Some Anticipatory Reminiscences

Abstract

For over three decades, Enric Trillas and I have had an ongoing dialogue about language and, derivatively, its place in logic and science. We have been doing so from different perspectives: Enric, being a mathematician and a logician, has no qualms about a formal approach to things or about the use of a priori models. I vividly remember a heated discussion at the Universidad Politécnica de Madrid (it must have been around 1990), when he summed up his position as: “We need models - there is no way around that!” This seemed like a reasonable statement to me, given the machine-oriented context in which approximate reasoning is usually cast. I also remember insisting, however, that, precisely because fuzzy logic actively attends to natural language, greater clarity about its workings is needed if we are to take the usual claims about its relevance to human reasoning at face value. So, my reply to him at that point was, “That depends on the kinds of models we build, what they are grounded in, and what we expect from them.” I would still abide by that.