In reply to Ian Michael's searching paper, I should like to consider what a historian of science might want to say about the study of scientific textbooks. A moment's reflection shows that there is a rich vein here to be tapped. For example, one question might be: what kind of history of science do we find in scientific textbooks? In his famous account of The Structure of Scientific Revolutions (1962) Thomas Kuhn gave a simple answer: bad. In his assault on the linear-accumulation model of scientific progress, text books were pilloried for their whiggish reduction of the history of science to stories of heroes and villains. It has, however, been suggested that Kuhn did professional historians of science a disservice in this way by providing them with too facile a target. To dissociate oneself from the kind of history enshrined within scientific textbooks became too easy a way to assert one's professional credentials.
A quite different question might be: how can we use the scientific textbooks of the past to reconstruct the then contemporary state of scientific theory? For Kuhn textbooks had a certain value because they do give us a window on the 'normal science' of a period. But they were not the obvious place to look for innovations. This may be a plausible starting position, but there are problems with it. In the first place, subsequent scholarship has shown that what Kuhn dubbed 'normal science' has not always been as consensual as he implied. In the field that 1 know best, that of organic chemistry in the nineteenth century, it may be tempting to speak of different paradigms separated by revolutions, but even within the periods of 'normal science', if they existed at all, there was intense competition between theoretical chemists, each seeking to put his own stamp on the science and each striving to promote his own formulas for different compounds. At one point in the middle of the nineteenth century there were close on twenty different formulas for acetic acid! Secondly, textbooks might actually be more useful than Kuhn sometimes implied. In cases where a textbook survived a 'revolution' one might be able to use its successive editions to illuminate the course of the revolution. A long established example would be the successive editions of a seventeenth-century Cartesian textbook by Jacques Rohault, which Samuel Clarke constantly updated with ever more Newtonian footnotes until the Cartesian text had been undermined.
These preliminary considerations might help to show why historians of science have something of a love-hate relationship with textbooks. One source of hate is that the textbook itself may tell one very little about how the science was actually taught, even at the site where the book was produced. As lan Michael points out in his paper, when one starts to study teaching practices the function of the textbook may recede from view. Another reason for hate is that one can hardly avoid exhausting debates about what counts as a scientific textbook. I have myself been privileged to participate in a European Science Foundation project on the history of chemical textbooks, during which it became clear that contributors were experiencing unease about definitions. For example, should one include works of popularization in the textbook category? Should one include systematic texts that nevertheless propound the idiosyncratic views of an author?
Underlying the problem of definition is the fact that the relation between the author's intentions and the actual audience for the book can be extremely complex. A well-known example might be the nineteenth-century texts of Mary Somerville who addressed her best known book on the Connexion of the Physical Sciences to other women but witnessed its success as a textbook. The work of Jonathan Topharn has shown how the Bridgewater Treatises, published during the 1830s, were read as texts transmitting politically 'safe' science, even though they were ostensibly works of natural theology and religious apologetics, Texts not written as textbooks could become reference works for students. Negotiations between author and publisher are particularly intriguing here.
Despite such complications, we should not draw too negative a conclusion about the value of studying 'textbooks'. They may, in some circumstances at least, illuminate important historical processes or at least help to identify them. A successful text might tell us something about the emergence of an audience for a particular form of science. For example, one scientific text by Nils Berlin sold some 450,000 copies in mid-nineteenth century Sweden, reflecting the introduction of compulsory elementary school education. Those texts explicitly geared to a preparation for examinations can tell us something about a changing syllabus. Comparing different editions can also show which new developments are deemed significant - unless of course the author was too lazy to revise them! Not that the result of such study necessarily conforms to a neat pattern. In nineteenth century chemistry, it is striking how different French authors, for example, took up John Dalton's atomic theory at different rates.
Comparing different national texts can be illuminating. In the ESF project to which 1 have referred, it emerged that Swedish chemical texts of the mid-nineteenth century were particularly dull and 'factual'. All controversy was avoided, despite the fact that the science was torn asunder with controversy at the time. But that , in a way, was the point: no-one in Sweden wished to take sides against Berzelius whose enormous authority was currently being questioned by a younger generation of French chemists disenchanted with his electrochemical theories. Comparing different translations of one and the same text may also raise fascinating questions. Why was a particular text translated in one country but not another? What do the trajectories of translation tell us about the centres of scientific activity and how these have changed over time? In Spain, in the early nineteenth century, chemical texts were almost all French, but indigenous products later began to compete. Translation can also introduce new meanings. In a Swedish translation of a textbook by Friedrich Wohler the word 'atom' was effectively deleted!
There is a further set of questions, of great interest to historians of science, concerning the images of scientific processes and epistemologies that textbooks purvey. The deliberate elimination of controversy can produce an artificial image of scientific objectivity. For the ESF project, Mary Jo Nye produced an instructive case-study of a text by Linus Pauling who believed in the value for students of picturing a concrete world of particles. The question is whether this might not have reinforced a realist rather than an instrumentalist view of theoretical entities. There were, of course, equivalents in nineteenth- century Britain, with ball and stick models for chemical molecules such as those devised by Edward Frankland and A.W.Hofmann. In such cases do the didactic purposes of textbooks not lead to a distortion rather than an enhancement of scientific understanding? This is a question that remains a challenge for those involved in curricular planning today. And for the historian of science there are associated questions that are unlikely to lose their fascination. How, for example, do textbook writers cope with the impossible problem of stabilising knowledge when that knowledge is itself in flux? Has the attempt to achieve stabilisation for teaching purposes had, on some occasions, a beneficial effect on contemporary scientific thinking? There are suggestive examples of precisely this process, whereby innovations proposed for their didactic value are seen to offer a deeper theoretical clarification that might otherwise be missed. A classic example would be the efforts of Cannizzaro at the Karlsruhe conference of 1860 to resuscitate Avogadro's hypothesis as a way of injecting coherence into the determination of atomic and molecular weights. At the very least, his promotion of Avogadro's hypothesis by showing how it would inform a course of chemical philosophy was an inspired piece of rhetoric. Nor is this an isolated case. Earlier in the nineteenth century, there had been the claim in William Henry's Epitome of Chemistry (1801) that one of the advantages of the new anti-phlogistic chemistry of Lavoisier was that it facilitated the learning of chemistry. It becomes, then, a nice question whether we yet have adequate resources for learning the history of science.
The editor has generously offered me a few lines after reading John Brooke's paper, which I was not present to hear. I am not offering a 'response' in the adversarial sense, as I feel only respect and admiration for it.
John Brooke makes me rethink the question of defining a textbook. I used to hold that we spent too much time on definition; that just as three socks, if they were hung on a gallery wall, could thereby claim the status of a work of art so the Yellow Pages, used by a class of sociology students, could be considered a textbook. I now think that we must distinguish more subtly between textbooks and (the wider category) teaching material. We must allow the writer's intentions a defining role; we must not blame the compiler of the telephone directory for failing to make his argument clear. In spite of my ignorance of science I wonder if academic subjects are the best framework within which to organise our thinking about textbooks. May not disciplines which conventionally differ from each other, as Arts and Science, nevertheless require similar styles of thinking and exposition --- rely on similar logical or associative processes --- and therefore need textbooks of a similar structure ?
Table of Contents - Joint Meeting between the Textbook Colloquium and The British Society for the History of Science held on January 10th 1998 at Leeds University.
- The TEXTBOOK COLLOQUIUM