The Idea of Mechanism

1. The Idea of Mechanism Stathis Psillos University of Athens Greece

2. What is its content? It does not have a fixed and definite (ahistorical) content. In the seventeenth century and beyond, the idea of mechanism had something to do with matter in motion subject to mechanical laws. A mechanism, nowadays, is virtually any relatively stable arrangement of entities such that, by engaging in certain interactions, a function is performed, or an effect is brought about. To call a structure a mechanism is simply to describe it in a certain way (relative to a function)—focusing on a certain decomposition of it and on the steps or processes through which the parts of a whole interact with each other and bring about an effect.

3. Aim • A philosophical critique of mechanism • MECHANISM as a methodologically useful concept—and just about that!

4. Two Conceptions of Mechanism—the first Mechanical Mechanism • Mechanisms are configurations of matter in motion subject to mechanical laws (the laws of mechanics). Margaret Wilson (Ideas and Mechanism, 1999, xiii, note1) : The mechanism characteristic of the new science of the seventeenth century may be briefly characterised as follows: Mechanists held that all macroscopic bodily phenomena result from the motions and impacts of submicroscopic particles, or corpuscles, each of which can be fully characterised in terms of a strictly limited range of (primary) properties: size, shape, motion and, perhaps, solidity and impenetrability. A narrow idea of mechanism

5. The second conception Non-mechanical mechanism • Any arrangement of parts into wholes in such a way that the behaviour of the whole depends on the properties of the parts and their mutual interactions • A concomitant conception of mechanical explanation as a kind of de-compositional explanation: an explanation of a whole in terms of its parts, their properties and their interactions. A broader characterisation of the idea of mechanism Kant (KU 5:408) : If we consider a material whole, as far as its form is concerned, as a product of the parts and of their forces and their capacity to combine by themselves … we represent a mechanical kind of generation.

6. The first conception—an overview and a problem Boyle—The mechanical hypothesis:all natural phenomena are produced by the mechanical interactions of the parts of matter according to mechanical laws. Descartes—Nature is mechanical and subject to mechanical explanation. Explanation should have a deductive structure, but the explanatory hypotheses should be mechanical, cast in terms of the shape, size, position and motion of particles of matter, and the selfsame mechanical principles should account for the whole of nature, both in the heavens and on the earth. Mechanism and unification become one. Unification under mechanical laws. But the specific principles of the mechanical conception of nature were heavily debated. For instance, some mechanical philosophers (notably Pierre Gassendi) subscribed to atomism, while others (notably Descartes) took the universe to be a plenum, with matter being infinitely divisible.

7. Newton and a new question • Newton—the content of the mechanical conception of nature is altered and broadened. A new category, force, was firmly introduced alongside the two traditional mechanical categories, matter and motion. But a new question arises, with the emergence of systematic theories of heat, electricity and magnetism How are these theories related to the theories of mechanics. In particular, do thermal, electrical and magnetic phenomena admit of mechanical explanations?

8. Two ways to address the new question One, developed mostly in Britain, was the building of mechanical models. These models were meant to show a) the realisability of the system under study (e.g., the electromagnetic field) by a mechanical system; (the system is not a free-floater) and b) the possible inner structure by means of which the physical system under study operates. The other way (developed mostly on the continental Europe) was the construction of abstract theories under which the phenomena under study were subsumed and explained. These theories were mechanical because they started with principles that embodied laws of mechanics and offered explanation by deductive subsumption. Even within the model-building tradition, especially in its post-Maxwellian period, mechanical models were taken to be, by and large, heuristic and illustrative devices—the focus being on the development of systematic theories under which the phenomena under study are subsumed and explained.

9. Joseph Larmor’s division (t)he division of the problem of the determination of the constitution of a partly concealed dynamical system, such as the aether, into two independent parts. The first part is the determination of some form of energy-function which will explain the recognised dynamical properties of the system, and which may be further tested by its application to the discovery of new properties. The second part is the building up in actuality or in imagination of some mechanical system which will serve as a model or illustration of a medium possessing such an energy function (1894, 417).

10. Poincaré’s problem Once the first part of Larmor’s division of labour is dealt with, the second part (the construction of possible configurations of matter in motion) takes care of itself. 1900, International Congress of Physics in Paris ‘Relations entre la Physique Expérimentale et de la Physique Mathématique’ Most theorists have had a constant predilection for explanations borrowed from mechanics. These attempts had taken two particular forms: they traced all phenomena back either to the motion of molecules acting-at-a-distance in accordance to central force-laws; or to the contiguous actions of molecules that depart from the rectilinear path only by collisions. “All [physicists] wish to bend nature into a certain form, and unless they can do this they cannot be satisfied” . “Is nature flexible enough for this?”

11. A positive but surprising answer Poincaré’s theorem: a necessary and sufficient condition for a complete mechanical explanation of a set of phenomena is that there are suitable experimental quantities that can be identified as the kinetic and the potential energy such that they satisfy the principle of conservation of energy. Given that such energy functions can be specified, Poincaré proved that there will be some configuration of matter in motion (that is, a configuration of particles with certain positions and momenta) that can underpin (or model) a set of phenomena. As he put it: In order to demonstrate the possibility of a mechanical explanation of electricity, we do not have to preoccupy ourselves with finding this explanation itself; it is sufficient to know the expressions of the two functions T and U which are the two parts of energy, to form with these two functions the equations of Lagrange and, afterwards, to compare these equations with the experimental laws (1890/1901, viii).

12. The irony A corollary: if there is one mechanical explanation of a set of phenomena, i.e., if there is a possible configuration of matter in motion that can underpin a set of phenomena, there is an infinity of them. And not just that. Another theorem proved by the French mathematician Gabriel Königs suggested that for any material system such that the motions of a set of masses (or material molecules) is described by a system of linear differential equations of the generalized co-ordinates of these masses, these differential equations (which are normally attributed to the existence of forces between the masses) would be satisfied even if one replaced all forces by a suitably chosen system of rigid connections between these masses. Heinrich Hertz (1894) had made use of this result to develop a system of mechanics that did away with forces altogether.

13. Setting limits to mechanistic ambitions Though the possibility of a mechanical explanation of electromagnetic phenomena is secured, the empirical facts alone could not dictate any choice between different mechanical configurations that satisfy the same differential equations of motion. How can scientists choose among these possible mechanical configurations? For Poincaré this was a misguided question. “The day will perhaps come when physicists will no longer concern themselves with questions which are inaccessible to positive methods and will leave them to the metaphysicians” (1902, 225). His advice to his fellow scientists was to content themselves with the possibility of a mechanical explanation of all conservative phenomena and to abandon hope of finding the true mechanical configuration that underlies a particular set of phenomena. He (1900a, 1173) stressed: We ought therefore to set limits to our ambition. Let us not seek to formulate a mechanical explanation; let us be content to show that we can always find one if we wish. In this we have succeeded.

14. Unity not mechanism According to Poincaré, the search for mechanical explanation (i.e., for a configuration of matter in motion) of a set of phenomena is of little value not just because this search is massively underdetermined by the phenomena under study but mainly because this search sets the wrong target. What matters is not the search of mechanism per se, but rather the search for unity of the phenomena under laws of conservation. Understanding is promoted by the unification of the phenomena and not by finding the mechanisms that bring them about. As he said “(...) The end we seek (...) is not the mechanism. The true and only aim is unity” (ibid.). Is the law of conservation of energy a mechanical principle? • No fixed characterisation of what counts as mechanical • But it does block vitalism

15. The first conception comes full circle! Poincaré’s problem was not that mechanisms are unavailable or non-existent. It was not that mechanical explanation—that is, explanation in terms of the laws of mechanisms—was impossible. Rather, his problem was that, mechanisms are very easy to get. Under certain plausible assumptions that involve the principle of conservation of energy, the call for mechanical explanation is so readily satisfiable that ceases to be informative. Phenomena subject to conservation laws are not free-floaters. When it comes to conservative phenomena, searching for mechanisms is uninteresting.

16. The second conception—Hegel’s critique Georg Wilhelm Friedrich Hegel--Science of Logic. Wissenschaft der Logik (1812-1816) He attacked the idea that all explanation must be mechanical. Making mechanism an absolute category—applicable to everything—obscures the distinction between explanation and description and hence undermines itself. A perfectly sensible and quite forceful objection to the view that all explanation is mechanical explanation; that the only mode of explanation is mechanical; that to explain X is to offer a mechanical explanation of it.

17. Hegel’s argument Mechanistic explanation proceeds in terms of breaking an object down to its parts and by showing its dependence on them and their properties and relations. Explanation, then, amounts to a certain de-composition of the explanandum, viz., of a composite object whose behaviour is the result of the properties of, and interactions among, its parts. But there are indefinitely many ways to decompose something to parts and to relate it and its behaviour to them. For the call for explanation to have any bite all, there must be some principled distinction between those de-compositions that are merely descriptions of the explanandum and those de-compositions that are genuinely explanatory. In particular, some decomposition—that which offers the mechanical explanation—must be privileged over the others, which might well reflect only pragmatic criteria or subjective interests. How is this distinction to be drawn within the view that all explanation is mechanical? If all explanation is indeed mechanical, and if mechanical explanation amounts to de-composition, no line can be drawn between explanation and description—no particular way to de-compose the explanandum is privileged over the others by being mechanical; mechanical as opposed to what?All decompositions will be equally mechanical and equally arbitrary. Hence, there will be no difference between explanation and description.

18. Hegel’s problem The unity of a mechanism is not just of matter of arranging a set of elements into a whole; nor is it just a matter of listing their properties and mutual relations. There are indefinitely many ways to do this and most of them will be arbitrary since they will not be explanatorily relevant. The unity of the mechanism comes from something external to it, viz., from its function—from what it is meant to be a mechanism for. The function that a mechanism performs is something external to the description of the mechanism. It is the function that fixes a criterion of explanatory relevance. Some descriptions of the mechanism (decompositions) are explanatorily relevant while others are not because the former and not the latter explain how the mechanism performs a certain function.

19. The lesson How the mechanisms are individuated is a matter external to them—what counts as a mechanism, where it starts and where it stops, what kind of parts are salient and what kind of properties are relevant depend on the function they are meant to perform. The unity of the mechanisms is not intrinsic but extrinsic to them. Nature, even if it is mechanical, does not fix the boundaries of mechanisms. When it comes to the search for mechanisms, anything can count as a mechanism provided it performs a function that it is meant to explain.

20. The resurrection of mechanism: causation Generic account of mechanism—mechanisms as processes of a certain sort • Mechanisms have been taken to be the tie that connects cause and effect and explains the productivity of the cause: how the cause brings about the effect. As Hume claimed, the alleged necessary tie between cause and effect is not observable. But as John Mackie (1974) argued, we might still hypothesise that there is such a tie, and then try to form an intelligible theory about what it might be. The tie consists in a “causal mechanism”, that is, “some continuous process connecting the antecedent in an observed (...) regularity with the consequent” (1974, 82). But processes of what sort?

21. Persistence • Mackie--the causal mechanism consists in the qualitative or structural continuity, or persistence, exhibited by certain processes, which can be deemed causal. There needn’t be some general feature (or structure) that persists in every causal process. What these features are will depend on the details of the actual “laws of working” that exist in nature. But: this account borders on vacuity. It’s not that nothing persists in processes. Rather, it is that something does persist always, even in typical pseudo-processes. The notion of persistence is unhelpful unless there is a suitable characterisation of the properties, or the features, that persist. An account of mechanism in terms of the persistence of some feature of a process suffers from the problem that Hegel identified, viz., that it cannot differentiate between explanation and description—it cannot tell us what kinds of persistence are explanatory and what are not.

22. Transference Salmon’s account of causation: Causal processes, causal interactions, and causal laws provide the mechanisms by which the world works; to understand why certain things happen, we need to see how they are produced by these mechanisms (1984, 132). But Salmon was aware of the fact that ‘persistence of structure’ is not enough to characterise a process as causal. Structure-transference (or more generally, mark-transmission) became the distinguishing characteristic of a mechanism.

23. The problems The very idea of mark-transmission cannot differentiate causal processes (and hence mechanisms) from non-causal ones, since any process whatever can be such that some modification of some feature of it gets transmitted after a single local interaction. Salmon strengthened his account of mark-transmission by requiring that in order for a process P to be causal it is necessary that “the process P would have continued to manifest the characteristic Q if the specific marking interaction had not occurred” (1984, 148). But this takes us back to persistence! In effect, the idea is that a process is causal if a mark made on it (a modification of some feature it) gets transmitted after the point of interaction and in the absence of this interaction, the relevant feature would have persisted, where the required persistence is counterfactual. In any case, we are in need of a theory as to which properties are such that their presence or modification marks a causal process and whose possession marks the presence of a causal mechanism. This is none other than Hegel’s problem.

24. Possession Dowe (2000, 89): It is the possession of a conserved quantity, rather than the ability to transmit a mark, that makes a process a causal process. Dowe fixes the characteristic that renders a process causal and, consequently, the characteristic that renders something a mechanism. A conserved quantity is “any quantity that is governed by a conservation law” (2000, 91), e.g., mass-energy, linear momentum and charge. Two objections: Too narrow—applicable only to physical mechanisms. Non-physical processes (biological, geological, medical, social) should be understood either in a reductive way or in non-mechanistic terms. Too close to Poincaré’s problem for comfort.

25. The new mechanists The New Mechanism is not tied to a mechanical conception of nature of the sort that has characterised the Old Mechanism. Nor it is committed to the view that mechanisms are simply suitably understood causal processes. The New Mechanism has learned the lessons that are connected to Hegel’s problem. A mechanism is a complex system that consists of some parts (its building blocks) and a certain organisation of these parts, which determines how the parts interact with each other to produce a certain output. The parts of the mechanism should be stable and robust, that is their properties must remain stable, in the absence of interventions. The organisation should also be stable, that is the complex system as a whole should have stable dispositions, which produce the behaviour of the mechanism. Thanks to the organisation of the parts, a mechanism is more than the sum of its parts: each of the parts contribute to the overall behaviour of the mechanism more than it would have achieved if it acted on its own.

26. Two major conceptions Stuart Glennan Machamer-Darden-Craver (MDC) The major difference: what makes the mechanism run? Top-down (laws) vs bottom-up (properties and activities) approaches.

27. No mechanism, simpliciter • No talk of mechanism simpliciter. As Glennan has put it, a mechanism is always a mechanism for a behaviour. The unity of the mechanism is determined by its function. The very same complex system may issue in different behaviours. What the mechanism does determines its boundaries, its division into parts and the relevant modes of interaction among these parts. This is coming to terms with Hegel’s problem. Glennan (2002, S344): (M) A mechanism for a behaviour is a complex system that produces that behaviour by the interaction of a number of parts, where the interactions between parts can be characterized by direct, invariant, change-relating generalizations. ‘direct, invariant, change-relating generalizations’: it connects mechanisms with Woodward’s replacement of laws with generalisations (or relations) that remain invariant under actual and counterfactual interventions. –See my 2004!

28. Causation iff mechanism? Glennan (1996, 56): “a relation between two events (other than fundamental physical events) is causal when and only when these events are connected in the appropriate way by a mechanism”. A quick problem: it follows that the parts of mechanism are not connected causally with each other since there is no mechanism that connects them—unless the whole idea of mechanism is trivialised (mechanisms all the way down). In any case, Glennan denies that it is mechanisms all the way down. Fundamental laws are not mechanically explicable—there is no mechanism that underlies them.

29. Not necessary! Only when? • Putative counterexamples—causation by disconnection; causation by absences. • But also: the sinking of the Titanic and the collision with the iceberg. Causation yes! But no mechanism (at least no non-trivial mechanism). • No causation at the fundamental level—where there are no mechanisms? The irony: fundamental laws might be mechanically explicable in Poincaré’s sense.

30. Sufficient, at least? Notice: ‘events are connected in the appropriate way by a mechanism. If this is omitted, anything admits of a mechanical explanation in terms of M.   E.g. Shadow and height of pole. There is a mechanism that connects the length of the shadow and the height, but the shadow does not cause the height. Presumably, the events are not connected in the appropriate way. But what makes a way appropriate?  Whatever it is (call it X), it is only mechanism + X  causation.

31. Causing behaviours Another problem: mechanism is always mechanism for behaviour X. But behaviours are not events. Mechanism M brings about (and hence causally explains) behaviour B Glennan (2002; 2005) (MM) mechanical model: (i) a description of the mechanism’s behaviour; and (ii) a description of the mechanism which accounts for that behaviour The external behaviour of mechanism is the explanandum, and the internal structure of the mechanism is the explanans.

32. But then again not? How the mechanism is wired depends on the behaviour to be explained. If causation is a matter of mechanism M bringing about behaviour B, whether or not M causes B will depend on how the mechanism is circumscribed and what the behaviour to be causally explained is (and how it is picked). Also M causes, typically, C too, where C is a side effect. But if M is not the mechanism for C (since C is a mere side effect), and if causation requires mechanism, then M does not cause C—which is absurd. Cf. heart-beating—circulation of blood; noise.

33. Necessity and mechanism Glennan (1996, 64) : “The necessity that distinguishes connections from accidental conjunctions is to be understood as deriving from a underlying mechanism” Qualms about necessity, but… • Causation1—necessary succession in time vs Causation2—mechanical I1: Reduction of complex wholes to aggregates of simpler elements I2: The necessary connection between two wholes completely reduces to necessary connections between the elements of the one and the elements of the other. • These two ideas are distinct. I2 does not follow from I1.

34. Dualism Machamer-Darden-Craver (2000, 3) “Mechanisms are entities and activities organised such that they are productive of regular changes from start or set-up to finish or termination conditions” MDC: a mechanism consists of two distinct kinds of building blocks—entities (organised in a stable way into a spatio-temporal pattern) andactivities. Activities are meant to explain how change is brought about, that is how entities interact with each other, and how, as a result of these interactions, the mechanism as a whole brings about an effect or performs a function. They are the ontic correlates of (transitive) verbs and are necessary for the grounding of the productivity of mechanisms.

35. Do we need activities? • They seem to be nothing but an elaborate characterisation of processes or a summing up of the capacities that characterise entities Without them, we cannot understand the productivity of the mechanism. • But productivity is an overarching activity! • And in any case, so be it. For many, there is nothing like productivity of causation—just some kind of robust dependence. Mechanisms, after all, add new loops in causal chains.

36. The ubiquity of mechanisms Even a cursory look at journals or web-pages returns titles like ‘the mechanism of nerve growth’, ‘How Antibiotics Work—the Mechanism of Action’, ‘the mechanism of muscle contraction’, ‘the mechanism that controls the formation of the vertebral column’, ‘the mechanism of malaria’, ‘the mechanism of plate tectonics’, ‘anabolic-androgenic steroids: mechanism of action and effects on performance’. All this proves is that MECHANISM is a methodologically useful concept— nothing more than that. We are interested in mechanisms because we are interested in patterns that exemplify unity, stability and predictability.

37. The philosophical questions • Are mechanisms the ultimate building blocks of the universe? The answer is YES and NO. Yes—given that the world is governed by conservation laws. No—given that there are many ways to skin the cat! • Is causation mechanistic? NO. In fact, there is no need for a metaphysical theory of causation. • Does search for mechanism improve understanding? YES—The description of a mechanism is a theoretical description and, as such, it tells a story as to how the phenomenon under study is brought about—if the story is true, our understanding of nature is enhanced.

38. Hume and Hegel Hume denied that causation involves “some secret mechanism or structure of parts, upon which the effect depends”. Nothing so far has proved him wrong. Hegel did think that mechanism is a form of objectivity, claimed that it is applicable to areas other than “the special physical department from which it derives its name” but denied that it is an “absolute category” that is constitutive of “rational cognition in general”. Nothing so far has proved him wrong.

39. Hence… Insofar as mechanisms are taken to be the local tie between cause and effect—the locus of necessity and the intrinsic basis for the relation between cause and effect—there is little, if any, hope that there is anything that can play this role; let alone anything along the lines of the diverse things that are identified as mechanisms by philosophers and scientists. Insofar as mechanisms are taken to be stable explanatory structures (whose exact content and scope may well vary with our best conception of the world) which enhance our understanding of how some effects are brought about or are the realisers of certain functions, mechanisms can play a useful role in the toolkit of explanation.