable modern explanation. High-speed photography showed that assistants would almost always unconsciously release the lighter weight, which they were gripping less tightly, an instant before the heavier. This of course gave the light weight a head start, which the would be magnified over the early part of the fall before friction overwhelmed it.

The fact that Galileo had observed this catching-up effect, but did not understand its origin, appeared to Settle overwhelming evidence that the falling-body experiment had indeed taken place (just as the result's "absurdity" had earlier been considered strong evidence to the contrary!).

And yet, one cannot help wondering a bit. Galileo continues: "Oh, how easily are true demonstrations derived from true principles! If ... mobiles [i.e. moving bodies], as they recede from a state of rest, recede with an amount of contrary impressed force equal to their heaviness, then those that are heavier will recede in conjunction with a greater contrary force: but if heavier ones must consume more force impelling them in the contrary direction, it will surely be necessary that heavier mobiles are moved more slowly, since they undergo a greater resistance. And if, on the other hand, these things are true, it follows that heavier things, after they have consumed so much of the contrary resistance that they are no longer hindered by as much as lighter things are, must go down more speedily: which, again, experience surely shows." In other words, he had an elaborate theoretical explanation in terms of (his own anti-Aristotelian version of) impulse theory ready at hand! Was this explanation made after the fact, or before?

From the diachronic point-of-view, then, we can say with confidence very little about Galileo's inner life, because he lived it in a world isolated from ours by the barriers of time. Repeating his experiments can perhaps help to "give a feel" for what his experience was like, but only if, first, the original experiments really took place, and second, we are already sufficiently in harmony with him to feel something of what he felt.

Adopting instead the anachronic point-of-view, we may take this imaginative sympathy for granted and trust the historical record rather as we trust our own memories; Galileo then remains in the past, like our own remembered selves, but one assumes the past is reconstructable. This is a common-sense attitude, and probably it works better and in more circumstances than professional historians care to admit. Nevertheless, its rejection by academia has not been without reason: the historical record is fragmentary and in places non-existant; the human mind fills in the gaps with its own speculations, hopes and fears; thus each era creates its own anachronic past. Is there a third way?

Here at last we come to what I have called (in Part One of this essay) the cenochronic approach to history. But this blog installment is already rather long; I will give a cenochronical interpretation next week.

Vivaldi: Concerto for Two Violins (excerpt)
Performed by the Tafelmusik Baroque Orchestra as part of their astronomy-inspired "Galileo Project". [4 min]