What Was Carnap Rejecting When He Rejected Metaphysics?

2.1 Mortimer J. Adler

I have already mentioned that Carnap took Bergson to be a prime example of a metaphysician in his sense. I will come back to Bergson in a moment. But first I consider a different case that Carnap discussed, Mortimer J. Adler, because Carnap is explicit there about his objections. Adler was a colleague of Carnap’s at the University of Chicago and later became the editor of The Great Books of the Western World and of the Encyclopedia Britannica and also a mainstay of the Aspen Institute. People will differ in their assessments of his success in these endeavors. Here I look at Carnap’s report (1963, 42) of a philosophy department seminar lecture by Adler and also at another lecture by Adler (Adler 1941).

Carnap does not give the date of Adler’s department seminar lecture but says this about it:

Adler . . . 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 accepted scientific theory. What I found startling was rather the kinds 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 on those observable facts in biology and paleontology, which are regarded by scientists as relevant and decisive for the theory in question. (Carnap 1963, 42)

Plainly, Carnap is objecting to a conception of philosophy rather than to a branch of philosophy. That way of philosophizing involves making a priori claims about the world that purport to be substantive—in other words, not about the language used or to be used—and to which the empirical science must conform.

The other lecture by Adler is “God and the Professors” of 1940, and it shows a view much like that criticized by Carnap in the previous quote.5 The paper is a Jeremiad against almost all professors, whether of science or of philosophy, in American academia. He says that almost all of these professors are “positivists” and goes on to hint at views that are like Carnap’s or are caricatures of them. Once again, Adler proceeds on an a priori basis to assert the following:

1. “Philosophy is superior to science . . .” because philosophy is “. . . knowledge of the being of things whereas science studies only their phenomenal manifestations . . .” (1941, 129).
2. “There are no systems of philosophy” only the one true one (1941, 129).
3. “Sacred theology is superior to philosophy . . . because it is more perfect knowledge of God and His creatures . . .” (1941, 131).
4. “Just as there are not systems of philosophy, . . . there is only one true religion, less or more embodied in the existing diversity of creeds” (1941, 131).
5. “Because God is its cause, faith is more certain than knowledge resulting from the purely natural action of the human faculties” (1941, 130).
6. “Science, philosophy, and theology cannot really disagree because they have different subject matters” (1941, 128–31).6

At the end of the same lecture, Adler (1941, 137–38) welcomes Hitler (this was 1940!) to cleanse America’s universities of their professors. Earlier he had spoken of “liquidating” the professors, scientists and philosophers alike (Adler 1941, 134).

This lecture, “God and the Professors,” clearly illustrates the features that Carnap found objectionable in the department seminar: a priori arguments that supposedly can give substantive results that overturn empirically established theories. If taken seriously, this can harm scientific progress. But I know of no evidence that Adler was taken seriously by anyone in the scientific community. He did have friends in high places, such as University of Chicago president Robert M. Hutchins and the publisher of Time Magazine, Henry Luce. Adler had a certain popular following as well. Certainly, inviting Hitler’s armies to make a clean sweep in American academia can hardly be considered science-friendly. Moreover, Adler’s claim that there are not many different systems of philosophy, but only the one true one, hardly squares with the history of philosophy, where there certainly seem to be multiple such systems, all impervious to his arguments. It is precisely such controversies that Carnap sought to sidestep in rejecting Adler’s metaphysical approach and by adopting the principle of tolerance.

2.2 Henri Bergson

Henri Bergson is a vastly more influential and subtler philosopher than Mortimer Adler. And unlike Adler, Bergson was interested in what scientists had to say and was, in turn, taken seriously by some in the scientific community, at least at the time. He was also prolific and difficult to analyze. So it is not really possible to look at the full breadth of Bergson’s views.7 What I can do is highlight a few general threads of those views to which Carnap might and did object. I will then go on to look at two more particular arguments that Bergson raised against Einstein. My aim in all this is not so much to assess the merits of Bergson’s views as to use his example to illuminate what Carnap means by ‘metaphysics’.

Bergson was, at least until the First World War, wildly popular, both in academic circles and outside them. It was thought that perhaps only the Paris Opera could hold the throngs of people who wanted to hear him speak (Gunter 1969, 16). So far, we have seen only that Bergson had some sort of clash with Einstein. Since I think Bergson’s texts are often unclear, let’s begin by looking at a brief passage from his An Introduction to Metaphysics:

Now it is easy to see that the ordinary function of positive science is analysis. Positive science works, then, above all with symbols. Even the most concrete of the natural sciences, those concerned with life, confine themselves to the visible form of living things, their organs, and anatomical elements. They make comparisons between these forms, they reduce the more complex to the more simple; in short, they study the workings of life in what is, so to speak, only its visual symbol. If there is any means of possessing a reality absolutely instead of knowing it relatively, of placing oneself within it instead of looking at it from the outside points of view, of having the intuition instead of making the analysis; in short, of seizing it without any expression or symbolic representation—metaphysics is that means. Metaphysics, then, is the science which claims to dispense with symbols. (Bergson 1903/2012, 8–9, italics in the original)

Carnap actually quoted from this passage in saying that he, Carnap, is using the term ‘metaphysics’ just as many who claim to be metaphysicians use it:

Other philosophers use the name “metaphysics” for the result of a non-rational, purely intuitive process; this seems to be the more appropriate usage:

REFERENCES. In referring metaphysics to the area of the non-rational, we are in agreement with many metaphysicians. Cf., for example, Bergson ([Metaphysik] 5): “That science that wants to get by without symbols.” This means that metaphysics does not wish to grasp its object by proceeding with concepts, which are symbols, but immediately through intuition. (Carnap 1928/1967, 295)

In identifying the nonconceptual with the nonrational, it may be that Carnap misunderstands the passage from Bergson that he has quoted.8 But Carnap is saying that he and Bergson are in agreement about what metaphysics is, that they are talking about the same approach to philosophizing (though of course they take different attitudes toward it). In any case, my concern in this chapter is with what Carnap understood himself to be rejecting, so how he saw the matter is what is relevant.

What Bergson suggests is that the mind is divided into two parts. The first is the rational, intellectual, conceptual, analytical side. This is the side where we find ordinary scientific theories and the ordinary empirical observations on which science is based. The second side of the mind is the nonrational/nonconceptual part. This is the home of philosophical intuition, which is to be a direct and better grasp of how things really are. As we shall see, Bergson holds that the first part of the mind, including science, inevitably distorts that, but philosophical intuition can correct that.

This is not just my (or Carnap’s) interpretation of Bergson. One of his more influential defenders, P. A. Y. Gunter (1969, 3–42 and esp. 29ff), says that Bergson holds that science is the product of the analytical intellect, which necessarily distorts reality.9 Philosophy, however, has its source in the philosophical intuition, namely philosophical insight that is not part of the conceptual world of the intellect. This intuition can correct the distortions of the scientific intellect. As Gunter (1969, 29) puts it, “The intellect, especially the scientific intellect, is for Bergson a pragmatic faculty that, rather than comprehending things, utilizes them, and in utilizing them spatializes, fragments, and materializes them beyond recognition.” Later, on the same page, Gunter rephrases this—four times: “intellectual analyses distort reality,” “intellectual analysis distorts,” “the intellect in most respects fragments, spatializes, and distorts reality,” and “intellectual analysis spatializes and distorts reality.”

This idea that science distorts, and it is the science of Newton or Darwin or Einstein that is referred to here by the phrase “intellectual analysis,” is not the idea that scientific claims are fallible because they can be challenged and replaced by further applications of scientific methods. It is, rather, the idea that science as a whole distorts, even in the long run. These distortions could only be discovered as such because there is a source of knowledge that does not distort and that is thus in a position to correct the errors of science. This source is philosophical intuition, a nonconceptual direct apprehension of the nature of reality.10

One persistent theme in Bergson is the visceral conviction that scientific theories and all purely intellectual works are bloodless, abstract things, but the real world pulses with life. This is why the conceptual side of the mind must distort. Bergson’s charge that the intellect/science distorts because it is abstract, however, misunderstands what science is supposed to do. Science is not supposed to reproduce the lived experience we have or to reproduce the world more generally. It is supposed to map it, to describe it, correctly one hopes. To say that the map does not turn green in the spring (Nelson Goodman’s phrase) is not to say that the map is incorrect or that the description is thereby a distortion (cf. Goodman 1963, esp. 552–54).

Bergson’s antidote to the distortions of science is philosophical intuition. The word ‘intuition’ has had a long and tortured history in philosophy. In the seventeenth and eighteenth centuries, ordinary observation was called intuition. This was a representation of individuals compared for sameness and difference. In this sense, Kant has the mind turn a manifold of intuitions into an intuition of a manifold. There is also a long tradition of platonic intuition to underwrite mathematics and logic. Gödel appeals to this, as do Russell and Moore. This is usually distinguished from sensory evidence because its objects are not individuals and not in the causal order—in other words, they cannot be observed, even in principle.

Carnap spoke of intuition in his own work only very early (when he is a neo-Kantian discussing visual space) or very late (when intuitions are simply grist for linguistic explications). Carnap would have sharply distinguished Bergson’s philosophic intuitions from the ordinary observations on which science relies, as does Bergson. Scientific observation, and more precisely observational judgments, are parts of the rational/symbolic/conceptual universe. And they are probative in science only because they are part of that domain. Bergson’s philosophic intuition by contrast has its roots in the nonrational/nonconceptual mind. Carnap can accept that philosophical intuition is “experience” in a suitably broad sense, like a feeling of ennui. He would say it is a subjective attitude toward life. But in this sense, it is not a judgment at all and not one that can help select between two purely descriptive (nonevaluative) accounts of the world. A judgment that one is feeling ennui, however, is a perfectly ordinary observation, albeit about the mind. Carnap from day one is interested in intersubjective science and scientific observation, even in the Aufbau and before. The track record of intuitions, philosophic and platonic, shows them to be highly subjective and unable to resolve the apparent disagreements that inevitably arise among alternative intuitions. These irresolvable disagreements yield the “wearisome controversies” in philosophy that it is the whole purpose of Carnap’s mature philosophy (from 1932 onward) to sidestep.

We have looked so far at Bergson’s general view. Now we can turn to look more closely at two general objections that Bergson lodges against Einstein. First, Bergson claims that to treat time as a fourth dimension like that of the three of space is to “spatialize” time by positing a static, block universe in which nothing ever changes. Einstein is trying to “stop time”—in other words, to bridle the vital, creative forces that are beyond the reach of science. It is hard to see what Bergson even could mean by ‘spatialize’ or what defect is involved. It is likely that Bergson’s claim that according to relativity theory nothing ever changes derives not from Einstein’s treating time as a fourth dimension but rather from his treating it as being composed of points. There is a popular, though I think mistaken, understanding of Zeno’s arrow paradox, according to which at every point, the arrow is not moving within that point.11 Hence, the arrow is not moving at all. The combination of Zeno’s idea with the idea that time is a series of points would explain the conclusion that in Einstein’s theory nothing ever changes. Bergson’s claim is not clear enough for us to be certain, but it is hard to see what else would explain his conclusion. Of course, the same criticism could be raised just as easily against Newton’s treatment of instantaneous velocity and acceleration. It is hard to see how the objection has any force at all, but Bergson concludes that the basic entities should be processes rather than punctiform events. Moreover, time embodies the vital/living forces that permeate what is real and important and can never be captured by science—because they are creative rather than subject to laws.

What in all this would Carnap object to? Not to Bergson’s process ontology. If scientists want to use a process ontology, that’s fine. If philosophers want to develop rigorously what a process ontology involves, that’s fine too. Does Carnap object that someone might disagree with a scientific theory? No. Does he object to Bergson’s claims that for Einstein nothing ever changes and that Einstein is trying to “stop time”? Carnap would certainly think that these claims are based on misunderstandings of Einstein, but this is not what makes them metaphysics. Would Carnap object to Bergson’s vitalism? Again, Carnap would not object to this if it were developed as an empirical theory or even as a well-worked-out conceptual framework, or at least he would not call either of these approaches to vitalism metaphysics.

Second, Bergson also objects specifically to the relativity of simultaneity embedded in the special theory of relativity. One might argue that present events are real in a way that future events are not yet real and that past events are no longer real. In this sequence of events, then, the “truly real” defines an objective simultaneity class. So Einstein’s physics must be at best about what we know rather than about what is objectively real. This argument about the objectivity of the “now” is not a good argument, though there is some evidence (Carnap 1963, 37f.) that Einstein was bothered by it (not convinced, but bothered).

But this is not the argument that Bergson gave in the 1922 exchange with Einstein. Bergson’s prose there is such that it is hard to see exactly what his argument is supposed to be. But he seems to argue that the relativity of simultaneity can be disproved on the basis of perfectly ordinary observations as follows:

1. I can have one experience of two nearby events such as a pair of flashes of light or a pair of notes, call them A and B.
2. Because this one experience is composed of experiences of A and B, my experiences of A and of B are simultaneous.
3. I can represent A and B as absolutely simultaneous, that is, as simultaneous independently of any inertial frame to which the events are measured. (A and B are taken as close to one another spatially but not exactly in the same spatial location.)
4. Therefore, the physical events A and B are absolutely simultaneous.
5. We can imagine a sequence of living beings observing a series of events, each near to the next, such that
   1. I observe A and B to be absolutely simultaneous, as noted.
   2. The second conscious being (Bergson suggests a “scientific microbe”) observes B and C to be absolutely simultaneous.
   3. The third conscious being (microbe) observes C and D to be absolutely simultaneous.
   4. And so on.
6. We can establish on this observational basis that events, however distant from one another, are absolutely simultaneous.

Note that none of the claims here is nonconceptual because claims cannot be nonconceptual. The only way to be wholly nonconceptual is to say nothing. But premises 1 and 3 are about first-person reports. Bergson may well believe that such reports can be underwritten by philosophical intuition. While there are many issues surrounding first-person reports, let’s ignore those issues and just grant these premises.

Claims 1–3 are also psychological, that is about mental states—mine. But 4 is about a physical state of affairs. Both Einstein and the third symposiast that day, Henri Piéron, challenged the legitimacy of such an inference in this case. Einstein began by noting the “the philosopher’s”—in other words, Bergson’s—concept of time is at once both psychological and physical. The former is about perceived time, while the latter is about events that are independent of us. Einstein took for granted that for perceived time there are no inertial frameworks to which temporal location and simultaneity even could be relativized. Thus, among psychological events there is no distinction to be drawn between absolute and relativized simultaneity. For physical events, however, there are alternative inertial frameworks. And relativity theory says that determinations of temporal location and simultaneity must be relativized to such frameworks. Einstein granted that drawing a temporal inference from the psychological to the physical often yields no conflict with the evidence, at least not in ordinary cases and for practical purposes, because, given the high velocity of light, the difference in temporal location from one inertial framework to another is too small to be observed. (As we shall see in a bit, this undermines Bergson’s response in two ways.) Nonetheless, there are such differences. And no psychological evidence permits the inference from psychological simultaneity to absolute physical simultaneity even for events that are near to each other.

Einstein (1922/1969, 133) concluded: “Hence, there is no philosopher’s time; there is only a psychological time different from the time of the physicist.” Given the gloss after the semicolon, what appears before it means only that there is no concept of time such as the one that Bergson tries to employ such that an inference from claims about the mind to claims about absolute simultaneity can be warranted. Psychological and physical times should be kept distinct. This conclusion is not, as some have suggested, either “scandalous” or “incendiary” (Canales 2015, 5) any more than it would be for an eighteenth-century chemist to say that there is no philosopher’s stone or for Lavoisier to say that there is no such thing as phlogiston.

Piéron was an empirical psychologist, and his comments on separating psychological and physical time are even more pointed. He gives essentially two arguments. First, he notes that his experimental results show that many factors other than physical temporal proximity can influence our perceptions of simultaneity. Hence any such inference is at best unreliable. Second, he notes that all observations are inexact, and so there can be no absolutely precise determinations of simultaneity even within the psychological realm. Piéron’s (1922/1969, 134–35) conclusion is “Thus determinations of psychological succession or simultaneity can in no case be utilized as a measurement of physical time. . . . And thus the Bergsonian duration seems to me to be obliged to remain a stranger to physical time in general and in particular to Einsteinian time.”

Bergson (1922/1969, 135) replied very briefly that he completely agreed that “the psychological establishing of a simultaneity is necessarily imprecise.” He went on to add that, nonetheless, the psychological determinations are basic, that is, presupposed, by any instrument reading.

Bergson’s concession that determinations of simultaneity are necessarily imprecise is fatal to the argument he gave against Einstein regardless of whether one is talking about psychological or physical simultaneity or about absolute or relativized simultaneity. The inference from the series of claims 5 to 6 has some chance of success if the simultaneities are exact because that relation is transitive. But the relation of almost simultaneity is not.

Bergson’s remark that psychological determinations are presupposed in any instrument reading can be doubted,12 but it does nothing to help his case, even if it is true. The inference from those psychological claims to Bergson’s desired conclusion requires far more than we are given here. The inference is not justified by pure logic. Nor is it warranted by the scientific facts. And Bergson does not intend it to be a mere linguistic choice. In fact, it seems doubtful that Bergson intends any of these alternatives. There seems to be little left but to suppose that Bergson believes the inference to claim 4 and beyond to be justified by some sort of philosophical intuition that sees deeper and corrects what physics has to say. Moreover, it is fair to say that Bergson is not trying to formulate an alternative theory to deal with the evidence that Einstein and other physicists take to be relevant to deciding whether to accept his relativity theory. Rather, Bergson is approaching the issue from the outside.

So what in Bergson’s argument would Carnap object to? Carnap would not object that it begins with first-person reports about experience or that Bergson wanted to use concepts other than those that Einstein wanted to use. If Bergson wants to use different concepts to formulate his own theory, that’s fine. That would not be to transcend science via philosophical intuition, but to do science, alternative science, in the familiar way. Carnap would object that the argument is a bad one, that Bergson has not understood Einstein or the argument as given ignores the approximative character of the observation of physical events. Carnap would object, but these features are not what makes the enterprise metaphysical. What makes it metaphysics is that Bergson believes that his nonconceptual (and thus to Carnap nonrational) philosophical intuition can see deeper than or behind what ordinary science can see and thus be in a position to overrule it. Bergson attempts to force science to turn away from the concepts and patterns of inference that it finds most helpful in organizing experience and in organizing our response to it. This attempt is what makes it metaphysics. Such a mode of philosophizing impedes scientific progress. And insofar as intuitive insights are used on behalf of conflicting philosophical claims, we lack a way to resolve the issues, and the result is endless controversies.

2.3 Cases That Carnap Did Not Call Metaphysical

We have seen two cases that Carnap did identify as metaphysics. There are other cases where one might expect Carnap to reach the same judgment but where he does not. These can be treated comparatively briefly. Hans Reichenbach was a vigorous supporter of scientific realism, the idea that the unobservable entities that science postulates, such as atoms and electrons, are really there and not just convenient fictions. And during much of his career, ontology was a central focus for W. V. Quine. Ontology is a central part of the field of metaphysics, and yet Carnap denies that either of these men are metaphysicians. Why? The answer is that both are trying to turn their ontological claims into empirical ones. Perhaps this is obvious in the case of Quine, who rejected a priori methods altogether and insisted that even logic and mathematics were empirical. Reichenbach didn’t go quite that far, but it did seem that he was defending scientific realism as an empirical claim. Certainly, Carnap (1963, 870) thought so.

This accords well with Carnap’s comments describing metaphysics as an attempt to gain knowledge that somehow transcends the knowledge that empirical science can aspire to. Having such deeper knowledge, metaphysics would be able, from some philosophical or intuitive perspective, to “correct” the results of empirical science. It is this that Carnap rejects.