> My impression was more "All models are wrong, therefore all models are valid". That's a misrepresentation. It's actually fairly interesting where that misre

> My impression was more "All models are wrong, therefore all models are valid". That's a misrepresentation. It's actually fairly interesting where that misrepresentation originated from (hint: it was propaganda from totalitarian movements). It's a reaction to modernist assumption that "absolute truth" is possible for any question, and in particular, how often that truth _just happened_ to be personal taste of some dude (brutalism (I personally like brutalism) and modernist architecture in general are probably the most obvious examples). For sure, there have been explorations into pretty much every field, with varying success. In science, though, that relativity is more oriented towards establishing what theories get investigated, than the actual result (for example, how many investigations of liberalism in political science completely ignored the racial aspect of it, and its rather dismal history in that respect, or various - thankfully largely defunct by now - excustions into phrenology, eugenics, and similar woo). So, modernism isn't "all models are valid." It's "you can't just declare that your model is the bestest because you like it/it's 'simple'/it's 'elegant'/it happens to declare Helvetica to be the objectively best font."