“Citizens, please think about this: do we learn to think the same way we learn to believe?
Believing is the easiest thing in the world. And thinking is the hardest thing in the world.
(Ferdinand Buisson 1841-1932)
Beyond the obvious and their appearances, the role of science is to seek objective arguments to verify the extent to which our observations reflect reality. The aim is not to claim to have access to reality, but to produce hypotheses and representations (models) that bring us closer to reality. These theories are never definitive, and must be guided by the experimentation that will enable them to continue to evolve. Georges Canguilhem describes Claude Bernard’s Introduction à l’étude de la médecine expérimentale as “a long plea for the use of ideas in research, it being understood that a scientific idea is a guiding idea, not a fixed idea. This complementarity between the two tools of observation, which seeks to establish facts, and theory, which seeks to predict facts, is the foundation of the experimental approach.
The chemist M-E Chevreul (1856) describes this search for “the truth” in similar terms: “A phenomenon strikes your senses; you observe it with the intention of discovering its cause, and to do this, you assume one whose verification you seek by instituting an experiment. The reasoning suggested by the observation of phenomena thus institutes experiments (…), and this reasoning constitutes the method I call experimental, because ultimately experiment is the control, the criterion of the accuracy of reasoning in the search for causes or truth”.
“The observer's mind must be passive, that is to say, silent”... “When you do research, you have to have a hypothesis in mind. Without it, you're walking on a wild goose chase. It's not having a preconceived idea : it's just having a plan”. Claude Bernard
According to Claude Bernard, we must know how to doubt, but this does not mean being skeptical: the scientific approach is rooted in confidence. Rigor stems from doubt: seeking reasons to doubt rather than reasons to believe means validating and relying on a set of reproducible procedures, applied with care and consistency…
Naive realism and the illusion of correlation
Naïve realism and the illusion of correlation are the two main pitfalls in our apprehension of reality.
The first impression given by our perceptual systems is that we perceive the external world as it is. We identify the contents of our perception with the objects we perceive. We therefore simply have the impression of perceiving reality, whereas we really only have access to the impressions the world produces on our sensory receptors and the sensations we construct from these bodily signals. This illusion of perceiving reality has been denounced by philosophers since Aristotle, and is referred to as Naive Realism. In the scientific context, this illusion can hinder our ability to design experiments to verify what we perceive as self-evident (e.g.: the sun revolves around the earth). In the context of communication, the impression of perceiving reality as it is is an obstacle to our ability to take into account the point of view of others. This illusion of perceiving reality is solidly supported by other cognitive biases, such as the illusions of objectivity or rationality.
Our thinking is also subject to coincidence detection: we are sensitive to synchronous events, and automatically generate causal associations. In the same way that we believe we perceive reality, we easily believe we directly perceive causal relationships.
A common example of this correlation illusion is the idea that childbirth is more frequent on full-moon days. This belief has no serious scientific basis (cf. http://lazarius-mirage/lune) and has never been backed up by serious statistics (e.g. Rossetti 1990), but continues to be irresistibly popular with maternity wards and the general public.
As positive associations (tonight we’re having a lot of births and it’s a full moon!) are more striking than negative ones (full moon evening without abundance of births, or abundance of births without a full moon), they dominate the overall impression and therefore contribute to generating or maintaining a misconception (cf. the family of confirmation and filtering biases).
And, of course, we all too easily transform coincidence into causality: we seek to understand the world by making sense of it. Since chance doesn’t always make sense, we tend to replace it with beliefs.
The correlation illusion consists in perceiving a relationship between two unrelated events, or in exaggerating a relationship that is actually weak. For example, associating a particular characteristic of a person with the fact that he or she belongs to a particular group, when the characteristic has nothing to do with the fact that he or she belongs to this group. The aim of the experimental approach is to extend observation by subjecting the hypothesis it has given rise to verification through an approach that seeks to establish the suspected causal relationship.