Chemistry and the Scientific Method

During the past decade, philosophers of science have been writing in detail about the practice of chemistry. Two new journals have been formed, several monographs have been written, and many papers have been presented at conferences focusing on the unique philosophical issues raised by chemistry. Much of this fruitful research has concentrated on what makes chemistry distinct from biology and physics, rather than on more general methodological issues. But this research also is concerned with the structure of theories, the relationship between theory and evidence, and the proper form of scientific explanations.

"CHEMICAL DISCOVERY" is his attempt to ameliorate this overemphasis on logic, bringing extant accounts from philosophy of science and chemical practice closer together. Its purpose is to teach philosophers more about the history of organic chemistry and to teach chemists about some of the important ideas and themes in 20th-century philosophy of science.

Berson's goal for this book is similar in many ways to his earlier book, "Chemical Creativity," which provided his reflections on the of ideas of Woodward, Hückel, Meerwein, Diels, Alder, and others to a wide audience of scientists, historians, and philosophers.

"Chemical Discovery" opens with two purely philosophical chapters that introduce influential accounts of theory testing. The first chapter explains the simple inductivist view of theory testing associated with Francis Bacon, while the second chapter outlines Karl Popper's falsificationist account. The subsequent chapters relate a series of interesting episodes from the past 150 years of organic chemistry, including Kekulé's proposed structure of benzene, attempts to understand the racemization of camphene, Wöhler's synthesis of urea, and research into the biogenesis of alkaloids.

These chapters are meant to provide case studies that "test the validity of philosophical proposals." Berson is aware that some readers will find the idea of testing philosophical proposals about how science ought to be practiced with case studies of how science actually is practiced to be a dubious philosophical method. If I understand it correctly, Berson's response to this worry follows philosophers such as Larry Laudan and Philip Kitcher, who advocate "normative naturalism"--norms should be drawn from paradigm examples of exemplary scientific practice. Berson's choice of historical episodes follows this pattern, as he highlights some of organic chemistry's most interesting moments.

Many of the philosophical threads running through Berson's book concern Popper and his account of theory testing. The emphasis on Popper reflects another unstated aim of the book. Berson apparently believes that Popper's ideas have had an undue influence on the philosophical thinking of some scientists. Many of Berson's examples are designed to show how Popper's ideas do not fit the basic facts of some of organic chemistry's greatest success stories.

Consider a very simple example of hypothesis testing. Say we hypothesize that all sodium salts burn with a yellow flame. In order to confirm this hypothesis, we would examine as many types of sodium salts as we can get our hands on and see if they all burn with a yellow flame. We can think of this process as one of conjecture and confirmation: We make a hypothesis, work out the logical consequences of the hypothesis, figure out what kind of evidence would have to be found if the hypothesis were true, then look for this evidence. If we repeatedly find such evidence, then we have confirmed our hypothesis and we are warranted in believing that it is probably true.

POPPER BELIEVED that this process reflects a dubious logic: induction. As many philosophers pointed out before him, it is impossible to give a deductive justification for the outcome of an inductive inference. Whereas others accepted this limitation of inductive forms of inference, Popper proposed a radical new way of looking at the relationship between theory and evidence. While seeing any number of sodium compounds burn with a yellow flame can only give inductive evidence for the hypothesis that all sodium compounds burn with a yellow flame, seeing just one compound burn with a purple flame will deductively falsify the hypothesis. Thus, Popper argued that science should be properly understood as consisting of conjecture and refutation, not conjecture and confirmation. In his view, scientists can never confirm theories; they can only try to refute them.

Popper's view might seem commonplace and obviously true until we attend to one important detail. Popper did not believe that immunity from repeated attempts at refutation rendered a theory more well confirmed than its rivals. He believed that scientists should always adopt the same tentative attitude toward theories that had not been refuted. Long-standing resistance to refutation is not a sign of truth; it is merely a sign that a theory has not yet been refuted.

Berson ably discusses this aspect of Popper's account and several others, including the requirement that theories have no internal contradiction, since this amounts to self-refutation. Berson's treatment of this topic is intriguing, especially because the requirement that theories have no internal contradictions is really quite general and not limited to Popper's philosophy.

Berson's discussion is also intriguing because the "no contradictions" doctrine has rarely been discussed in the philosophical literature. In what I think is the most fascinating chapter of the book, Berson argues that, if followed seriously, the requirement against logical contradictions in a hypothesis would have prevented Kekulé's correct depiction of the structure of benzene.

"The history of organic chemistry ... shows that even though [Kekulé's] theory was not really understood by most organic chemists of the 19th century, it was applied nearly everywhere."

KEKULÉ PROPOSED the cyclohexatriene structure for benzene in several publications in 1865 and 1866. He argued that this structure is a perfectly symmetrical hexagon and that all hydrogen atoms are equivalent. Kekulé believed that the hexagonal structure explained how benzene forms only three disubstituted isomers. While this explanation was received favorably by much of the chemical community, three of Kekulé's former students--Baeyer, Körner, and Ladenburg--remained skeptical. Ladenburg pointed out that while the hydrogen atoms themselves are equivalent in cyclohexatriene, pairs of hydrogen atoms are not structurally equivalent. Some of the atoms are separated by a C5C double bond and some are separated by a C2C single bond. Unlike chemists today who are used to drawing aromatic benzene rings by clicking their mouse, a cyclohexatriene structure did not automatically suggest delocalization in 1866.

In recounting the story, Berson observes the following: "[Kekulé] was not interested in a compromise which gave up equivalence of pairs of hydrogens in order to retain cyclohexatriene. In fact, the unswerving directionality and force of his presentation leave no doubt that he considered both parts of his theory to be essential."

Berson's narrative goes on to recount how, in 1872, Kekulé developed a response to his critics. But this response appeared highly ad hoc, thus violating another of Popper's strictures about theory construction.

Commentating on the reception of Kekulé's ideas, Berson makes the following observations: "The history of organic chemistry ... shows that even though this theory was not really understood by most organic chemists of the 19th century, it was applied nearly everywhere. Chemists of the time quickly suppressed any remaining distaste, swallowed this awkward bolus, and pressed ahead. Their subsequent achievements under the aegis of the theory vindicated their action."

Through insightful analysis of this and other historical examples, Berson's important book will undoubtedly serve as a useful introduction to some of the core philosophical issues about theory testing to students and practitioners of chemistry. It may also serve to widen the horizon of scientific doctrines and activities discussed by philosophers of science. Possibilities for the latter are especially worth mentioning.

In the chapter titled "Some Non-Refutative Motivations in Science," Berson maintains that there is more going on in organic chemistry than attempts to falsify theories. One motivation highlighted in this chapter is organic synthesis. Berson is certainly correct to observe that there is an overemphasis on theory testing among philosophers, but I think that he focuses his criticism too narrowly against Popperians.

He writes: "Whatever one's opinion of Popper's methodological system, that it is incomplete seems beyond dispute. Many advances in science do not originate in attempts to falsify a theory." While I completely agree with this claim, it should be noted that philosophers of science from nearly every school of thought have focused far too much on attempts to confirm or falsify scientific theories. No philosophical account of the nature and practice of chemistry could be complete without accounting for the roles of synthesis and analysis, as well as theory and hypothesis testing.

Despite its many good qualities, "Chemical Discovery" is marred by an unconventional choice about which philosophical positions to discuss. Berson admits that he is not producing an account of the "up-to-the-minute themes that occupy philosophers today." I don't think it would be appropriate to expect that of this book. However, given the topics Berson focuses on, such as theory testing, his choice about which philosophers and which philosophical accounts to cover is somewhat idiosyncratic.

For example, in his discussion of accounts of theory testing, Berson draws a contrast between simple inductivist views and falsificationist views. Simple inductivist views would be like the one attributed to Bacon--a good scientist or community gathers an enormous amount of evidence, then looks for patterns in this evidence. Usually, Bacon's position is contrasted with what philosophers of science call the "hypothetico-deductive" account of theory testing. According to this account, a hypothesis is conjectured, its logical consequences are worked out, and then evidence is gathered to test these logical consequences. When data match the logical consequences, we take the theory to be (partially) confirmed.

POPPER'S RADICAL POSITION is best understood as an alternative to both of these positions. Yet Berson doesn't mention the hypothetico-deductive account of theory testing at all. His discussion of theory testing could have made more contact with the contemporary literature had he discussed this view because many of the accounts of confirmation currently defended by philosophers of science are closer to hypothetico-deductivism than to either Bacon's or Popper's view.

"Chemical Discovery" is a fascinating book that should be of interest to both chemists and philosophers. The book could profitably play a key role in a seminar about the methodology or philosophy of chemistry, a use to which I hope to put the book myself. It is philosophically stimulating and, despite its sometimes unusual emphasis, makes many valuable points. I hope that the book will stimulate similar future projects, perhaps collaborations between chemists and philosophers of science.

Michael Weisberg is an assistant professor of philosophy at the University of Pennsylvania, where his research focuses on philosophy of science in the fields of chemistry and biology.