Thought experiments in science

Thought experiments are commonly defined as devices of the imagination used to investigate the nature of things. Such an experiment is a logical reasoning about an hypothetical situation where the results can be explored by imagination, physics or mathematics. It seems that this kind of methodology has been used at least since classical antiquity, however, it's not completely right to talk about ancient reasoning in terms of experiment, as this kind of verifiability didn't exist yet. We are going to focus, then, on the use of this method in the last 150 years, more concretely in the field of science.

The English term of thought experiment has its roots on the German word Gedankenexperiment (lit. experiment conducted in the thoughts), which was first used by Hans Christian Ørsted, nevertheless, it was Ernst Mach, physicist and philosopher of science, who later gave the word its actual sense (or at least an approach to it), and whose use of the word was translated into English.

According to Mach's paper On Though Experiments, it is an observable fact that experimenting is innate for human beings, we just need to check how children behave while they're growing. As people become adults, another kind of experiment is added to the physical one, “which is conducted extensively and on a higher intellectual level, namely, the thought experiment” (Mach 1905: 451). For Mach, “it shouldn't surprise us that, oftentimes, the though experiment precedes the physical experiment and prepares the way for it” (Mach 1905: 452), because, in fact, it is much easier to resort to than physical facts. Furthermore, Mach think that the thought experiment is a necessary pre-condition for a physical experiment and he argues: “Even the beginners learns in experimenting that insufficient preliminary estimate, or nonobservance of sources of error has for him no less tragic-comic result than the proverbial 'Look before you leap' has in practical life” (Mach 1905: 452). In accordance to these statements we could claim that thought experiments are a kind of tool with which we can write the scheme of the verification of our prior hypothesis and which is also a way to make a good selection between postulates. In some cases, the results of thought experiments are so clear that there will be no need for later physical experiment.

It is remarkable that “the possibility of thought experiments rests upon our ideas being the more or less copy of fact”(Mach 1905: 452), as well as the obvious statement that claims: “the most uncertain and more indefinite the results are, however, the more the though experiment necessitates the physical experiment (..)” (Mach 1905: 452). The former condition of the method is which let us distinguishing between simple hypothetical reasoning and thought experiments; this distinction rests mainly on the fact that thought experiments need to have a concrete observational/theoretical context (empirical observation and particular background theory) (Horowitz and Massey 1991: 158). Thus, not every hypothetical reasoning about nature will be considered a thought experiment.

The biggest advantage of this scientific method is that “it is free from practical limitations of the laboratory” ( Horowitz and Massey 1991: 160), as particular material or machines are not needed. They are also able to reach conclusions regarding some hypothetical situations that could be unfeasible in practice, which helps to establish some parameters of investigation. Moreover, they have been proven to be a powerful tool for the advancement of science. But how can this be possible? How can a hypothetical or even counter-factual reasoning concerning ideal limits be informative about the physical laws of the real world? We have already explained that thought experiments must stand in a privileged relationship to past observational data, which shows us that they are, indeed, grounded in the real world. Actually, imagining ideal limits in the parameters of a thought experiment means just ignoring some of the factors of an experiment that are considered irrelevant for the aim of it.

Anyway, there are also some objections regarding thought experiments, even though for the most part they have been accepted in science. For Pierre Duhem this kind of method has nothing to do with real experimenting in science and should even be forbidden (in this field, as scientific method). However, it seems he is alone in his view, at least on this field, since most critics come from philosophy, a discipline where thought experiments have been present for its whole history with an extensive use. From this side comes the critique of Daniel Dennett, who thinks they use folk concepts and that they are intuition pumps, which fail when carefully analyzed. As I pointed before, many thought experiments have been done in philosophy, but we could state that arguably the results are not so defining in this field as they are in science.

To finish with this paper I think we should bring here the answer that Mach gives to the question of how thought experiments originate: “just as every moment before it can become a voluntary one had to begin by chance, perhaps as a reflex action, likewise, it will also depend upon simply producing thought experiments through appropriate circumstances for it to become a permanent habit. For this purpose, the presentation of paradoxes is exceptionally appropriate. Not only does one learn by means of a paradox to best perceive the nature of a problem in which, indeed, even the paradoxical content is problematic, but the conflicting elements of a paradox permit thoughts no longer to come to rest. These elements produce the process which is characterized as a thought experiment” (Mach 1905: 455-456).

So, here we have another scientific method that could enlighten us a little bit regarding to the issue about verifiability in science. Nevertheless, there is still no way to deal with skepticism in this sphere.