The scientific method refers to the set of canons guiding or having to guide the production process of scientific knowledge, whether observations, experiments, reasoning, or theoretical calculations. Very often, the term “method” engages the implicit idea of its uniqueness, both with the general public and some researchers, who moreover sometimes confuse it with the only hypothetico-deductive method. The study of researchers’ practices reveals, however, such a great diversity of approaches and scientific disciplines that the idea of a unity of method is made very problematic.
This observation, however, should not be understood as a form of epistemological anarchism. If the question of the unity of the method is problematic (and this problem will be addressed in more detail below), it does not call into question the existence of a plurality of methodological canons which are imposed on the researchers in their scientific practices.
Discovery and theory
This brief introduction situates the basic process of the scientific method during the transition from one theory to another. This scenario is detailed in Thomas Kuhn’s The Structure of Scientific Revolutions.
In the framework of an established theory, a researcher may observe an anomaly or explore new experimental conditions, for example by using other instruments. He realizes his own experiences and repeats them first to validate them himself, then to document and publish them. Each of these scientific publications constitutes a basic observation. This is the experimental method, the beginning of a scientific discovery.
(The scientific method.)
When several researchers have repeated experiments on the same phenomenon with various variations (of experimental conditions, measuring instruments, types of evidence …) these basic observations mutually confirm each other without any limit. neither precise nor of a particular moment which validates them, it is the appreciation of several researchers which leads to a progressive consensus. The experiments and elementary observations then form a confirmed body of evidence of the existence of the phenomenon.
As a result of this scientific discovery, or at the same time, researchers are trying to explain the phenomenon through hypotheses. A hypothesis, to be scientifically admissible, must be rebuttable, that is, must allow experiments that corroborate it (confirm it) or refute it (do not know it).
These are the repeated and confirmed proofs by other researchers, various and varied, that support a hypothesis. It is its acceptance by many researchers that leads to a consensus on the explanation of the phenomenon. The acceptance of the hypothesis can be manifested by the quotation of previous works which often serve as validation markers. It thus becomes the new consensual theory on the phenomenon considered and enriches or replaces a theory previously admitted (or several, or in part).
Anomalies will appear little by little and a new cycle will begin.
Evolution of the notion
The scientific method, that is, the way to access knowledge, has been the object of philosophical attention since ancient times. In most cases, it is a matter of deciding on the right scientific method, which then becomes a normative notion.
It is important to distinguish these philosophical reflections from the actual practices of scientists. However, some are not always without influence on others. The canons enacted by Aristotle were thus for centuries at the heart of the “scientific” approach (if we accept the anachronism emphasized by the quotation marks).
Aristotle: Aristotle (384 BC, 322 BC) is the first to reflect on the development of a scientific method: “We suppose ourselves to possess unqualified scientific knowledge of a thing,” he writes, ”as opposed to knowing it in the accidental way in which the sophist knows, when we think that we know the cause on which the fact depends, as the cause of that fact and of no other, and, further, that the fact could not be other than it is.” (Posterior Analytics, I, 2). If he favors the idea of a deductive science, he recognizes a place for induction: that does not mean that by the repeated observation of this event, we can not, by pursuing the universal, arrive to a demonstration, for it is from a plurality of particular cases that the universal is disengaged.
Alhazen: Ibn Al Haytham (965-1039), better known in the West under his Latinized name Alhazen, considered the modern father of optics, experimental physics and the scientific method. He can be seen as the first theoretical physicist.
Roger Bacon: A Latin translation of part of Alhazen’s work, Kitab al-Manazir (optical book), has had a great influence on Western science. Including Roger Bacon (1220-1294), a famous English scholar, who took over and quoted his work.
René Descartes: In 1637, Descartes published the Discourse on Method which contains his explanation of the scientific method, that is to say, a step-by-step approach to reach a truth. In interpreting its approach, it can be divided into four stages:
- Obvious object (subject of the study, problem to be solved & hypotheses)
- Divide as much as possible
- Revise (global view, confirm or refute hypotheses)
He believed that not all the knowledge he had gained during his education was clear, safe, and useful. He therefore claimed that his method made it possible to arrive at knowledge having these characteristics. In other words, to arrive at an absolute truth (to explain a phenomenon, to understand its functioning, etc.).
Conventionalism: Conventionalism is a doctrine that stipulates a fundamental separation between the data of intuition and the senses, and the intellectual constructs that make it possible to found scientific or mathematical theories. This notion was first created by H. Poincaré, then developed by Pierre Duhem and Édouard Le Roy, in rather different forms, at the border of the nineteenth and twentieth century (although none of these authors used the term “conventionalism”). It finds its origin deep in the Kantian separation between intuition and concept.
Falsifiability (Karl Popper): Falsifiability is presented by Karl Popper in his book The Logic of Scientific Discovery. He criticizes inductivism and verificationism, which he says are not valid from a logical point of view or from an epistemological point of view to produce reliable scientific knowledge. According to Popper, rather than looking for verifiable propositions, the scientist must produce falsifiable statements. It is this falsifiability that must constitute the criterion of demarcation between a scientific hypothesis and a pseudo-hypothesis. It is based on such a criterion that Popper criticizes Marxism or psychoanalysis, which in his opinion do not meet this requirement of falsifiability, these theories being based on ad hoc hypotheses that would immunize them against criticism. It is on this basis that Popper develops his critical method, which consists in testing in all possible ways the theoretical systems.
Scientific pluralism: In Scientific pluralism, Stephen Kellert, Helen Longino and Kenneth Waters explain that scientific pluralism is a new approach that is defined first of all as a skepticism, or agnosticism, of scientific monism (which, for example, adhered to the Vienna Circle), which argues that:
- the goal of science is to establish a unique, complete and comprehensive description of the natural world based on a single set of principles;
- the nature of the world is such that, at least in principle, it can be described and explained by means of this description;
- there are, at least in principle, research methods to produce this description;
- research methods must be evaluated in terms of their ability to produce such a description;
- scientific theories and models must be evaluated largely based on their ability to provide such a description.
That there is not (necessarily) a single scientific method and a single theory to access a single set of principles does not mean that there are, as opposed to radical relativism, so many approaches and truths that there are points of view. Scientific pluralism considers that there are constraints that limit the number of classification and explanation schemes.