Alan Chalmers wrote this book (first published in 1976) because there was no suitable introductory text for undergraduate studies in the history and philosophy of science. It became something of a publishing phenomenon, translated into numerous languages, with regular reprints and a revised edition in 1981. This permitted the author to afford some of the affectations of the country squire, including Australian bush hats, as depicted on the back cover of this second revised edition.
Several generations of science students in New South Wales risked permanent spinal deformity by carrying the massive science "Messell" science texts to school. In addition to cricked backs, it seems that Harry Messell attempted to implant an equally crabbed view on the way that science works. The text advised that "Science advances in a definite pattern. First and foremost scientists must make observations. These observations must be careful and accurate; and the results of more and more observations accumulate."
The students and their posture have been emancipated from the Messell text by the educational powers that be, and Alan Chalmers, following Popper and others, has attempted to do the same for their attitudes to the role of facts in the scientific enterprise. At the same time, if students learn what Chalmers has to teach, they should be immune to the sophistry of the Creation Scientists and to the siren song of the dreaded post-modernists and deconstructionists.
Chalmers arrived in Sydney as a more or less orthodox Popperian but since then he has become more critical and he has directed some blows at the old master. One of the aims of this review is to suggest that these blows missed the mark but they have contributed to the misreading and under-appreciation of Popper's chievement. The question has to be asked, whether the label "falsificationism" is appropriate for Popper's views.
Demolition of the pre-Popperian orthodoxy
The preliminary chapters are devoted to a close scrutiny and demolition of the old orthodoxy in the pre-Popperian philosophy of science,
Chalmers notes the common belief that the authority of science depends on the way that it 'is derived from the facts.' However, as many research students discover when they turn from the bench to start writing up their results, it is very misleading to hope that accumulated observations (a la Messell) will turn into general principles or theories (or a thesis). The facts turned up by 'nature study' and collectors are virtually worthless unless they are handled by someone like Charles Darwin who tirelessly and imaginatively sought explanations for the forms and varieties of organisms that he encountered. The facts need to be located in a theoretical framework, and it is general theories which give science its explanatory and predictive power.
Still, scientific theories are supposed to be based on facts and confirmed by facts, and for a long time the official scientific method was an alleged process of induction, whereby scientific knowledge starts with the unbiased observation of the regularities which exist in the world around us and is finally warranted or verified by inductive proof. Chalmers explains with meticulous care how and why inductive verification and warranting does not work. Moreover he explains that it is not necessary to account for the growth or rationality of scientific knowledge.
One of the problems with the observational origin of theories is the abstract nature of advanced scientific theories. Electrons, wave currents and force fields are simply not accessible to observation, nor are the principles of natural selection or the laws of supply and demand. Equally embarrassing is the logical problem of induction. However many black ravens you observe, there is no way to prove that all ravens (in the universe) are black. Popper is probably the arch-enemy of induction in recent times, though it was all said by Duhem at the turn of the century and by Hume a long time before that.
Popper's contribution, sadly misread
Popper's contribution to advance the debate was a revamped version of the hypothetico-deductive method foreshadowed by Jevons, Whewell, Pierce and the French physiologist Bernard. Knowledge advances by a problem-oriented process of conjecture, followed by rigorous testing of tentative solutions. Then we select the best among the competing theories by its ability to survive the process of testing. Falsified theories need not be discarded because they may stage a revival, they may have instrumental value and they may persist as components of a larger structure. This approach is most usefully called "evolutionary epistemology" and it has been revived in modern times by Popper and others to liberate the philosophy of science from its 20th century obsession with physics. Peter Munz was a helpful commentator. Popper's theory in the large is a theory of conjectural objective knowledge and not a theory of justified belief. His views on falsification relate in a specific and narrow manner to the way that evidence can be used to contribute to the critical appraisal of theories. "Falsification" is not a full-fledged theory of epistemology or methodology, merely a critical corrective to the errors of "verificationism".
Unfortunately, by a serious error that is perpetuated by Chalmers, Popper's contribution was labelled "falsificationism" because it was advanced in the 1930s as a rejoinder to the logical positivists who wanted to use factual verification as a criterion of meaning. The label was unfortunate because when perceived difficulties emerged with falsification (for example it could not be decisive due to the Duhem problem and the uncertainties of observation), they were used as the rationale to prematurely eliminate Popper from the main game in the 1970s.
That was partly due to the raging success of Thomas Kuhn's The Structure of Scientific Revolutions which captured the sociological spirit of the age with a beguiling account of the diffusion of intellectual innovations in the scientific community. It can be argued that Kuhn's paradigm theory formed a symbiotic relationship with inductivism because inductivism tends to produce "normal scientists" and Kuhn's theory legitimates normal science. Thus while paradigm theory appeared to be radical (and gained a lot of support for that reason) it is really conservative because it gives no indication of the need for rational criticism of framework assumptions and no guidance as to how such a project of critical revision might be pursued.
Structures, programs and paradigms
On the account provided by Chalmers, the problems with falsification revealed some limitations of examining theories in isolation because observations, and especially experimental results, involve numerous assumptions, including assumption about the function of the equipment. Chalmers next proceeds to examine the treatment of theories as structures by way of Kuhn's paradigms and the 'methodology of scientific research programmes' proposed by Popper's junior colleague, Lakatos. It is worthy of note that Popper developed a theory of metaphysical research programmes during the 1950s though it was not published until Quantum Theory and the Schism in Physics appeared in 1982. So far very little work has been done to exploit the potential of Popper's work on programmes but it challenges the self-mutilating ban on the serious discussion of metaphysical (untestable) ideas that has prevailed among positivists from the time of Hume. Because untestable ideas of a programmatic nature cannot be eliminated from science the main effect of the ban was to place such ideas beyond criticism and to make positivists the slaves of whatever metaphysical ideas they picked up along the way.
One of the most valuable sections of this book is the scrutiny of Kuhn and here Chalmers identifies profound ambiguity. Kuhn insists that there is evolutionary progress in science, on the other it is hard to reconcile this view with the 'gestalt switching' process that is supposed to occur as people move from one paradigm to the next. Chalmers points out that the "gestalt switching' and the notion of incommensurability of paradigms can be put aside if one makes some fairly uncontroversial assumptions about the objectivity of scientific theories. In a previous publication, Science and its Fabrication, Chalmers has argued in a similar vein against the strong proponents of the sociology of science.
Turning to Lakatos, with his notion of a 'hard core' of a research program which has to be protected from falsification by deflecting criticism to other 'non core' elements of the program, Chalmers finds that there is no satisfactory guide to the selection of theories to be protected from the rigors of criticism. The radical views of Feyerabend are carefully dissected and Chalmers concedes that there are probably no universal and timeless standards in the philosophy of science. However this is not a concession that 'anything goes' because it simply mirrors the situation in science itself.
"If we have a conception of science as an open-ended [unended?] quest to improve our knowledge, then why cannot there be room for us to improve our methods and adapt and refine our standards in the light of what we learn" (p.162). Chalmers illustrates this with a beautiful account of the arguments used by Galileo to convince his contemporaries that the telescope might legitimately be used to augment the evidence of the naked eye.
The Bayesian turn and the new experimentalism
This edition has new material treating two interesting developments that have attracted attention in recent years. One is a form of Bayesian probability theory. This starts from the recognition that the long running program of the positivists to pin objective probabilities onto theories is a dead duck (or perhaps a dead raven). This was attempted in works by Carnap and Hempel which for David Stove "represent far more progress, in an area of the first intellectual importance, than the entire history of the human race can show before" (Anything Goes, p. 168). The Bayesians use information about the subjective assessment of rival theories by the scientific community, add the most recent and relevant information that is available and then employ a formula to calculate revised probabilities for the rivals. This approach has huge attraction for people of with a numerical turn of mind but there are many complications to work out before its claims can be taken seriously.
The other is the 'new experimentalism'. In some ways it is not all that new because it was stated in a rudimentary form by Dr Cheesman in his Chemistry I lectures at the University of Tasmania in the mid 1960s. (He was notorious on the campus for his pioneering studies in human olfaction. His assistants would accost students in the vicinity of the chemistry department and offer boiled sweets in return for poking their noses over a row of test tubes to record if they could detect any smell). After walking us through the evolution of ideas on the structure of the atom, past the friendly and helpful solar system atom to the rather darker and stranger quantum atom he pointed out that all these theories may come and go but some things do not change. Mix chemical 'a' with chemical 'b' and you obtain a blue precipitate. Place various metals in the bunsen flame and you will see their characteristic colours. The dehydrated salts of this particular group turn green when water is added. And so on.
Alan Chalmers explains how the role of experimentation has been rehabilitated in a very impressive fashion by recent writers who convincingly argue that 'experimentation has a life of its own' (due to the reproducibility of certain types of results and phenomena) against those who celebrate the primacy of theory and relegate the experimental side of science to a secondary role. This has two benefits for people who have a realist view of the world (it is actually out there, it is not a figment of our imagination) and who consider that the progress in science consists of better explanatory theories. The new experimentalism can be used to support the realist view of the world (by demonstrating our capacity to manipulate it) and it also provides the kind of observational base that is required to make testing effective, contra the critics who say "if every damn thing is up for grabs, how can you have a convincing refutation any more than a convincing verification?".
Even allowing for my concerns about the representation of Popper's contribution, this book that can be highly recommended for anyone who wants to obtain a firmer grasp of one of the most important yet simultaneously least understood developments of all time. That is, the spectacular successes of scientific research. The writing style is clear, engaging and unpretentious. The book is packed with episodes from the history of science so that there is a great deal to be learned about science itself in addition to the other lessons that Chalmers has to convey.
What is this Thing called Falsificationism?
Alan F Chalmers, What is This Thing Called Science, University of Queensland Press, 1999. Second revision.
