Meaning rationalism is the assumption that you can tell by pure reflection whether your thoughts are clear, coherent and non-contradictory, whether your apparent beliefs actually represent anything, hence could have truth-values, and whether your inferences are valid. It assumes that further experience, observation or experiment never bear on these questions. Without doubt, meaning rationalism is pure common sense, but occasionally common sense is a repository for obdurate error. What is so terrible about meaning rationalism is that lying at the very root of almost all classical and contemporary thought on language and thought, it is at once nearly invisible and deeply mistaken.
“Hilary is a woman, isn’t she?” My three-year-old daughter Hannah and I are watching a video that briefly features Hilary. In the recent months, Hannah had a lot of fun subsuming things and especially animals, both human and non-human, under concepts. Given how she applies them, I am fairly confident that she has mastered quite a number of concepts. Despite my protests, “stupid” is a concept that obviously applies to me. And Hannah’s smiles tell me that she enjoys my responses. Mastering concepts in interaction with the world and other people is a great part of what makes us rational. What often puzzles me is the widespread confidence in this kind of rationality. Isn’t it reasonable to assume that our rationality fails every now and then? Yet, especially philosophers seem fairly confident that they are rational, at least many of their waking hours, while they are equally confident in being able to track irrational moves in others. But it’s one thing to be rational, and quite another thing to assume that one is rational. If rationality and monitoring rationality can come apart, it’s vital to understand how we actually monitor rationality. In what follows, I’d like to suggest that we sometimes don’t merely fail to be rational but fail in our ways of monitoring rationality.
Often when I hear someone calling out someone else for failing to be consistent, unambiguous, clear or what have you, I am reminded of Ruth Millikan’s critique of what she calls meaning rationalism. Let’s look at this idea for a moment. Crucial for the mastery of concepts is identifying something as the same (kind of) thing again. If you identify Hilary as a woman, this normally comes with a number of other abilities of identifying. Identifying Hilary as a woman, for instance, precludes identifying her as a toddler. At some point, my daughter Hannah occasionally did not seem to make this kind of distinction, but somehow she grew out of it. If a grown-up were to assert without irony that some women are toddlers, we would probably assume they had an impaired rationality. More precisely, we would assume that they couldn’t apply the concept “women” or “toddler” coherently. Tracking coherence in our speech acts and concepts is something we do all the time.
But what makes us so confident that we monitor others’ and our own consistency correctly? The answer is: language. That’s a very old idea, of course. Bacon, Spinoza, Locke and others famously criticised that we confuse “thinking in words” with thinking. Wittgenstein cast further doubt on the assumption that thinking or speaking is under the control of the thinker. What Millikan’s critique adds to these worries and uncovers is, inter alia, that many philosophers still assume that we have a means to monitor thoughts reliably. While she targets interlocutors in philosophical debates, I think that her point deserves general attention. How does language create this illusion? Well, we can literally see that people who say p and then say not-p contradict themselves. What a powerful tool! In any case, it seems to foster the assumption that we can monitor rationality a priori. Language seems to allow us to spot contradictions in virtue of its form, without us having to actually figure out what one thinks. Millikan’s point is that this confidence is overblown and not justified. If her critique is applicable, then language does not provide a reliable means to figure out what one thinks and whether someone is contradicting themselves. One concequence of this view is that we might say seemingly contradictory things without contradicting ourselves, and vice versa. Rationality is not “in the head”; nor can it be read off the surface of linguistic units, as it were. Rather, it is to be found in the world, and that includes of course the social (and linguistic) world. But the stability that we seem to find in the overt structure of language is not necessarily corroborated by the world.
When we suspect an incoherent thought, we tend to call out the error. “You said p, now you seem to say not-p”. “Oops! Did I mean to say that?” All of this strikes me as more elusive than language makes us believe. And sometimes I genuinely don’t know what I think. In this context, the distinction between the confidence in monitoring our own and others’ thoughts is really telling. I am called out way more often than (I think) I am at fault. Reviewer 2 can confirm this. But if I contradict myself so often, why should you think that you’re in a better position? If this mismatch between monitoring one’s own and other’s thoughts shows anything it is that language is not a reliable means of monitoring rationality. As I see it, that is not a problem. It still seems to work often enough. But it does turn into a problem when we are lured into assuming that we have a reliable means to monitor rationality. Give yourselves and others the benefit of the doubt.
So what about Hilary? Is Hannah right? I don’t know, and I say so. The moment passes, but the question remains. Perhaps more importantly, it is not set in stone what determines the answer to this kind of question: the world, the social world? Panta rhei.
Happy Twenties, dear fellow readers.