Ontological reduction and molecular structure

Robin Findlay Hendry

Department of Philosophy, University of Durham, 50 Old Elvet, Durham DH1 3HN, UK
a r t i c l e i n f o
Article history:
Received 17 January 2009
Received in revised form
16 March 2010
Keywords:
Chemistry
Quantum mechanics
Molecular structure
Reduction
Emergence
a b s t r a c t
In this paper I outline how the debate concerning the intertheoretic reduction of chemistry reaches a
stalemate. One way forward is to switch discussion to the issue of ontological reduction and emergence,
so I present a counternomic criterion of emergence that should be acceptable to both sides of the
discussion. I then examine the bearing on this debate of the symmetry problem in molecular quantum
mechanics, as presented by Woolley and Sutcliffe (1977). I conclude by addressing some objections to
emergentist positions: that they posit miraculous violations of physical laws; that emergence is obscure
and of doubtful coherence; that causal theories of property identity render emergence, under the
counternomic criterion, metaphysically impossible.
& 2010 Elsevier Ltd. All rights reserved.
When citing this paper, please use the full journal title Studies in History and Philosophy of Modern Physics
1. An impasse
Since the early twentieth century, there has been a close
explanatory interaction between physical theories about structure
and dynamics at the atomic scale and chemical theories concerning
the structure of molecules. Clearly this interaction is the source of
the widespread view that chemistry is reducible to physics:
molecular structure is so central to chemical explanation that to
explain molecular structure is pretty much to explain the
whole of chemistry. The central question in what follows is
whether the relevant explanations, examined in detail, really do
support reductionism.
Quantum chemistry is the interdisciplinary eld that applies
quantum mechanics to explain the structure and bonding of
atoms and molecules. For any isolated atom or molecule, there is a
Schr odinger equation, which is determined by enumerating the
electrons and nuclei in the system, and the forces by which they
interact. Classical intertheoretic reduction would require the
derivation of the properties of atoms and molecules from their
Schr odinger equations. The non-relativistic Schr odinger equation
for the hydrogen atom can be solved analytically: its solutions
correspond to the electronic energy states, providing a beautiful
explanation of hydrogens atomic emission spectra. Some
philosophers make eloquent use of the elegance and precision
of this explanation (see for instance Carl Hoefer, 2003, p. 1404),
but the hydrogen atom is a special case: the simplicity and
symmetry properties of the problem are central to the elegance of
its solution. Caution is required in drawing any consequences for
how quantum mechanics applies to molecular systems more
generally. The Schr odinger equation for the next simplest atom,
helium, is not soluble analytically, although accurate numerical
methods are available. To solve the Schr odinger equations for
more complex atoms, or for moleculeswhich cannot share the
symmetry properties of the hydrogen atomquantum chemists
apply a battery of approximate methods and models. Whether
they address the electronic structure of atoms or the structure and
bonding of molecules, these models are calibrated by an array of
theoretical assumptions drawn from chemistry itself. In short,
quantum chemistry turns out not to meet the strict demands of
classical reductionism, because its models bear only a loose
relationship to exact atomic and molecular Schr odinger equa-
tions, and its explanations seem to rely on just the sort of
chemical information that, in a classical reduction, ought to be
derived (see for instance Bogaard, 1978; Hendry, 1998; Hofmann,
1990; Primas, 1983; Scerri, 2007, Chapters 79; Woolley, 1976,
1991, 1998; Woolley & Sutcliffe, 1977).
Now reductionists can always see a failure of reduction either
as requiring physicists to adjust their theories, or as a theoretical
argument for the revision or elimination of special-science
theories or concepts that, they argue, correspond to nothing real
at the fundamental level. Non-reductionists will see either of
these diagnoses as adopting too easily the hubris of physics:
either in the continuing myth that physical theories have some
ARTICLE IN PRESS
Contents lists available at ScienceDirect
journal homepage: www.elsevier.com/locate/shpsb
Studies in History and Philosophy
of Modern Physics
1355-2198/$ - see front matter & 2010 Elsevier Ltd. All rights reserved.
doi:10.1016/j.shpsb.2010.03.005
E-mail address: r.f.hendry@durham.ac.uk
Studies in History and Philosophy of Modern Physics 41 (2010) 183191
ARTICLE IN PRESS
special duty of full coverage, when the historical evidence is that
physics often develops its theories in blessed innocence of how
they square with the results of other sciences, or in the airy
dismissal of special-science discoveries like molecular structure
(see Hendry, 2006, pp. 185187). All this leads to an impasse-
temperamental reductionists and non-reductionists can agree
that classical intertheoretic reductions of chemistry are not
currently available, but will differ in how they interpret the
situation. As long as reduction is seen as a dated intertheoretic
achievement, however, the issue is essentially future directed-
both sides must wait and see, even if they would bet different
ways. But why do the two sides make different bets?
Perhaps the answer concerns their different underlying
metaphysical views. To identify those views requires that we
separate the intertheoretic and the metaphysical aspects of the
reduction debate: the former concern the explanatory relation-
ships between theories and the latter the relationships between
their subject matter. This separation is necessary because there
are reasons why the intertheoretic reduction of a special science
may fail that are quite independent of any metaphysical relation-
ship between physics and the special sciences. The reasons are
twofold, and can be illustrated by brief reections on (i) how
physical theories are applied to complex physical systems and (ii)
the nature of scientic disciplines.
First consider the kind of physical theory, like quantum
mechanics, whose explanatory successes ground reductionist
arguments. The principles of quantum mechanics are expressed
mathematically, at very abstract levels, and applying them to
particular kinds of system involves what Nancy Cartwright calls
theory entry (1983, Chapter 7). Molecules are complex, and the
quantum-mechanical equations describing them insoluble. This is
a familiar situation, faced by scientists ever since Newton
grappled with the moons orbit, and can be addressed by
introducing approximate and semi-empirical models. Reduction-
ists see the situation as follows: the exact equations are insoluble,
but the semi-empirical models are approximations to rigorous
treatments, standing in for them in explanations of special-
science phenomena. Hence the explanatory link between physics
and the special-science phenomena can be forged indirectlycall
this the proxy defence of reductionism (see Hendry, 1998,
Section 2). This defence, however, relies on the assumption that
any explanatorily relevant features of the semi-empirical proxies
are shared by exact solutions to the rigorous equations. The truth
of this assumption needs to be demonstrated case by case.
Secondly consider the nature of disciplines. Given some
acquaintance with the history of science, even the most robust
scientic realist will accept that scientic theories are partly
human creations, the contingent products of disciplines whose
existence and development are themselves contingent. The
development of a discipline is the work of a community of
scientists who may be relatively isolated from the work of
neighbouring disciplines. Each discipline may have its particular
theoretical concepts, styles of explanation, and judgements of
theoretical plausibility, so there can be no guarantee that physics
and chemistry will mesh even if, ultimately, they are in some
sense talking about the same things. Thus for instance the concept
of a chemical bond was developed by organic chemists in the
1860s as part of an informal and qualitative body of theory
concerning the structure of molecules. Yet the development of
quantum chemistry in the twentieth century provided no clear
identication of what chemical bonds are, from the point of view
of fundamental physical theory (see Hendry, 2008, 2010). Once
again, reductionists and non-reductionists will interpret this
situation differently. Reductionists will say that, even if physical
laws undergird chemical explanation to the extent that for every
chemical explanation there is a complex physical reason why it is
correct, the physical counterparts to chemical explanations may
be too complex to replace them. Where a special-science law
cannot be eliminated from chemical explanation, this reects
chemistrys human componentits historically contingent con-
cepts or styles of explanationrather than how things ultimately
are. This amounts to a concession on the issue of what can be
achieved in respect of intertheoretic explanation, but with an
intimation that it does not matter for reductionism more
generally. Some non-reductionists accept the implied anti-
realism, though they argue that physics should not be privileged,
and is just one perspective among many (see for instance van
Brakel, 2000, Chapter 8). Either way, there is an anti-realist
reading of chemical theories and concepts, because their non-
reducibility arises only from limitations of the explanatory
concepts and practices of human scientists.
Realists about chemical theories will not want to leave things
there. On the one hand, the realist reductionist should be
pressedwhy does it not matter that chemistry resists inter-
theoretic reduction? Answering that question requires a distinc-
tion between two different kinds of failed intertheoretic
reduction: those that do not matter because they involve only
our limited mathematical powers, or failures of t between the
explanatory and conceptual schemes associated with different
disciplines; and those that arise from a disunity in nature that is
deeper than the failure of intertheoretic reduction. A positive
argument is needed that the chemical cases fall into the former
category. On the other hand, the realist non-reductionist should
regard the reductionists concession of the failure of intertheoretic
reduction on the grounds of intractable mathematics or con-
ceptual mismatch as a pyrrhic victory, and articulate a more
robustly ontological non-reductionism.
2. Ontological reduction and emergence
If the reduction debate is to develop beyond the impasse over
intertheoretic reduction, it must turn to the ontological relation-
ships between the entities, processes, and laws studied by
different sciences, which are fallibly and provisionally described
by their theories. One obvious requirement on a criterion of
ontological reduction is that whether or not it obtains must be a
substantive metaphysical issue that transcends the question of
what explanatory relationships exist between theories now, or
might exist in the future, even though intertheoretic relationships
must continue to be relevant evidence. It should also be
acceptable to both reductionists and non-reductionists.
Ontological reducibility is plausibly regarded as one end of a
spectrum of relations that includes weak existential dependence
close to the other end. Weak existential dependence (x could not
exist without y) is close to one end of the spectrum because a
version of physicalism according to which everything (concrete) is
weakly existentially dependent on the physical can be accepted
even by those who think that the ontology of science is far from
unied (see for instance Crane & Mellor, 1995, p. 85; Dupre , 1993,
pp. 9192). Reducibility is at the strong end of the spectrum
because it is the limiting case that denies the distinct existence of
what is dependentthe reductionist slogan is that x is reducible
to y just in case x is nothing but its reduction base, y. One can
imagine many ways to cash out this slogan, depending on the
aspect under which the reduced is held to be nothing but its
reduction base, but a consensus has emerged in recent philosophy
of mind that the relevant aspect should be causal. Alexanders
dictum is the principle, often cited by Kim (1998, p. 119, 2005,
p. 159), according to which being real requires having causal
powers. Construed as a biconditional principle, Alexanders
dictum makes bearing causal powers both necessary and
R.F. Hendry / Studies in History and Philosophy of Modern Physics 41 (2010) 183191 184
ARTICLE IN PRESS
sufcient for a propertys being counted as real, providing a
criterion of ontological reduction that cuts both ways. The
ontological reductionist thinks that special-science properties
are no more than their physical bases because the causal powers
they confer are a subset of those conferred by their physical bases;
the emergentist sees them as distinct and non-reducible just
because the causal powers they confer are not exhausted by those
conferred by their physical bases. The additional causal powers
are exerted in downward causation.
Consider the familiar dialectical situation of supervenience in
Kims work, which was introduced into the mindbody problem
by way of formulating non-reductive physicalism (see Kim, 1998,
Chapter 1). As Kim succinctly puts it Supervenience is Not a
MindBody Theory (1998, p. 9), because supervenience is just
covariance between groups of properties (though it comes in
different modal strengths). If a special science property or group
of properties supervenes on the physical, there is a (possibly
messy) correlation between the special-science properties and the
physical properties. Just like statistical correlation, supervenience
requires explanation by deeper relationships, and can be
explained in one of two ways. Emergentism invokes downward
causationthe special-science properties sometimes push their
physical supervenience bases around. Ontological reductionism
assumes the causal closure, or completeness, of the physical-
physical effects are brought about solely by physical causes via
physical laws (see Papineau, 2002, pp. 233234). The physical
base properties are immune to intervention from above. Hence
supervenience is inadequate to articulate a mindbody theory
because it is compatible with, and explainable by, either
emergentism or reductive physicalism. Emergentism allows that
supervenience can be explained either by upward determination
(plus physical causation) or downward causation, while reductive
physicalism appeals only to upward determination (plus physical
causation).
C. D. Broads book The mind and its place in nature (Broad, 1925)
provides an account of emergence from which a model of
downward causation is readily extracted (see Hendry, 2006,
pp. 177180; McLaughlin, 1992, pp. 7589). First consider the
position that Broad calls pure mechanism, according to which
every material object is made from the particles of one
fundamental kind of stuff (1925, p. 44). One physical law governs
the interactions between the particles, and according to pure
mechanism, this law determines the behaviour of every material
object. According to Broad, the existence of irreducibly macro-
scopic qualities like colours and temperatures (1925, pp. 5051)
shows that pure mechanism must fail. Such qualities are
associated lawfully with microscopic states (through trans-
physical laws) and cannot be accounted for by pure mechanism.
In addition to the necessary emergence of properties involved in
trans-physical laws, Broad also countenanced emergent intra-
physical laws relating only physical properties. Breathing, for
instance, is a type of movement, and if it is not determined to
occur by interactions between the particles from which breathing
systems are formed as governed by basic laws, then it is emergent
(1925, p. 81). Broad saw obvious epistemic differences between
trans-physical and emergent intra-physical laws, because a failure
to account mechanistically for an intra-physical law may arise
either from incomplete knowledge or from its being genuinely
emergent (1925, pp. 8081). That an intra-physical law is
emergent must always remain a hypothesis.
Broads account of the disagreement between pure mechanism
and emergentism is easily formulated within quantum
mechanics, in which motions are governed by Hamiltonian
operators determined by the forces acting within a system. Pure
mechanism expects only one force term to appear in the
Hamiltonians that govern the behaviour of complex systems. By
analogy with the well-known terminology for forces, call the
appropriate one-force-term Hamiltonian for a system its resul-
tant Hamiltonian. The emergentist posits that non-resultant, or
congurational Hamiltonians govern the behaviour of at least
some complex systems. Broads pure mechanism is an extreme, of
course, because it countenances only one basic kind of interaction,
and physics countenances more. But the point is usefully made
that any account of the disagreement between emergentism and
pure mechanism (and, more generally, reductive physicalism)
involves essential reference to some set of fundamental physical
properties and laws.
Where does downward causation t into this? For the
emergentist, every complex system is composed of the same
basic stuff, but some complex systems are covered by
non-resultant or congurational Hamiltonians. In an emergent
complex system, the behaviour of the basic stuff of which it is
made is governed by a congurational Hamiltonian, which is
different from what it would be were its behaviour governed by
the resultant Hamiltonian. Since the Hamiltonian of a system
determines the precise nature of the physical law that governs its
behaviour, to say that some system exhibits downward causation
is to make a counternomic claim about itthat its behaviour
would be different were it determined only by the more basic
laws governing the stuff of which it is made. The emergentist and
the reductionist can agree that a unied framework of physical
law (quantum mechanics) governs how forces act, but disagree
on the extent to which the range of forces invoked in quantum-
mechanical explanations is unied from a dynamical point of
view, that is, on how many independent kinds of Hamiltonian
operate in the world. Note that the truth of emergentism would
be no barrier to genuine quantum-mechanical explanations of
chemical structure and bonding. The difference between the
emergentist and the reductive physicalist concerns only the form
of those explanations: the emergentist expects that they will
involve congurational Hamiltonians, the reductive physicalist
that they will involve only resultant Hamiltonians (see Hendry,
2006, pp. 176185).
3. The symmetry problem
The reductionist is committed to seeing the apparent diversity
of molecular structures as arising in a unied way from just a few
basic interactions. As we have seen, the failure of quantum-
mechanical explanations in chemistry to meet the strict demands
of classical reduction tends not to impress reductive physicalists.
Recognising that quantum-mechanical calculations put molecular
geometries in by hand, they have three main lines of defence:
(i) to claim that the semi-empirical models are merely proxies
for more rigorous treatments; (ii) to wait for new theoretical
developments; or (iii) to accept the failure of intertheoretic
reduction, cite mathematical intractability as the reason, and
retreat to ontological reducibility. In what follows I will argue that
there are difculties with any of these strategies with regard to
molecular Schr odinger equations. In contrast I will make no claim
about the non-reducibility of multi-electron atoms, because the
application of models that are motivated by empirical information
from the higher science is not in itself enough to ground an
argument against ontological (as opposed to intertheoretic)
reduction. There must also be principled reasons why inter-
theoretic reduction fails. There are such reasons in molecular
cases, but none that I know of in the atomic cases.
Although molecular structures cannot be derived directly from
exact molecular Schr odinger equations, quantum-mechanical
explanations employ models in which molecules are assumed to
have them. Molecular structures are justied as the starting point
R.F. Hendry / Studies in History and Philosophy of Modern Physics 41 (2010) 183191 185
ARTICLE IN PRESS
of approximate solutions to the exact Schr odinger equation in a
way that does sound as if it will allow a proxy defence, via the
BornOppenheimer approximation. The justication is as follows.
The nuclei within a molecule are thousands of times more
massive than the electrons, and so they can be regarded as
approximately at rest when the electronic motions are consid-
ered. The trick is to solve a Schr odinger equation just for the
electrons, in which a xed nuclear geometry appears as a
parameter. In principle, the electronic Schr odinger equation could
be solved for many different arrangements of the nuclei to see
how the electronic energy depends on nuclear geometry. In
practice, it may be enough to consider only an equilibrium
geometry that is known empirically, and small oscillations around
it. The justication for this approximation is that using the Born
Oppenheimer solution instead of the exact solution makes only a
small difference to the energy.
The justication is limited, however. The BornOppenheimer
approximation makes only a small difference to the calculated
energy of the molecule, but it makes a big difference to its
symmetry properties (Woolley & Sutcliffe, 1977). Although
Schr odinger equations for complex polyatomic molecules cannot
be solved analytically, much can be known about their solutions
by considering the nature of the forces that appear in them. Of the
four fundamental forces, three (gravitational, weak, and strong
nuclear) can be neglected in calculating the quantum-mechanical
states governing molecular structure. Hence physics itself tells us
that the Coulomb (electrostatic) force is the overwhelming
determinant of molecular structure, which should arise from the
quantum mechanics of systems of charged particles moving under
electrostatic forces.
Now arbitrary solutions to exact Coulombic Schr odinger
equations should be spherically symmetrical, but polyatomic
molecules cannot be spherically symmetrical, for their lower
symmetries are important in explaining their behaviour. Consider
for example the hydrogen chloride molecule, which has an
asymmetrical charge distribution that explains its acidic beha-
viour and its boiling point. In contrast, the expectation value for
the dipole moment of a molecule in an arbitrary eigenstate of the
full molecular Hamiltonian must be zero. More generally it
follows that according to quantum mechanics, no directional
properties will be possessed by an isolated molecule in a general
energy eigenstate (Hendry, 1998, p. 131). In the BornOppenhei-
mer approximation, the spherical symmetry that is expected of
exact solutions to the full Schr odinger equation is simply replaced
by a less symmetrical structure that is compatible with the
asymmetrical charge distribution. Molecular structures cannot be
recovered from the Coulomb Schr odinger equations, but not
because of any mathematical intractability. The problem is that
they are not there to begin with.
The Coulomb Schr odinger equations describe mere assemblages
of electrons and nuclei rather than molecules, which are structured
entities (Woolley, 1991, p. 26). This is illustrated by the fact that
isomers, which are distinct molecules sharing the same molecular
formula, share the same Coulomb Schr odinger equation (Woolley,
1998, p. 11). Recall that the proxy defence of the use of approximate
models in explanations relies on the models and the exact solutions
sharing explanatorily relevant features. So it must fail in this case
because the BornOppenheimer models fail to share explanatorily
relevant features of the exact treatments, namely their symmetry
properties. Without a quantum-mechanical justication for the
attributions of structure (and the lower symmetry) within the
BornOppenheimer models, they simply assume the facts about
molecular structure that ought to be explained. Of course the
Coulomb potential is not the only determinant of molecular
structure, for other forces are involved, however weakly. But
molecular quantum mechanics has, for good reasons, based its
explanations on Coulombic Hamiltonians, and it is entirely unclear
how introducing (for instance) the weak nuclear force into
consideration would account for the complex and varied symmetry
properties of molecules.
The spherically symmetrical states could perhaps be regarded
as superpositions of asymmetrical states with opposite orienta-
tions, just as the spin states of a silver atom may be regarded as
superpositions of spin-up and spin-down, or the quantum state of
Schr odingers cat can be regarded as a superposition of cat-alive
and cat-dead states. If anything, this makes the symmetry
problem look more intractable, and a proxy defence based on the
BornOppenheimer approximation more obviously inadequate. In
the quantum-mechanical measurement problem, the difculty is
in explaining why, when we have suitably entangled the cats
quantum state with a radioactive source, which itself is in a
superposed state with respect to a radioactive decay event, we
nd on measurement of the cats state one of the determinate
states (cat-alive or cat-dead), rather than their superposition.
Imagine the following solution to the measurement problem:
although, according to quantum mechanics, the cat ought to be
found in a superposition once its quantum state interacts with the
radioactive source, this can be ignored. It makes little difference to
the overall energy of the system which of the three states the cat
exhibits (cat-alive, cat-dead, superposition). Whether the cat is
found to be dead or alive, one can explain the measurement
result, and thereby solve the measurement problem, by invoking
the superposition approximation in which the actual measure-
ment result is replaced by the superposition, on the grounds that
the substitution makes little difference to the overall energy.
Therefore it is no problem that, according to quantum mechanics,
the cat is in a superposition. I submit that it is just as much a
mistake to invoke the BornOppenheimer approximation as part
of the proxy defence of model Hamiltonians that presume a
determinate molecular structure.
What of the second response? It seems unpromising to wait for
new developments. The wait is for new kinds of molecular model
meeting two conditions: they should have the right symmetry
properties to explain the structure and bonding of molecules, and
also be defensible as approximations to exact quantum mechanics
in a way that allows a new version of the proxy defence. However,
Woolley and Sutcliffes symmetry problem arises from founda-
tional features of how exact quantum mechanics is generated
using the Coulomb force, and shows that the two demands cannot
simultaneously be met. Hence the wait must either be for an
entirely new framework for exact quantum mechanics (or the
replacement of quantum mechanics itself, but I take it that this
involves a retreat to ontological reducibility, see below). The
symmetry problem arises in the rst instance by considering the
Schr odinger equation for an isolated molecule, and the only
obvious solution is to appeal to the molecules interaction with its
environment, which would be represented by a symmetry-
breaking non-Coulomb term in the molecules Schr odinger
equation. The particular form of the symmetry-breaking addition
must be justied however, and it is quite mysterious how that
could work if all one has in the environment are more molecules
described by Coulombic Hamiltonians. The Coulomb Schr odinger
equation for an n-molecule ensemble of hydrogen chloride
molecules has precisely the same symmetry properties as a
Coulomb Schr odinger equation for a 1-molecule system. If the
particular form of the symmetry-breaking addition is not justied,
then it is just ad hoc: a deus ex machina.
The nal option is to wait for a replacement for quantum
mechanics, and (for the moment) retreat to the ontological
reducibility of chemistry, hoping that some future physical theory
will achieve an intertheoretic reduction to reect this metaphy-
sical fact. It might be thought that this is unproblematicthe
R.F. Hendry / Studies in History and Philosophy of Modern Physics 41 (2010) 183191 186
ARTICLE IN PRESS
above issues bear only on intertheoretic reduction and not
ontological reduction, because they concern the explanatory
relationship between a physical theory, quantum mechanics,
and chemical theories of molecular structure. But that would be
too quick. We saw in the last section that if molecules are
ontologically reducible to their physical bases, then they ought to
have no causal powers beyond those that are conferred by those
physical bases. That much follows if ontological reduction is
committed to the causal completeness of physics, as physicalists
generally require it to be (Horgan, 1993; Kim, 1998; McLaughlin,
1992; Papineau, 2002).
The symmetry problem impacts on ontological reduction via
its commitment to the completeness of physics in two ways. The
rst is directif the acidic behaviour of the hydrogen chloride
molecule is conferred by its asymmetry, and the asymmetry is not
conferred by the molecules physical basis according to physical
laws, then surely there is a prima facie argument that ontological
reduction fails. Of course future physics and chemistry may
support ontological reduction, just as it may solve the related
quantum-mechanical measurement problem, but proponents of
ontological reduction are not entitled to presume that it will. On
any conservative amendment to quantum mechanics, the ex-
planation of why molecules exhibit the lower symmetries they do
would appear to be holistic, explaining the molecules broken
symmetry on the basis of its being a subsystem of a supersystem
(molecule plus environment). This supersystem has the power to
break the symmetry of the states of its subsystems without
acquiring that power from its subsystems in any obvious way.
This looks like downwards causation. As for a non-conservative
amendment to quantum mechanics, the bets are off, but there is
no particular reason to think the successor to quantum mechanics
will exclude downward causation. In fact the inductive evidence
is that it will not, because its immediate predecessor, quantum
mechanics, does not. This is not just a contentious appeal to the
results of the foregoing symmetry argumentit may well be that
quantum-mechanical entanglement should be interpreted more
generally as indicating a failure of ontological reducibility
(Humphreys, 1997, Section 6).
The second way that the symmetry problem impacts on the
completeness of physics is indirect. If the ontological reducibility
of chemistry is not a default position that can be established by
appeal to intuition, then its most important element, the causal
completeness of physics, must be a substantive thesis requiring
empirical support. The symmetry problem removes much of the
empirical support that is claimed for the principle. Here is why.
The completeness of physics involves the claim that the general
framework of mechanics is able to unify the motion of any
physical system by seeing it as arising from just a few forces that
apply very generally. Is there any evidence for this principle? In an
interesting account of the history of the completeness of physics,
David Papineau (2002, pp. 232256) sets out to explain what he
sees as the consensus in 20th-century science that physics is
complete. Papineau explains that consensus as arising from the
acceptance of two interlocking arguments, the argument from
fundamental forces (2002, p. 250), and the argument from
physiology (2002, p. 254). The argument from physiology may be
relevant to Papineaus intended application of the completeness
of physics to the mindbody problem, but only the argument
from fundamental forces is relevant to the reduction of chemistry.
In any case I am highly sceptical about how closely physics
(as opposed to chemistry) has unied, or even been involved in
physiological explanation at the cellular level. Hence I will
concentrate on the argument from fundamental forces. The
conclusion of this argument is that all apparently special forces
characteristically reduce to a small stock of basic physical forces
which conserve energy (2002, p. 250). This conclusion, Papineau
argues, was available to Hermann von Helmholtz because in
order for the principle of the conservation of energy to hold,
non-conservative, dissipative forces like friction must be
reducible to conservative forces, which is ensured if dissipative
forces all arise from a few basic forces. The reducibility of all
special forces is, he argues, a natural generalization of this
reduction.
Whether or not this is the right psychological explanation for
the views of Helmholtz or any other 19th-century scientist, to
have held that all special-science forces are reducible at that time
would have been grossly premature, since there were precisely no
detailed physical explanations of the special forces invoked even
by physics neighbouring science, chemistry. (The more general
reduction is in any case not required for conservation of energy
but merely suggested by it, as Papineau himself points out.) In
the 19th century, there were only speculative (and unsuccessful)
attempts to explain the microstructure of chemical substances in
terms of physical theory, partly because the microphysical
structure of chemical substances at that time was simply unclear.
Only in the 20th century was there any detailed and successful
application of physics to the explanation of chemical structure
and bonding that could ground an argument for the reduction
of chemistry. But I am dubious of the suggestion that even in
20th-century physics and chemistry there was any consensus
in favour of the completeness of physics. To invoke one important
constituency, among the chemists who founded quantum chem-
istry, there were many temperamental non-reductionists, who
saw the quantum-mechanical explanation of chemical structure
and bonding as a process that drew equally on physical principles
and chemical knowledge, adapting quantum mechanics signi-
cantly in the process (see Hendry, 2004).
The application of quantum mechanics to chemistry is appealed
to by Brian McLaughlin in his account (1992) of why the currency of
emergentism declined in the 20th century. McLaughlin sympathe-
tically sets out the doctrines of emergentism, including the existence
of congurational forces and downwards causation, and argues that
the failure of any congurational forces to turn up in quantum-
mechanical explanations of chemical structure and bonding
undercuts emergentism. Now Papineau points out (2002, p. 254)
that the reduction of chemistry hardly enforces the reduction of life
or mind, but if McLaughlin is right then his argument is at least
relevant to chemical emergence. However, we have already found
that actual quantum-mechanical explanations of chemical structure
and bonding seem to presuppose unexplained symmetry-breaking,
and so do, in effect, employ congurational forces. This undercuts
any empirical support they could offer to the completeness of
physics with regard to chemical systems, and with it, ontological
reduction (see Hendry, 2006 for a more detailed discussion
of this point).
Both emergentism and reductive ontological physicalism are
consistent with the successful application of quantum mechanics
to explain chemical structure and bonding. The situation is not,
however, symmetrical between the two. Emergentism is at least
as well supported as reductive physicalism by the data of the
explanatory interface between physics and chemistry, for
two reasons. If emergentism were true, and congurational
Hamiltonians really did govern the behaviour of molecules, then
the disunied structure of quantum-mechanical models explain-
ing molecular structure and bonding, including the unexplained
symmetry-breaking through the imposition of determinate
molecular structure by hand, is just what one would expect.
Although it can be made consistent with this situation, reductive
physicalism has to appeal to independent factors to explain the
disunied structure of quantum chemistry, and must posit a
mechanism for symmetry-breaking, for which there is no
independent evidence. In short, the overall philosophical package
R.F. Hendry / Studies in History and Philosophy of Modern Physics 41 (2010) 183191 187
ARTICLE IN PRESS
of emergentism gives the more unied explanation of the
interaction between physical and chemical theories, and also that
between physical and chemical properties. On any evidential
principle under which the hypothesis that gives the more unied
explanation of the phenomena is better supported, emergentism
wins out. Secondly, reductive physicalism embodies a logically
stronger version of the priority of physics, and there are no good
arguments for the excess content of the stronger version. Since
the probability of a logically weaker principle is an upper bound
on that of a logically stronger principle, whatever evidence there
is for completeness would support ubiquity at least as well.
4. Understanding emergence
In this section I will try to respond to three kinds of objection
to emergentism. The rst is that emergentism involves miracu-
lous violations of physical laws for which there is no evidence. The
second group of objections concern the coherence of downward
causationKim (1999, Sections VI and VII) questions the clarity
with which downward causation is articulated, and worries that,
when fully articulated, it admits of two sorts of case: one sort is
rather unremarkable while in the other, downward causation
cannot be given a coherent sense (1999, p. 31). Although Kim
does not address his arguments to chemistry in any detail, he does
think that chemistry is reducible (1998, p. 100, 1999, p. 9 both
these claims citing McLaughlin, 1992). However, it will be helpful
in further developing the emergentist position I have sketched to
see how it can resist these arguments. The third group of worries
concern Alexanders dictum and the connection between the
identity of properties and the causal powers they bestowdoes
Alexanders dictum render counternomic discourse incoherent?
4.1. Emergentism does not involve the violation of physical laws
Emergentism is clearly incompatible with one (very strong)
principle of the universality of physical lawthat physical
principles completely determine the behaviour of the systems to
which they apply. It is, however, compatible with an alternative
principle of universality, the ubiquity of physics. Under the
ubiquity of physics, physical principles constrain the motions of
particular systems though they may not fully determine them.
Some physical principles are naturally understood this way,
even by physiciststhe second law of thermodynamics, and
(ironically) the conservation of energy are obvious examples.
Taken individually, the various force laws that dictate the form of
the potential terms in quantum-mechanical Hamiltonians seem
also be understood this way even within physics, since they
operate together to produce the potential term, which governs the
overall motion of a system. The Coulomb law, for instance, is not
violated in systems in which (say) gravitational forces also act.
Robert Bishop (2006, Section 2) makes a closely related point,
arguing that Physics itself does not imply its own causal closure
(2006, p. 45)the completeness of physics is a metaphysical
principle, an important point to note since completeness is a
necessary premise in any physicalist argument for the causal
exclusion of the non-physical. In short, one may accept that
physical principles apply universally without accepting that they
fully determine the motions of the systems they govern. The
universal applicability of physical principles does of course imply
that, when acting on their own, they fully determine the motions
of any system they govern. But this leaves open what happens
when acting in concert with other forces. The difference between
the two principles is subtle and easily missedin defending
fundamentalism against Cartwrights arguments against it
(Cartwright, 1999), Carl Hoefer (2003) rightly distinguishes
fundamentalism, the claim that there are universal fundamental
laws with which all phenomena are in accord (2003, p. 1403)
from the stronger thesis of the reducibility of biology, chemistry,
or meteorology to physics (2003, p. 1408). As formulated,
Hoefers fundamentalism seems to express ubiquity very
nicelyhe does not explicitly distinguish intertheoretic and
ontological reduction, though he clearly intends to deny only
intertheoretic reducibility. Therein lies the problem. He commits
fundamentalism to the causal completeness of physics
(and therefore ontological reducibility) although he defends only
ubiquity against Cartwrights arguments. He endorses a
quote from Richard Feynman clearly expressing ontological
reductionism, glossing it as ubiquity:
What the fundamentalist believes in is a sort of no-conicts
thesis, between physical laws and higher-level phenomena.
Feynman y expresses it nicely: For example, life itself is
supposedly understandable in principle from the movements
of atoms, and those atoms are made out of neutrons, protons,
and electrons. I must immediately say that when we state that
we understand it in principle, we only mean that we think
that, if we could gure everything out, we would nd that
there is nothing new in physics which needs to be discovered
in order to understand the phenomena of life. (Hoefer, 2003,
note 3, pp. 14081409, quoting Feynman, 1965, p. 151)
In the argument between Cartwright and Hoefers fundamentalist
there are now two versions of fundamentalism in play: one is
committed to completeness, the other only to ubiquity. Hoefers
arguments defend only the latter against Cartwright. There are
two corresponding versions of Cartwrights alternative metaphy-
sical picture of a patchwork of laws, one denying only
completeness, the other denying ubiquity too. The former version,
I think, is close to the nomological formulation of emergentism I
have set out here, although Cartwrights intentions are primarily
sceptical, and so she eschews the metaphysical terms required to
formulate emergentism.
Emergentism is compatible with ubiquity, though not with
completeness, understood so as to apply to (a specied list of)
force laws. Completeness implies reducibility, as we have seen,
while the ubiquity of physics does notthis is not surprising, as
completeness is a logically stronger principle of the priority of
physics than ubiquity. The mere applicability of physical
principles to chemical bonding requires only ubiquity, and does
not rule out downward causation. Emergentism implies no
violation of physical laws.
4.2. Kims worries: the coherence of downward causation
Let us turn now to Kims arguments concerning the coherence
of downward causation. It is a long-standing worry for tempera-
mental reductionists that the apparent plausibility and interest of
emergence depend on its being presented in imprecise terms. In
particular, criteria for emergence may fail to distinguish between
causal processes in complex systems that are uninteresting
because unproblematic for reductionism (whence the apparent
plausibility), and cases that appear to be of genuine interest but
are not, however, coherent on closer scrutiny.
Kim rightly points out that the interest and distinctiveness of
emergentism as a form of materialism derives from downward
causation, understood as higher-level entities and properties
having a causal inuence on the ow of events at the lower levels,
levels from which they emerge (1999, 24). For Kim, the question
is how they can coherently be supposed to do this. Following Kim,
let W be a whole that has some property M, which is realized by
Ws microstructural properties, and yet is emergent with respect
R.F. Hendry / Studies in History and Philosophy of Modern Physics 41 (2010) 183191 188
ARTICLE IN PRESS
to them. First distinguish reexive from non-reexive downward
causationin the former, the whole W is a cause of, or has a
causal inuence on, the events involving its own microconstitu-
ents (1999, p. 26). Then distinguish synchronic and diachronic
downward causation, depending on whether or not Ws emergent
property helps to bring about its own realisers (1999, p. 28). These
two distinctions result in a four-way division of downward
causation into synchronic reexive, diachronic reexive, synchro-
nic non-reexive, and diachronic non-reexive cases. Kim argues
that only one category, synchronic reexive downward causation,
is interesting. Non-reexive downward causation, Kim argues, is
mundane (1999, p. 26), encompassing legion cases like that of a
vase falling and bringing about all sorts of changes to nearby
molecules as it cuts a swathe through the air and smashes on the
ground (1999, pp. 2526). Emergentists, he rightly points out,
would not count this as an interesting case of downward
causation even though neither the causally active property of
the vase, its mass, nor the causal powers it confers, is possessed by
any of its microconstituents. This is because the mass of the whole
vase, and the causal powers it confers, arise in perfectly explicable
ways from properties of its microconstituents, hence both are
resultant (1999, p. 26). Diachronic downward causation is
similarly ubiquitous and uninteresting:
I fall from the ladder and break my arm. I walk to the kitchen
for a drink of water and ten seconds later, all my limbs and
organs have displaced from my study to the kitchen. y It
doesnt seem to me that these cases present us with any
special mysteries rooted in self-reexivity, or that they show
emergent causation to be something special and unique.
(1999, p. 30)
I think that, for the purposes of judging whether they are
compatible with reductive physicalism, these cases of downward
causation have been under-described, but we can get on to that in
a moment. First note that Kim considers only one category of
downward causation to be interesting, and troubling for reduc-
tionismsynchronic reexive downward causation. This is just
the category that must be empty, however, given what Kim calls
the causal-power actuality principle (1999, p. 29) according to
which, in order to exercise some causal power, an object must at
that time possess the property that confers it. The causal-power
actuality principle, Kim argues, renders it incoherent to suppose
that a higher-level property could be part of the process that
brings about its realization base. Diachronic reexive downward
causation is not troubling because it allows for a time delay
between an objects coming to have the emergent property and its
exercising the causal powers conferred by that property. The
upshot of Kims argument is worth quoting in full:
We must conclude then that of the two types of reexive
downward causation, the diachronic variety poses no special
problems but perhaps for that reason [is] rather unremarkable as
a type of causation, but that the synchronic kind is problematic
and it is doubtful that it can be given a coherent sense. This may
be due to its violation of what I called the causal-power actuality
principle, but apart from any recondite metaphysical principle
that might be involved, one cannot escape the uneasy feeling that
there is something circular and incoherent about this variety of
downward causation. (1999, pp. 3031)
Now the causal-power actuality principle looks plausible, and
whether or not it fully captures Kims uneasy feeling
about synchronic reexive downward causation, I think that the
emergentist could accept it as a constraint on their understanding of
downward causation without undermining the metaphysical
interest of downward causation, or its incompatibility with
reductive physicalism. Here is why. Kims four way categorization
of downward causation cuts across what is important in judging
whether a particular kind of case is interesting or mundane. In
particular, that categorization does not consider the structure of the
laws by which a property confers the causal powers it does. In
Section 2, however, we saw that the crucial disagreement between
the emergentist and the reductive physicalist concerns their
differing views of how just such laws are structured. Hence Kim is
premature in dismissing cases of non-reexive downward causation
and diachronic downward causation as mundane or unremarkable.
Consider the vasethis is an uninteresting case of downward
causation just in case the causal powers conferred by a composite
bodys having a particular mass are fully explicable in terms of
properties of its microconstituents and the laws governing their
causal powers. In this case, causal processes involving the mass of
the vase fail the counternomic test of Section 2the behaviour of
the vase would be no different were it determined by the more basic
laws governing the stuff of which it is made. How does this relate to
chemistry? In Section 2 we saw that the emergentist interprets the
symmetry problem as showing that the structures of molecules are
not conferred by more basic laws governing the particles of which
they are made. Since a molecules causal powers depend on its
structure, its behaviour would be different were it determined by
more basic (quantum-mechanical) laws governing the particles of
which the molecule is made. Hence there is a clear difference
between the molecule and the vase, whether the molecules causal
powers are exerted internally or externally (that is, reexively or
non-reexively). Nor is there anything spooky or circular about the
molecules acquiring its structure or the causal powers it con-
fersthe powers are fully determined by the structure, and so the
molecule acquires them exactly when it acquires the structure.
Hence there is no obvious violation of Kims causal-power actuality
principle, it is just that the structure it has is not determined by
deeper laws, namely the quantum mechanics of systems of charged
masses interacting electrostatically. Those deeper physical laws are
not violated by the molecules structureas we saw in Section 2, the
deeper physical laws do determine that the molecular system
(it cannot yet be considered a molecule) has a state that can be
regarded as a superposition of many different molecular structures.
These laws do not, however, seem to determine which particular
molecular structure it has. In short, physical law determines only
that the system has a determinable property of which its particular
molecular shape is a determinate, and the molecules causal powers
are conferred only by the determinate. This, I submit, is an
interesting form of downward causation, and is incompatible with
reductive physicalism.
4.3. Causation and property identity
Turning now to a third and last group of worries, there is a
possible tension in the emergentist position I have sketched,
which arises as follows. To focus the discussion of whether or not
chemical properties are reducible toor nothing buttheir
physical bases, I invoked Alexanders dictum, according to which a
property is to be counted as (distinctly) real just in case having
that property makes a difference to an objects causal powers. I
then explicated emergence in terms of downward causation,
giving a counternomic criterion for downward causation
according to which a system exhibits downward causation if its
behaviour would be different were it determined by the more
basic laws governing the stuff of which it is made. Is this
not inconsistent? Alexanders dictum might be interpreted as
requiring that the identity of a property is determined by the
causal powers it bestows. Hence a propertys causal powers
are metaphysically necessary to it. But as formulated, the
R.F. Hendry / Studies in History and Philosophy of Modern Physics 41 (2010) 183191 189
ARTICLE IN PRESS
counternomic criterion for downward causation requires compar-
isons that assume that a propertys identity is independent of the
causal powers it bestows. There are two obvious responses to this
tension. The rst is to show that in formulating criteria of
reduction and emergence, Alexanders dictum need not be
interpreted as involving a criterion of property identity, but
instead something weaker. The second is to reformulate the
counternomic criterion for property identity.
The rst option is the most obvious and attractive, unless one
is already committed to a causal theory of property identity. There
seem to be two interpretations of Alexanders dictum that allow it
to play the required role in focusing the distinction between
reducibility and emergence without constituting a criterion for
property identity. As invoked by Kim, Alexanders dictum some-
times appears to be an epistemic principle, applied within
ontology, governing the conditions under which a property ought
to be counted as real:
As Samuel Alexander said, something that has nothing to do,
no purpose to servethat is, something with no causal
powermight as well, and undoubtedly would in time, be
abolished. (Kim, 1998, p. 119, quoting Alexander, 1920, p. 8)
Under this interpretation, Alexanders dictum is quite plausible: a
property ought to be counted as real just in case it has a role to
play in explaining the worldly course of events, and it can have
such a role just in case having that property confers causal
powers. Applied to reduction and emergence, the principle tells us
that a property ought to be considered to have a reality distinct
from its putative reduction base (and hence be irreducible) just in
case it confers causal powers that are distinct from those
conferred by its reduction base.
At other times, Kims formulation sounds more robustly
ontological:
[O]ne might argue that epiphenomenalism is a fate no better
than irrealism and in fact indistinguishable from itSamuel
Alexander urged that to deprive something of causal powers is
to deprive it of existence. (Kim, 2005, p. 159)
Under this interpretation, Alexanders dictum is a criterion for
property existence. One way to cash it out is as follows: for every
possible world at which a property exists, there is some causal
power that is conferred by possession of that property. This allows
that the causal powers conferred by a property vary across the
different possible worlds at which it exists. Hence this version of
the principle involves no causal criterion for property identity,
which would require the logically stronger principle obtained
when the order of the two quantiers is swappedthere is some
causal power that is conferred by possession of a property for
every possible world at which that property exists. The weaker
version of the ontological principle is sufcient to focus the
difference between reducibility and emergence. In an argument
for reduction, the principle dictates that if a property confers no
causal powers in addition to those conferred by its reduction base,
then it has no reality distinct from that of its reduction base. In an
argument for emergence, it dictates that if a property confers
causal powers in addition to those conferred by its reduction base,
then it has a reality distinct from that of its reduction base.
In conclusion, there are two ways to interpret Alexanders
dictum that allow it to focus the difference between reducibility
and emergence, and yet fall short of implying that the causal
powers bestowed by a property are metaphysically necessary to
it. This is only a proof of independence, however. What if one is
already convinced of the causal theory of property identity? In
fact, even a causal theory of property identity need not entail the
necessity of laws (see Hendry and Rowbottom, 2009; for another
recent defence of the metaphysical contingency of laws see Lowe,
2006, Chapter 9). But even if one is already committed to the
metaphysical necessity of laws, one might allow that there are
objective similarity relations between properties in possible
worlds in which different laws operate. In that case the counter-
nomic criterion for downward causation could be reformulated so
as to involve a comparison between the causal powers conferred
by a candidate emergent property in the actual world, and those
conferred by a similar property in a world governed by unied
laws of nature as envisaged by the reductive physicalist. In effect
one merely replaces the transworld relationship of identity
between the comparator properties with that of similarity. There
are independent reasons for allowing counternomic comparisons
in discourse about chemical substances, since they seem to be
involved in chemical explanation. For instance, water has an
anomalously high boiling point in comparison with the analogous
hydrides of other elements in its group, hydrogen sulphide (H
2
S),
hydrogen selenide (H
2
Se), and hydrogen telluride (H
2
Te)when
compared with these other substances, water ought to be a gas at
room temperature (see Greenwood & Earnshaw, 1997, pp. 5354).
Chemists explain waters high boiling point by invoking hydrogen
bonds between its molecules, yet its boiling point arises (like
that of any substance) as the overall effect of many different
interactions between its molecules. The anomalousness of waters
boiling point, the demand for explanation, and the explanatory
salience of the hydrogen bonding all arise from a comparison
between waters actual boiling point and what its (or its
counterparts) boiling point would be without hydrogen bonding,
that is, were its molecules to interact only via the weaker
attractive forces, which dominate in the case of H
2
S, H
2
Se,
and H
2
Te.
5. Conclusion
In this paper I have developed a counternomic criterion for
ontological emergence as a way out of the impasse that often
develops between reductionists and non-reductionists. I then
defended the resulting emergentist position as a viable account of
the relationship between molecular structure and the physical
states on which it depends. The disagreement between the
ontological reductionist and the emergentist comes down to a
disagreement over the causal completeness of microphysics. Since
the evidence for the causal completeness of microphysics is weak,
and better supports a logically weaker principle of the ubiquity of
physical laws that is compatible with emergentism, I have argued
that the ontological emergence of molecular structure with
respect to quantum-mechanical systems of nuclei and electrons
interacting via Coulomb forces is at least as well supported by the
available scientic evidence as its ontological reducibility. I then
responded to three objections: that emergentism posits miracu-
lous violations of general physical laws; that it is either vacuous
or incoherent; that the counternomic criterion is incoherent if the
identity of a property is determined by its causal role.
References
Alexander, S. (1920). Space, time and deity, Vol. II. London: Macmillan.
Bishop, Robert (2006). The hidden premiss in the causal argument for physicalism.
Analysis, 66, 4452.
Bogaard, P. (1981). The limitations of physics as a chemical reducing agent. PSA 1978,
Proceedings of the biennial meeting of the philosophy of science association, Vol. II.
East Lansing, MI: Philosophy of Science Association. p. 345356.
Broad, C. D. (1925). The mind and its place in nature. London: Kegan Paul, Trench
and Trubner.
Cartwright, N. (1983). How the laws of physics lie. Oxford: Clarendon Press.
Cartwright, N. (1999). The dappled world: A study of the boundaries of science.
Cambridge: Cambridge University Press.
R.F. Hendry / Studies in History and Philosophy of Modern Physics 41 (2010) 183191 190
ARTICLE IN PRESS
Crane, T., & Mellor, D. H. (1995). There is no question of physicalism. In P. K. Moser,
& J. D. Trout (Eds.), Contemporary materialism (pp. 6589). London: Routledge.
Dupre , J. (1993). The disorder of things: Metaphysical foundations of the disunity of
science. Cambridge, MA: Harvard University Press.
Feynman, R. (1965). The character of physical law. Cambridge, MA: MIT Press.
Greenwood, N. N., & Earnshaw, A. (1997). Chemistry of the elements (2nd ed.).
Oxford: Butterworth-Heinemann.
Hendry, R. F. (1998). Models and approximations in quantum chemistry. In
N. Shanks (Ed.), Idealization in contemporary physics (pp. 123142). Amster-
dam/Atlanta: Rodopi.
Hendry, R. F. (2004). The physicists, the chemists and the pragmatics of
explanation. Philosophy of science, 71, 10481059.
Hendry, R. F. (2006). Is there downward causation in chemistry?. In D. Baird L.
McIntyre, & E. Scerri (Eds.), Philosophy of chemistry: Synthesis of a new discipline
(pp. 173189). Berlin: Springer.
Hendry, R. F. (2008). Two conceptions of the chemical bond. Philosophy of Science,
75, 909920.
Hendry, R. F. (2010). The chemical bond: Structure, energy and explanation. In
M. Sua rez, M. Dorato, & M. R edei (Eds.), EPSA Philosophical Issues in the Sciences.
Berlin: Springer. forthcoming.
Hendry, R. F., & Rowbottom, D. (2009). Dispositional essentialism and the necessity
of laws. Analysis, 69, 668677.
Hoefer, C. (2003). For fundamentalism. Philosophy of Science, 70, 14011412.
Hofmann, J. R. (1990). How the models of chemistry vie. PSA 1990, Proceedings of the
biennial meeting of the philosophy of science association, Vol. I. East Lansing, MI:
Philosophy of Science Association pp. 405419.
Horgan, T. (1993). From supervenience to superdupervenience: Meeting the
demands of a material world. Mind, 102, 555586.
Humphreys, P. (1997). How properties emerge. Philosophy of Science, 64,
117.
Kim, J. (1998). Mind in a physical world. Cambridge, MA: MIT Press.
Kim, J. (1999). Making sense of emergence. Philosophical Studies, 95, 336.
Kim, J. (2005). Physicalism, or something near enough. Cambridge, MA: MIT Press.
Lowe, E. J. (2006). The four-category ontology: A metaphysical foundation for natural
science. Oxford: Clarendon Press.
McLaughlin, B. (1992). The rise and fall of British emergentism. In A. Beckermann,
H. Flohr, & J. Kim (Eds.), Emergence or reduction? Essays on the prospects for non-
reductive physicalism (pp. 4993). Berlin: Walter de Gruyter.
Papineau, D. (2002). Thinking about consciousness. Oxford: Clarendon Press.
Primas, H. (1983). Chemistry, quantum mechanics and reductionism (2nd ed.).
Berlin: Springer.
Scerri, E. (2007). The periodic table: Its story and its signicance. Oxford: Oxford
University Press.
van Brakel, J. (2000). Philosophy of chemistry. Leuven: Leuven University Press.
Woolley, R. G. (1976). Quantum theory and molecular structure. Advances in
physics, 25, 2752.
Woolley, R. G. (1991). Quantum chemistry beyond the BornOppenheimer
approximation. Journal of Molecular Structure (Theochem), 230, 1746.
Woolley, R.G (1998). Is there a quantum denition of a molecule?. Journal of
Mathematical Chemistry 23, 312.
Woolley, R. G., & Sutcliffe, B. (1977). Molecular structure and the BornOppenheimer
approximation. Chemical Physics Letters, 45, 393398.
R.F. Hendry / Studies in History and Philosophy of Modern Physics 41 (2010) 183191 191