Log in / Register
Home arrow History arrow Integrated History and Philosophy of Science: Problems, Perspectives, and Case Studies


Carnap’s rejection of metaphysics, especially in the early 1930s, is, to say the least, vigorous. He wants us to understand that he is not calling it doubtful, or false, or even logically false. He is not saying that it is of no earthly value for any practical purpose, though presumably this would follow from what he does say of it. Instead, he says that metaphysics is utterly meaningless. It is without cognitive content of any kind; it is gibberish. Unsurprisingly, many contemporary philosophers who count themselves as metaphysicians feel slighted by this. They take him to be rejecting the entire field and everything in it as gibberish. In fact, Carnap is not rejecting the whole field - quite. And the way in which he is not rejecting the whole field may be surprising. He might still reject many specific things that metaphysicians have written. But then what metaphysician doesn’t?

If it is a mistake to think that Carnap’s criticism must apply to the whole field, it is also a mistake to think that it applies only to work of long ago. It is true that Carnap’s most notorious paper against metaphysics, “Overcoming Metaphysics Through the Logical Analysis of Language” (Carnap 1932a/1959), prominently featured Heidegger as an example of the sins of metaphysics. So one might mistakenly reason: OK, perhaps some of Heidegger’s claims like “Nothing noths.” Or “Anxiety reveals the nothing.” are really meaningless. But what does that have to do with us? We are completely unlike Heidegger. We are trained in logic, and even our metaphysics is highly technical. We would never make those mistakes. It is true that contemporary analytic metaphysicians are unlike Heidegger in several respects that Carnap takes to be important. But it would be premature to think oneself safe from whatever criticism Carnap might have of metaphysics simply on the grounds of dissimilarity to Heidegger. We need to articulate what those criticisms were, and to that we now turn.

For Carnap, metaphysics is an enterprise based on the idea that by philosophic or other non-empirical means one can know truths about the world that are prior to or lie behind or are deeper than the truths to which empirical science can aspire. The idea that there are such deeper truths that are inaccessible to science perhaps goes back to Kant’s discussion of things-in-themselves. Kant said that science could never know these things-in-themselves (except that there were such), and so we were to refrain from speculating about them. Well, German Idealists rush in where Kantians fear to tread. There aren’t things-in-themselves exactly (It was the “in- themselves” part that they objected to.), but there was a domain of deeper truths about which we could have genuine knowledge. And the road to this knowledge runs through philosophy, i.e., metaphysics, rather than through empirical science.

Besides the German Idealists, Fichte, Schelling, and Hegel, Carnap also names Bergson and, notoriously, Heidegger under the heading of “metaphysicians”. Actually, though we need not go into this here, he has much more extensive lists of metaphysical philosophers, even indicating their individual degrees of guilt. In his “Intellectual Autobiography” he adds one more name to the list, his sometime colleague, Mortimer J. Adler. Carnap tells the story of a lecture that Adler gave in which

[h]e declared that he could demonstrate on the basis of purely metaphysical principles the impossibility of man’s descent from “brute”, i.e., subhuman forms of animals. I had, of course, no objection to someone’s challenging a widely held scientific theory. What I found startling was rather the kind of arguments used. They were claimed to provide with complete certainty an answer to the question of the validity or invalidity of a biological theory, without making this answer dependent upon those observable facts in biology and paleontology, which are regarded by scientists as relevant and decisive for the theory in question. (Carnap 1963, 42)

Carnap uses this as an example of a kind of “cultural lag” that he found at the University of Chicago. But in its confidence in philosophic methods and its dismissal of empirical ones, it is hard not to see Adler as the perfect example of what Carnap means by ‘metaphysics’.

And what Carnap meant is largely what these metaphysicians themselves meant. The writers that Carnap was talking about defined their own approach in opposition to empirical science, both in terms of its methods and in terms of its results: Metaphysical methods are not at all empirical, and the resultant knowledge is deeper than and concerns a reality that lies behind the world of empirical science. Some metaphysicians, like Heidegger, devalue not only empirical science but logic as well. But see also (Naess 1965/1968).

Against this Carnap argues that with a metaphysical approach each writer builds his or her own system among which there are “wearisome controversies” that never get anywhere. One system may become temporarily fashionable, but it doesn’t last. There is nothing that would count as a significant test between such metaphysical systems as Realism, Materialism, Idealism and so on. Science by contrast, is cooperative and progressive. So science gets somewhere, and metaphysics doesn’t. This, by the way, is a thoroughly historical argument.

In the early 1930s Carnap saw empirical science as a neutral core to which all metaphysicians could agree and the various metaphysical positions as various additions to that. Carnap illustrates this by giving a parable in Pseudoproblems (Carnap 1928b/1967, 333) about the Realist and the Idealist geographers who go off on an expedition to find out whether a certain mountain exists. They come back with the same report about the mountain’s height, shape, and location. This is hardly surprising. They used the same instruments for determining latitude, longitude, and elevation. The two geographers agree on the report, but the Realist adds that the mountain that they have located and measured is also real. The Idealist, by contrast, insists that the mountain itself is not real, and only our perceptions are real. No evidence could settle this dispute. All the evidence they have is summarized by the neutral core. Where they disagree is over what to infer from that evidence, and this is not an empirical question. If this is the metaphysicians’ own conception of their enterprise, then (to remodel a remark by Quine): Why all this metaphysical make believe? Why not settle for empirical science?

There was to be a subtle but important change in his attitude toward metaphysics as he moved to his mature philosophy in the mid-1930s. In the new regime Carnap recommends that we reconstrue erstwhile metaphysical positions, such as Realism and Idealism, as proposals for structuring the language of science. Philosophic work would thus comprise the engineering tasks of explicating our current linguistic framework, devising new ones, and exploring their virtues and vices for various specific practical purposes. The surprise is that Carnap thus shows us how to be a Realist or Idealist or Platonist or nominalist, that is, how to be a metaphysician, in a sense that he finds unobjectionable. This is what I meant earlier that Carnap did not dismiss the field of metaphysics entirely, just one dominant conception of it. Just as there can be legitimate analogs of the old metaphysical positions, it is, alas, possible to make mistakes that are analogs of the old ones: One might think that there is one uniquely correct linguistic framework, and one might think that we have some power of metaphysical insight that reveals which framework is correct.

So what then would be Carnap’s assessment of contemporary analytic metaphysics? The answer undoubtedly would be mixed. Contemporary analytic metaphysics generally lacks the dismissive attitude toward empirical science that characterized much of nineteenth and some early twentieth century metaphysics. And happily much current metaphysics is highly logic based, and it is often uneasy with claims about deep insight. But too often this unease is soothed by disguising the claims to insight by hiding them under a different name, like ‘intuition’ that covers both problematic uses and completely innocuous ones. Giving metaphysical insight a different name does not change anything. So some of the old errors can get smuggled in with the result of controversies that never get anywhere. But Carnap’s point is that these errors don’t have to be repeated.

Carnap’s mature advice for a remedy can be summarized in a single word: clarity. State clearly whatever you propose in the way of rules for inference, for observation, and for anything else. State them in full explicit detail so that we can get them out in the open so we can see how they work.1 The historical record of the approach Carnap calls metaphysical is not good. The historical record of judgments made directly on the basis of publically accessible observations is much better. There is wide and durable agreement about the observational judgments themselves, and there is an emerging theory explaining the connections between the events observed and the observational judgments. These judgments also form the basis of broad and useful theories. And we can specify the respects in which they have improved. This, of course, gives no absolute guarantee. But it does give pragmatic grounds for using those observations for now, for pushing the theorizing and testing further, and for using the best of our empirical theories as a basis for practical action. I don’t think we can, or should, ask for more.

But what does all this have to do with the unity of science? A lot, as it happens. To this we now turn.

This, of course is the Principle of Tolerance, first stated as such in (Carnap 1934/1937, 51).

Found a mistake? Please highlight the word and press Shift + Enter  
< Prev   CONTENTS   Next >
Business & Finance
Computer Science
Language & Literature
Political science