Home » Cohen (W.) » Marschall (B.) » Metaphysics » Ontology » Values » Voluntarism » Carnap against analytic metaphysics

Carnap against analytic metaphysics

Wouter Cohen and Benjamin Marschall, two graduate students at Cambridge (one of my long-ago almae matres), have a terrific new paper in the latest issue of The Monist (the issue whose theme is “Against Metaphysical Grounding”), arguing that Carnap was not only — as everyone knows — against German idealism and the various metaphysical schools current in Germany between the wars (including the wilder and woolier outgrowths of phenomenology such as Heidegger), but would have been just as opposed to the current metaphysics emanating from analytic philosophers.

This might seem totally obvious, and not worth writing a paper about, but actually, if you look at analytic philosophy right now, not only is it once again in the grip of metaphysics, but many of those so gripped think their metaphysics is entirely reconcileable with some not-too-nitpicky version of logical empiricism, or of Carnap anyway. Theirs is a chastened metaphysics, they believe, and escapes the strictures pronounced back then. Many of those I’ve criticized on this blog, over the years, are of this persuasion, as are many I haven’t criticized. (Which means that being of this persuasion isn’t a sufficient condition to get yourself criticized on this blog.) Unfortunately, the situation isn’t very straightforward, though; there is no bright line separating the side of the angels from the dark side, and the toleration of analytic metaphysics ranges from zero to 100, with most people somewhere in the middle. So to explain why I think this new Cohen and Marschall paper is so terrific, I need to situate it in a larger picture of the place of metaphysics (and of Carnap) in current analytic philosophy.

Start with a broad, coarse-grained distinction between those who actually do what Carnap would have classified as metaphysics (and claim that it — more or less — escapes logical empiricist strictures) and those who do things that Carnap would not have classified as metaphysics, but which they call metaphysics. A paradigmatic example of the latter are the writings (esp. the book All Things Must Go) of Ladyman and Ross (which I’ve discussed here repeatedly, beginning with this post). They call their own efforts (the direct descendents of Carnap’s and Neurath’s efforts) at unification of the sciences “naturalized metaphysics.” Just to be sure no one confuses them with the first category, though, they begin their book with a comprehensive assault on the whole edifice of analytic metaphysics, portraying its project of supplementing physics with metaphysical “foundations” or “ontologies” or whatever as fundamentally misleading, since it seeks to “domesticate” scientific knowledge by assimilating it into folk terms of ordinary discourse and of human intuitions adapted to the surroundings we’ve evolved to survive in.

I think it’s too bad that Ladyman and Ross call their project “metaphysics” as it tends to empty that word of any meaning at all. It’s long been used in the sciences (and elsewhere) as a term of abuse, to refer to high-falutin’ talk that has lost contact with the shared physical and social environment we are trying to grapple with. The logical empiricists tried to explicate this distinction — between discourse that maintains contact with some part of that shared environment and discourse that has lost any such contact (while dispensing with any substantive notion of “reality”) — in a number of different ways we all know about (actually, only the first, Tractarian version became really notorious). These early attempts failed, as is well known. But the distinction is still worth making; the explicandum is still worth trying to explicate; philosophy is (and always has been, from the beginning) uniquely exposed to the danger of erring unintentionally into what Hume called “fairyland.” And yes, of course we can make that distinction without using the actual word “metaphysics”; it just makes it clunkier and forces us to use circumlocutions like “what Carnap classified as metaphysics.” But that particular circumlocution then raises the historical question of what exactly falls within its scope — and that’s the first reason why the Cohen and Marschall paper is so valuable.

(I have a small quibble with them on this historical question. They do point out that Carnap’s early critique of metaphysics differed from his later one, but they miss out the syntactic step in between, where “metaphysical” meant roughly “metalinguistic without being translatable into the formal mode of speech.” For their purposes, it’s fine to sweep that episode under the rug, I guess, but a footnote would have been reassuring.)

Anyway, the second reason their paper matters is that it shows the ultimate ground of Carnap’s later critique of metaphysics to have been not cognitive (descriptive) but normative (practical). “Cognitively meaningless” terms can of course be given a meaning by constructing a framework on the fly to which those terms are “internal.” (When Carnap informally classified something as “cognitively meaningless” he meant it wasn’t internal to any frameworks in current scientific or practical use.) But as Cohen and Marschall point out, this ad hoc framework-constructing move is pointless if the framework constructed has no intended purpose, i.e. if no one has an actual use for it, or a recognizeable location in a shared mental economy. This “use” needn’t of course be immediately practical, but even if it is purely theoretical it needs to have some relevance to some larger body of theory or practice, either completely detached from anything in the world (i.e. formal or mathematical) or pertaining to some aspect of the world, with the hope that eventually some way can be found to bring empirical observations or practical relevance to bear on it. The “needs to” in that sentence is normative; there is no fact of the matter about this!

Wouters and Marschall are particularly persuasive in their diagnosis of a certain “ontological anxiety” at the basis of much analytical metaphysics, an obsession with reducing the ontological commitments of some discourse to a minimum (a burden inherited from Quine, no doubt, though they don’t go into its genealogy). Again, they argue that Carnap would have wondered what the point was, and that this skepticism would have stemmed from normative rather than cognitive considerations.

I think they’re completely right about this, but they don’t take it far enough. They make Carnap sound a bit like some soulless Gradgrind of a utilitarian; in their portrayal, what motivates this normative rejection of metaphysics is that it diverts human time and attention from profitable to low-return preoccupations, it is a “questionable use of time” (p. 338). But that obscures Carnap’s positive vision, which I once described in print as follows:

From his earliest writings, it was Carnap’s deep conviction that human kind had a responsibility to choose and shape its own institutions — including its scientific language and scientific practices — rather than to accept them passively (Jeffrey 1994).  Philosophy, for him, including especially the entire program of explication, was highest-level linguistic or conceptual engineering, the planning and optimization of the future of the species.  Metaphysical arguments about “what really exists”, or about the true “substance” or “ontology” of the universe, were to be discouraged precisely because they undermined such planning.  They took for granted that the important choices are not choices; they are not a matter of what we want but of what is.  To accept metaphysical terms was to capitulate to our existing, traditional language and the traditional institutions it supported; it was to be overwhelmed by the weight of the past.

Nothing in the Wouters and Marschall paper contradicts that portrayal; they just stop short of attributing any positive views to Carnap at all, so one could easily get the impression that he is concerned only with optimizing human productivity, or something vaguely utilitarian along those lines. But this impression, especially reinforced by his late preoccupation with inductive logic and decision theory, is highly misleading; I think it’s important to stress (as I think that 1958 fragment on “Value Concepts” makes very clear) that this wasn’t anything like the whole picture for him; in fact his conception of ultimate values was radically pluralistic and open-ended. And his own, somewhat utopian values were more about giving full scope to human creativity and imagination than about discouraging people from indulging their metaphysical longings. His point was that these tended to be dead ends, and did nothing, in the end, to free us in any way from the weight of the past, from the burden of human conformity and convention that constrains creativity and imagination. And that normative stance would have applied equally to most of the analytic metaphysics of recent decades.

2 thoughts on “Carnap against analytic metaphysics

  1. Thanks for discussing our paper! I see the danger of our Carnap coming across as slightly joyless, and the positive vision of Carnapian philosophy you sketch is a welcome antidote to this impression.

    Some brief remarks on the origins of “ontological anxiety” and the role of Quine, since this is a topic that interests me greatly. Analytic metaphysicians who take ontology seriously often see themselves as following Quine’s lead. And to an extent they have a point. Quine’s early passion for nominalism that Carnap responds to in “Empiricism, Semantics, and Ontology” does seem to betray something like an aversion to abstract objects based on their allegedly problematic metaphysical status. But the picture changes once we look at Quine’s “mature” position towards ontology, by which I mean Word and Object and beyond. As times goes on Quine downplays the importance of ontology more and more, and the 1992 edition of Pursuit of Truth contains a section called “Ontology defused” which states that “our ontol­ogy, like grammar, is part of our own conceptual contribu­tion to our theory of the world”. Which is precisely the opposite of the ontological realism that, according to Ted Sider’s Writing the Book of the World, “meshes with” and is an “unacknowledged presupposition” of “the Quinean methodology for ontology”!

    It therefore seems to me that, over time, Quine’s views on ontology moved closer to Carnap’s and away from those of analytic metaphysicians. In the also otherwise interesting In Conversation panel with Dreben, Quine for instance says that reflecting on the nature of reference (and hence also the nature of objects and ontology) made him lose his earlier desire for nominalism:

    “My old hankering for a nominalistic basis for science and for everything has certainly dwindled. First of all, still there was the hankering but I decided it was impossible. But then I even lost that because of the indeterminacy of reference.”
    (https://youtu.be/KJlnJPYst44?t=2220, the remark starts around minute 37)

    Of course much more needs to be said about this whole development, for instance about the remaining value of ontological economy after ontology has suffered its “humiliating demotion”. But I think it is important to stress that Quine was not as pro-metaphysics as he sometimes appears in the presentations of contemporary practitioners.

    1. Sorry about the delay in replying — this is the week both our sons have birthdays, so things tend to get a little backed up.

      First, the business about “slightly joyless.” My point was less about rhetoric and marketing than about doctrine; Carnap and the Vienna Circle are known mainly for their negative agenda, the elimination of “metaphysics.” Each of them, and Carnap in particular, was mainly motivated by positive goals, though, to which metaphysics was an obstacle and a distraction. Carnap recognized, in his later years, that not only had he failed to draw a bright line between the cognitively meaningless and meaningful, but that substantial bits of what the Vienna Circle would have dismissed as “metaphysics” (e.g. much of Aristotle, Kant, Peirce, even Whitehead) could be interpreted more charitably as constructive work on proposed languages of (and metalanguages for) science. The demarcation between the bad stuff and the good stuff becomes at least partly a matter of what you want to call metaphysics.

      Which brings me, second, to your point about Quine. Huw Price, in his contribution to the Metametaphysics volume, makes a case that Quine, despite having given rise to the obsession with ontology in subsequent generations by making it so central to his own view back in the 40s and around 1950, was actually more or less continuous with Carnap in his attitudes to metaphysics. Recently I read a paper by Fraser Macbride and Frederique Janssen-Lauret (“Metaontology, epistemology, and essence: on the empirical deduction of the categories” in The Monist 2015) that takes a different view and seeks to refute Price. They argue that Quine in fact articulated a positive program of metaphysics, albeit a naturalized metaphysics, aligned with science. On the other hand, they then go on (in a different paper five years later) to tell the story of Quine’s flirtation with Lewis’s supposed reduction of set theory using mereology in the late 1980s. Quine’s excitement about this project (to which he actually contrbuted, though declining to be listed as a co-author) was due to its potential for dispensing with set theory (in ontological principle, at least) in favor of a more perfect nominalism of the sort he and Goodman had proposed back in the 1940s. So Janssen-Lauret and Macbride claim, anyway (their paper is in the book on Quine edited by Janssen-Lauret, OUP 2020).

      My own view, before I read that last paper, was basically what I said in the post on which you commented; “metaphysics” is a matter of classification, probably also in the discussion between Carnap and Quine. But now I wonder whether Quine didn’t after all harbor some residual metaphysical impulses/prejudices that (pace Macbride and Janssen-Lauret) really were fully metaphysical in the old sense (or in any old sense), i.e. nominalistic ones, which I don’t think even Macbride and Janssen-Lauret would want to classify as “empirical.” This would fit also into the picture that Michael Friedman paints in his Howard Stein Lecture (later published in Philosophical Topics 2006 as “Carnap and Quine: 20th-century echoes of Kant and Hume”), with which I generally sympathize. I’m undecided. The video you link to is a few years after the episode with Lewis, so perhaps he’d recovered from it again by then. But the indeterminacy of reference didn’t prevent him from getting excited about a possible new basis for radical nominalism in the late 1980s, so who knows what was going on. Sounds like a good project to sort this out; there seem to be plenty of relevant documents.

Comment (I will respond, however old the post!):