[This is a guest article posted for purposes of constructive discussion between a "naturalist" philosopher and Catholic philosophers. For myself, there isn't much in its content that I can object to, even as an opponent of metaphysical naturalism. Perhaps that means I've misunderstood David. If so, that question would be a good way to start the combox. —ML]
by David Hirst
Metaphysical naturalism and scientific realism
Methodological naturalism – that is, the assumption, for the ends of scientific investigation, that explanations of physical phenomena are only acceptable when they postulate ‘natural’ causes – is a generally accepted element of the scientific method. There is, of course, no particular requirement that a scientific theory should not have recourse to ‘supernatural’ evidence; even so, the criteria of observation, prediction, and experimentation leading to reproducible results are not congenial to the introduction of such evidence. Indeed, the very success of the scientific method, and the concomitant paucity of evidence for any supernatural phenomena, is seen by some as providing strong evidence for metaphysical naturalism – the thesis that the arrangement of matter and energy in spacetime exhausts ‘what there is’. Nonetheless, metaphysical naturalism is not as easily defended as some of its more vocal contemporary supporters might wish to believe.
Scientific theories do not, in general, make metaphysical claims; a given scientist can conform to methodological naturalism in the exercise of her discipline (in the generation and testing of hypotheses) while remaining entirely agnostic concerning the corresponding metaphysical view. While – once again – the scientific method in no way requires that she should, it is common for the scientific practitioner to proceed as if the physical world were both real and mind-independent; this corresponds largely to the naïve or commonsense (pre-Humean) view of the objective reality of the physical world in which many of the practical or technological aspects of science have their origins. To take the well-known puzzle: If a tree falls in the forest and there is no-one to hear it, the naïve proto-scientific view is that the ‘physical nature’ of its fall is such that, if there were an observer present, the observer would hear it. This, of course, is where what we might call the ‘proto-scientific’ and the ‘proto-philosophical’ accounts diverge. While a proto-philosophical account would depart from the phenomenal experience – from the event as ‘sound’ – the proto-scientific account postulates that, as a kind of physical event, the fall of any tree has the same general characteristics (including the characteristic of displacing a certain quantity of air). The account is completed by suggesting an elementary causal relation: the supplementary proto-law that, given the right conditions, any such event would be audible.
Insofar as they correspond to a first level of ‘theoretical speculation’ about the relations obtaining between observable phenomena and the observation sentences used to describe them, these are ‘proto-refinements’ of our naïve or pre-theoretical intuitions; I shall return later to a more sophisticated account of the process of refinement itself. Suffice it to say for the present that the proto-scientific account allows for the metaphysical theses that the physical world is real and mind-independent by suggesting that certain conditions obtain whether or not they are observed (or even, by extension, observable). While a philosophically-refined account of the proto-scientific view would undoubtedly point out that ‘allowing for’ is not ‘showing that’, the scientifically-refined account simply evacuates further metaphysical speculation (thus, scientifically-refined accounts are sometimes philosophically naïve). The proto-philosophical view is more sophisticated in that it departs from the given of the phenomenon – the experience of ‘a sound’. Such a view has already questioned the naïve realism of the proto-scientific view: whatever the physical characteristic of the falling of some tree, if there is no-one to hear it, there is no experience of ‘sound’. While the proto-scientific view postulates regularities in the world, all that we effectively know about these regularities is that they correspond to certain sequences of phenomenal experience. Thus, while it is the case that the falling of a tree makes a sound in the presence of a (human) observer, we are unable to affirm that it effectively ‘makes a sound’ when there is no human observer.
So, in a certain sense, metaphysical naturalism might seem philosophically naïve, as it departs from the view that the physical universe is effectively real and mind-independent. In philosophical terms, it gives precedence to what we experience over the way in which we experience it; it differs from naïve realism in postulating that the phenomena we experience are determined not by the object of perception itself but by the more general ‘causal structure’ of the physical universe – of which we are also a part (indeed, much of its sophistication is borrowed from scientific refinements of naturalism; this finds its full expression in the thesis that our perceptual apparatus has developed in response to the physical characteristics of our environment). Now, while it would be philosophically naïve to accept it too hastily, the view that the physical universe is effectively real and mind-independent is commonsensical enough, insofar as it conforms to certain of our most basic pre-theoretical intuitions. Of itself, the view doesn’t preclude dualism, and is even compatible with certain supernaturalist accounts. Metaphysical naturalism, however, draws from this ‘reasonable’ or ‘commonsensical’ view the entirely more substantial metaphysical conclusion that only the physical world has objective reality, and that non-physical phenomena can either be reduced to physical phenomena, or simply eliminated from our account.
The how of this conclusion is evident: it satisfies the principle of parsimony. While an idealist could argue that his view is perhaps a more parsimonious response to the mind-independence (and thus, by elimination, the objective reality) of the physical world, the thesis that only the physical universe is real allows for parsimony while also preserving our very basic pre-theoretical intuition that physical things actually exist (thus avoiding the multiplication of unobservable, supplementary, or ad-hoc entities which idealism implies). The why of the conclusion is perhaps more pertinent to our immediate common interests. A preference for parsimony is, for many of us, a question of aesthetics – a preference for Quine’s ‘desert landscapes’; this is perhaps no more than a matter of taste, though it evidently reflects a heuristic common to much of naturalist thinking, and one which reflects the common historical origins of both philosophical empiricism and the scientific method. As naïve realism is evidently irreceivable, physicalism is only meaningful if it can been shown to correspond to some reasonably coherent and consistent body of knowledge about the physical world. Unsurprisingly, the majority of contemporary physicalists therefore defend some form of scientific naturalism (to such a degree, indeed, that in certain undergraduate-level presentations of naturalism, ‘metaphysical naturalism’ and ‘scientific naturalism’ are given as equivalent terms). The scientific naturalist’s appeal to parsimony applies the scientific method to methodological naturalism itself, arguing that if in scientific enquiry we adopt as a methodological postulate the view that the physical universe is real and mind-independent, then in metaphysical enquiry we should adopt the postulate that only the physical universe is real.
Scientific realism – or scientism?
Naïve acceptance of supernaturalism within a philosophical position does not seem intellectually viable for those of us who have no pre-theoretical reasons for accepting supernaturalism (and, at this point, I would underline that there are no arguments for supernaturalism which are convincing to one who is not already predisposed to a supernaturalist account). Yet the naturalist’s appeal to, and faith in, the theories and findings of contemporary science might – to one holding the opposing view – seem equally naïve.
A number of naturalists – and many of those most present in the popular eye – are scientific realists. Yet scientific realism, like supernaturalism, has various forms. Generally speaking, the scientific realist can hold either that scientific theories are largely true (the semantic commitment); that we have reason to believe them, or a significant part of them (the epistemological commitment); or that they describe objective, mind-independent aspects of reality (the metaphysical commitment). How these commitments are organised, and what priority relations obtain between them, depends largely on the individual commentator; for present purposes we are more concerned with metaphysical commitment. A strong metaphysical commitment would, for example, hold that the hypothetical entities of physical science have effective existence; such a view is largely open to the argument from ‘pessimistic induction’, to the problem of the underdetermination of theory by evidence, and to the objections formulated by instrumentalism and by ‘confirmation holism’ (unsurprisingly, the majority of objections depart from epistemic considerations); it is further weakened by the partial nature of scientific theory and by unreconciled incompatibilities between our best-candidate physical theories. Indeed, given the present state of scientific knowledge, strong metaphysical commitment to scientific realism is as great a leap of faith as is supernaturalism; as such, it would be reprehensibly scientistic.
The ‘neo-Quinean’ view is perhaps more acceptably modest: for the neo-Quinean, the adoption of metaphysical naturalism in philosophy reflects the adoption of methodological naturalism in the sciences; nonetheless, her approach is more concerned with the methods of scientific enquiry than with its immediate findings. As Smart1 has remarked, Quine held “that philosophy and science are continuous with one another”; this is not a matter of subordinating philosophy to science, but rather of aligning scientific and philosophical investigation, and particularly in matters of ‘first philosophy’. Now, the methodology of scientific investigation is independent of any metaphysical commitment – science has little to say about what ‘exists’. The scientific investigator can thus proceed ‘as if’ naturalism were true while avoiding any deeper questions concerning the actual ontological status of the entities he postulates (indeed, his approach is compatible with an instrumentalist view). The naturalist philosopher, on the other hand, cannot so easily evacuate such questions, and particularly if her field is metaphysics; however, for the neo-Quinean, scientific naturalism is not an matter of scientific realism, but rather a ‘working hypothesis’ from which we can evaluate and investigate certain metaphysical puzzles.
Science, philosophy, and scientific materialism
Taken to the letter, scientific realism is naïvely, even quaintly, optimistic about the contribution scientific accounts can make to metaphysical speculation. Indeed, while science can have metaphysical implications, we must – as Sklar2 reminds us – be wary of “reading off a metaphysics” from the “overt appearance” of some given theory; likewise, confirmation holism requires that we take into account the metaphysical assumptions – and the pre-theoretical intuitions – which underlay the initial formulation of the theory. Above all, we should bear in mind that a scientifically-refined notion might bear no relation either to our intuitive notions, or to our philosophical refinements of those notions. In the Stanford entry3 on ‘causal determinism’, Carl Hoefer reminds us that “neither philosophers’ nor laymen’s conceptions of events have any correlate in any modern physical theory. The same goes for the notions of cause and sufficient cause” and, in a note, that “some philosophers are misled on this point by the fact that some now-defunct presentations of Special Relativity theory seem to be grounded on an ontology of events. But Special Relativity does not need to be so presented, nor were the “events” used anything like common sense events”.
Hoefer’s remarks are justified, but nonetheless require comment and elucidation. In a scientifically-refined theory (here, the contemporary theory or family of theories that are themselves refinements of Einstein’s theory), it is indeed the case that there are no direct correlates for our intuitive or traditional philosophical understanding of either ‘events’ or ‘causation’. For example, in special relativity, and given the limitations imposed by the speed of light, a light-cone shows what points lie in the ‘past’ of a point p, what points lie in its ‘future’, and what points are in its ‘absolute elsewhere’ (that is, no light from them can reach p). As such, this is nothing more than an informal description of the geometry of spacetime (not allowing, of course, for gravitational effects); it corresponds rather to a mathematical model than to a ‘physical description of spacetime’. Similar observations found what Katherine Hawley4 has called ‘radical pessimism’ concerning the validity of scientific findings in metaphysical enquiry: the (primarily mathematical) models with which the scientist works are at best generalisations or approximations of conditions obtaining in the physical universe, and as such, are incommensurable with the entities and relations postulated in metaphysics. Yet are we really dealing with notions that are entirely incompatible?
Evidently, when working within a given theory, on its implications for associated theories, and in peer-to-peer discussions within his speciality, the scientist tends to use a mathematical formalisation of spacetime. Nonetheless, in informal or semi-formal descriptions of spacetime (such as those used in teaching, in certain reflections on the interpretation of theory, and in presentations of the specialist’s field when addressing scientists in other disciplines), the refined view is frequently related to (a semi-refined version of) the intuitive notion from which it is – however distantly – derived. Thus, the notion of light-cone is frequently introduced by discussing the ‘events’ which can ‘influence’ or ‘be influenced by’ some ‘point-event’5. Likewise, the series of pairwise B-series orderings between point-events (that is, the series of events as ordered by the relations ‘lies in the past light-cone of’ or ‘lies in the future lightcone of’) is said to determine the ‘causal structure’ of spacetime. Evidently, the adoption of intuitively-recognisable entities and relations in what have been called ‘lies-to-children6‘ should not mislead us concerning the probable incommensurability of these intuitively-recognisable elements with the notions adopted in the refined theory. Nevertheless, the characteristic of a ‘lie-to-children’ is that it is ‘wrong for the right reasons’ – frequently, in that it takes a local, but well-known, case of some more general feature and proceeds ‘as if’ that case held generally. The ‘weight/mass’ example illustrates this well. ‘Weight’ is a local semi-constant under common conditions on Earth which is entirely eliminated from the Newtonian account; nevertheless, we continue to employ ‘weight’ as a measurable quantity in our everyday discussions of the mass of commonplace terrestrial things such as gold funerary masks, sacks of potatoes, or American film actresses.
Proto-scientific theories are often generalisations from entirely local features of some given set of phenomena; these local features are in many cases causally related to the regularities more properly described by one, or several, more refined theories (and in some cases by theories from widely differing disciplines). While it is certainly the case that the regularities underlying the distribution of mass and energy in spacetime bear little resemblance to the ’causes’ and ‘effects’, or the ‘objects’ and ‘events’, of either our commonsense or our philosophically refined notions, the naturalistic view would nonetheless suggest that such everyday phenomena depend upon, or are determined by, the more fundamental regularities. While we cannot simply cite a given scientific theory in defence of some pre-established metaphysical position, is it perhaps possible that examination of the informal or semi-formal notions employed in the interpretation and exposition of the theory might nonetheless answer certain requirements of philosophical investigation into such areas as events or causation?
The philosophical and scientific refinement of commonsense notions
In a recent paper7, Roberto Casati and Achille Varzi attempt to account for the “plurality of concepts” subtending the general category of ‘event’ as it is understood in our pre-theoretical intuitions; as it is understood in scientific revisions of our pre-theoretical intuitions; and as it is understood in the various philosophical discussions of mind, action, causation, change, time, and the like. Casati & Varzi refer the specific problem to the more general question of how our intuitive notions are taken up, refined – and, in some cases, superseded by – philosophical and scientific enquiry. Science, they suggest, departs from the phenomena postulated by our commonsense notions and examines them “with respect to exogenous empirical considerations”; it has no particular interest in saving or accommodating our pre-theoretical intuitions. Philosophy also departs from our phenomenal experience of the world; however, philosophical refinement of our intuitive or commonsense notions is dictated by “endogenous a priori considerations”, such as certain internal inconsistencies in some commonsense notion. Unlike scientists, philosophers are generally sensitive to the charge that their refinements run counter to our pre-theoretical intuitions about some phenomenon, and will scrupulously weigh the advantages of any given refinement against its cost in terms of subverting some commonly-held intuition. Many philosophers (and as many – if not more – scientists) would say that this divergence does indeed represent a fundamental difference in orientation between the two disciplines, and that the possibilities of dialogue are minimal. In philosophy, this is the view represented by – for example – Peter Hacker8, for whom “whatever utility such language [that of contemporary physics] might have for specialized purposes, it cannot incorporate our concepts of event or material object, for they are firmly embedded in our conceptual scheme, and intimately bound up with our ordinary notions of space and time” (cf. Hoefer’s remarks above).
Such objections would be justifiable were we to follow a strong scientifically realist programme, and attempt to subsume or to reduce philosophically-refined notions to scientifically-refined notions (as in the naïve hypothesis that mental phenomena can be subsumed or reduced to brain activity). Yet such simplistic positivism is a relic of another age. We have seen that, while a naturalist approach in philosophy can be seen as prolonging and developing on the scientific project, we are not justified in simply grafting the theoretical notions of physical science onto our metaphysics. If we follow Casati & Varzi’s characterisations, the role of metaphysics should be to clarify our intuitive notions about the world and our place in it, and to investigate any inconsistencies we might find either within some intuitive notion, or between associated notions. It follows that any such project would also require that we compare our intuitive notions with our knowledge of the world, and it is perhaps here that we can find an area of fruitful exchange between scientific and philosophical modes of investigation. In an unpublished critique9 of contemporary philosophy of time, Craig Callender suggests an explanatory or mediating role for such investigation: philosophy should seek “to refine our description of what needs to be explained, carefully examine science and the way it treats time, compare the two, and then try to account for any explanatory gap that arises”. Callender’s suggestion strikes me as eminently reasonable. While the scientific refinement of a certain notion might evacuate all semblance of, or correlation with, the pre-theoretical or naïve intuitions with which the notion is associated in our everyday lives, there remains a philosophical requirement to explain any residual lacunae between, on the one hand, our best candidate for knowledge about a certain phenomenon and, on the other hand, our experience of that phenomenon.
While scientific realism is philosophically naïve, physical science can and should guide the metaphysician when she is addressing the (actual) physical world and any ‘possible worlds’ derived from or referring to some physical theory (such as the ‘many-worlds’ interpretation in quantum mechanics). We should indeed be wary of conflating scientific and philosophical accounts of this or that phenomenon when the accounts address the world at entirely different orders of magnitude. But a philosophical account of the physical world should, nonetheless, be compatible with our knowledge of the physical world at an equivalent order of magnitude. Should we say “such-and-such a condition holds at such-and-such an order of magnitude”, then we are constrained to defend its claim at that order of magnitude. A similar remark applies to any metaphysical notion concerning the physical world and our relation to it which generalises from local conditions (a fictive example would be some notion of ‘absolute weight’; in the philosophy of time, there is the very real example of the ‘absolute present’).
In these respects, and perhaps in these respects alone, physical science can be a guide for metaphysics (similar remarks apply to the biological sciences and – with the evident provisos – to the human sciences). Undoubtedly, the naturalist metaphysician’s self-elected role consists in part in showing up incompatibilities and inconsistencies between scientific and philosophical notions about the physical world, or in identifying inappropriate generalisations from local conditions. A further part consists in tracing the limits of what science can tell us about the world of our everyday experience. But the naturalist philosopher is not restrained to a merely critical role: by following Callender’s recommendations, she can play an active part in tracing and clarifying the various parallels and divergences between our phenomenal experience and our knowledge of the physical world.
1 “Quine’s philosophy of science”, Synthese, 19 (1968).
2 1981, “Time, Reality, and Relativity”, in Healey (ed.), Reduction, Time, and Reality, Cambridge: Cambridge University Press.
3 “Causal Determinism”, http://plato.stanford.edu/entries/determinism-causal.html
4 “Science as a Guide to Metaphysics?”, Synthese, 149 (2006).
5 Indeed, in their 2002 article “Ephemeral point-events: is there a last remnant of physical objectivity?” (Dialogos, 37), Pauri and Vallisneri remark on “…an unfortunate ambiguity in the usage of the term space-time points in the literature: sometimes it refers to an element of the mathematical structure that is the first layer of the space-time model, sometimes to the points interpreted as physical events”.
7 Casati, R. & Varzi, A. (2008) “Event Concepts” in T. F. Shipley & J. Zacks (eds.), Understanding Events: How Humans See, Represent, and Act on Events, New York: Oxford University Press, forthcoming.
8 Hacker, P. (1982), “Events and Objects in Space and Time”, Mind, 91.
David Hirst is British, but lives and works in France. He earned his Master’s in philosophy at the Sorbonne and took his PhD with the Sorbonne and the French Institute for the Philosophy of Science; he also holds degrees in linguistics and semiotics. Since 2001, and to meet the financial requirements of a young family, he has been working in executive training, but he plans both to reorient himself towards popular science education and to return to academia (probably by taking a PhD with an Anglophone university).
David has worked primarily in metaphysics, language, and the philosophy of science. At present, his main areas of interest are events and the philosophy of space and time; he is also interested in the philosophy of mind (and particularly the divergence between our best-candidate theories of spacetime and our phenomenal experience of time).
David is a “Bright,” but would distance himself from certain of his more ‘bulldoggish’ fellows: his interests as a Bright lie rather in combating ‘magical thinking’ and in promoting better popular understanding of science. While he defends both metaphysical and scientific naturalism, he considers that the term ‘atheism’ has no place in philosophical debate, and would argue for more debate between believers and naturalists.