At the beginning of the eighteenth century a variety of different versions of an argument based on the analogy between mental phenomena and the functions of a machine circulated in print and in manuscripts. These arguments tended to point out that the way in which mental functions are related to the mechanisms of the body can be compared to the way in which the functional properties of machines are related to the material mechanisms through which they are realized. Insofar as we are now used to seeing the use of mechanical metaphors in physiology and psychology as symptoms of mechanism or of reductionism, these arguments are fascinating since they tend to use these metaphors against mechanism rather than for it. Here is one version of the argument:
Because clear and evident ideas are not always the mark of truth; it is to be presumed that Descartes, when he deprives matter of sentiment in all states, is wrong, like a man that would represent brass and who would know that it is fusible, malleable and that one can make filings of it: but who, however, on account of the idea of these properties would deny that it would be capable of marking the hours; because all that he knew and conceived of this metal would not appear to him have anything to do with hours. The thought of this man is even more comparable to that of Descartes, if he denies that the watch that the watchmaker presents him marks the hours, because he cannot conceive how brass could produce such an effect. Because Descartes denies also that animals would have sentiment solely on the grounds that he does not conceive how the matter of which they are made could have it. However that may be, one must consult nature closely and listen to her language, which is always very real. If she says that matter, although insensible as it is naturally, can after changes and certain constructions acquire sentiment, one must believe her, although one does not understand how this is possible. (Gaultier Reference Gaultier1714, 85-86; my translation and my emphasis)Footnote 1
Excerpts from this text circulated in the form of clandestine manuscripts. Two versions of the text, titled Nouvelle philosophie sceptique and Parité de la vie et de la mort, can be found in the libraries of Paris (in the following volumes consisting of several manuscripts: Arsenal ms. 2239 and Mazarine ms. 1192). A printed version of the manuscript appeared in Pieces philosophiques (3 vols, Amsterdam, 1771, vol. I).Footnote 2 Variants of the argument can also be found, for example, in the manuscript Dissertation sur la formation du Monde (Mazarine, ms. 1168, no. 2, 212-213); in Voltaire's Dictionnaire philosophique, article “Ame” (Voltaire OC, Reference Voltaire1855-1875, XVII, 150); and in the Médicine de l'esprit of Antoine Le Camus (Reference Le Camus1755, vol. I, 174-175). Condillac denies the validity of the argument in his Essai sur l'origine des connoissances humaines (Condillac [Reference Condillac and Roy1746] 1947–1951, I, I, i, § 7, n.1). La Mettrie obviously also refers to this argument when discussing Locke's suggestion of “thinking matter” when he observes that asking whether matter can think is tantamount to asking whether matter can mark the hours (La Mettrie [Reference La Mettrie and Vartanian1747] 1960, 149).
The wide circulation of this argument shows that it was a standard move in the discussions on materialism. We can see that it is directed against Cartesian dualism and that it can be used for defending materialism. The point of the argument is that we should not hesitate to allow material entities’ vital or mental properties, such as sentience, even if we are unable to explain them mechanically. It points out that such properties observed in some material wholes may be – to use modern jargon – emergent, in the sense that we could not predict them merely on the basis of the knowledge we can acquire by studying the properties of their parts.
Motoichi Térada (Reference Térada1998) has already made a good case for showing that Gaultier's materialism was an early example of emergent materialism. However, Térada contrasts this emergentism with the vitalist tradition represented by the Montpellierian school and the materialism of Diderot. Insofar as we can find similar arguments referring to the emergent nature of vital properties in the writings of Diderot, I would, instead of opposing Gaultier's materialism to their views, emphasize the emergent aspects of vital forces or properties in the writings of the vitalists and of Diderot. Thus I would see Gaultier's arguments for emergence as paving the way for vitalist materialism instead of moving towards an alternative conception concerning the ontology of life. The vitalists, however, usually avoided making strong metaphysical pronouncements about vital properties, including their emergence, and thus they often preferred ambiguous or disjunctive hypothetical statements when speaking about the relationship of emergent properties to those on which they supervene, as we would put it today. Furthermore, it is characteristic of Diderot that he took seriously the philosophical problems associated with the idea of novel properties emerging from the more elemental properties. This led him to defend the complementary hypothesis according to which emergence alone cannot explain the apparition of vital properties and that we must allow that matter primitively possesses certain properties that enable it to form living organisms whose functioning cannot be reduced to the mechanical laws that govern matter in motion.
In any case, the argument quoted above claims that mental phenomena and the functional properties of machines are analogous in the sense that they are both irreducible to the mechanical or physical properties of matter. Therefore, Locke's suggestion that matter could think becomes trivial: there are no a priori reasons why it could not acquire this capacity. Although the previously mentioned occurrences of the argument concern primarily psychological properties or functions, similar arguments could also be used for defending the postulation of various organic or vital properties in living organisms. Thus eighteenth-century vitalism and materialism had a common enemy in Cartesian mechanism, which tended to regard all properties of matter that are not reducible to extension and movement as illegitimate “occult qualities” (see Clarke Reference Clarke1989).
On Mechanical Models and Analogies
We are used to relating mechanical models and analogies to reductionism and mechanism. One should however make a distinction between analogies and metaphors referring to technological artefacts and mechanical explanatory models. In the eighteenth century both of these were popular. But they were usually used for quite distinct purposes. And these distinct purposes correspond to distinct ontological commitments. In the Cartesian tradition, mechanism was concerned with providing explanations. The Cartesian animal-machine hypothesis and the Cartesian mechanistic physiology purported to explain all of animal physiology and a lot of human physiology and the psychology of passions in terms of mechanistic explanatory models. Later eighteenth-century dualists like Charles Bonnet tended to explain complex mental activities by referring to the mechanical workings of the nervous system: thus for example, association and reasoning were explained by mechanical neural models (Kaitaro Reference Kaitaro1999, Reference Kaitaro2004, and Reference Kaitaro, Whitaker, Smith and Finger2007).
But the mechanical analogy presented in the argument quoted is entirely different from the explanatory hypothesis of Cartesianism: it does not provide any explanation of mental phenomena. Instead, it clearly states that we don't understand how sentience arises, or even how it could arise from organized matter. But the argument not only makes a sceptical epistemological claim about the limits of our knowledge; it also makes the positive claim that material entities can have properties that its parts do not have. This fact could permit one to allow material entities vital or psychic properties without being obliged to postulate non-material entities like souls to account for them nor to reduce them to the properties of unorganized matter or to the movements of material particles. This in turn would permit one to defend the autonomy of physiology, which could thus be construed as the study of physiological functions. These functions would be analogous to the functions of technological artefacts. And just as knowing the properties and physical interactions between the constituent parts of an artefact is not sufficient in order to understand its function, so one needs something more than physics and chemistry to understand biological and physiological functions. In both cases one needs to examine the role of the artefact or the organ in a larger context: human practices or the vital functions of an organism.
This contrasts with the iatromechanistic tradition. In this tradition, which relied on Cartesian mechanistic models, physiology and physics were hardly distinguished; physiology was merely applied mechanics. Physiological phenomena were explained as mechanical movements of inert matter. Only the human soul was supposed to lie outside the domain of mechanics. During the eighteenth century this dualistic word view was no longer taken for granted. For example, the Montpellierian vitalist school of medicine emphasized the autonomy of the life sciences and considered life as a phenomenon sui generis not reducible to physical phenomena or mechanics (Roger Reference Roger1963; Rey Reference Rey1987/2000). One should not, however, let the term “vitalism” mislead: their vitalism does not resemble the nineteenth-century spiritualistic varieties of vitalism, which tended to postulate metaphysical or immaterial “vital forces” or “vital principles.” Their vitalism was compatible with materialism as its influence on Diderot's biological materialism shows (Kaitaro Reference Kaitaro1997). Thus, their theories could as well be characterized by such terms as “holistic” or “organicistic” (see Rey Reference Rey1987, vol. I, 1, 9–11, 28, 34–35 and vol. III, 155–156). The Montpellierians criticized both iatromechanism and the animistic physiology of Georg Ernst Stahl, who explained physiological processes by the guiding action of the soul. In fact, one of the Montpellierian vitalists, Ménuret du Chambaud dismisses mechanism exactly because it is favorable to animism: if the human body would be an inorganic machine, one would be have to postulate some other agency to guide its movements (“Inflammation” in Encyclopédie VIII, 713a). It is true that taking mechanistic metaphors literally, as theoretical explanatory models, is plausible insofar as they can be supposed to have, as machines always do, a designer (God) and a user (the soul). Thus it is natural that mechanism was more popular among dualists than among materialists (Kaitaro Reference Kaitaro2001 and Reference Kaitaro2004).
Vital Properties
During the eighteenth century the life sciences gained autonomy, which allowed for theoretical concepts to be introduced even if they were not reducible to physics: one could refer to non-physical properties like irritability or to functions like sensation without being obliged to be able to give a mechanistic account of them – or, as Gaultier's text cited above observes, even without being able to understand how such properties or functions were possible. Thus, for example, Albrecht von Haller introduced irritability as a property of muscle tissue, without being able to explain its causes – just as Newton had done in postulating gravitation as a general property of matter. The Swiss physiologist emphasized that irritability was distinct from other known properties of matter. According to him it was not reducible to mechanical properties such as elasticity, although he supposed that it depended on the “primordial fabrication” of the parts that are susceptible to it. The postulation of such unexplainable properties was considered perfectly legitimate insofar as they – or their effects – were experientially observable: so Haller provided “operational” definitions of both irritability and sensibility (Haller [Reference Haller1753] 1892, 11–71).Footnote 3 By considering irritability to be a property of organic tissue and by distinguishing it from sensibility or sentience, which involved the soul (at least in human organisms), he opened the way to attributing vital properties to matter itself, without needing recourse to soul-like entities. The Montpellierian vitalists and Diderot tended to neglect the Hallerian distinction between irritability and sensibility and to conflate the two in the notion of sensibilité, which was considered to be the property that makes organs and organisms living. In so far as the vital properties of organisms were regarded as properties of organized matter, the Montpellierians and Diderot escaped both mechanism and dualism.
The Montpellierian doctor Théophile de Bordeu – who appears as a protagonist in Diderot's dialogue Le rêve de d'Alembert – thought that the force that guides the manifold movements of the human body could be the essence of a portion of matter or a necessary attribute of its combinations. He admits that he cannot explain the ultimate nature of this force and that some of his ideas – such as sentience attributed to separate organs – are presented as “manners of conceiving things,” as metaphorical expressions and comparisons (Bordeu Reference Bordeu1818 (OC), I, 163, n.1, and 166). Paul-Joseph Barthez, the eighteenth-century vitalist who did use the term “vital principle,” defended the introduction of this principle (which according to him is not reducible to the laws of physics and mechanics) by claiming that it is not illegitimate to introduce “occult” faculties, principles, or forces as long as these are based on experimentally verified regularities (Barthez Reference Barthez1778, “Discours Préliminare”). And in fact, the vitalist criticism of iatromechanistic theories was not merely philosophical but also experimental: for example, in his Recherches anatomiques sur la position des glandes et sur leur action (1751) Bordeu produced anatomical observations and experimental evidence in order to falsify mechanistic theories concerning the secretion of glands (see Rey Reference Rey1991).
Interestingly, Newton's gravitation was widely used as a precedent when one wanted to postulate new vital properties to matter (Kaitaro Reference Kaitaro1997, 83–86). And such “occult” forces, postulated but not explainable by mechanical interactions between inert particles of matter, were just what the Cartesians mechanists considered suspect. Thus, instead of mechanizing the world-view of the Enlightenment, at least in France, Newton's influence counteracted mechanistic speculations and mechanistic explanatory models. When La Mettrie claimed that man was a machine, he was not so much developing Descartes’ mechanistic view into its logical conclusions as calling into question its metaphysical postulates, which opposed inert and passive matter to the activity of the soul. Despite the polemical title of his L'homme machine, La Mettrie was not a mechanist in the sense that Descartes was in his explanations of animal behaviorFootnote 4: instead of proposing mechanistic explanatory models in this work, he emphasized that matter can have inherent organic or vital properties (Thomson Reference Thomson1988; Kaitaro Reference Kaitaro2001). In fact, one could summarize the distinction between La Mettrie and the Cartesian mechanists by saying that while the former relied on mechanistic metaphors the latter proposed mechanical explanations to be taken literally. Also Diderot's materialism stems from the vitalist background. He supposed that sentience or sensibilité is a (potential) property of matter activated in living organisms (Kaitaro Reference Kaitaro1997, 101–106).
That the life sciences (natural history, physiology, and medicine) gained their autonomy during the eighteenth century by freeing themselves from the excessive reliance on mechanical explanations meant that phenomena like sensation and perception were now seen as material events belonging to the domain of these sciences – as phenomena that did not, despite that fact that they were not reducible to physics, involve theological considerations of the existence of an immaterial soul. There is, however, one ambiguity in the theoretical positions defending the specificity of life. The authors we have cited tend often to speak of properties instead of functions. Strictly speaking properties are intrinsic to the entities to which they belong, whereas functions, although they depend on the properties and structure of the things that have them, are extrinsic: parts of body or of machinery have functions only by being integrated in a larger structure in which they have their specific role. This ambiguity is revealed in Diderot's hesitations concerning the nature of sentience: “a general property of matter or the product of organization” (propriété générale de la matière, ou produit de l'organisation) (Diderot OC/AT II, 116). In fact, the functions of tissues or organs are not really properties. Nor are they merely products of the internal organization or tissues or organs. A sense organ does not sense unless it is part of a whole organism (although this function naturally poses some restraints on its internal structure). Thus there is a constant tension in Diderot's writings between emphasis on the autonomous life of organs and organic molecules and emphasis on the life of the whole organism (ibid., 114–117). This tension results at least in part from not distinguishing clearly between the function of a part and the properties this part must have in order to be able to carry out this function. The distinction corresponds to that between “upward” and “downward” explanations of functional systems. Diderot seems to want to keep both dimensions present in his analyses.
However, Diderot's disjunctive formulation concerning the nature of sentience probably reflects more than the mere ambiguity of the notion of property. He had realized that it may in fact be extremely difficult to conceive how inert matter could – albeit after some “changes” or “constructions” as Gaultier indicates – acquire sentience. In his writings Diderot expresses serious doubts that organization alone could be capable of producing new vital properties like sentience. In a letter to Sophie Volland, dated 15th October 1759, he writes that however one arranges, disposes, or organizes particles of inert matter, they remain what they are, inert matter. Supposing that their combinations could result in a living body is, according to Diderot, simply absurd (Diderot Reference Diderot and Babelon1938, I, 69). Likewise in the article “Naître” in the Encyclopédie, Diderot presents the same argument and concludes that “life cannot be the result of organization” (La vie ne peut être le résultat de l'organisation) (Encyclopédie, XI, 10a–b). Thus Diderot is led to suppose that sentience is somehow present, although in an “inert” form (comparable to movement in heavy bodies immobilized by an obstacle) in all matter, as a property which is activated when this matter is assimilated into a living organism (Diderot Reference Diderot and Roth1955-1979, Corr. V, 138-142). Consequently, Diderot's disjunctive formulation can be taken to mean that insofar as matter can acquire vital properties when it is organized, it can do so only if we suppose that it already possesses a primitive property or disposition that allows it to acquire these novel properties. In this way, he could as well have written that sentience is a general property of matter and the result of organization.
Diderot was also well aware of the fact that in speculating on the vital properties of matter he was dealing with mere suppositions. In the Réfutation de d'Helvétius where he repeats the idea that the mere organization of inert parts is not sufficient to produce sentience, he also admits that the general sentience of matter is merely a supposition (Diderot, OC/AT II, 302). His position, which reminds us of the likewise carefully formulated disjunction of Bordeu on the nature of vital force, is characterized by a combination of hypothetical speculation and a healthy dose of epistemological scepticism. Thus insofar as Diderot's or Bordeu's position could not unambiguously be characterized as “emergentist” (see Térada Reference Térada1998), this is evidently due to their recognizing the limits – the explanatory vagueness or even emptiness – of organization in explaining emergent properties. But here they are in fact in line with Gaultier, who openly admits his ignorance: if there are phenomena that we cannot explain by the mechanical laws governing inert matter in motion, we cannot explain their emergence either, for the very reason that they cannot be predicted on the basis of our knowledge of the physical properties of matter. Seen in this light, the postulation of such emergent properties would, despite having the appearance of a scientific supposition, perhaps be better understood as an argument for methodological pluralism and for the autonomy of sciences dealing with functions, rather than as a scientific hypothesis concerning the properties of matter.
Varieties of Vitalism
We have already hinted that one should make distinctions between different forms of vitalism. Indeed, as E. Benton (Reference Benton1974) has pertinently observed, sometimes the issues dividing vitalists from one another can be more significant than those which divided vitalists from non-vitalists and that in some cases the latter issues were little more than terminological disputes. To remedy the situation Benton proposes some useful classificatory dimensions. Benton's first dimension concerns the epistemological scepticism of the vitalists: the phenomenalist vitalists considered vital powers to be “general relationships arrived at through observation of phenomena” whereas the realist vitalists hypothesized “the existence of a non-physico-chemical agency or entity underlying observable vital phenomena which are its effects.” Benton's second dimension distinguishes vitalist theories in terms of the formal character of the explanation they propose: the explanation can be either teleological or nomological. Thus the teleological vitalists conceived vital powers “not as minds or souls, but as analogous to them in that they were supposed to operate rationally in pursuit of some aim or goal.” The nomological vitalists thought of their vital powers as being exercised according to laws of blind necessity, by “laws which would be discovered by observation and experiment, but which would not be reducible to physics and chemistry.” Then there were also non-nomologial vitalists who “stressed the variability of vital powers in such a way as to invite the interpretation that the exercise of the vital powers was not subject to law-like regularity of any sort” (neither teleological nor nomological).
We can now see that the eighteenth-century vitalists were mostly of the phenomenalist variety. It can be disputed whether they were teleological or nomological. Benton's distinction seems to presume that the laws that are not reducible to physics and chemistry have to be teleological in the rather strong sense that they operate rationally in pursuit of some aim. The teleology implied in Benton's formulation seems a bit too strong to apply to the vitalism of the Montpellierians. Anyway, it is at least clear that they were less teleological than the Cartesian mechanists such as Descartes, Malebranche, or Bonnet, who constantly referred to the ultimate aims of physiological mechanisms, the conservation or the survival of the organism, planned by the ultimate final cause, God (Descartes AT IX, Méditations 64–70 & Principes 64–65; Malebranche Œuvres I, 87, 95, 297, and passim; and Bonnet Reference Bonnet1760, § 119). For the vitalists, vital properties seem to be inherent in living tissues or organisms. This tendency to view them as properties of matter or of material structures constitutes, as we have already observed, perhaps one of the conceptual weaknesses of these doctrines. But from the materialist point of view this strategy could be motivated by the need to avoid evoking final causes. Apart from materialist philosophers such as Diderot, such a motive can also be attributed to medical writers who were anxious to follow the contemporary requirements posed for properly scientific explanations and who thus tended to view teleology referring to aims and ends with suspicion. Thus, insofar as some kind of teleology would be involved it would be different from that involved in artefacts, which are made for a purpose and which require a separate entity to plan and guide their movements. And although the eighteenth-century French vitalists were of the opinion that the regularities observed in living organisms cannot be reduced to physical laws, they did not deny the existence of law-like regularities concerning living organisms. As we have seen, the existence of such experimentally verified regularities are what motivated Barthez to postulate vital properties.
But what really connects this kind of vitalism to the materialist arguments concerning mental functions presented at the beginning of this paper is the emphasis on the fact that one can admit the existence of vital or mental properties without being obliged to be able to explain their presence mechanically or otherwise. In this way these physiologists could postulate physiological unknowns, such as Haller's “irritability,” which were considered to be properties of certain tissues, but properties that one need not be able to assign a cause. Such speculation on the causes of vital properties would have been at least premature. But it is not necessarily merely a question of waiting for science to progress. Speculating on their causes might also have been useless insofar as there need not be any common causes for these phenomena. Hall (1969) observed pertinently on this matter: “Had a step-by-step analysis been essayed, to show the working out of such properties and powers in physiological systems, the analysis would have failed because behavior is dependent, in them, on complexities of organization which differ from system to system and render the search for common causes futile.” The eighteenth-century vitalists studied and named physiological phenomena without being able to give physical or mechanistic explanations for them. In addition to the fact that such speculations on reductionist explanations may have been premature, i.e. impossible at that stage of scientific development, they may also be futile due to the multiple realizability of the functions in question. Although irritability, sensibility, or contractility of certain tissues depends on their structure and on the arrangement of their constituent parts, there need not be any common physical features involved in tissues or organs taking care of such functions that would explain these properties or functions.
Conclusions
To sum up, when modern physiology constituted itself in the eighteenth century as an autonomous science by distinguishing itself from physics and mechanics, this was done by accepting theoretical concepts that could not be reduced to mechanics.Footnote 5 We can see a similar logic at work when new sciences or epistemic practices reach their autonomy from models borrowed from other sciences. For example, a decisive step in the constitution of a science of social facts was taken when Comte gave up previous Saint-Simonian pretensions to deduce social sciences from the natural sciences or physiology and considered the possibility of an autonomous social science, “social physics,” or sociology (Bénichou Reference Bénichou2004, 680, 690, and 694). It seems that Freud's predicament was similar in relation to physiology or neurology: he founded a new kind of clinical psychiatry by insisting on theoretical concepts that were not reducible to physiology or neurology. By doing this he founded a new autonomous field of study.
Although we now know more about the mechanisms by which different organs realize their functions, there is no evidence that the sciences studying functions – whether they are of cultural artefacts, of organs, of tissues, of cells, or of mental representations – could be made obsolete by the study of the laws that govern the behavior of the matter which the objects having a function are made of. Of course we can usually explain dysfunction by referring to the properties of this matter, for example, when we explain why a clock made of brass stops when the temperature gets too high or why we lose consciousness if our brain does not get enough oxygen. But here we are dealing with pathology and not with physiology. We cannot find out the function of an unknown artefact found in a jungle without knowing something of the society in which the artefact was used. Likewise we cannot find out the function of a muscle sold in the butcher shop without knowing something about the physiology of the whole animal from which it was removed. Tools and organs are not only etymologically related – the ancient Greek word “organ” means a “tool” – but they also share the feature that finding out what they actually are requires looking for the larger context rather than analyzing their parts. Insofar as vitalism does not require the postulation of metaphysical and non-material entities or vital forces, which it did not necessarily do in the eighteenth-century,Footnote 6 but is satisfied with the more modest claims concerning the autonomy of physiology and the emergent uniqueness of life,Footnote 7 vitalism is not necessarily dead. What has definitely changed since is our understanding of what vitalism was about. The historical varieties of vitalism have unfortunately been replaced with an image of a reactionary spiritualist doctrine, an easy target for biological reductionists.Footnote 8 But we can hope that recent work done on eighteenth-century French vitalism by scholars like Roselyn Rey (Reference Rey1987 and 2000) and Elizabeth Williams (Reference Williams2003) will replace this obsolete image of a carcass with something with a more vital appearance.
