One of the main themes of my book about Carnap is that a decisive component of the original motivation first to write the Aufbau and then to push forward to the radical pluralism of the Syntax (and beyond) was Carnap’s diagnosis of the political situation immediately after the First World War in Germany. Amidst revolution and upheaval, Carnap saw very clearly that the German intelligentsia had contributed to the outbreak and continuation of the war by failing to get politically involved in the 19th century as French and British intellectuals had done, and thus failing to restrain the German elite’s war-oriented Betonköpfe (yes, that’s an anachronism — to make a point). So he advocated greater political involvement (and got involved himself), but also thought that to prevent future wars, the chaotic development of societies over the past century or two of rapid industrialization had to be steered by reason to a greater extent. But there existed no satisfactory account of reason, no system in which all the knowledge accumulated by the sciences over the past century could be seen to fit together. In contrast to the traditional Enlightenment system of knowledge (e.g. the Encyclopédie), where the various parts and components of knowledge were anchored in the various human cognitive, practical, and other mental faculties — so that the system was ultimately grounded in human psychology — Carnap (having studied with Frege) thought, like Leibniz, that the system of knowledge should be deductive. And so on from there.
To show that there was really something to this, I attempted in my concluding chapter, very briefly, to suggest how one might apply the later Carnap’s perspective to actual political discourse. This was suggested to me by Michael Friedman, who was on my dissertation committee, and who said that it might help people to get what I was talking about if I applied it to a controversy they were familiar with, such as the Rawls-Habermas debate. Well, I did exactly that, but no one really noticed, and when they did, they ridiculed the idea that such a rarefied nerd as Carnap, who sticks to the “icy slopes of logic,” could possibly have had a political motivation for the intricacies of his formalistic obsessions — or that his perspective could possibly contribute anything to political deliberation. Until now. James Pearson has taken pity on this explicitly political dimension of my book and devoted a paper to it, in which he tries to apply my Carnapian suggestions to a couple of concrete cases.
Before I comment on his discussion, I should point out that I was never very happy with that part of the book. I was forced to be far too brief, since CUP had a strict word limit. This meant, among other things, that I simply didn’t have the space to give concrete examples, which is what Pearson now sets out to do. He concludes, in the light of his examples, that Carnapian explication can be applied to political discussion only retrospectively, but not in real time — i.e. only by historians of political processes, not by the actors in them. I’m not sure I follow his reasons for arriving at this conclusion, but realistically, of course, it is undoubtedly true that it is easier to understand what went on in the past than what is going on right now. He appears to have more in mind than this platitude, but I think he has a tendency to project certain limitations of liberalism itself onto the framework of explication I was suggesting could reconcile certain tensions within it. Also I think his examples, while undoubtedly concrete and specific, are treated in too much isolation from the surrounding institutional structure to do the job he expects of them.
Some details. His examples, are, first, the institution of marriage and, second, the institution of adoption (of children). After giving quite a fair summary of my proposal of seeing political deliberation in the framework of Carnapian explication, he argues that there will be participants in political debate who will reject any proposal of a common agreed language to be used for the negotiation of disagreements about the shaping of future institutions. But isn’t this inherent in liberalism? I define liberalism in my book as the separation of content values from framework values, as illustrated by the famous quotation attributed to Voltaire, “I despise what you are saying but will die for your right to say it.” Liberalism is the idea that you need two systems of values, one by which to lead your own life (content values), and one by which you can live productively and harmoniously in a society in which people of many conflicting content values have to co-exist somehow (framework values). Of course this idea is one that most people in the world (including perhaps most people who might self-identify as liberals) vehemently reject. Most people think there can only be one set of values, that they know what that set is, and that we all should be living by them, period.
The attempt to define some minimal canon of framework values that everyone should in principle be able to accept goes back at least to Grotius, Hobbes, and Mandeville, if not to the medieval philosophers. Locke, Pufendorf, Hume, Rousseau, Rawls, Hayek, and many others have tried different approaches. But none of them would have claimed that there aren’t going to be people, even “rational” people, who will reject their proposed minimal canon. Rawls specifically clarified, in his later work, that his principles of justice as fairness will be accepted behind the veil of ignorance only by people whose content values (my term, not his) are “reasonable” by certain fairly uncontentious standards. Uncontentious (to us), yes, but Rawls would certainly have acknowledged that if you are not acculturated from childhood to a society where the distinction between content values and framework values is at least nominally embraced, you will not feel inclined to accept that distinction as a matter of course, and might well be viscerally opposed to it. (Even within such societies, as we are seeing, there can be an increasing subgroup who are opposed to it.) So I don’t think this problem has anything to do with putting the agreement on framework values in the context of Carnapian explication; it’s built in to liberalism itself.
Pearson goes through the thought experiment of a kind of social convention to agree on a language in which all the members of a society could negotiate about how to shape the institution of marriage. This is not in itself absurd, but it takes for granted that a lot of supporting institutional infrastructure would already have to be in place, or at least a lot of metalinguistic agreement about how to talk about it, and Pearson nowhere mentions this or betrays any awareness of it. In social science, those who think about institutions tend to think of them as intertwined and dovetailed, and hard to separate as independent entities from a comprehensive institutional (or cultural and institutional) system in which they’re embedded; that sort of idea is familiar to philosophers from later Wittgenstein. But even ignoring such complications inherent in the subject matter, explicating “marriage” without first having a linguistic framework in which to talk about its institutional context is a bit like imagining that you could explicate “simultaneity” without an agreed and robustly-tested framework of Newtonian mechanics already in place. It just makes no sense; you can’t be more precise than the linguistic framework in which you’re locating the explication.
In the last part of his paper, Pearson argues that explication can play a role in what he calls the “social history” of institutions. It happens that I’ve written a paper on how to do the history of institutional systems (with Sheilagh Ogilvie at Cambridge), and Pearson has actually done his homework and spotted it (and detected its Carnapian inspiration), but he wants to apply explication in a different direction, which I have to confess I don’t quite understand. He wants to simplify the complexity of social systems into readily graspable models or “ideal types” (along Max Weber’s lines) rather than to stick as closely as possible to empirical facts (as Sheilagh and I recommend), letting the data shape the concepts in the first instance. His proposal seems regressive to me — the opposite of explication, which is the replacement of vague ordinary-language terms by more precise ones, preferably in a constructed language. Unless understood in Stan Engerman’s way (who in his contribution to the Cambridge Companion to Weber regards the — precisely specified — simplifications of “homo economicus” as what Weber meant by an “ideal type”), it seems that an “ideal type” in Pearson’s sense is the replacement of an empirically-identified (or identifiable) concept by an ordinary-language one. Perhaps I’m missing something. If so I’m happy to be corrected.
PS: It turns out that Pearson doesn’t have a copy of his paper on his website or even on Academia.edu or some such thing. So the above link will work only for those who have access to the journal through their institutions. If anyone wants to look at the paper and can’t get to it, e-mail me and I’ll send it. Sorry; didn’t realize this before I wrote the post.