1. Introduction
There is a wide-spread and often-repeated assumption in the literature that Werner Heisenberg is essentially a naive positivist or instrumentalist (see, e.g., Popper Reference Popper[1963] 1989, 113). This reading of Heisenberg has two primary sources: first, his association with the so-called Copenhagen interpretation,Footnote 1 which is traditionally taken to view quantum theory simply as a ‘black box’ tool for making predictions; and second, the emphasis that Heisenberg himself places on observable quantities in his 1925 matrix mechanics paper and 1927 uncertainty paper.Footnote 2 I shall argue, however, that this common reading of Heisenberg is mistaken. Not only does Heisenberg explicitly reject positivism (see, e.g., Heisenberg Reference Heisenberg1971b), but also, as I shall show, he in fact endorses a version of scientific realism.
Throughout his career, Heisenberg held a highly original view in the philosophy of science, centered on his notion of a closed theory. Very briefly, a closed theory is a tightly knit system of axioms, definitions, and laws that provides a perfectly accurate and final description of a certain limited domain of phenomena. This notion has profound implications for Heisenberg's understanding of scientific methodology, theory change, intertheoretic relations, and realism. Recognizing the central role that the concept of closed theories plays in Heisenberg's thinking reveals that he had a much more considered and consistent philosophy of science than has hitherto been assumed.Footnote 3
Further insight into Heisenberg's philosophy of science can be gained through a comparison of his views with those of Thomas Kuhn. As Mara Beller briefly notes in her book Quantum Dialogue, there is a striking similarity between Heisenberg's account of closed theories and Kuhn's model of scientific revolutions (Reference Beller1999, 288).Footnote 4 In this paper, I explore this interesting parallel between Heisenberg and Kuhn in greater detail. My analysis draws on a little-known interview of Heisenberg by Kuhn conducted in 1963 for the Archive for the History of Quantum Physics, in which Heisenberg mentions that he has just read Kuhn's recently published Structure of Scientific Revolutions. At this point, Kuhn breaks with his role as interviewer and begins to compare Heisenberg's views with his own. Their argument in this interview over whether Ptolemaic astronomy is to count as a closed theory helps to illustrate precisely where the similarities and dissimilarities between their views lie. By making use of this interview and Heisenberg's other writings, I show that while Heisenberg and Kuhn share a holistic account of theories, a revolutionary model of theory change, and even a notion of incommensurability, their views diverge fundamentally when it comes to the issue of scientific realism.
My aim in this paper is not to defend Heisenberg's philosophy of science, but is rather, first, to draw attention to the surprising prima facie similarities between his views and Kuhn's, and second, to use this comparison to correct a number of long-standing misconceptions about Heisenberg's philosophy.
2. Heisenberg's Closed Theories
Arguably the most central idea in Heisenberg's philosophy of science is his concept of a ‘closed theory’. Heisenberg first introduces the concept of a closed theory in connection with quantum theory as early as 1927: In the conclusion of his Solvay Conference paper with Max Born they write, “We wish to emphasize that … we consider quantum mechanics to be a closed theory, whose fundamental physical and mathematical assumptions are no longer susceptible to any modification” (Born and Heisenberg Reference Born and Heisenberg[1927] 1928, 178).Footnote 5 From this point on, Heisenberg discusses closed theories regularly in his writings, up until his death in 1976. Throughout these forty some years his views on this topic seem to have changed very little.
The German phrase that Heisenberg uses is abgeschlossene Theorie, where abgeschlossene can be translated as ‘closed’, ‘locked’, ‘isolated’, or ‘self-contained’. Heisenberg defines a closed theory as a system of axioms, definitions, and laws whereby a field of phenomena can be described in a correct and noncontradictory fashion. Heisenberg had attended Hilbert's lectures in Göttingen during the period 1922–1924, and as Mélanie Frappier (Reference Frappier2004) has shown in detail, Heisenberg's understanding of closed theories was greatly influenced by David Hilbert's axiomatic program. Heisenberg describes the relation between the empirical world and the axiomatic system of a closed theory as follows: “It must be possible for the concepts initially stemming from experience to be so specified and fixed in their relations by definitions and axioms that mathematical symbols can be annexed to these concepts, among which symbols a consistent system of equations evolves” (Heisenberg Reference Heisenberg[1948] 1974, 43). As Frappier notes, however, “Heisenberg's claim that the relations existing between physical concepts are defined in systems of axioms and definitions … [might suggest] that he did not completely understand Hilbert's idea that a theory's axioms implicitly define its theoretical terms, and that consequently, a concept can change reference across interpretations” (Frappier Reference Frappier2004, 165). Frappier, however, offers an alternative interpretation that by “definitions” Heisenberg really means something closer to “interpretative principles.”Footnote 6 Exactly how Heisenberg understands an axiomatic system and its elements remains an issue of debate.
Heisenberg identifies four theories of physics that have achieved the status of being closed: Newtonian mechanics, the nineteenth century theory of heat (which he also calls statistical thermodynamics), Maxwell's electromagnetism together with optics and special relativity, and perhaps not surprisingly, quantum theory. He notes that apart from these four areas, there remain other parts of physics, such as general relativity and elementary particle physics, that are still “open.”Footnote 7 Heisenberg justifies the choice of these four theories on the basis that “for each of these realms there is a precisely formulated system of concepts and axioms, whose propositions are strictly valid within the particular realm of experience they describe” (Reference Heisenberg1971a, 98).
Heisenberg makes three key claims about the nature of closed theories: first, that a closed theory covers a certain limited or bounded domain of phenomena; second, that it is a perfectly accurate description of that domain of phenomena, and third, that it is a final description—a permanent achievement that cannot be called into question by any future developments of science.Footnote 8 For example, in his 1948 Dialectica article, he explains:
What, then, finally, is the truth content of a closed-off theory? … The closed-off theory holds for all time; wherever experience can be described by means of the concepts of this theory, even in the most distant future, the laws of this theory will always prove correct. (Heisenberg Reference Heisenberg[1948] 1974, 45)
He reiterates these points again in his February 27, 1963, interview with Kuhn:
Newtonian mechanics is a limited description of nature and in that limited field it is perfectly accurate. It can never be improved. All attempts to improve Newtonian mechanics are just fruitless. … Since it is a closed axiomatic system I think it should be left as it is. … Of course it doesn't cover the whole of physics. There are other schemes. Already Maxwell theory is entirely different from it and again that is a closed scheme and cannot be improved. (Heisenberg Reference HeisenbergFebruary 27, 1963, 21–22)
Regarding his belief that closed theories are not universal, but rather only describe a limited domain, Heisenberg offers the following justification: The domain of application of a closed theory is determined by where the theory's concepts apply, and the process of axiomatizing the theory necessarily restricts the domain of applicability of the concepts. He describes the process by which axiomatization restricts the domain as follows:
For so long as concepts stem directly from experience … they remain firmly linked to the phenomena and change along with them; they are compliant, as it were, toward nature. As soon as they are axiomatized, they become rigid and detach themselves from experience. To be sure, the system of concepts rendered precise by axioms is still very well adapted to a wide range of experiences; but we can never know in advance how far a concept established through definitions and relations will take us in our dealings with nature. Thus the axiomatization of concepts simultaneously sets a decisive limit to their field of application. (Heisenberg Reference Heisenberg[1948] 1974, 44)Footnote 9
The boundaries of a particular closed theory are not given a priori, but rather discovered empirically. Heisenberg goes on to give the example of how some of the limits of the concepts of classical mechanics were found with Michael Faraday's work on the concept of a ‘field of force’. Heisenberg's conclusion is that new developments in physics, such as quantum mechanics, infringe not on the validity of classical mechanics, but only on its scope.
Specifying the domain of applicability of the concepts of a closed theory is also important for his claim of accuracy. For example, Heisenberg frequently writes that wherever the concepts of Newtonian mechanics apply, there the laws of Newtonian mechanics will always prove to be correct (see, e.g., Heisenberg Reference Heisenberg[1935] 1979, 41–42; Reference Heisenberg1971a, 96). He does not, however, provide an independent way for determining where the concepts apply, other than the criterion that the laws of the theory accurately describe the phenomena. Heisenberg combines the empirical accuracy of closed theories with the fact that they can be given an axiomatic formulation to justify his third claim of finality. He writes that the finality of closed theories is secured by “their compactness and manifold confirmation by experiment” (Heisenberg Reference Heisenberg[1972] 1983, 124), where by ‘compactness’ he means that a wide range of phenomena can be accounted for by means of a small number of postulates and axioms.
Further insight into Heisenberg's philosophy of closed theories can be gained through a detailed comparison with Thomas Kuhn's paradigms. Both exhibit a sort of holism, both lead to a revolutionary model of theory change, and both involve a notion of incommensurability. The real value of this comparison, however, lies not in the similarities, but in the differences. In particular, a detailed comparison with Kuhn reveals that Heisenberg is not an instrumentalist as is commonly presupposed, but rather is what I call a pluralistic realist.Footnote 10
3. Holism and the Revolutionary Model of Theory Change
Although the accuracy and finality claimed for closed theories suggests a disanalogy with Kuhnian paradigms, the two concepts do share a number of surprising similarities. For example, both Heisenberg's closed theories and Kuhn's paradigms exhibit a sort of holism—an interconnectedness of elements that resists small changes or modifications. In a 1958 celebration of the centenary of Max Planck's birth, Heisenberg explains the holistic nature of his closed theories by means of the following analogy:
One finds [in closed theories] structures so linked and entangled with each other that it is really impossible to make further changes at any point without calling all the connections into question … . We are reminded here of the artistic ribbon decorations of an Arab mosque, in which so many symmetries are realized all at once that it would be impossible to alter a single leaf without crucially disturbing the connection of the whole. (Heisenberg Reference Heisenberg[1958b] 1974, 28)
Holism, on this view, amounts to the claim that the structures and elements of a closed theory exhibit such a strong interdependence that one cannot modify a single element without modifying the whole. Kuhn espouses a similar holism, as he describes a paradigm as being characterized by “a strong network of commitments—conceptual, theoretical, instrumental, and methodological” (Kuhn Reference Kuhn[1962] 1996, 42).
While Heisenberg and Kuhn share a holistic view of theories, they do so for quite different reasons. Recall that the holistic nature of paradigms, for Kuhn, is due to their constitutive nature, that is, the fact that paradigms determine the relevant concepts, methods, problems and solutions—even the very data and observations that a scientist makes (see, e.g., Kuhn Reference Kuhn[1962] 1996, 110). For Heisenberg, however, the holistic nature of a closed theory arises out of the fact that it can be given an axiomatic formulation. He emphasizes this point in his interview with Kuhn:
When you have a number of axioms as Newton had in the first pages of his famous book, Principia Mathematica, then the words are not only defined by the ordinary use of the language, but they are defined by their connections. That is, you cannot change one word without ruining the whole thing. Everything is bound up. (Heisenberg Reference HeisenbergFebruary 27, 1963, 24)
The holism that Heisenberg describes here appears to involve an even greater degree of interdependence than Kuhn's. While Kuhn's holism significantly restricts the sort of modifications that a scientist can make while still remaining within a given paradigm, he does allow for modifications that fall under the headings of paradigm articulation and paradigm support. For example, he writes, “No part of the aim of normal science is to call forth new sorts of phenomena … . Nor do scientists normally aim to invent new theories. … Instead, normal scientific research is directed to the articulation of those phenomena and theories that the paradigm already supplies” (Kuhn Reference Kuhn[1962] 1996, 24). However, a change in either the concepts, methods, or theoretical commitments of a particular paradigm, would take one outside of the domain of that paradigm, and hence are not permitted within normal science.Footnote 11
Despite the different sources and strengths of their views, both Heisenberg's and Kuhn's commitment to holism leads them to adopt a revolutionary and anti-gradualist model of theory change. For Heisenberg, the holistic nature of closed theories means that when a genuinely new phenomenon is discovered, it cannot be accommodated by simply modifying an existing closed theory. Heisenberg describes this situation as follows:
The transition in science from previously investigated fields of experience to new ones will never consist simply of the application of already known laws to these new fields. On the contrary, a really new field of experience will always lead to the crystallization of a new system of scientific concepts and laws … . The advance from the parts already completed to those newly discovered, or to be newly erected, demands each time an intellectual jump, which cannot be achieved through the simple development of already existing knowledge. (Heisenberg Reference Heisenberg and Hayes[1934] 1979, 25)
In other words, new phenomena require the formation of a new closed theory. One can compare this claim to Kuhn's view that “the transition from a paradigm in crisis to a new one … is far from a cumulative process, one achieved by an articulation or extension of the old paradigm. Rather it is a reconstruction of the field from new fundamentals” (Kuhn Reference Kuhn[1962] 1996, 84–85). In both cases the need for a wholesale revolutionary change is tied to the holistic nature of the predecessor theory, which cannot be modified or extended in such a way as to account for the new phenomena.
The fact that these two share a revolutionary and anti-gradualist model of theory change quickly becomes clear in the course of the 1963 interview. Here Heisenberg mentions that he has just read Kuhn's recently published book, and notes this parallel between Kuhn's views and his own:
I have studied your book already and it gave me great pleasure to see the way you use the term paradigm. … Yes, the necessity is to break away … . One always, in such a situation, is forced to cut the branch on which one is sitting. (Heisenberg Reference HeisenbergFebruary 27, 1963, 22)
At the conclusion of their discussion of theory change, Kuhn agrees with Heisenberg that, “with regard to evolutionary views about the development of science, you and I say very much the same thing” (Kuhn Reference HeisenbergFebruary 28, 1963, 3).
Heisenberg explicitly connects his revolutionary model of theory change with his belief that, in order to make progress in science, one must abandon the concepts of the predecessor theory. In the interview with Kuhn he goes on to criticize many of the other founders of quantum theory who “were not able to do the other step which would have been absolutely necessary to come further and that is, to throw away the old physics, that is throw away all the classical concepts and replace them with new ones” (Heisenberg Reference HeisenbergFebruary 28, 1963, 13). He emphasizes this point again in 1968 when he writes:
The decisive step is always a rather discontinuous step. You can never hope to go by small steps nearer and nearer to the real theory; at one point you are bound to jump, you must really leave the old concepts and try something new … in any case you can't keep the old concepts. (Heisenberg Reference Heisenberg1968, 39)
It is particularly interesting to note here Heisenberg's emphasis on the necessity in quantum theory of abandoning the concepts of classical mechanics. From comments like these one begins to get a sense of how Heisenberg's views differ from those of Niels Bohr, who, by contrast, infamously argued for the indispensability of classical concepts in quantum theory.Footnote 12
4. Incommensurability
Heisenberg's view that each closed theory makes use of its own distinct closely knit system of concepts leads him not only to an anti-gradualist model of theory change, but also to a position very close to Kuhnian incommensurability. Kuhn describes a central component of his incommensurability thesis as follows:
Within the new paradigm, old terms, concepts, and experiments fall into new relationships with each other … . To make the transition [from Newton's] to Einstein's universe, the whole conceptual web whose strands are space, time, matter, force, and so on, had to be shifted and laid down again on nature whole. (Kuhn Reference Kuhn[1962] 1996, 149)
Four years before Kuhn, Heisenberg expressed a remarkably similar view:
New phenomena that had been observed could only be understood by new concepts which were adapted to the new phenomena. … These new concepts again could be connected in a closed system. … This problem arose at once when the theory of special relativity had been discovered. The concepts of space and time belonged to both Newtonian mechanics and to the theory of relativity. But space and time in Newtonian mechanics were independent; in the theory of relativity they were connected. … From this one could conclude that the concepts of Newtonian mechanics could not be applied to events in which there occurred velocities comparable to the velocity of light. (Heisenberg Reference Heisenberg1958a, 97–98)
Again in a later article, in which Heisenberg is discussing revolutions in the history of science, or what he prefers to call “changes of thought pattern,” he writes, “Quite generally we may say that a change of thought pattern becomes apparent when words acquire meanings different from those they had before and when new questions are asked” (Heisenberg Reference Heisenberg[1969] 1990, 156). Although Heisenberg never uses the term ‘incommensurability’, one can identify in quotations such as these two key elements of the incommensurability thesis. First, although different theories may employ the same terms, the meanings of those terms can change fundamentally. Second, each theory has its own system of concepts, and these concepts are related to each other and the world in such a way that one cannot import them into the context of a different theory.Footnote 13
Fortunately one need not speculate on how Kuhn would have reacted to Heisenberg's notion of incommensurability, for Heisenberg brings up this very issue in the course of the 1963 interview. When explaining to Kuhn his view of closed theories, Heisenberg introduces the following example:
As soon as you come to velocities, near the velocity of light, then it is not only so that Newtonian physics does not apply, but the point is that you even don't know what you mean by ‘velocity.’ Well, as you well know, you cannot add two velocities and so on, so just the word ‘velocity’ loses its immediate meaning. That, I think is a very characteristic feature of what I mean by close[d] system; that is, when you have such a system and you get disagreement with the facts; then it means that you can't use the words anymore. You just don't know how to talk. (Heisenberg Reference HeisenbergFebruary 27, 1963, 24)
Kuhn's immediate response at this point in the interview is to say: “O.K. That I entirely agree with. But I think you look too narrowly at the group of systems within which one can have that experience” (Kuhn Reference KuhnFebruary 27, 1963, 24). Kuhn clearly recognizes in Heisenberg's account elements close to his own view, though he seems to criticize Heisenberg for not applying this incommensurability thesis more widely. However, in order to better understand Kuhn's criticism here (which I will return to and discuss in Section 6), one must first take a closer look at Heisenberg's views on realism.
5. Realism: Closed Theories vs. Phenomenological Theories
There is a general consensus that Heisenberg is an instrumentalist (see, e.g., Cassidy Reference Cassidy1992; Popper Reference Popper[1963] 1989, 113; Jammer Reference Jammer1974, 58); however, a more careful reading of Heisenberg reveals that this view is mistaken. Part of the difficulty in interpreting Heisenberg on this issue is that he does not explicitly make use of the realism-instrumentalism terminology; hence some care needs to be exercised in drawing out of his statements the right implications for the realism debate. One explanation for the frequent misreadings of Heisenberg on this issue is that he often makes positivistic sounding statements when he is discussing his scientific methodology, that is, his views concerning how one should go about constructing new scientific theories.Footnote 14 It is important, however, to distinguish Heisenberg's methodological views from his views regarding the status of the final product of that methodology. In arguing that Heisenberg is in fact a realist, I do not mean to imply that he espoused some sort of hidden-variable interpretation—a realist view need not be committed to hidden-variables.Footnote 15 In calling Heisenberg a realist about quantum theory, what I mean is that he believed that quantum mechanics is an accurate, final, and true description of the way the world really is.
An important part of Heisenberg's defense of realism is the sharp distinction he draws between, on the one hand, phenomenological theories and ‘the pragmatic philosophy’, and, on the other hand, closed theories and his own philosophy. Throughout his writings, Heisenberg is quite critical of phenomenological theories.
[Phenomenological theories] can be extremely successful insofar as they can sometimes give the exact results and consequently agree extremely well with experiments. Still, at the same time they do not give any real information about the physical content of the phenomenon, about those things which really happen. (Heisenberg Reference Heisenberg1968, 33)
Phenomenological theories, then, are understood as just calculational tools, theories that ‘save the phenomena’ while failing to give us any true insight into nature.
In a short article on the role of phenomenological theories in physics, Heisenberg notes how the realist and the instrumentalist will evaluate phenomenological theories differently. He somewhat misleadingly has the Platonist stand in for the part of the realist.Footnote 16
One could at this point note that the physicist or astronomer is going to unconsciously evaluate phenomenological theories quite differently depending on whether he has had his philosophical views formed by pragmatism, or by another way of thinking, like Plato's theory of ideas. For one who has grown up in pragmatism, a phenomenological theory will be valued more highly the more successes it can exhibit and the more exact predictions it can make. However, one who has from early on been seized by the persuasiveness of Platonic thought will, above all else, judge the phenomenological theories by whether and to what extent they can lead to an understanding of the true relations. (Heisenberg Reference Heisenberg, De-Shalit, Feshbach and Hove1966, 168, my translation)
In his recollections of a 1929 conversation with Barton Hoag, an experimental physicist from Chicago, Heisenberg elaborates on the differences between his view and that of what he calls “the pragmatic philosophy.” Heisenberg has Hoag, as the spokesperson for the pragmatic view, describe the position as follows:
Perhaps you make the mistake of treating the laws of nature as absolutes, and you are therefore surprised when they have to be changed. To my mind [that is Hoag's mind], even the term ‘natural law’ is a glorification or sanctification of what is basically nothing but a practical prescription for dealing with nature. (Heisenberg Reference Heisenberg1971a, 95, paraphrasing Hoag)
The pragmatic philosophy that Heisenberg has Hoag describe here is quite clearly an instrumentalist interpretation of scientific laws. Heisenberg responds that he finds this view entirely unsatisfactory and goes on to try to convince Hoag of his own alternative view of closed theories. It is here that he presents one of his clearest endorsements of realism:
If, as we must always do as a first step in theoretical physics, we combine the results of experiments and formulae and arrive at a phenomenological description of the processes involved, we gain the impression that we have invented the formulae ourselves. If, however, we chance upon one of those very simple, wide relationships that must later be incorporated into the axiom system. … then we are quite suddenly brought face to face with a relationship that has always existed, and that was quite obviously not invented by us or by anyone else. Such relationships are probably the real content of our science. (Heisenberg Reference Heisenberg1971a, 99)
We see once again here the sharp contrast that he wants to draw between phenomenological theories and closed theories. While Heisenberg is willing to adopt an instrumentalist attitude toward phenomenological theories, he believes that closed theories should be interpreted realistically.
It is important to note that Heisenberg takes neither quantum mechanics nor classical mechanics to be a phenomenological theory. That is, he adopts a realist interpretation not only of quantum theory, but also of Newtonian mechanics. In his book Physics and Philosophy, he writes that in his view of Newtonian mechanics, “the system of definitions and axioms … is considered as describing an eternal structure of nature” (Heisenberg Reference Heisenberg1958a, 93). In light of the standard view that quantum mechanics is the theory that falsified and replaced classical mechanics, it is difficult to understand how Heisenberg could hold a realist interpretation of both these theories. In a 1935 article he tries to explain: “When one considers the basis of modern physics, one finds that it really does not infringe on the validity of classical physics. … It is not the validity but only the applicability of the classical laws which is restricted by modern physics” (Heisenberg Reference Heisenberg[1935] 1979, 42). This suggests that the way in which he maintains a realist interpretation of both classical and quantum mechanics is by denying the universality (but not validity) of these theories’ laws; this view fits with his understanding of a closed theory as the accurate and final description of a certain limited domain of phenomena.
The picture that starts to emerge from Heisenberg's views seems to be something like what Nancy Cartwright has called metaphysical nomological pluralism, which she defines as the view that “nature is governed in different domains by different systems of laws not necessarily related to each other in any systematic or uniform way” (Reference Cartwright1999, 31). The similarity between her view and Heisenberg's become even more evident in her defense of classical mechanics:
It is generally assumed that we have discovered that quantum mechanics is true … and hence classical mechanics is false. … All evidence points to the conclusion that … Nature is not reductive and single minded … and is happily running both classical and quantum mechanics side-by-side. (Reference Cartwright1999, 361)
Nomological pluralism seems to capture quite naturally the way in which Heisenberg seeks to combine the theoretical pluralism of his closed theories with a sort of realism.
This interpretation of Heisenberg finds further support in his 1942 unpublished manuscript “The Order of Reality.” In this manuscript, he introduces the idea of various distinct regions of reality corresponding to his closed theories:
By the expression ‘region of reality’ we mean a collection of nomological connections. Such a collection [of laws] must on the one hand form a solid unity, because otherwise one could not legitimately speak of a ‘region,’ while on the other hand, it must be able to demarcate itself clearly from other collections [of laws] in order to render possible a division of reality. (Heisenberg Reference Heisenberg[1942] 1998, 273, my translation)
Later, he explicitly ties these various regions of reality to his closed theories: “The collection of nomological connections that we describe using classical mechanics is closed in itself. One means by this that the collection [of laws] can be explicated in a system of concepts and axioms that contain no contradictions and can be said to be in a certain sense complete” (Reference Heisenberg[1942] 1998, 284; my translation). Heisenberg's concern here with the idea of various regions of reality speaks once again to his realist—not instrumentalist—interpretation of closed theories.
6. The Case of Ptolemaic Astronomy
In the 1963 interview, these four themes of holism, revolutionary changes, incommensurability, and realism come together. Recall (from the discussion at the end of Section 4) that while Kuhn embraces Heisenberg's notion of incommensurability, he criticizes him for not applying this notion more widely. As Heisenberg finishes describing his account of closed theories, Kuhn makes the following comment: “Now what strikes me is that there are many more systems to be found in the history of physics which would satisfy your criteria. You could have Ptolemaic astronomy …” (Kuhn Reference KuhnFebruary 27, 1963, 23). Heisenberg interjects here that Ptolemaic astronomy is merely a phenomenological theory that fails to meet his criteria for closedness. Kuhn then raises the very natural question of why Newtonian mechanics would not also count as a phenomenological theory. Heisenberg responds that Newtonian mechanics is importantly different because it is a closed axiomatic system—it exhibits a tight interconnected system of concepts and axioms, such that a single element cannot be changed without destroying the entire system. He continues:
And just because of this an experiment which cannot be described by this axiomatic system means something more radical than for Ptolemaic astronomy. In Ptolemy's case, if the orbit didn't fit, he could add other epicycles. But if an experiment does not fit in Newtonian physics, you don't know what you mean by the words. (Heisenberg Reference HeisenbergFebruary 27, 1963, 24)
Heisenberg then goes on to give the example of the change in the meaning of ‘velocity’ from classical mechanics to relativity that I quoted earlier. Notice that Heisenberg is here trying to distinguish Newton's theory from Ptolemy's by appealing to the holistic nature of Newton's theory and the concomitant idea of incommensurability.Footnote 17 Kuhn, by contrast wants to extend these notions of holism and incommensurability to the case of Ptolemaic astronomy as well. Kuhn replies,
I am perfectly sure that this is what happened to people who had trouble with the Ptolemaic system and that when you are now looking back and … seeing it as a phenomenological theory, this is in large part simply because you are too far now away from it. Just as you say, the problem one runs into in Newtonian mechanics when the velocity gets too high is … that one doesn't know what velocity is anymore. I want to say with respect to a nice little phrase like ‘the stability of the earth’ exactly that same sort of problem happens in the transition out of the Ptolemaic system. (Kuhn Reference HeisenbergFebruary 27, 1963, 24)
For Kuhn, there is, fundamentally, no dividing line between phenomenological theories and theories that should be interpreted realistically. The central challenge of Structure is to realism—the idea that through these changes in paradigm, scientists are getting closer to the truth. All theories, for Kuhn, are essentially just tools for puzzle solving.
Kuhn goes on in this interview to note another important difference between Heisenberg's views and his own:
As you perhaps see in the book [Structure], I am not very happy about taking out a sub-group of these important discarded theories and saying, ‘Now these few are still with us.’ … In the long run quantum mechanics must be the substitute for Newton in the sense that Newton is the substitute for Ptolemy and Aristotle. (Kuhn Reference KuhnFebruary 27, 1963, 25)
In this quotation we see quite clearly how Heisenberg's theoretical pluralism is at odds with what we might call Kuhn's serial theoretical monogamy.
Kuhn concludes the second day of his interview with Heisenberg on this topic by saying:
Most of the things you've said I thoroughly agree with and I'm delighted. There is one area in which I think we're in quite essential disagreement. … I find it very hard to talk about this permanent content of the closed scientific theory. … So much of what you say are things that I think of myself as also wanting to emphasize in this book. But on this sort of point I would find that this undercuts the whole point of view for me. (Kuhn Reference KuhnFebruary 28, 1963, 3)
Thus, while Heisenberg and Kuhn somewhat surprisingly share a holistic account of theories, a revolutionary model of theory change, and a notion of incommensurability, they disagree fundamentally over the epistemological status of these theories and the issue of scientific realism.
7. Conclusion
The similarities between Heisenberg and Kuhn on these issues of holism, revolutionary change, and incommensurability naturally lead one to wonder whether Kuhn found in Heisenberg's writings a source for his own views. In this key 1963 interview that I have discussed in some detail, however, there is no evidence of a direct causal influence. Kuhn does not express a prior familiarity with Heisenberg's account of closed theories, nor does he take this opportunity to discuss the origin of his own views on these topics. Beller (Reference Beller1999), on the other hand, has argued for an indirect influence of Heisenberg on Kuhn via the work of N. R. Hanson. My aim in this paper, however, has not been to pursue this thesis, nor to speculate on the sources or origins of Kuhn's views. Instead, I have tried to show that a careful comparison of Heisenberg's and Kuhn's views—facilitated by this remarkable interview in which they had the opportunity to sit down and discuss their views face to face—can bring us to a deeper understanding of Heisenberg's philosophy of science.
In particular, this comparison provides a foundation from which to correct several widespread and persistent misunderstandings about Heisenberg's views. First, Heisenberg is neither a naive positivist nor an instrumentalist. Instead, he espouses an interesting and novel form of pluralistic realism. Second, this deeper understanding of Heisenberg reveals a striking contrast between his views and Bohr's, calling into further question the unity of the so-called Copenhagen interpretation. Finally, these parallels between Heisenberg's and Kuhn's views show that there is a richness in Heisenberg's thinking about the nature of theories, realism, and theory change worthy of more careful consideration. An examination of Heisenberg's account of closed theories leads not only to a new understanding of Heisenberg's philosophy, but also to an appreciation of the potential relevance of his philosophy of science for contemporary debates.
