I still get copies of Open Court books sent to me, for some reason, and I recently received the Library of Living Philosophers volume on Hilary Putnam. I could have sworn there already was one, but evidently I was wrong. Like most of these volumes, it’s huge, and I’ll obviously be looking at it for a while, but I have some immediate responses especially to the many Carnap-related remarks in Putnam’s autobiography, which are very interesting (see for instance section XVII, “Becoming a Philosopher: Carnap” in which Putnam attributes his beginning in philosophy to Carnap). Today I want to focus on a section entitled “The Story of Carnap’s Wire Recorder,” which addresses the very subject I posted on a few days ago, from a different, almost opposite, angle.
The way Carnap wrote things, in his later years, was to scribble down some notes, sometimes quite extensive ones, in his German (Stolze-Schrey) shorthand, and then to dictate something based on those notes in English into a machine. Someone (often Ina) would then type up a rough script from the recording, and Carnap would edit it further by hand, often circulating carbon copies for comments and suggestions on improving the English. By the time Carnap was at Princeton in 1953-4 he had evidently already been doing this for a while, as he got himself a new machine at that point, and gave his old one to Putnam. This old one was a “wire recorder,” something I’ve never heard of, and neither had Putnam, even at the time:
I would not have known there was such a thing, if Carnap had not given it to me — it literally recorded onto a metal wire. The quality might not have been good enough for music, but, as I recall, it was fine for recording dictation. When I turned it on for the first time, I realized that I was hearing one of Carnap’s drafts — a draft of a piece that I do not believe Carnap ever published, perhaps just remarks for his own use, although they were certainly polished enough to be published. Perhaps it was improper of me to do so, but I of course listened to the whole draft before erasing it. The subject was precisely what I was discussing the other junior faculty members : the merits and demerits of “rational reconstruction”. . . and “ordinary language philosophy.” (pp. 40-41)
The young Putnam was surprised by what he heard. He himself had been, at the time, “scornful” of ordinary-language philosophy, he reports, but “that was not Carnap’s attitude at all.” The recording began by describing Goethe’s theory of color, and while of course admitting that the theory, qua theory, has long been rejected along with Goethe’s polemical attack on Newton, Carnap was full of admiration for Goethe’s subtle and precise observations of the phenomenology of colors.
And after praising Goethe’s work as a description and Newton’s theory of color as a theory, he went on to make an astounding comparison (at least I found it astounding at the time, and, in a way, I am still astonished that Carnap made it). Carnap said that Newton’s theory, like many theories, was an oversimplified description, and needed to be an oversimplified description in order to bring out the essential structure of the phenomena but that it needed to be corrected by just the kind of detailed observation of those phenomena represented by Goethe’s Farbenlehre. In Carnap’s analogy, “rational reconstructionists” like himself were building the oversimplified but invaluable models, and “ordinary language philosophers” were supplying the needed corrective, by pointing out all the phenomena that the oversimplified models failed to capture. A truly generous attitude! (pp. 41-2)
Putnam was (and continues to be) surprised by this, but it seems entirely consistent with everything else we know about Carnap, e.g. his remarks on Goethe in the 1966 Martin Gardner book based on Carnap’s UCLA seminar (pp. 109-112) or his reflections on the complementarity of “constructivist” and ordinary-language philosophy in his reply to Strawson (pp. 939-40 of Schilpp volume), where he largely agrees with what Strawson says (pp. 510-14). James Justus has fleshed out this reconciliation of Carnap and Strawson in more detail in a very nice paper that is quite upfront about being much more on Carnap’s wavelength than Strawson’s (and does, in the end, assign Strawson’s mode of “connective analysis” a rather subordinate role in the overall system of explication — the role of helping to identify and clarify the explicandum).
That said, there is a problem with Carnap’s Goethe analogy here. It suffers from the same defect, in my view, as Carnap’s analogy between semantics and geometry on p. 12 of the Introduction to Semantics, where he says:
Both in semantics and in syntax, the relation between the descriptive and the pure field is perfectly analogous to the relation between pure or mathematical geometry, which is a part of mathematics and hence analytic, and physical geometry, which is a part of physics, and hence empirical.
Actually it’s not “perfectly analogous” at all, and this has misled many people, including Wilfrid Sellars; even Tom Ricketts, who alludes to this geometrical analogy in his “Languages and Calculi” paper (Minnesota Studies 18), p. 261, is thrown off by it about Carnap’s intended relation between pure and descriptive syntax in Logical Syntax of Language (there’s a rather dry and pedantic discussion of this in my book, pp. 245-50). There’s undoubtedly an analogy — a relation of theory to model in each case. But the purpose of geometrical and semantic theories is completely different; in geometry, empirically interpreted theories have the purpose of explaining and accounting for phenomena. In syntax and semantics, empirically interpreted theories can have this purpose, with respect to a used language, but that’s highly atypical. Usually (and entirely irrelevantly to such empirical use), pure semantic or syntactic theories, with or without empirical models, have the purpose of explicating vague concepts, such as the logical words of ordinary language. An empirically interpreted semantic theory may succeed or fail as an empirical hypothesis about some used language, but this has no bearing on the appropriateness of its corresponding syntactic theory (or a purely logical model of that theory) as a candidate for explicating some vague concept.
The same issue arises in the Goethe analogy Putnam encountered on Carnap’s wire recorder. Newton’s theory is intended to account for (many and various) phenomena, while Goethe’s observations point out phenomena of color perception that Newton’s theory doesn’t encompass. The relation between explication and ordinary language philosophy is completely different. Ordinary-language philosophy may certainly be regarded, in some of its guises, as pointing out phenomena that more sweeping theories of language and its functions or structure might be missing; but explication is not attempting a unified theoretical explanation of any (range of) phenomena; explication isn’t seeking any sort of knowledge at all, it’s a form of engineering. (It’s invention, not discovery.) The analogy breaks down.
I would like to hope that Carnap saw this problem, and didn’t go any further with the text he’d dictated onto the wire because he saw the problem, and thought it would be too much trouble to spell out something like the later reconstruction by Justus, while still retaining the attractive analogy between Goethe and ordinary-language philosophy. Who knows? The geometry analogy in the Introduction to Semantics indicates, though, that he was capable of a certain carelessness.