4 Assemblages and Scientific Practice

[W]hat becomes apparent in the rivalry between [minor and royal sciences] is that the [minor] sciences do not destin science to take on an autonomous power, or even to have an autonomous development. They do not have the means for that because they subordinate all their operations to the sensible conditions of intuition and construction – following the flow of matter . . . However refined or rigorous, ‘approximate knowledge’ is still dependent upon sensitive and sensible evaluations that pose more problems than they solve: problematics is still its only mode. In contrast, what is proper to royal science, to its theorematic or axiomatic power, is to isolate all operations from the conditions of intuition, making them true intrinsic concepts, or ‘categories.’ That is precisely why deterritorialization, in this kind of science, implies a reterritorialization in the conceptual apparatus. Without this categorical, apodictic apparatus, the differential operations would be constrained to follow the evolution of a phenomenon . . .

Deleuze and Guattari, A Thousand Plateaus1

[W]hen Galilei experimented with balls of a definite weight on the inclined plane . . . a light broke upon all natural philosophers. They learned that reason only perceives that which it produces after its own design; that it must not be content to follow, as it were, the leading-strings of nature, but must proceed in advance with principles of judgement according to unvarying laws, and compel nature to reply to its questions . . . Reason must approach nature with the view . . . of receiving information from it, not, however, in the character of a pupil who listens to all that his master wishes to tell him, but in that of a judge that compels the witness to reply to those questions which he himself thinks fit to propose. To this single idea must the revolution be ascribed, by which after groping in the dark for so many centuries, natural science was at length conducted into the path of certain progress.

Immanuel Kant, Critique of Pure Reason2

Most philosophers of science define the object of their research using the ideas that Kant sketched two centuries ago. The products of science must consist of necessary and general laws, derived not from messy laboratory phenomena but from ideal phenomena created in advance by the intellect using the language of mathematics. Since only one scientific field fully complies with this requirement, the field of physics, the latter is viewed as the only real representative of science, the other fields relegated to a lower status. And this conception is not just dominant in philosophical circles but often affects the practitioners of minor fields who tend to succumb to the temptation of thinking that the discovery of universal laws is the only way to achieve success. This does not imply that these practitioners do not discover regularities in the behaviour of laboratory phenomena, but these are not synthetic a priori truths derived deductively from an idealised model created ahead of experimental interventions. Rather, they are derived inductively from many acts of observation and measurement as a phenomenon’s evolution is followed, and this has the consequence that the ‘laws’ are not exceptionless but subject to variation, implying a lack of apodictic or uncontroversial certainty.

In the opening quotation above, Deleuze and Guattari acknowledge the existence of both kinds of scientific practice, major (or royal) and minor. The former not only manages to produce necessary and general statements but to arrange them in the form of an axiomatic, the exemplary case of the synthetic a priori. The latter break with Kant’s conditions because they follow a phenomenon and allow the latter to pose problems, rather than sitting it in the witness stand and confronting it with questions prepared in advance. Unfortunately, the only case of a minor science that the authors analysed was metallurgy, but metallurgy is a craft not a scientific field. And in later publications they revert to a monolithic treatment of science, in which the latter is modelled on physics.3 Nevertheless, their powerful insights on the matter can be fully recovered by framing the minor/major distinction in terms of assemblage theory. As usual, the very first move in this task is to replace a reified generality, science, with a population of individual scientific fields, each with its own methods, procedures, and instrumentation, a population that is not converging on a final field – as all become more like physics and are eventually reduced to it – but diverging as new fields are created and as sub-fields proliferate through specialisation and hybridisation.4

A scientific field can be modelled as the assemblage formed by a domain of objective phenomena, a community of practitioners, and the laboratory instruments and machines that allow the latter to interact with the former. The whole formed by these three components must possess irreducible properties and dispositions of its own, and the identity of the components must not be determined by their relations. These are, of course, the two minimal requirements of emergence and exteriority. The concept of a domain was relatively recently introduced into the philosophy of science.5 Rather than viewing scientists as interrogating Nature, another reified generality, we must model their practices as oriented towards the contents of a historically constituted domain of laboratory phenomena, many of which are not naturally given but artificially created. The concept of an objective phenomenon has itself been recently revived, to refer to laboratory effects that can either arise spontaneously or, on the contrary, that may involve a lot of work to be refined and stabilised. Given the variety of fields and their domains, we can expect objective phenomena to be wildly diverse, sharing only a few characteristics in common: all phenomena must be public, recurrent, and noteworthy.6 A good example of a phenomenon meeting these conditions is the chemical reaction between an acid and an alkali. The furious effervescence characterising this interaction had been a recurrent, publicly witnessed effect for centuries before chemistry became a field, being remarkable, or worthy of notice, because it suggested a battle or an intense struggle.

The human component of the assemblage is a community of practitioners, although not one as tightly knit as those discussed in previous chapters. In these communities, reputations still matter, and ridicule and ostracism can still be used to enforce community norms, but what really binds the community together is the set of cognitive tools that govern the personal practices of its members: the concepts they use to refer to the properties and dispositions of phenomena; the statements about phenomena they believe to be true; the problems posed by phenomena, problems that can be modelled as ‘Why’ questions (Why do acids and alkalis react violently instead of just mixing peacefully?); the reasoning strategies developed to tackle those problems; and the classification schemas used to order the domain.7 This list should be left open because it is impossible to tell in advance what novel cognitive tools may be added in the future. In the late eighteenth century the personal practice of chemists did not include the use of empirical chemical formulas like H2O, and as recently as the 1870s mathematical models were an alien cognitive artifact to most chemists. Yet by the start of the twentieth century different types of chemical formulas had been added to the repertoire, and mathematical models were on their way to general acceptance. A century later, the list now contains several novel tools, such as computer simulations. Thus, the cognitive tools that are available to practitioners at any particular time form an open set, and must be conceived as related to one another in exteriority. This implies a rejection of holism, that is, the idea that all cognitive tools are fused into a monolithic theory or paradigm which must be accepted or rejected as a whole.

In addition to characterising personal practices, we need to consider the emergent properties of the community as a whole: the consensus practices that slowly become distilled as personal differences are partially worked out and the repertoire acquires a more impersonal nature. Evidence for the emergence of a partial consensus can be produced by checking textbooks at regular intervals (say, fifty years apart) and analysing their contents comparatively: what cognitive products are included in a textbook compared with another half a century apart? Textbooks are notoriously unreliable when it comes to either the history of the field (they invariably include mythologised versions) or when discussing general methodological issues, such as the nature of the Scientific Method. But when it comes to individual cognitive tools, textbooks contain a good record of what has become consensual, and more importantly, being teaching tools themselves, they are evidence of what was transmitted from master to disciple at any given time, and hence, of what can be considered to be shared by a particular generation of practitioners. On the other hand, the convergent effect of consensus formation (a territorialising effect) must be complemented by the variation in personal practices, as each practitioner confronts new phenomena, finds new uses for old tools, or is forced to adopt new, unfamiliar ones. At any given time, different members of the community may accept different tools as valid while rejecting others. But as long as there is enough overlap among their repertoires (enough common concepts, acknowledged statements, recognised problems) there will not be any danger of a breakdown in communication, as in those models that postulate an inevitable incommensurability between practices belonging to different holistic paradigms.8

There is one more element that must be added to the assemblage to complete the model. A field is not just composed of a community of practitioners but also of one or more organisations. The simplest of these is the laboratory, a social assemblage with an authority structure, even if its hierarchy includes only two levels: the master and the disciple or assistant. But larger organisations may also be included, such as the Royal Academies that formed in the national capitals of various countries to promote scientific agendas in education, to gain legitimacy in the eyes of government and ecclesiastical officials, and to provide the required services to keep the community bound together: most Royal Academies had a staff member dedicated to reading letters from other similar organisations, translating them, and responding to them. They also offered other services, like organising national (and later on, international) conferences and publishing the proceedings. Adding the organisational infrastructure constituted by laboratories, academies, and university departments allows us to view the community of practitioners as being socially situated without having to bring in the reified generality Culture. This way, with individual domains replacing Nature, and a population of individual fields replacing Science, we can finally avoid the absurd questions that dominated the discussion in the closing decades of the twentieth century, questions like ‘Are controversies in Science settled by Nature or Culture?’.

We are in a position now to tackle the distinction between minor and major scientific fields. Only two distinguishing characteristics are mentioned in the opening quotation, but more are listed elsewhere. The first is that while minor fields stick to the sensual materiality of the phenomena, following wherever they lead, major fields seek universal truths that can be used as axioms to mechanically derive further truths (theorems). Because, unlike inductive logic, deductive logic is not ampliative (that is, it does not add any truth that is not already in the axioms), the resulting axiomatic structure is entirely self-contained and divorced from the phenomena in the domain. The second characteristic is closely related to the first: major fields interrogate laboratory phenomena using questions derived from the regularities displayed by ideal phenomena (regularities encapsulated into synthetic a priori statements referred to as ‘laws’) while minor fields allow phenomena in their domain to pose problems, and indeed, are forced by changes in the domain to constantly face new problems. In this chapter we will use chemistry as our main example of a minor field.

The domain of chemistry since 1700 was made up of substances and their transformations. This statement must be qualified, because while substances were there from the start – as part of the material culture of pharmacists, metallurgists, and alchemists, from which chemistry evolved – chemical transformations were at first only instruments, having to wait until the middle of the eighteenth century to become objects of study in their own right.9 Replacing Nature with an individual domain is particularly illuminating when the field under examination practises synthesis, in addition to analysis, because the former creates new phenomena that may not exist naturally. As one philosopher of chemistry puts it:

On the very experimental level chemistry reveals a momentum to multiply the number of its objects, because every chemical experiment possibly generates new chemical substances for new chemical experiments . . . As a result, the number of chemical substances has increased to more than 16 million in 1995 and about 1 million (!) new ones are made a year now. There is actually no comparable natural science with such a productive power concerning its own classificatory objects.10

When the size of a domain increases due to the very activities of a community, the only possible course of action is to surrender to the phenomena, following them as they accumulate, trying out their ever more numerous combinations – the number of possible chemical reactions increases much faster than that of substances – and giving up any notion of a final account that could be neatly encapsulated into an axiomatic. These ‘population explosions’ are veritable deterritorialisations of the domain, generating new lines of research the end of which cannot be predicted in advance.

Historically, the first population explosion was produced not by synthesis but by chemical analysis, and it did not just deterritorialise the domain but decoded concepts and statements. From the inception of their field, chemists distinguished between elementary and compound substances. Compounds, like acids and alkalis, can be synthesised, so chemists understood that there might not be an upper limit to their number. But elementary substances were constrained by an inherited code that demanded that their number be small. In the eighteenth century chemists had a choice of what substances to count as elementary from those coded by Aristotle – earth, water, fire, and air – and those coded by Paracelsus, including sulphur and mercury. These were not the ordinary substances brimstone and quicksilver, but hypothetical principles assumed to compose those ordinary substances. Different practitioners adopted various combinations of the abstract principles, more or less dogmatically, but also guided their research by pragmatic guidelines that were more directly in touch with the phenomena: they tended to regard a substance as elementary if it was impossible to decompose it any further using only chemical means. The coded dogma and the pragmatic guideline coexisted peacefully for a while, but as the chemical means available to isolate and decompose substances increased, code and guideline began to clash. Air was the first casualty of this confrontation. Up to 1727, chemists let the gases produced during a reaction escape, but in that year a simple apparatus was created that allowed their capture, after washing them through a liquid like water.11 This made those ‘airs’ subject to chemical analysis and some (carbon dioxide) were found to be compounds, while others (oxygen, hydrogen) showed themselves to be different elements.

Water was the next casualty, not by being analysed, but by sparking together two of the recently isolated airs. Despite the evidence from their senses – water had clearly been synthesised from the two airs so it had to be considered a compound – some chemists still tried to save the old code.12 But all resistance crumbled when the availability of constant electric current from Volta’s recently invented pile, a primitive form of a battery, massively increased the power of chemical analysis. A single chemist, using the Volta pile, isolated sodium, calcium, boron, potassium, strontium, and magnesium in 1807, a list to which he added chlorine, iodine, and bromine three years later.13 As a response to this dramatic proliferation of novel elements, the abstract principles were abandoned, and chemists were forced to follow wherever analysis took them. They had to accept that there was no a priori upper limit to the possible elementary substances, and that the actual number had to be determined empirically. For a while this threw the portion of the domain containing elements into disarray, textbook writers struggling to explain to students what was going on, but then the famous Periodic Table was used to bring order back on the domain, in effect reterritorialising it.

Let’s move on to illustrate the second characteristic of a minor field: that phenomena, far from being passive witnesses to be interrogated, actively pose problems to practitioners. For most of the eighteenth century chemists sought to solve problems related to composition. The most active part of the domain were the neutral salts that resulted from a reaction between acids and alkalis (or more generally, bases). They knew that when the reaction was over the properties and dispositions of acids and bases had disappeared, replaced by those of the neutral salt. But they also knew that if the salt was analysed, its components could be recovered. In other words, synthesis could be used to check the validity of analysis, and vice versa. Thus a problem like ‘Why does substance X have these properties rather than other properties?’ could be solved by using analysis and synthesis to give the answer ‘Because it is composed of substances Y and Z’. For most of the century this qualitative statement sufficed, but in its closing decades chemists began to add quantitative information, as they realised that not just the composing substances mattered, but also their relative proportions: two compounds could be made out of the same components and yet have different properties. However, the techniques to accurately measure the relative quantities of a substance present in another took decades to be perfected and trusted, because chemical reactions had to be performed in closed form to avoid losing any component; substances had to be accurately weighed before and after a reaction; and a standard system to express proportions had to be developed. The slow pace of development of a quantitative approach was also caused by lack of urgency: the relative rarity of inorganic compounds with identical components but different properties meant that the phenomenon, although clearly problematic, was not pressing.

As organic chemistry began to develop in the following century, however, the problem became urgent, because analysis of organic compounds revealed that they were all made out of the same components: carbon, oxygen, hydrogen, and nitrogen. Yet compound substances of organic origin displayed an enormous variety of different properties and dispositions. Thus, the organic portion of the domain posed a problem in a way that could not be ignored, forcing practitioners to devise ways of solving it, including the development of chemical formulas, as well as various hypothetical explanations of how the synthesis of organic compounds was achieved in the bodies of plants and animals. Then, starting in 1823, chemists had to confront an even more problematic phenomenon, isomeric substances, organic compounds that have identical composition and proportions but different properties.14 The only available solution seemed to be to postulate that in addition to the nature and relative quantity of components, their spatial arrangement was needed for a full explanation. But unlike nature and quantity, both of which could be established by dealing with phenomena at the macroscopic scale, using spatial structure as part of the explanation seemed to demand moving to the microscopic scale, a very speculative move at the time and one about which chemists were deeply divided. Nevertheless, the problem raised by isomers eventually forced the development of new chemical formulas, structural formulas, and a more serious consideration of microscopic spatial factors.

These two examples do seem to indicate that chemistry can be considered a minor field, fitting perfectly Deleuze and Guattari’s definition. But as we have done throughout this book, it will prove useful to replace two static categories (minor and major) by two settings of the parameters of the assemblage, allowing us to treat a scientific field as a historical entity undergoing episodes of ‘becoming minor’ and ‘becoming major’. To return to our first example, the explosion in the number of elementary substances was a deterritorialisation, followed six decades later by a reterritorialisation, as patterns in the properties of the phenomena were first noticed, then actively searched for, culminating in the creation of the Periodic Table.15 Similarly, the proliferation of elements caused a decoding to occur, the demise of ancient principles, but this was followed almost immediately by a recoding, as what used to be an implicit guideline became the very definition of an elementary substance. Lavoisier was the first to make explicit the idea that an element was simply a substance that had not yet been decomposed, and thus he gave us a new codification. Since both territorialisation and coding vary by degree we can compare the settings of the parameters at the start and the end of the century: the old code used a preconceived schema, accepted a priori, while the new one made the status of elementary substance analytically contingent, keeping the practice closer to the materiality and expressivity of the phenomena.16

But if this is so, then why is the period between 1780 and 1800 considered the time when chemistry came close to achieving the status of physics? This was partly due to the way in which chemical reactions had to be territorialised – kept tightly closed, with the reactants and products carefully weighed – in order to produce quantitative statements about composition. This clearly increased the degree of control that practitioners had on phenomena. But it was also related to the replacement of the old names for substances with a uniform, highly coded nomenclature. The old names for chemical substances were coined following a variety of criteria: the sensible properties of a substance (its colour, smell, taste, consistency); its method of preparation; the name of its discoverer or the place where it was discovered; or even symbolic associations, like that between metals and the planets. This variation was exorcised by the new nomenclature in which names for compounds were derived in a systematic way from their composition.17 The new nomenclature allowed chemistry to present itself to other fields as a maturing field on its way to become a major science.

Deleuze and Guattari include other characteristics of minor and major fields in addition to the contrast between problematic and axiomatic approaches, and between the practice of following phenomena rather than interrogating them using predefined categories or laws. The authors argue that while minor science concerns itself with flows, major science treats fluids as a special case of a theory of solids; that while minor science deals with becoming, major science concerns itself with what is stable, eternal, identical, and constant; and that while major science prefers uniform or laminar flows, minor science is fascinated by the spirals and vortices that form when an inclined plane makes a fluid cross intensive thresholds, giving rise to convection and turbulence. As stated by the authors, these characteristics seem to apply to only one sub-field: hydrodynamics, a sub-field of physics that was indeed less subject to axiomatisation, and much less prestigious, for most of the nineteenth century.18 But they can be rephrased to be applicable to chemistry. Let’s start with the second distinction, that between variable becoming and eternal constancy. Chemical reactions illustrate a process of becoming in which one set of substances is transformed into another set. But in order for this to count towards the minor status of chemistry, reactions had to cease being mere instruments of analysis or synthesis, and start being treated as phenomena in their own right, that is, as something in need of explanation. This was achieved when regularities in these transformations were first classified in tabular form.19

The chemical reactions in question, referred to as displacement reactions, confronted chemists with a remarkable phenomenon. When they added powdered silver to a liquid solvent like nitric acid, the metallic substance dissolved, uniting with the acid; if they added powdered copper to this solution, the metal united with the solvent displacing the silver which now precipitated to the bottom of the container; if they added iron next, the copper itself was displaced and forced to precipitate; finally, adding zinc displaced the dissolved iron.20 This well-defined sequence of displacements was a striking display of the fact that the disposition of metallic substances to unite with nitric acid, the affinity of metals for the solvent, was a matter of degree: copper had a stronger disposition than silver; iron stronger than copper; zinc stronger than iron, and so on. And similarly for the affinities of different acids for different alkalis (and other bases). When it came to explain these dispositions to combine, some chemists tried to make an analogy with gravity. But although affinity was in fact a force of attraction, it was, unlike gravity, a selective force, and this selectivity had no counterpart in physics. Also, unlike physicists who reduced matter to mass, and were therefore unconcerned with understanding variation, chemists had to take the latter into account, even if just to classify the variants into what came to be known as ‘affinity tables’. Thus, by the middle of the eighteenth century it had become clear that temperature caused variations in selectivity, and that different tables had to be compiled for reactions carried on with the use of heat (distillation) and those carried on in solution. Later on, the degree of concentration of reactants was shown capable of reversing the direction of a chemical reaction as determined by the affinities of the substances involved.21 Thus, although chemists did search for a ‘law of affinity’ modelled on the eternal and immutable law of gravity, and performed this search with the goal of giving their field the status of a major science, the phenomena in their domain resisted this assimilation and they were forced to incorporate becoming and variation as non-eliminable features.

Like any other classification scheme, the different affinity tables brought order to the domain, or in other words, they territorialised it. However, to the extent that an ordered domain was required to be able to view chemical reactions as phenomena in their own right, and that as the different factors that could affect reactions were subsequently discovered they made previous tables obsolete, this classification scheme also help unleash decoding movements. At any rate, affinity tables were useful only for the part of the domain constituted by inorganic neutral salts (and the acids, alkalis, and metals that composed them) but were powerless when confronted with the most chaotic portion of it: the resins, gums, waxes, syrups, oils, sugars, and alcohols that constituted the organic part of the domain. Although these substances were of animal or vegetable origin, they were not taken from Nature but were common materials used by crafts and sold in markets.22 As it turned out, a tabular form of classification had to be replaced by a serial method to territorialise these problematic substances. Series of kindred substances had to be created and laid out over the relatively undifferentiated mass of raw materials, as if lines of latitude and longitude were being drawn over an unknown stretch of land to be able to record the gains from previous journeys and to guide further exploration. To produce the series, certain chemical reactions were recruited that performed a substitution of a single component (or group of components) of a substance by another, yielding a close relative to the original. The series negotiated a compromise between constancy and variation, each member sharing something with the others but also varying in a definite way.

The first series to be fully worked out was made up of the family of substances we know as methane, ethane, propane, butane, pentane, and so on. Using simple chemical formulas for these substances, the internal structure of the series could be clearly displayed: CH4, C2H6, C3H8, C4H10, C5H12, C6H14, C7H16, C8H18. Even a cursory examination of this sequence shows that it increases by a modular amount, CH2, and that the entire family can be generated from the formula CnH2n+2.23 This series had entered the consensus by the middle of the nineteenth century, allowing chemists to get a foothold on the confusing mass of disparate plant and animal substances, creating a small but secure territory from which to launch other series. By 1900 many more sequences containing a constant part and a variable part were crisscrossing the organic domain: the series of saturated and unsaturated alcohols, the formulas for which are CnH2n+1OH and CnH2n-3OH, respectively; the series of saturated ethers, expressed as CnH2n+2O; the series of aldehydes and ketones, with the formula CnH2nO; and the series of saturated acids, CnH2nO2.24 It may be argued that series like these subordinate variation to constancy, but the opposite effect could also be achieved: by substituting hydrogens in each substance with other elements (like chlorine, bromine, or iodine), an entire series could be set in variation.25

We may conclude from the preceding analysis of the cognitive content of chemistry that the field meets the requirements of a minor science. On the other hand, because the organisational part of the field (laboratories, academy sections, university departments) was affected by a variety of non-cognitive factors, the manner in which that content was presented to non-chemists could also be a source of territorialisations and codings that made chemistry seem closer to a major field. Among these non-cognitive factors are legitimacy and prestige. There is no doubt that chemists were always impressed by the much higher level of prestige of physics – even if many recoiled from the idea that a chemical phenomenon could be reduced to a physical one – and wanted to increase the social standing of their discipline by borrowing from their more prestigious relative. A symptom of this sense of social inferiority is that, on occasion, famous chemists felt the need to apologise when speaking to an audience of scholarly professors, because their subject was not mathematical, and because their laboratories were filled with strange smells, potentially poisonous airs, caustic liquids, fire, and smoke.26 In addition, chemists shared the belief that the legitimacy of a field was linked to the discovery of immutable laws, so they never stopped searching for them. The Periodic Table, for example, was at first referred to as the ‘Periodic Law’, even though it is a very different cognitive object, capturing the rhythmic variations in real elementary substances, not the constant dependencies among the properties of ideal phenomena. And whenever chemists discovered important regularities in the composition of substances they referred to these as ‘laws’ (like the law of definite proportions and the law of multiple proportions), even though these ‘laws’ were empirical regularities derived using inductive reasoning (not deduced from ideal phenomena) and, more importantly, they were not exceptionless.27 On the other hand, it may be argued that the image of science that attracted those who wanted to make chemistry more prestigious or legitimate is a rhetorical crust, a hardened stereotype, constituting the most rigidly coded form of the real cognitive content of physics. No doubt, physicists themselves promoted this stereotype to governmental and ecclesiastical authorities as part of their own search for legitimation, and once this superficial image of its content was appropriated by philosophers, it came to be identified as the essence of science.28

But if it is true that the content of physics, as understood by outsiders, is mostly rhetorical, then the image of physics as a major science must be questioned. Or rather, we must dig deeper into its many heterogeneous sub-fields to find mobile distributions of becoming minor and becoming major, just as in any other field. Thus, although the domain of chemistry was the first to include transformations (variable becomings), by the nineteenth century these were also part of thermodynamics, the sub-field of physics that studies the various ways in which one form of energy can become another. It is true that this sub-field began by studying transformations as if they had already taken place and the various forms of energy had all become heat. And it is also true that the emphasis was originally on the constant state achieved at the end, the state of equilibrium, and that the process through which this state was reached was conceived in terms of laws (the second law of thermodynamics). We will argue in the following chapter that in the twentieth century the study of energy transformations became deterritorialised, progressively moving further away from equilibrium, but even in its classical form thermodynamics may be said to be a minor science concerned with becomings because of the concept of time that it used. The time of macroscopic energy transformations is irreversible, one form of energy turning into another but always involving a degradation, until all the energy involved has acquired one homogeneous form. By contrast, the concept of time in classical and relativistic physics is fully reversible. This does not mean, of course, that time can be made to run backwards, from the future to the past. Rather, if we think of a physical process as a series of events, what the reversibility of time amounts to is that there is no difference in outcome if the series runs from the first event to the last, or if the series begins with the last event and ends with the first one.

Clearly, physicists did not arrive at this conception of time by keeping close to the materiality and expressivity of phenomena, since most physical phenomena do not exhibit this reversible character. If we took a video camera and recorded many series of spontaneous events, and then projected the video in reverse, the majority of the shots would look wrong: a broken plate would spontaneously put itself together again; a recently deceased animal would come back to life; a diver would be spat out of the swimming pool to fly into the air and land neatly on the diving board. In other words, following the phenomena provides evidence against reversibility. So the strong conviction that most physicists have that their concept of time is correct must come from ideal phenomena, since the equations used to model the latter do indeed remain invariant under a transformation that inverts time. If we follow Kant into thinking that the path to scientific progress depends on proceeding in advance with rational principles derived from laws, principles that should be imposed on a phenomenon as we interrogate it, then the fact that most classical laws display time reversibility carries enormous conviction, more so than all the evidence from thermodynamics. Nevertheless, the tension between a minor and a major conception of time is there at the heart of physics. Deleuze and Guattari acknowledge these internal tensions when they write that: ‘What we have . . . are two formally different conceptions of science, and, ontologically, a single field of interaction in which royal science continuously appropriates the contents of [minor] or nomad science while nomad science continually cuts the contents of royal science loose. At the limit, all that counts is the constantly shifting borderline.’29

This shifting borderline not only passes through different sub-fields, but also right through the middle of classical physics, although it is normally hidden from view by the hardened rhetorical crust. Historically, the cognitive content of this field has been variously coded in axiomatic form, thus matching a criterion for a major science. Yet when checking a modern textbook to see how this cognitive content is being transmitted to a new generation, we do not find a logical system of statements divided into self-evident first principles and derived theorems, but a master differential equation (called a Hamiltonian) from which entire families of other equations can be constructed as models for various ideal phenomena. The process of constructing such models is not a logical derivation, and demands creativity, as does extending each model to cover less idealised phenomena. Thus, the structure of the body of knowledge of classical physics revealed by an open-ended family of Hamiltonians is very different from that presented by a closed axiomatisation of the field.30 We can push this line of argument even further. When we examine an actual axiomatisation, particularly one created by an experimentalist – such as the one created by Heinrich Hertz in 1894 – we find that the author understood both the benefits and the limitations of the approach. For Hertz, casting a theory in this form was useful because it revealed logical relations between statements and concepts, displaying which statements were presupposed by others, and which concepts were more basic than others. In this way, he was able to show that the concepts of time, space, and mass were more basic than that of force, a useful result considering that the latter was the most puzzling, particularly in the case of action at a distance. He clearly saw his own work as foundational, but did not think of these foundations as immutable: the experimentalist in him realised that an axiomatic only systematises past knowledge and has nothing to say about future discoveries.31 Thus, the image of an axiomatised theory as capturing the essence of a field for all times, and in a way that is entirely separated from phenomena, is an image mostly confined to outsiders, including philosophers of science. We may conclude from this that major science in its ‘pure form’ exists only as part of the rhetoric of physics.32

If this is indeed the case, then we have a strong motivation to replace the categories ‘minor’ and ‘major’ by different phases determined by the settings of the assemblage’s parameters. To do this we must keep in mind that a scientific field is an assemblage of assemblages. As stated, the field as a whole is composed of a domain, a community, and the methods and instruments connecting one to the other, but each of these components is itself an assemblage: of phenomena, of practitioners, of machines and tools. Each assemblage in this nested set must be conceived as having its own parameters, so any claim that a deterritorialisation or a decoding has taken place must identify the level at which such an event occurred. Let’s begin with the extreme case in which the two parameters affecting the entire cognitive content of a field have high values. This yields the extreme case of an assemblage designated as a ‘stratum’. We saw in Chapter 1 that strata are the result of a process of double articulation, the first loosely sorting the raw materials into sedimentary layers, the second cementing those layers into a more enduring whole. Taking the example of classical physics, the first articulation involves interactions between practitioners (controversial or amicable), as well as interactions between practitioners and phenomena, leading to the sorting of the accumulated cognitive tools into sets with different degrees of consensus. Then the second articulation is an operation of retrospective consolidation, performed by the axiomatisation of the cognitive content.33 Without this second articulation, cognitive tools undergo only the territorialisation of consensus formation, that is, they remain an assemblage, in the original sense of the term (a decoded stratum).

This case is the easiest to analyse and comprehend, because it involves only one level. But other examples show that this need not be the case. When Lavoisier perfected his method for establishing proportions in chemical composition he had to tightly enclose the chemical reactions used as instruments of analysis, so that every reaction product could be captured and weighed. This territorialisation of one particular instrument (reactions) was coupled with a highly coded cognitive tool, the principle of conservation of weight, to perform the transformation of qualitative statements into quantitative ones. Conservation principles are one of the constants prized by major science, and are used by Kant as exemplary cases of the synthetic a priori.34 But it would be incorrect to conclude that ‘chemistry became a royal science only by virtue of a whole theoretical elaboration of the notion of weight’.35 This conclusion plays into the hands of those who see the controversies at the end of the eighteenth century as having created a new monolithic paradigm. But the increased values of the parameters did not affect the entire field. They did not affect the concept of affinity (which was used as common currency during the controversies) nor the phenomena that the concept referred to, phenomena that continued to pose problems, like the one expressed by the question ‘Why is the constancy of affinity set in variation by temperature, concentration, and other factors?’. Moreover, whatever degree of territorialisation and coding was achieved by closed chemical reactions and the new nomenclature, it affected mostly the inorganic portion of the domain, and was counteracted by the deterritorialisation it was simultaneously undergoing by the magnification of the power of analysis caused by the availability of continuous electric current, and the consequent population explosion of new elementary substances.

As all this was happening, the organic part of the domain was still a wild territory. The elaboration of the notion of weight, or more precisely, the creation of comparative methods to establish relations between weights, did help bring order to it, but this order was always partial and did not go unchallenged for long. Thus, by picking a particular acid substance as a fixed reference point and measuring the parts per weight of an alkali that exactly neutralised that acid, a standard scale of weights (referred to as ‘equivalent weights’) could be established, and the relative parts per weight could be used as combinatorial units to express proportions. This practice came to be known as ‘stoichiometry’, a practice that not only provided the units for empirical formulas like H2O, but that also allowed chemists to sharpen the borders of the organic domain by excluding any substance that did not posses a well-defined stoichiometric identity.36 But the phenomena continued to pose problems and challenge the new codings. Thus, formulas were forced to change (from empirical to rational) as chemical analysis revealed that certain clusters of components (called ‘radicals’) tended to stay together more often than other clusters, introducing variation at the level of groupings. Then isomeric substances forced another change, by showing that the nature and proportions of components were not enough to define identity: formulas now needed to be recoded, becoming structural formulas, to reflect the variation introduced by the connectivity between components. Overall, this line of development was a deterritorialisation because it pushed chemists towards the invisible reality of atoms and the bonds that assemble these into molecules. Although chemists never abandoned the molar level of macro-substances, they were forced by different phenomena (isomers, radiation, spectrographic signatures) to confront the underlying molecular reality.

If we are to follow the development of the chemical field (or any other field) rather than to interrogate its history using preconceived categories, we must track all these local movements of deterritorialisation and decoding (as well as those that result in territorialisations and codings), using the appropriate settings of the parameters at the right level of scale: this instrument; that portion of the domain; this cognitive component of personal or consensus practice; the whole field. Only then will the reified generality Science be finally vanished from a materialist philosophy.

Notes