When the historian of science studies a product of scientific work - a published paper, a laboratory record, a transcript of an interview - he is dealing first of all with an event.  A number of different facets of the event can engage his attention.  One can distinguish at least eight such facets, corresponding to different types of interesting questions:

First is of course the understanding of the scientific content of the event (E) at a given time, both in contemporaneous terms and, separately, in terms of what we now believe to be the case.  What did the scientist claim was at issue?  What was he in fact confronted with?  For this we need to establish the awareness, within the area of public scientific knowledge at the time of the event, of the so-called scientific facts, data, laws, theories, techniques, lore.  I would include under this heading the larger part of historical research on what are called scientific world views, paradigms, and research programs; chiefly, however, historians and scientists are still concerned with digging out the concepts and propositions embodied in the event studied and with rendering them in empirical and analytical language.

This tracing of conceptual development and of the “context of justification” is the most frequent and strongest activity of historians of science and historically inclined science educators.

Third is the more ephemeral personal aspect of the activity in which E is embedded.  Here we are in the context of discovery, trying to understand the “nascent moment,” which may be poorly documented and not necessarily appreciated or understood by the agent himself.  Except for work on a few such figures as Kepler or Einstein, scientists until recently have been rather impatient with such studies, and so have philosophers.  The very institutions of science - the methods of publication, the meetings, the selection and training of young scientists - are designed to minimize attention to this element.  The success of science itself as a shareable activity seems to be connected with this systematic neglect of what Einstein called the “private struggle.”  Moreover, the apparent contradiction between the often “illogical” nature of actual discovery and the logical nature of well-developed physical concepts is perceived by some as a threat to the very foundations of science and rationality itself.

The alternative path is not easy.  In one of his interviews, Einstein urged historians of science to concentrate on comprehending what scientists were aiming at, “how they thought and wrestled with their problems.”  But he pointed out that the scholar would have to have sufficient insight, a kind of Fingerspitzengefühl both for the content of science and for the process of scientific research, as solid facts about the creative phase are likely to be few; and that, as in physics itself, the solution to historical problems may have to come by very indirect means, the best outcome to be hoped for being not certainty but only a good “probability” of being “correct anyway.”

One of the nicest testimonials on the contrary orientations of scientists and of historians of science was provided by P. A. M. Dirac not long ago.  He agreed to lecture about his work at the summer school on the History of 20th-Century Physics at Varenna in 1972.

I have learned a great deal here, not only individual facts about the history• of science, which I have picked up from various lectures, but I have learned to appreciate the point of view of the historian of science.  It is really a very different point of view from that of the research physicist.  The research physicist, if he has made a discovery, is then concerned with standing on the new vantage point which he has gained and surveying the field in front of him.  His question is, Where do we go from here?  What are the applications of this new discovery?  How far will it go in elucidating the problems which are still before us?  What will be the prime problems now facing us?

He wants rather to forget the way by which he attained this discovery.  He proceeded along a tortuous path, followed various false trails, and he doesn’t want to think of these.  He feels perhaps a bit ashamed, disgusted with himself, that he took so long.  He says to himself, What a lot of time I wasted following this particular track when I should have seen at once that it will lead nowhere.  When a discovery has been made, it usually seems so obvious that one is surprised that no one had thought of it previously.  With that point of view, one doesn’t want to remember all the work that led up to the making of the discovery.

Now, that is just the opposite to what the historian of science wants.  He wants to know the various influences at work, the various intermediate steps, and he may have some interest in the false trails.  These are rather contradictory points of view, and most of my life has been spent with the point of view of the research physicist, and that involves forgetting as quickly as possible the various intermediate steps.

However, with the understanding of what the historians of science are concerned with, I have tried to think over the past, and have done my best to remember the various incidents, things that happened 50 years ago.  I have tried to figure out the influences, the effect of the various teachers that I had and the training that I received, to see how these things led me to the style of work which I followed in later life.

These eight areas of study are not separated by hard barriers.  To be sure, each has invited its own specialization and thereby its own operational self-definition.  For each we could quickly put forward the names of heroes and the shape of future hopes of development - though we might now all agree, with various intensities of regret, that the resolution of a real case in the history of science in all its ambiguities and interdisciplinary connections into separable components is, after all, a reductionistic strategy which our human limitations force or doom us to employ.

As we now know, Salviati was grossly exaggerating.  Resolving the motion of the falling object into two components in order to understand motion and its causes is only the first step in an essentially infinite chain of resolutions.  If one wants more detail about the motion, other laws enter.  The appearance of the Coriolis force is responsible for an eastward deflection of the object.  The laws for falling bodies in real media at various Reynolds numbers have to enter to calculate the effect of friction and turbulence.  The more detail one wants to know, the more resolutions become necessary.  The process would have become infinitely regressive if an Occam’s Razor had not been invented in our century for cutting off all side effects below a certain limit.  Quantum physics did give us a way to stop, owing to the uncertainty principle and the finite size of Planck’s constant; they extinguish the meaningfulness of all further questions.

And there is another lesson.  The two components Salviati chose, while they were plausible enough and even turned out to be useful, were not endowed with any provable necessity over any other set of two or more components of motion that might have been imagined.  I mention this to acknowledge that my list of components is not to be taken as the recital of an unchangeable, sacred Eightfold Way.  On the contrary, one reason for making the list was to be able to conclude that it is incomplete in an important respect.  That is to say, there remains a set of questions which are irresistible (to me, at any rate); which cannot be handled naturally in this eightfold scheme at all; and which lay bare a link between scientific activity and humanistic studies, a link that few have studied so far.

Any listing of such questions will have to include these: What is constant in the ever-shifting theory and practice of science - what makes it one continuing enterprise, despite the apparently radical changes of detail and focus of attention?  What element remains valuable in theories long after they have been disproved?  What are the sources of energy that keep certain scientific debates alive for decades?  Why do scientists - and for that matter also historians, philosophers, and sociologists of science - with good access to the same information often come to hold so fundamentally different models of explanation?  Why do some scientists hold on to models of explanation that are contrary to the evidence, sometimes at enormous risks?

Why do scientists privately not accept a dichotomy between the context of verification and that of discovery, and publicly often accept it?  If it is true, as Einstein believed, that the process of formulating laws purely by deduction is “far beyond the capacity of human thinking,” what may be guiding the leap across the chasm between experience and basic principle?  What is behind the obviously quasi-esthetic choices which some scientists make - for example, in rejecting as merely “ad hoc” a hypothesis that to other scientists may appear to be a necessary doctrine?  Are the grounds from which such choices spring confined to the scientific imagination, or do they extend beyond it?

To handle such questions I have proposed a ninth component in the analysis of a scientific work.  I have used for it a time-honored term, “thematic analysis,” familiar from somewhat related uses in anthropology, art criticism, musicology, and other fields.  In many (perhaps most) past and present concepts, methods, and propositions or hypotheses of science, there are elements that function as themata, constraining or motivating the individual and sometimes guiding (normalizing) or polarizing the scientific community.  In the scientists’ own public presentations of their work, and during any ensuing scientific controversy, these elements are usually not explicitly at issue.  The discussion seems to concern chiefly the empirical content and the analytical content, that is, the repeatable phenomena and the propositions concerning logic and mathematics.  By way of a very rough analogy, I have suggested that those two elements be considered the x and y coordinates of a plane within which the discussion seems chiefly to proceed since the “meaningfulness” of concepts is tested by the resolution of concepts or propositions into those elements - “meaningful” in the sense that agreed-upon rules generally exist for verifiability or falsifiability of the statements made in that language.

Thus, in R. A. Millikan’s famous oil drop experiment the question whether or not the electric charges on small objects always come in multiples of some fundamental constant, called the charge of the electron, could in principle have been resolved quickly by coming to terms on how and what was being observed through the telescope or ultramicroscope when a particle was seen to move in the view field, and whether and how to amend the equation for Stokes’ law for the fall of small objects by extrapolation of a correction term.  If that were all, the lengthy fight about the existence of a postulated “subelectron” would never have happened.  But in 1910, and continuing for some years afterward, the controversy between Millikan and Felix Ehrenhaft was joined - at the intersection, as it were, of two sets of world lines.

Analysis of the published research reports, of the expressed motivations, and of the ever-hardening attitudes of the protagonists on opposite sides of the question shows here, as in other cases, the strong role of an early, unshakable commitment to opposite themata on the part of the opponents; the one to the thema of discreteness as the fundamental explanatory principle operating in electricity, the other to the thema of the continuum and hence antiatomism.  In Millikan’s case the commitment to an atomistic explanation of electricity predated his experimental verification and indeed helped him to pick his way through initially indifferent data to support his contention.  In the other, the growing dedication to the antithema led to a veritable flood of counter-experiments.

The themata that appear in science can, in our very rough analogy, be presented as lying along a dimension orthogonal to the (xy) plane in which verification and falsification can take place, hence somewhat like a z axis rising from it.  While the xy plane does suffice for most discourse within science in the sense of a public, consensual activity, it is the three-dimensional (xyz) space which is required for a more complete analysis—whether historical,

The attitude I have taken in the task of identifying, ordering, and categorizing thematic elements in scientific discussions is to some degree analogous to that of a folklorist or anthropologist who listens to the epic stories for their underlying thematic structure and recurrences.  While the analogy leaves much to be desired, there are more than superficial relations.  For example, the awareness of themata which are sometimes held with obstinate loyalty helps one to explain the character of the discussion between antagonists far better than do scientific content and social surroundings alone.  The attachment of physicists such as H. A. Lorentz, Henri Poincaré, and Max Abraham to the old electromagnetic world view and their discomfort with Einstein’s relativity theory become a good deal more understandable when the ether is thought of as operating as the embodiment of thematic concepts (for example, of the absolute and the plenum).  Thus in their obituary for Abraham, Max von Laue and Max Born wrote perceptively (3):