Sunday, 18 September 2011

The thin line between science fact and science friction: Antisocial meateaters, Psi and other forms of empiarrhea

There is a beautiful paragraph in the history of science Michael Seevinck introduced me to. I immediately incorporated it into talks and papers. It is a letter by Albert Einstein he wrote to his friend Maurice Solovine on the 7th of April 1954 in which he explains his philosophy of science. Einstein distinguishes between two types of theories: Theories of construction and theories of principles. Theories of construction are found in a science that is immature. Those theories basically just describe a phenomenon in the empirical record (the data collected in experiments) using fancy language. My favorite example is the scientific notion about reading aloud written text. This is involves converting graphemes into phonemes according to scientific theory. What are graphemes? Well, they represent letters. What are phonemes? They represent sounds. So reading aloud is actually converting letters into sounds... Duh! Of course it is! Why do you need graphemes and phonemes to explain it?

Figure adapted from van Dongen (2010)

A theory of principles is completely different. It embarks from a set of postulates and axioms (A). It derives laws from those claims (S) and seeks to test how accurate those laws are. Einstein understood that a theory of principles “pays for its higher logical unity by having elementary concepts […], which are no longer directly connected with complexes of sense experiences.”. To create a theory of principles means sacrificing explanatory power of a theory. We cannot explain the way the universe works in terms of sensory experience. This was not just a coincidental oddity for Einstein, he firmly believed the fundamental notions of physics cannot be induced from experience and they cannot be justified a priori on the basis of our faculty of knowledge. (This strongly opposes Kant’s notion of incomensurability between Biology and Physics ).

How do scientist get to those systems of axioms and postulates then? Einstein is very clear about this and claims there is no logical route to arrive at A directly from E (direct experience). A formalism is created due to an intuitive leap, a creative act in which a conjecture is raised to a postulate. The arrow that doesn’t quite touch E and jumps upwards to A represents this creative act. The only route that is deducible by logic and reason flows from the formalism in A to the derived laws S and from those laws back to E. Theories of construction on the other hand infer “laws” from complexes of immediate sensations (sensory phenomena) in E. These laws are crudely glued together to construct ad-hoc theories that, in a way, save those phenomena for the constructed theory so far.

Well, what can I say... Sometimes it just doesn't feel entirely right to say "I told you so".

I am not saying that I expected anything like the fraud committed by social psychologist prof. Diederik Stapel, but the entourage, the context, the circumstances that allowed such fraud to go unnoticed and the responses of the universities and his colleagues... well let's just say that I am not entirely surprised it happened.

The reason for me not being surprised is due to the fact that I have been trying to figure out what kind of philosophy of science social psychologists actually adhere to, by studying the argumentation in the papers they write and the way theories are evaluated by the social psychology community. I have for instance been analyzing the paper by Darryl Bem who claims to have found evidence for the existence of Psi. More importantly perhaps, with the help of some bright students I have been trying to replicate some of the famous social psychology experiments. We failed to do so, but we also think we know why they did find something and we didn't, you know, what science is all about. No reason to be suspicious. I also didn't expect fraud because I know social psychologists are masters at inventing ingenious experiments. Really, they generate beautiful datasets by tweaking and fine tuning their stimuli and experimental procedures and they will continue to do so until they find the right conditions to support their hypotheses.


Yes, indeed, that's not exactly the way science should work. I thought that was the problem. They have no real theories, but just produce an ever increasing empirical record of interesting and sometimes very cute phenomena that have no say whatsoever on the level of the major psychological theories about cognition, behavior or development. I call this enormous structureless record of so called scientific facts: empiarrhea (see Caught in the undertow).

Now when I say "them" or "they", to refer to social psychologists, of course I am generalizing: No I do not believe all social psychologists are pretesting designs until they work. What I do believe however, is that psychologists in general, but social psychologists in particular, should change the way they theorize about reality. "It is the theory that decides what we may observe" said Albert Einstein to Werner Heisenberg and he was right.