Harry Collins is a sociologist of science who, first from the University of Bath and now from the University of Cardiff, has spent his entire life analysing the processes of generating scientific knowledge. Collins has demonstrated how scientific data can be subject to interpretative flexibility and how the «non-scientific» social context (political, cultural, economic, academic, etc.) contributes to closing scientific controversies, regardless of the empirical or explanatory value of the abandoned lines of research. Collins is usually positioned, incorrectly in my opinion, in a current of the sociology of science called relativist.
With his colleague Trevor Pinch (Cornell University) he has written a series of three informative books, joyful to read («The Golem: What You Should Know about Science», «The Golem at Large: What You Should Know about Technology» and «Dr. Golem: How to Think about Medicine»), where he explains the historical details that have involved different scientific milestones. These texts have a mythical figure in common: the golem. In the introduction to the first book in the series, the authors explain what the golem is and why they have chosen it:
«It seems that science is good or bad at all. For some, science is a crossed knight who is harassed by mystics with few lights while other figures, more sinister, wait to found a new fascism after the victory of ignorance. For others, science is the enemy; our sweet planet, our feeling of the just, the poetic or the beautiful are being attacked by a technological bureaucracy controlled by capitalists who are concerned with nothing but profit.»
And it continues:
«These two ideas of science are erroneous and dangerous. The personality of science is neither that of the noble knight nor that of a ruthless ogre. Science is a golem».
The golem is a creature of Jewish mythology, a man-made humanoid, made of clay and water, with soaps and spells. It is powerful. Every day it is a little more powerful. He will obey orders, do your work and protect you from the ever-threatening enemy. But it is clumsy and dangerous. If not controlled, the golem can kill his owners with his crushing vigor.
The purpose of the authors is to show that science is like a golem, a foolish monster, but saving the monster, that is, making it clear that the errors of the golem are our errors.
While they describe historical scientific controversies (such as the one surrounding the possibility of the chemical transfer of memory in the 1960s, the empirical weakness of the experiments that validated the theory of relativity, the history of cold fusion, or the controversy over the origins of life between Pasteur and Pouchet), the authors point to the non-scientific strategies used by researchers to defend experiments that confirmed their hypotheses or refuted those of others.
For example, the greater the number of variables considered in the experiment (hilarious is the episode describing how the chemical transfer of memory between worms was demonstrated), the easier it is to accuse those who fail to replicate it of failures in the process, i.e., «the more difficult it is to decide whether one experiment actually reproduces the conditions of another.»
For this reason, the replicability criterion for rejecting a hypothesis is always attackable by those who accepted it when they obtained their positive results in the original experiment: the experiment will be defended, it has not been replicated due to lack of skill, bad technique, changes in conditions, etc. In the case of the worms it was something like «you don’t understand the worms».
The case of the chemical transfer of memory between worms also illustrates how the number of experiments reproduced is not enough to persuade the scientific community of a heterodox finding: a negative experiment carried out by influential scientists was enough to reject the hypothesis supported by a larger number of experiments that confirmed it.
Collins and Pinch conclude that, despite the widespread current rejection of the theory of the chemical transfer of memory (the controversy occupied the covers of Nature until 1972), there is no published refutation based on decisive technical evidence. During the controversy explanations were found to discard each negative result but not to explain the positive results:
«We no longer believe in memory transfer, but if we do not, it is only because we have grown tired of it, because more interesting problems have arisen, and because the main experimenters lost credibility. The transfer of memory was never perfectly refuted; it simply ceased to occupy the scientific imagination. The gaze of the Golem turned elsewhere.»
Science is progressive in its methodological foundations and values but very conservative in its real practices. Once an idea is established in the scientific community it is very difficult to change it in spite of the existence of experimental data that contradict it.
That is why the pharmaceutical industry dedicates so much effort today to the construction of consensus through experts, protocols, Clinical Practice Guides, etc: it almost assures its impugnability by many data contradictory to those consensus that appear.
The relativity of the anomaly
In the chapter dedicated to the inconclusive experiments that were given as sufficient to accept the theory of relativity, the authors deal with how the concept of anomaly is interpreted in science:
«The notion of anomaly is used in science in two ways. It serves to describe an inconvenience («we will ignore it; it is nothing more than an anomaly») and to refer to a serious problem («in current theory there are disturbing anomalies»)».
In other words, if the new finding contradicts a theory firmly rooted in the scientific community, the anomaly is merely an inconvenience that «will have an explanation». When that same finding challenges a theory that, for other than strictly scientific reasons, is collapsing, then that finding «is a serious problem».
Thus, when some researchers ignore experimental discrepancies, that decision, always taken for non-strictly scientific reasons, will be used as scientific evidence to continue ignoring those or other discrepancies by other researchers. This process of self-censorship finally creates a scientific consensus that corners the small minority of dissenters.
«Compare this procedure of quasipolitical consensus with the idealized notion of the scientific method,» say the authors, who continue:
«We have no reason to think that relativity is but truth – and a very beautiful, delightful and surprising truth – only that, if that truth came to be, it was the result of decisions on how to endorse our observations; it was a truth born of agreement to agree on new things. It was not a truth imposed upon us by the inexorable logic of a series of crucial experiments.»
And they conclude by comparing science with history:
«In history, as in science, facts do not speak for themselves, at least not exactly.»
No scientific proposal is incontrovertible, that’s what scientific methodology says, but the truth is that when a proposal reaches the status of truth accepted by the scientific community, the «broad consensus» is very difficult to contradict. For this to happen, we will have to change our interpretation of the anomaly from an «inconvenience that will have an explanation» to a «serious problem». This qualitative step never depends on facts but on other non-scientific aspects.
That is why the «truths» that dominate contemporary medicine are so difficult to dismantle: scientific consensuses are increasingly dependent on the economic and political context that they reinforce with their support the leaders, experts and institutions that align themselves; but they tend to centrifuge the «rebels» (as we have seen in the case of Peter Gotzsche).
The controversy surrounding the papillomavirus vaccine is a perfect example of how difficult it is to break down a scientific consensus carefully constructed by the parties that benefit most from that consensus: the pharmaceutical industry. The expulsion of Peter C. Gøtzsche is a political decision that seeks to close some cracks that were beginning to open in the consensus. That is why it is a serious decision in terms of the credibility of science. Gøtzsche distanced from the institution that underpins consensus today, the Cochrane, these will be easier to achieve and defend.
The vicious circle of the experimenter
Collins has been following the scientific process behind gravitational waves for more than 40 years and has published a reference book in 2017. In 1993, the date of publication of the book we are commenting on, «The Golem: what you should know about science», he dedicated a chapter to the waves that Joseph Weber detected in 1969, by the way, with no credibility.
This chapter in the history of science serves to explain to us what the «vicious circle of the experimenter» consists of:
«The experimental work can only serve as a contrast if in some way the vicious circle of the experimenter is broken. In most scientific activity the circle is not broken because the appropriate range of results is known beforehand, providing a universally accepted criterion of experimental quality. Where such a clear criterion is not available, the vicious circle of the experimenter can only be avoided if there is some other means of defining the quality of the experiment; and the criterion must be independent of the result of the experiment itself.»
That is to say, in contradictory novel investigations, as no one knows what the correct result is, it is not easy to discern who has done a good experiment. The result that is accepted will depend on both scientific and social aspects.
Weber, who said he had detected gravitational waves, did not have that support at the time: although no one could reproduce his findings no one could rule out that they were true, but the scientific community finally «closed» the debate in the mid-1970s: large flows of gravity waves do not exist and only incompetent scientists think they see them.
Cochrane, for example, is currently in a kind of vicious circle of the experimenter: the quality criteria of their scientific papers accept as valid only the review of published papers. It is a question, as Jefferson and Jorgensen say, of changing what means the E of «evidence» and that decision, not to accept as valid meta-analysis that they have not access to the raw data of all the experiments, whether these have been published or not, is fundamentally political.
While this is happening, every scientist who points out gruiettes in the consensus will be accused of bad scientist.
Recently, the result of an experiment validating Weber was published worldwide:
«That faint rising tone, physicists say, is the first direct evidence of gravitational waves, the ripples in the fabric of space-time that Einstein predicted a century ago»
What does it mean to be unscientific if science is relative?
Collins and Pinch conclude that «there is no logic to scientific discovery. Or, rather, that if there is, it is only the logic of everyday life. It is not possible to separate science from society».
When something goes wrong in science, the authors remember, there is a tendency to sacrifice those responsible for human error:
«But human error is at the very heart of science, because that heart is made of human activity. When things go wrong, it is not because of human error that could have been avoided, but because there will always be things that go wrong in any human activity. Scientists and technicians cannot be asked to stop being human.»
The disproportionate claims of authority by scientists and technologists are «offensive and unjustified» and, the authors warn:
«The likely reaction born of broken promises could precipitate an even worse anti-scientific movement. Scientists should promise less; they could then better deliver on their promises.»
What do they mean that unfulfilled promises can provoke an unscientific reaction?
The unscientific reaction would have three interpretations:
(1) Attempting perfection in science and rejecting anything that is not scientific is unscientific.
Collins and Printch do not intend to change the way scientists function:
«The social vision of science is of no use to scientists… because it would destroy a science incapable of living up to an ideal.»
That is to say, science works as it works and there is no point in trying to make it work in a different way because that would mean that nothing would work as long as an impossible ideal is not fulfilled.
(2) To doubt everything because science is not perfect is unscientific.
Since things are the way they are, it is stupid to attack science just because they reveal scientific practices that do not fall within the canon of scientific purity.
It is obvious that scientists are not able to settle their controversies and disagreements through better experiments, more knowledge, more advanced theories or clearer thinking. In other words, the processes of scientific advancement are far from being a mere question of experimental facts. But these arguments cannot serve to feed an anti-scientific movement.
(3) Defending the purity of science to attack those who criticize some of its findings is unscientific.
As science is as it is, defending the findings of science as true in the name of its methodological purity and values is also unscientific:
«These defenders of a supposed purity can become a different kind of golem that could destroy science itself»
Rebuilding the relationship between science and society
How then can society evaluate the activity of scientists and the value of science itself?
With what criteria can we establish a science-society relationship capable of assuming all this uncertainty?
Fundamentally assuming that:
(1) It is not possible to separate science and society:
Science is a human activity that functions according to a methodological and procedural ideal that is essential to approach the truth but never sufficient since this tentative truth is finally accepted thanks to non-scientific aspects, that is, political, cultural, economic, etc…
(2) Scientists are neither gods nor charlatans; only experts:
«They have, of course, their area of personal experience but their knowledge is no more immaculate than that of plumbers. Plumbers are not perfect, far from it, but society is not besieged by antiplumbers because we don’t have the option of antiplumbing».
Anti-science is not possible
(3) Science actually works by producing agreements between experts.
Giving scientific legitimacy to any scientist, regardless of the quality validation systems that exist in science, is as bad as allowing the opinions of a group of scientists to become hegemonic and, through political or economic procedures, i.e. power, to prevent dissenting voices from being heard.
We recall once again the terrible damage that the expulsion of Peter C. Gøtzsche has done to understood science in all its political complexity. The cover of his new book is sufficiently explanatory
(4) The best science never produces experiments that make it possible to make categorical public decisions.
Use science to defend political decisions is unscientific
The final phrase of the book is very good:
«All the scientists we have mentioned in this book had clean coats and the title of doctor before their names. They all had overwhelmingly divergent results. There are theorists hovering everywhere who explain and try to reconcile. In the end, however, it is the scientific community that brings order to this chaos and transmutes the clumsy buffoons of the collective science of the Golem into a clear and orderly scientific myth. There is nothing wrong with it; the only sin in not knowing that it always was like this.»
Despite what it looks like, Collins is not a relativist. And that has a lot of merit knowing how he knows his secrets. His vital work, from that very detailed knowledge of how science really works, is a very deep defense of science. In his latest book, he points out ways for science to continue to be one of the fundamental tools of democracy and to resist the attacks of both anti-scientists and scientists. It will be for the next entry.
Abel Novoa is a family doctor and president of NoGracias.