Objections and Replies
In this section I respond to some common objections to the idea that chemical substances and molecular structure could be strongly emergent.
Objection 1: The Chemical Bond Is “Just a Model”
One response to the foregoing arguments, available to the philosopher or scientist who is temperamentally disposed to physicalism, is to deny the reality of anything which is irreducible to physics, arguing that anything which cannot be reduced to the physical is of dubious physicalistic respectability. To the emergentist, this is a cheap move, since it begs the question by declaring unreal anything that doesn’t fit within the physicalist’s philosophy. And a cheap move it is, unless it is backed up with independent grounds for denying the reality of the items in question. In the case of abstract objects such as numbers, independent grounds for questioning them might be that they are not located in time and space, raising the question of how we can know anything of them. A more detailed and interesting version of this kind of response would draw upon independent scientific considerations in the case against the dubious items. Alexander Rosenberg (1994) has argued that biology studies properties and processes which have been honed by natural selection. They are highly complex and multiply realised, so it is beyond human cognition to grasp the underlying (chemical and physical) reasons why these processes work in just the ways they do. Instead, biology must fashion functional explanations which Rosenberg proposes to interpret instrumentally, since they do not latch on to the fundamental (physical) forces that drive things. Even in this case, the independent grounds for doubting the reality of functionally characterised properties and processes are only semi-scientific. There have long been worries within biology that function has a whiff of teleology, but Rosenberg’s argument also depends on the presumption that the underlying reasons why biological processes work the way they do can only be found in the chemical and physical realisers.8 That is a different objection (see below, Objection 3).
In the case of chemistry, the scientific pedigree of instrumentalism about structure is quite as long as that of structural explanation itself. Around the mid-nineteenth century, chemists were divided on whether or not chemical formulae should be used; they were divided on whether or not chemical formulae should be given an atomistic interpretation, and what this involved; and they were divided on whether, under an atomistic interpretation, chemical formulae should be interpreted literally (see Rocke 1984). This can be read partly as reasonable caution. Structural explanation was purely hypothetical in the 1860s and 1870s. Chemists constructed a range of possible structures which both respected the elemental composition of the substance, and were “ ‘legal’ by valence rules,”9 to borrow a phrase from Alan Rocke (2010: 132). They then selected from among these possibilities on the basis of chemical evidence. It was only in the twentieth century that X-ray crystallography and various kinds of spectroscopy allowed structural theory and experiment to become more closely integrated, with the measurement of (for instance) bond lengths and (vibrational) force constants. A second problem was that molecular structures consisted of atoms connected by bonds, but the bond was no more than an explanatory role in a theory. There was no account of what bonds were, or how they attached one atom to another. G. N. Lewis identified paired electrons as the realisers of this role, but given that Lewis’s atom was static while physics seemed to demand constant motion, it was far from clear how the physicists’ and the chemists’ models could both be true (see Arabatzis (2006: chapter 7)), with Lewis even querying whether chemistry would require a revision of Coulomb’s law at short distances (Lewis 1917). By the mid-1920s, quantum mechanics had come into being, a theory which seemed to pose severe difficulties for Lewis’s conception of structure, because electrons ought to be delocalised: smeared out across the whole molecule, rather than held static between two atoms.
As we have already seen, the advent of quantum mechanics in the 1920s brought in its train fundamental equations describing molecules which could not be solved exactly. The chemists and physicists who faced this situation in the early days of quantum mechanics developed semi-empirical models. They interpreted the situation in quite different ways (see Hendry 2003). Linus Pauling saw quantum chemistry as a synthesis of quantum mechanics and autonomous structural insights provided by chemistry. John Clarke Slater, who was, with Pauling, one of the founders of the valence- bond method for constructing semi-empirical models of bonding, saw that method instead as something that stood proxy for the exact equations which Dirac had recognised to be “much too complicated to be soluble”. On this view, quantum chemistry should be much less autonomous, with every explanatory step justified as one that could also be made in the exact theory. The problem is that these strictures have only rarely ever been met. The explanatory and predictive successes of quantum mechanics in chemistry, including the novel predictions provided by the Woodward-Hoffmann rules (see Brush 1999) were achieved through simplified models which neglected the quantum-mechanical character of parts of the molecules whose behaviour they predicted, assuming them instead to be classical. Quantum chemistry seems more like Pauling’s synthesis than Slater’s reduction.
All this motivates the following argument, which I have heard in different forms from chemists, physicists and philosophers. The chemical bond is a theoretical figment. It was useful in the 1860s and remains useful now, for predictive and heuristic purposes. But bonds are not real. Quantum mechanics, which provides the best description of the world at the atomic level, has shown the structural theories of the 1860s to be at best naive portrayals of molecular reality.
I do not find this argument convincing, and more generally I am at a loss to understand why philosophers and scientists alike are so ready to approach the results of the special sciences in a spirit of ontological nonseriousness, yet the craziest ideas from physics are taken much more seriously.10 The argument for instrumentalism about molecular structure can be resisted in a number of different ways. Firstly, structural theory has been around for a very long time—some sixty years longer than non-relativistic quantum mechanics—and its development has been cumulative: the theory itself, and the structures assigned to substances within it, have been retained or extended, our understanding of them deepened by the interaction with physics. I cite two scientific authorities in support of this claim. In a systematic presentation of his views on structure and bonding, Lewis said that
No generalization of science, even if we include those capable of exact mathematical statement, has ever achieved a greater success in assembling in simple form a multitude of heterogeneous observations than this group of ideas which we call structural theory.
(Lewis 1923: 20-21)
In his presidential address to the Annual General Meeting of the Chemical Society (later to become the Royal Society of Chemistry) in April 1936, Nevil Sidgwick rejected the idea that new scientific theories must always overthrow the conceptions of their predecessors (Sidgwick 1936). A detailed examination of the development of chemistry, he argued, revealed that although “the progress of knowledge does indeed correct certain details in our ideas”, the structural theory of Kekule, laid down in the 1860s, had “undergone no serious modifications” (Sidgwick 1936: 533). As we have seen, the chemists of the 1860s had assigned molecular structures to substances so as to account for isomerism, and Sidgwick confidently asserted that “[a]mong the hundreds of thousands of known substances, there are never more isomeric forms than the theory permits” (Sidgwick 1936: 533). Subsequent developments had clearly enriched the theory, in two ways. On the one hand, Kekule’s theory “assumes that the molecule is held together by links between one atom and the next,” but in that theory “[n]o assumption whatsoever is made as to the mechanism of the linkage” (Sidgwick 1936: 533). A proposal as to how molecular structure is realised came only later, in Lewis’s theory of the electron-pair bond. On the other hand, later developments enriched structures with detail:
To Kekule the links had no properties beyond that of linking; but we now know their lengths, their heats of formation, their resistance to deformation, and the electrostatic disturbance which they involve.
(Sidgwick 1936: 533-534)
158 Robin F. Hendry He concluded:
I hope I have said enough to show that the modern development of the structural theory, far from destroying the older doctrine, has given it a longer and a fuller life.
(Sidgwick 1936: 538)
A second argument against an instrumentalist interpretation of structural theory is that its development has been extremely fruitful from an empirical point of view: structural theory has underwritten the design and synthesis of many thousands of new substances; the theories of reaction mechanisms developed from the 1920s onwards depended on Lewis’s insight that the chemical bond is realised by pairs of electrons. If longevity, theoretical continuity and fruitfulness are hallmarks of the real, then structure has a claim on our commitment, and perhaps a stronger one than quantum mechanics, on the basis of which it is called into question. Thirdly, the conflicts with quantum mechanics are often overstated. To be sure, Lewis’s static electron pairs seem naive, but within theoretical chemistry the attempt to recover different aspects of ‘classical’ structure and Lewis’s account of bonding remain important (see, for instance, Bader 1990), for the explanatory successes of these theories must be accommodated somehow within the theory which replaces them. I think these considerations should give us pause before we sweep away these intellectual achievements of chemistry with the wave of an instrumentalist hand.