Structural realism
The structures realism (SR) is that our scientific theories provide a label for a variety of epistemological positions that one in the thesis us no objective but a structural approach to the world. A distinction is made between an epistemic SR (ESR) and an ontological variant of the SR (OSR): The ESR states that scientific theories refer to structures in the world, whereby the objects on which the structures are based are epistemically inaccessible to us. The ontic variant shares the position of structural access to the world, but, in contrast to the ESR, claims that there are only structures (eliminative OSR) or objects characterized except for relations (the moderate version of the OSR) and at the same time denies the existence of objects as Support of structures. The SR thus represents a special variant of scientific realism , according to which the best and most mature theories in science are approximately true and the terms contained in those theories represent objects that exist but cannot be directly observed or - as in the case of the SR - existing structures.
What is a structure?
In a purely formal view, structures can be explained as follows: Given is an (unstructured) set a = [a1, a2 ... an] of n objects (objects) ai, where the n objects are related to each other. The tuple T = [a, R (a)] is then a structure. The objects ai can only be individuated via the relations R (a). Structures are determined by relations and are therefore explained relationally. In a broader sense, a structure is a network of physical relations between objects (see Lyre 2006).
Recent debate about structural realism
Arguments against Scientific Realism
Scientific realism asserts, on the one hand, the existence of an external world that is independent in the epistemological sense; Furthermore, on the other hand, the central thesis is represented that our best and most mature scientific theories describe the nature of that external world correctly and that the theoretical terms essentially refer to the characteristic features of that world. The current strongest argument in favor of scientific realism is the so-called " no-miracle " argument, according to which scientific realism offers the best explanation for the predictive success of the sciences (Putnam 1975). The prediction success is guaranteed precisely because scientific theories accomplish the achievement of delivering approximately true descriptions of the world and its nature. This means that theoretical terms refer to actual conditions in the world, otherwise the prediction success would be a miracle. On the one hand, the argument of pessimistic meta-induction (PMI) and, on the other hand, the argument of underdetermination of theories (TUB) speak against scientific realism.
The discussion about the SR in recent philosophy of science goes back to an essay by John Worrall (1989). In his work, Worrall formulates a possibility to represent a scientific realism that tries to plausibly explain the predictive success of scientific theories without succumbing to the common anti-realistic objections: on the one hand the argument that theories are underdetermined (TUB) and on the other the pessimistic meta-induction (PMI) argument. These anti-realist objections agree that they question the real status of the entities that are not directly observable as formulated in a certain theory and thus also question the ontology on which a theory is based (as a true reference to the world).
The Pessimistic Meta-Induction (PMI) Argument
The PMI argument essentially states that the conceptions of the ontological nature of entities postulated in the respective accepted scientific theories change over time. In the past there were a large number of theories that were accepted and mature in their time, which in the further course of scientific development have proven to be fundamentally wrong. The terms postulated in these theories referred to objects which, however, did not actually exist and the meaning of those terms thus turned out to be empty. Phlogiston or ether theories are cited by anti-realistic representatives as exemplary cases. It follows from the meta-induction inference that there is no guarantee for our today's best theories that they refer to actual entities, the theories beyond that contain nothing that can be absolutely interpreted as an approximation to a truth (on this, Lyre 2004 & 2006).
The theory underdetermination argument
Another common argument against scientific realism is the TUB argument. It says that even an abundance of data leaves the theory based on it in principle under-determined. For every possible theory there would and could therefore be empirically equivalent competition theories, the competition resulting from the fact that the same data can be traced back to different and mutually incompatible theoretical assumptions and thus can have different ontologies. In this form, this thesis is stronger than the confirmation holism , which is also called the “Duhem-Quine thesis” after Quine and Duhem , according to which not individual sentences, but only comprehensive sentence systems can be empirically tested. Quine exaggerates this idea by claiming that any anomaly in an observation can be embedded in any theory by changing possible other parts of the theory. In such a theory, formulated as a sentence system, there is always an indefinite space that is reserved for methodological and pragmatic criteria (Lyre 2004 & 2006).
Structural Realism: The best of both worlds?
Epistemic structural realism
According to Worrall, the problem of scientific realism is that all anti-realist objections are correct, provided that scientific realism continues to adhere to an ontology that formulates the existence of entities of an objective nature. Structural realism, on the other hand, says that our best theories refer to structural and non-representational entities. Worrall understands structures as networks of relations between (physical) objects. He assumes that the objects are individuated by intrinsic properties, so that these also underlie the relationships and thus also the structures imprinted on them. However, according to Worrall, we cannot see those intrinsic properties. Structural realism thus contains an epistemological thesis: all we can grasp are the relations emanating from the objects and the structure imprinted on them. Only vague ideas are formulated about the ontology underlying the structures. Thus, the objects are assigned intrinsic properties, but these are beyond any possibility of knowledge. Worrall is silent about how something that is epistemically closed to us can at the same time be presupposed as an ontological carrier. But this circumstance is initially not a problem for his structural realism: Because it is certainly possible, for example, to recognize a fishing net as a net without having to know anything about the material nature of an individual node at the same time. The ontology on which it is based does not initially play a role in recognizing a structure.
A structurally realistic approach also enables the objections of TUB and PMI to be avoided. Thus, as formulated in the TUB's objection, competing theories may be characterized by a different ontology, but this ontological difference only exists with regard to the alleged object-like entities, but not with regard to their structure. The history of science makes it clear that theories change over time, show breaks and incommensurabilities, but these show up primarily in relation to the significant objects, but less clearly in relation to their structural properties.
A structural-realistic consideration therefore enables a more conclusive reconstruction of a constant and cumulative scientific development if one wishes to defend the idea of a scientific development as a history of the progress of scientific knowledge. Structural realism thus holds out the prospect of “the best of both worlds” according to Worrall: On the one hand, the explanatory power of the “no-miracle” argument is emphasized and thus the reference between the theoretical term and the world is postulated as actually given. On the other hand, structural realism does not succumb to the objections of TUB and PMI (see: Worrall 1989).
Worrall has formulated a moderate epistemic structural realism (ESR): Our best and most mature scientific theories describe the structural properties (e.g. symmetries) of a support that lies beyond our knowledge, which regardless of non-structural properties (e.g. . Spacetime points as non-observable entities).
Ontological structural realism
The discussion about structural realism received a new impetus from an essay by James Ladyman (1998). Ladyman distinguishes between an epistemic variant (ESR) and an ontic variant (OSR) of structural realism. While ESR representatives like Worrall assume that both the relations between the objects and the structure imprinted on them on the one hand and the intrinsic properties on which the objects are based exist on the other hand, we have no epistemic access to these, OSR representatives claim that it there are no such objects as supports of a structure. In an eliminative sense, only structures exist, namely structures as physical relations, but no objects that function as relata in the relations and, through their intrinsic properties, justify the relations and thus the structural relationships between them (Ladyman 1998).
In contrast to this eliminative view, a non-eliminative view of the existence of objects has developed among the OSR representatives: According to this, those objects - from which the eliminative variant assumes that they do not exist - exist here as individuated relata in a relation. Objects are therefore a place in a structure. Those objects are constituted by their relational properties. Nevertheless, they do not exist independently and have no intrinsic properties, but are conditioned by their role in the relational structure. This view is also a criticism of the eliminative variant of the OSR. Chakravartty (1998, 399) says: "One cannot intelligibly subscribe to the reality of relations unless one is also committed to the fact that some things are related". An analogy: How can one speak of groups in a meaningful way if one denies the existence of the individual group members? (See Esfeld and Lam 2008, Lyre 2004, and Stachel 2006).
Ladyman tries to illustrate the plausibility of his approach of differentiating between ESR and OSR using an example from the problem area of individuality in quantum theory based on the work of Steven French (French / Ladyman 2003). Lyre (2006):
“According to the common understanding, the quantum statistics of indistinguishable particles violate the Leibniz principle of the identity of the indistinguishable. In the absence of an alternative individuation principle, one speaks of the non-individuality of quantum objects. In contrast, French in particular emphasized that quantum theory is definitely compatible with an ontology of individuals, insofar as one can hold on to the metaphysical assumption of a Haecceitas of objects, although the properties that can be operationally ascertainable in them do not allow any distinction. According to French, there is rather an elementary “metaphysical underdetermination” (in contrast to the usual TUB), insofar as the fundamental building blocks even leave us in the dark with regard to their individual character. According to French and Ladyman, this is to be seen as direct evidence for OSR, since an underdetermination of the subject ontology in this sense amounts to a dissolution of the subject concept. "
This example illustrates the shift in approaches in the more recent SR debate. While Worrall referred to the common anti-realist objections TUB and PMI to support his argument for an epistemic, structural-realistic view, French and Ladyman try to construct their arguments for their ontic variant from a direct treatise on fundamental physics.
literature
- Martin Carrier : Realism, more scientific. , in: Jürgen Mittelstraß (Hrsg.): Encyclopedia Philosophy and Philosophy of Science. 2nd Edition. Volume 7: Re - Te. Stuttgart, Metzler 2018, ISBN 978-3-476-02106-9 , pp. 13–15 (detailed references)
- A. Chakravartty: Semirealism. In: Studies in History and Philosophy of Modern Science. 29 (1), pp. 391-408
- A. Chakravartty: The Structuralist Conception of Objects. In: Philosophy of Science. 70, pp. 867-878
- A. Chakravartty: Structuralism as a Form of Scientific Realism. In: International Studies in Philosophy of Science. 18, pp. 151-171
- M. Esfeld, V. Lam: Moderate structural realism about space-time. In: Synthesis. 160, pp. 27-46
- M. Dorato: Substantivalism, relationism and structural spacetime realism. In: Foundations of Physics. 30 (10), pp. 1605-1628
- St. French, James Ladyman: Remodeling Structural Realism. Quantum Physics and the Metaphysics of Structure. In: Synthesis. , 136 (1) 2003, pp. 31-56
- James Ladyman: What is Structural Realism? In: Studies in History and Philosophy of Science. 29 (3), pp. 409-424
- James Ladyman: Entry Structural Realism in the SEP (with bibliography)
- H. Lyre: Local Symmetries and Reality. mentis, Paderborn, 2004
- H. Lyre: Structural Realism (PDF; 98 kB)
- D. Papineau (Ed.): The Philosophy of Science. Oxford University Press, Oxford 1996, (A collection of classic essays on scientific realism - including Worrall (1989))
- H. Putnam: Mathematics, Matter and Method. Cambridge 1975
- S. Saunders: Structural realism again. In: Synthesis 136, pp. 127-133
- BC Van Fraassen: Structure: Its shadow and substance. In: The British Journal for the Philosophy of Science 57, pp. 275-307
- BC Van Fraassen: Structuralism (s) About Science: Some Common Problems. In: Proceedings of the Aristotelian Society LXXXI, pp. 45-61
- I. Votsis: Is Structure Not Enough? In: Philosophy of Science. 70, pp. 879-890
- I. Votsis: The upward path to structural realism. In: Philosophy of Science 72, pp. 1361-1372
- J. Worrall: Structural realism: The best of both worlds? In: Dialectica 43, pp. 99-124. [Republished and printed in D. Papineau 1996 (see above)]