Misunderstanding idealization, truth, and understanding (1)

By Jesús Zamora Bonilla June 17, 2024 0 comments Print

Angela Potochnik’s Idealization and the Aims of Science is probably one of the most interesting and ambitious philosophy of science books of the last years. It offers a picture of scientific knowledge and of its production that elaborates on, and wisely amends, some of the best literature in the recent ‘pragmatist’ tradition of philosophy of science, and that takes roots on the less ‘formalism-lovers’ of the positivist tradition (like Otto Neurath). Just as a summary, these are some of the book’s main points: Philosophical theories of science should reflect how science in fact proceeds (p. 7); we inhabit in an extremely complex world, woven by incredibly complex causal webs, which force us to search for simplified causal patterns (p. 19); on the face of this complexity, idealization in science is totally unavoidable and widespread (p. 2); many philosophical accounts do not adequately reflect how science is shaped by the need to grapple with complex phenomena (p.17); science has many and very different aims, amongst which there can be tensions and conflicts (p. 110); ‘levels’ have to be understood pragmatically, rather than metaphysically (p. 185); etc. I shall concentrate, however, on what for me are the two most controversial aspects of Potochnik’s views: her theory about the role and working of idealization in science (ch. 2), and her claim that ‘science isn’t after the truth’, but that, instead, science’s epistemic goal is understanding (ch. 4).

According to Potochnik, we call ‘idealizations’ any “assumptions made without regard for whether they are true and often with full knowledge they are false”. The complexity of the real systems we try to study forces us to concentrate only in simplified, partial representations of the causal patterns they embody. Hence, idealization is ‘rampant and unchecked’ in science. By ‘rampant’, she means that idealization is universally widespread, it exists ‘throughout our best scientific representations’, and ‘they stand in even for crucial causal influences’, and by ‘unchecked’, she means that ‘little effort is put toward eliminating or even controlling these idealizations’. I don’t find very felicitous neither Potochnik’s concept of idealization, nor the two main properties (‘rampantness’ and ‘uncheckness’) she attributes to it, perhaps because the book doesn’t show any familiarity with the most important philosophical works about idealization (e.g., the Polish school of philosophy of science –Nowak, Krajewski…– and authors working under their influence –Niiniluoto, Moulines, Kuipers, Mäki… –; the prestigious collection Poznan Studies in Philosophy of Science published between 1990 and 2016 a series of nothing less than fourteen volumes under the title ‘Idealization’), and displays instead a strong disdain towards any formal philosophy of science whatsoever, what seems to be denying for philosophy what Potochnik grants for science: that it (philosophy of science) also necessarily works with idealizations. Having devoted my academic life to construct ‘idealized models’ of how science works, I cannot but feel a little upset by that.

More specifically, anyone familiar with these formal work on idealization (and related concepts), admits that, contrarily to what Potochnik seems to assume, not every ‘departure from the truth’ counts as an idealization. For example:

-Abstraction (‘isolation’): ignoring the effects of some real factors; (e.g., considering the movements of earth and moon as a two-body system)

–Ceteris paribus assumptions.

-Simplification: using simpler equations instead of more complex ones

-Approximation: being content with some more or less non-null errors

-As-if models (e.g., 19th century mechanical models of the electromagnetic aether).

-Provisional conjectures in a research programme; etc.

In a nutshell, idealization, as a technical term in the philosophy of science, usually refers to only one specific type of ‘departure from the truth’: those cases in which the idealized system can be described as a limiting case in which some magnitude, property of factor of the real system (or of other models describing the system) are progressively approximated to that ‘ideal’. In those cases, there is usually a more or less fixed procedure to ‘de-idealize’ the model, theory, etc., by a process called ‘concretization’. So, we might use as a rule of thumb to know whether we are in front of a case of idealization, the following: “No concretization, no idealization”.

Of course, perhaps Potochnik could argue that she is using ‘idealization’ in a broader, less technical sense, that includes by stipulation all cases of ‘departure from the truth’. But in this case, the ‘rampantness’ of idealization in science becomes a trivial, platitudinous claim, something everyone learns at Philosophy of Science 101: ‘Scientific laws are strictly speaking false’. The thing does not improve if one understands ‘rampant’ in the more etymological sense of ‘growing’, for, are scientific theories becoming more and more idealized as science progresses? It doesn’t seem clearly so. But I doubt that this is the meaning with which Potochnik uses the term, nevertheless.

Regarding the other feature of idealization, its ‘uncheckedness’, it seems that what Potochnik’s means by this (by using also expressions ‘keeping them in check’ and ‘holding them in check’, i.e., ‘having them under control’) is something like that scientists just don’t care about many idealizations (i.e., they could de-idealize them, but prefer not to). I think this is also a misfortunate metaphor, because (obviously) scientists regularly have their models’ idealizations ‘under control’ in the sense that they know those ‘idealizations’ exist, how are they to be use, what are their justifications, and what problems they may cause… I.e., those assumptions are perfectly ‘checked’ in an obvious sense. At least, scientists often know it very well because all this are points routinely levelled against their models by the colleagues in the course of scientific debate. So, perhaps the most charitable interpretation would be that scientists know about their idealizations contained in their models and hypotheses, but very often don’t feel pressed to de-idealize them, even when it is very well known how that could be carried out (and what consequences -usually not positive- doing that would have). But one thing is that every model always contains some strong idealization, and a different thing is that in the historical process of science some of these idealizations are not replaced for other that render the models ‘more accurate’. But this leads us directly to the question of truth, which will be the topic of our next entry.