This paper examines Claude Bernard's arguments against the misuse of statistics in nineteenth-century medicine and his advocacy for rigorous scientific experimentation. Drawing on Bernard's An Introduction to the Study of Experimental Medicine, the paper outlines his distinction between "determinate" and "indeterminate" sciences, his critique of the law of large numbers as applied to clinical practice, and his principle of experimental determinism. The paper also reviews two secondary sources — Steven Darian's Understanding the Language of Science and a journal article by LaFollette and Shanks — that contextualize Bernard's influence on medical science through the twentieth and twenty-first centuries, including his controversial reliance on animal experimentation.
This paper demonstrates the technique of source triangulation: presenting a thinker's own arguments from a primary text, then layering in secondary scholarship that both supports and complicates those arguments. The LaFollette and Shanks section, in particular, shows how to engage critically with a legacy rather than simply endorsing it.
The paper opens with an exposition of Bernard's core critique of statistical medicine, then develops his positive alternative (experimental determinism). It then turns to two annotated secondary sources — one affirming Bernard's contributions and one raising ethical concerns about animal experimentation — each introduced under its own heading. The bibliography follows standard Chicago footnote formatting throughout.
Claude Bernard is regarded as one of the first physicians to embrace scientific experimentation as a means of defining medicine. He believed that people who conducted statistical experimentation and stated statistically derived numbers without a definite purpose were in error. His belief was that there should always be a clear goal that an experiment was designed to pursue. He gave several examples to illustrate the ridiculousness of merely producing numbers for their own sake. He cited the study of spinal root nerves, which found that the nerves were sometimes sensitive and other times were not. He argued that this experimentation yielded nothing of value because it specified nothing.1 He next raised the example of an individual who conducted a series of operations for the same condition and reported a 40% mortality rate.2 Bernard again contended that this result meant nothing because it did not indicate under what conditions a patient would be expected to live or die upon receiving the operation. His central complaint, therefore, was that so-called medical scientists did not provide true scientific findings that could be either verified or refuted — their method was fundamentally lacking.
It is the tendency of people to believe in their own observations and to trust the statistics they have compiled. The doctor who reported a 40% fatality rate for his surgery may well have understood what made one patient live and another die, but he did not share this understanding with the broader medical community — at least not according to Bernard. This misuse of statistics occurs today just as much as it did in the past. One researcher offered the example of a recent United States census that found a "high correlation between location of churches and violent crime."3 While most people intuitively recognize that no causal relationship exists, the statistical correlation can still be drawn. The reason is that people sometimes assume that because two things correlate they must be related. It is poor science not to look further and discover exactly what underlies a seemingly meaningful statistic. In reality, churches are so numerous that many different phenomena — including violent crime locations — will appear to correlate with them simply because of their prevalence. Bernard was making much the same point. He examined the so-called scientific experiments being conducted with clinical cases and realized that because physicians were not applying the correct rules of scientific inquiry, they were reaching false conclusions, or at the very least conclusions that meant nothing when applied to practice.
Bernard's first argument is that statistics cannot be properly applied to medicine because medicine should be what he calls a "determinate" rather than an "indeterminate" science.4 He acknowledges that much of medicine is conjectural, but insists that physicians should be working to make it less so — and the way to accomplish this is through scientific experimentation rather than by treating medical practice as an art form.5 He also addresses the use of the law of large numbers.6 His position is that this principle is appropriate in an indeterminate science, where large numbers are often necessary to make sense of the data. The law of large numbers functions as an argument against small sample sizes, and it applies to most scientific disciplines where large datasets are accessible. However, Bernard argues that it does not apply to medicine to any great degree. Most of the time, a physician is not dealing with a large number of cases. The physician must examine what is available and must not shy away from experimentation simply because there are not enough patients with a particular disease or disorder to constitute a statistically large sample.
Bernard makes the case that statistics are useful to physicians insofar as they reveal the indeterminate, but that this cannot be a stopping point.7 The solution is to identify these indeterminate problems and, through experimentation, transform them into determinate ones. He based his entire book An Introduction to the Study of Experimental Medicine on the principle of "experimental determinism" rather than "statistical conjecture."8 His conviction in the superiority of this system rests on the belief that it is the only method capable of yielding "absolute law" in medicine.9 He held that once a law has been established through experimentation, one can no longer retreat to conjecture — it must be accepted as a certainty. In a meaningful sense, his concern was ultimately for the patient: his belief in rigorous experimentation made medicine more exact and almost certainly saved many lives. He was striving to move the profession closer to scientific enlightenment and out of what he saw as the dark ages of medical practice.
Always verify citation format against your institution’s current style guide requirements.