Jean Ladrière gives a definition of scientific rationality: “A rational approach, in the cognitive as in the order of action, is an approach that is accompanied by the demonstration of its validity or its legitimacy, in accordance with criteria which can themselves be accepted as acceptable in the light of a possible criticism “. The fundamental requirement of rationality is the need to justify the reason for its judgments.
“The search for rationality is a timeless approach, but the forms of reason are […] historical and therefore contingent,” says Michel Morange.
In these various questions, we find descriptive and normative aspects.
Context of discovery and context of justification: for a long time, the question of discovery does not belong to epistemology, but to the best of psychology (research of intentions, pre-thought … researcher).
Things have changed gradually: modern epistemology re-questions the body of scientific knowledge acquired and questions the contexts of discovery, validation, communication and teaching of science and research being done.
Production of scientific knowledge
Epistemological questions relate for example to:
- What place should be accorded to intuition, creativity, imagination, the analogy between disciplines, serendipity?
- Which methods? The question of deduction, induction …
- What forms of validation? … We find here the question of explanation, validation …
- There is also the question of the unity of science.
One example that is readily quoted is the astonishment of Greek mathematicians in the fact that the diagonal of the square can not correspond to any irreducible fraction p/q, at a time when only rational numbers were imagined (the irrationality of pi was still unknown). Indeed, we would have then (p/q) ² = 2, that is p² = 2q². This would imply that p² is even, or p = 2k; but in this case p² would have been worth 4k² and the fraction p/q would not have been irreducible, which was contrary to the hypothesis.
The hypothetico-deductive method is regularly regarded as the scientific output par excellence, especially since science is part of the paradigm of applied research, which consists in working to solve problems identified in advance, according to the problem-solving method. However, the approach implemented by the discoverers regularly escapes this very rationalist approach.
Induction consists in relying on the observation of singular cases to justify a general theory; it is the operation which consists in passing from the particular to the general. The problem is to know if it can be epistemologically valid to believe that the universal theories are justified or even verified by the only taking into account of a large number of singular past observations. For example, we have observed that the sun so far rises in the morning. But nothing seems to justify our belief that it will rise again tomorrow. This problem had been deemed unsolvable by Hume, for whom our belief was based on the habit of seeing such cause causing such an effect, which does not presume that this is the case in reality. This unrealistic position was criticized by Kant and Popper, who thought it possible to reach a certain objectivity in empirical theories. Ernest Mach also criticized induction.
There are very varied forms of induction theories ranging from the most naive to the most sophisticated (just like the theory of refutation).
Validation of scientific knowledge
This is the problem of the foundations of scientific knowledge:
- the nature of knowledge: scientific or general knowledge, exclusion from the metaphysics of science … It is in particular the question of the demarcation,
- the validation of knowledge, the question of realism / antirealism, and of course the question of the relation to truth.
Which also leads to the question of relativism.
Nature of knowledge
Historically, this epistemological question concerns more directly the question of how to identify or demarcate the scientific theories of metaphysical theories. In the twenty-first century, there is also the separation between knowledge in general and truly scientific knowledge.
The founding positivist philosophers of the Vienna Circle thought that the only criterion of demarcation that could be valid, (in order to eliminate metaphysics), was the verifiability of singular statements, the only data of the senses capable of allowing the verification of the general theories of science, provided that they are sufficiently numerous and well observed.
For Karl Popper, philosopher of science of the twentieth century and opponent of the theses and the project of the Circle of Vienna, no general scientific theory could ever be established by any form of induction, so be verified. He criticizes the reasoning by induction: The latter has, for him certainly, a psychological value but not a logical value. Many coherent observations are not enough to prove that the theory we are trying to prove is true. On the contrary, a single unexpected observation is enough to refute a theory. Thus, a thousand white swans are not enough to prove that all swans are white; but a single black swan is enough to prove that not all swans are white.
Karl Popper thinks that scientific theories can not be justified, even on the basis of a very large number of empirical observations, they can only be evaluated from tests whose logic consists in trying to test the scientific knowledge (refutation). As a result, a theory can not be “proven” but only considered unvalidated until proven otherwise. From there, we can distinguish:
- theories that can not be refuted (by observation or experience)
- theories that can be invalidated.
On the other hand, he thinks that no scientific theory is logically or even empirically verifiable if one accepts under this term the notion of certainty or verification with certainty. Karl Popper even argues that a theory can only be scientific if it is potentially false (falsifiable), and even false in comparison with the certain truth to which it would claim to be closer. Only potentially falsifiable theories (those associated with experiments whose failure would prove the error of the theory) are part of the scientific field; it is the “criterion of demarcation of the sciences“.
The problem of demarcation (identified as Kant’s problem by Karl Popper) is articulated with that of the justification of theories:
- either according to an inductive method,
- either by a hypothetico-deductive method.
In the field of empirical science, the verification should rather be assimilated to corroboration (Karl Popper), that is to say to a relative and non-absolute form of truth, always dependent on the scientific tests that have been carried out by a community of researchers. Thus, in science, the verification of theories would therefore always relate to tests themselves relating to other previous tests and always improvable, and never absolute.
Made famous by the work of Karl Popper, this term implies the possibility of empirically evaluating the general statements of science through tests. Only theories formulated in such a way as to allow the logical deduction of a particular utterance with the potential to refute them can, for Karl Popper, be considered as scientific and not metaphysical.
But Popper proposes that there are two levels of falsifiability. “Logical” falsifiability and “empirical” falsifiability; knowing that a falsifiable statement from a logical point of view may not be from an empirical point of view. For example, the statement “all men are mortal” is logically falsifiable, but empirically non-falsifiable since no human being could live long enough to verify that a man is immortal.
Karl Popper has always maintained that no empirical falsifiability can be certain, because it is always possible to save a theory of falsifiability by adopting ad hoc ploys. Consequently, for Popper, the criterion of demarcation based on falsifiability must first and foremost be a methodological criterion, since everything would ultimately rest on the decisions of the scientific community to accept or reject the value of a test, a falsifiability or corroboration.
Paul Feyerabend observed the example of the birth of quantum mechanics that often scientific advancement does not follow strict rules. Thus, according to him, the only principle that does not prevent the advancement of science is “a priori everything can be good” (which defines epistemological anarchism – to distinguish from “everything is good” (anything goes), that Feyerabend himself challenged). He therefore criticizes the reductive aspect of the theory of falsifiability and defends methodological pluralism. According to him, there is a very wide variety of different methods adapted to different scientific and social contexts.
Moreover, it challenges the place that the theory of falsifiability places on science, making it the sole source of legitimate knowledge, and the foundation of universal knowledge that goes beyond cultural and community divides. Finally, Feyerabend criticizes the lack of relevance to correctly describe the reality of the scientific world and changes in scientific discourse and practices.
His main work, Against the method. Sketch of an anarchist theory of knowledge, was received very negatively by the scientific community, because it accused the scientific method of being a dogma and raised the question of whether the community should be as critical of the scientific method as compared to the theories that result.
The epistemological question concerns:
- the organization of knowledge in different corpora, in different disciplines …
- interdisciplinary, transdisciplinary links …
- principles chosen at all organizational levels
- Theories, models …
- Hypotheses, thought exercises …
- Theorems, laws, principles …
Evolution and dynamics of knowledge
The epistemological question concerns the nature of the dynamic process of scientific change:
- Does science advance in leaps? Continuism and discontinuism,
- Is science progressing only from the inside? or do non-scientists advance Science? internalist and externalism.
- This brings back to the problem of relativism.
Continuism and discontinuism
Bachelard and the “epistemological obstacle”: Gaston Bachelard defines, in 1934, in an article entitled The formation of the scientific spirit, the latter as being “the rectification of the knowledge, the extension of the frames of the knowledge”. For him, the scientist must strip himself of all that constitutes the “internal epistemological obstacles”, by submitting to an interior preparation so that his research progresses towards the truth. The notion of “epistemological obstacle” is what makes it possible to pose the problem of scientific knowledge: it is from the moment when it is overcome, giving rise to an “epistemological break”, that one reaches the purpose sought. Obstacles are, for Bachelard, not only inevitable, but also indispensable to know the truth. This one indeed never appears by a sudden illumination, but on the contrary, after long gropings, “a long history of errors and wanderings surmounted”.
Bachelard denounces the opinion left by empirical experience and its influence on scientific knowledge: “the real is never what we could believe, it is still what we should have thought,” he says. “Science formally opposes opinion: opinion does not think, it reflects the need for knowledge”. Scientific knowledge will consist in constantly returning to the already discovered.
Putting emphasis on discontinuity in the process of scientific construction, Thomas Samuel Kuhn discerns relatively long periods during which the research is described as “normal”, that is to say, it is part of the lineage dominant theoretical paradigms, periods during which brief and inexplicable changes constitute a true “scientific revolution“. The choice between paradigms is not rational. This posture implies that each paradigm solves some problems, and hence the paradigms would be immeasurable.
Internalism and externalism
The internalist vision takes into account only the history of scientific ideas, from discovery to discovery, regardless of any context: scientists are a world apart, progressing independently of the rest. Science feeds on itself. It is thus possible to understand the history of science without referring to the historical, social, cultural context. In this vision the important thing is the stages of progression of scientific history.
On the contrary, the externalist vision makes the science dependent on economics, psychology, and so on. This leads to different consequences depending on the context.