Introduction: the ‘absolute existence’ of phlogiston

The author of this paragraph, James St John, was anything but a militant advocate of phlogiston theory. In fact, as Crosland puts it, he was ‘convinced of the superiority of the oxygen theory and the new nomenclature’,Footnote ² and in this passage of 1788 he was prefacing his own English translation of the Lavoisians’ seminal Method of Chymical Nomenclature, the work that is customarily taken to have all but concluded the takeover by the new French chemistry. Writing only one year before the publication of Lavoisier's Traité élémentaire de chimie (published in English in 1790 as Elements of Chemistry), St John grants that ‘experiments demonstrate’ that combustion and calcination processes cannot be accounted for in traditional phlogistic terms. Moreover, St John's words embed Lavoisier's operative view of elementary constituents (of composites) as the endpoint of chemical analysis: since phlogiston could be obtained from neither metals nor sulphur it was not one of their constituents. Even his definition of what phlogiston actually is – ‘the matter of fire, of flame, of light, and of heat which is liberated in combustion’ – could have just as easily been captured under Lavoisier's new term ‘caloric’. Yet St John insists on the need to ‘still allow [for] the absolute existence of a phlogiston’.

Why? What did St John find amiss in the chemistry he was presenting to the English reader that required clinging to the existence of an entity whose elimination was one of its main aims and achievements? Perhaps the most surprising, hence most telling, aspect of St John's remark was this insistence on the ‘absolute existence’ of phlogiston over and above the hypothetical properties that gave this entity its theoretical import and sustained it for the better part of the eighteenth century. Found in the preface to the Method this is all the more striking, since one of the core motivations of this work was to assure that each chemical term connoted the precise composition of the substance it denoted. This, Lavoisier boasted, meant that

The lack of such analytical precision was one of Lavoisier's main allegations against phlogiston and phlogistic chemistry. A decade earlier, in his memoir Réflexions sur le phlogistique (1777), Lavoisier famously framed phlogiston as a ‘vague principle, lacking a rigorous definition, and which is, consequently, adaptable to all explanations … It is a veritable Proteus that changes its form at each instance’.Footnote ⁴

The claim that ‘it is still phlogiston’ was not peculiar to St John or to the English-speaking parts of the world of chemistry. Antoine François Fourcroy, one of the co-authors of the Method and Lavoisier's close collaborator, expressed a similar sentiment:

Like St John, Fourcroy assumes that phlogiston had already been stripped of its traditional theoretical functions and properties, but insists on its existence, which is allegedly agreed upon by all chemists. Nor was this expression only an early hesitation by Fourcroy or a reflection of his well-known slow transition into Lavoisier's camp. In 1790 and again in 1796 he still insisted that

Even staunch supporters of the ‘anti-phlogistic theory’ – a term coined by the Anglo-Irish chemist Richard KirwanFootnote ⁷ – found it seemingly hard to depart from this beleaguered entity.

Phlogiston, of course, survived long after Lavoisier's experimentum crucis, as even Joseph Priestley referred to the calcination of metals.Footnote ⁸ Nor was it laid to rest with Lavoisier's ensuing publications, beginning with the 1775 ‘Easter memoir’ and culminating with the 1789 Traité.Footnote ⁹ ‘Phlogistians’, to use Kirwan's terminology again – those still working and reasoning within phlogistic frameworks after Lavoisier's memoir – could convincingly claim that although the nature and existence of phlogiston were in question, they were not more questionable than the nature and existence of the new theoretical entities proposed in its stead. How was phlogiston inferior to Fourcroy's ‘matter of fire or of heat’, or to St John's ‘matter of fire, of flame, of light, and of heat’? Lavoisier himself professed openly in the Traité his lack of a theoretical grasp of the nature of caloric, or that of light, as well as the relationship between the two, conceding ‘the existence of this fluid to be hypothetical’ and admitting that ‘we are not obliged to suppose [caloric] to be a real substance’.Footnote ¹⁰ In his 1792 First Principles of Chemistry, William Nicholson could therefore package old and new entities together in a sceptical remark: ‘the existence of heat, light, and phlogiston, as chemical principles of bodies, is not yet incontrovertibly established’.Footnote ¹¹ Phlogiston survived also on its own: in the 1780 Treatise on Air and Fire, Scheele reasoned exclusively within a phlogistic framework and dedicated an entry to phlogiston.Footnote ¹² Bergman's ‘Table of affinities’, published in his 1785 Treatise on Elective Affinities, comprised a column for phlogiston and a lengthy entry explored its nature and chemical functions.Footnote ¹³ In 1787 Kirwan published his much-anticipated Essay on Phlogiston, mentioned above, in which he presented his own phlogistic experiments and analysed Lavoisier's experiments in phlogistic terms. Phlogiston was also accorded an entry in the 1795 Dictionary of Chemistry of Nicholson, the English translator of the Lavoisians’ replies to Kirwan's Essay,Footnote ¹⁴ and as late as 1800 Fourcroy still enumerated the friends and foes of ‘the new doctrine of the French chemists’.Footnote ¹⁵ Priestley, among the foes, was to remain a lifelong vocal supporter of the phlogistic cause.Footnote ¹⁶

Our present interests, however, are not in the fact that the demise of phlogiston was a protracted process, contrary to the once common view of the Chemical Revolution. Nor are we interested in the relative strength of the phlogistian and anti-phlogistian arguments and especially not in the various theories of phlogiston still in play in the 1780s and beyond. It is well established that throughout the eighteenth century, and more so toward its end, there were as many such theories as there were theoreticians, thriving within different cultural and national traditions.Footnote ¹⁷

What we will study in the following is the insistence on the ‘absolute existence’ of phlogiston, particularly when the demise of the theories in which it featured became all but a foregone conclusion. This insistence on existence is conspicuous especially given the acknowledged fundamental disagreements about phlogiston's theoretical status, about its constitution, about its properties and even, as the words of Fourcroy and Nicholson above suggest, about the very meaning of ‘existence’ of a chemical substance.

Increasingly defensive, the late phlogistians rarely questioned the merits of the new theory and the experiments adduced in its support; nor did they attempt to undermine the competence of its promulgators. What their arguments reflect is rather a clear sense of loss. For phlogiston's dwindling supporters, and even for some Lavoisians like St John and Fourcroy, the new chemistry was evidently coming at too high a price. In exchange for what Kirwan dubbed a ‘false shew of simplicity’, the traditional chemist was forced to abandon ontological assumptions and epistemological mores that were essential to what comprised the science of chemistry.Footnote ¹⁸

Our paper is therefore an inquiry into what late eighteenth-century chemists reveal as their most cherished presuppositions, when the deep changes to their discipline force them to reflect on, reformulate and argue for what used to be self-evident. These presuppositions are ontological in the sense of determining what a chemical substance is, what are its ways of engaging with other substances and what are its relations to its properties. They are epistemological in the sense of defining the ways such a substance was expected to present itself to experience; in dictating the criteria for successful experiment, convincing argument and tenable theory. Yet these are not philosophical assumptions in the sense of being elaborated in philosophical tracts or reflecting a reasoned allegiance to a philosophical school.Footnote ¹⁹ Rather, the ontology and epistemology that the late defenders of phlogiston were trying to rescue were embedded in the practices of producing chemical knowledge and in the criteria by which this knowledge is assessed; they reveal themselves most prominently when these practices are endangered and when their proponents come under threat.Footnote ²⁰ This is the reason why the very ‘existence’ of phlogiston was defended, while the important theoretical functions it fulfilled were surrendered in face of the theoretical and experimental prowess of ‘the new system of chemistry’.Footnote ²¹

‘Red vapours’ versus ‘absolute facts’

It was the existence of phlogiston, rather than any of its hypothetical qualities, which was being defended. According to Nicholson, by 1792 the claim that during calcination metals unite with ‘the vital part of the air’ ceased to be a matter of debate. All that phlogistians still insisted upon was that this process also involved ‘phlogiston [being] disengaged’.Footnote ²² Fourcroy made a similar claim from the opposite perspective. ‘All chemists are of course agreed’, he exclaimed, that the ‘phlogiston of Stahl … exists’, yet the phlogistians are wrong in everything else they maintain about this entity:

Even Priestley, commonly considered the ‘last important defender of phlogiston’ or ‘the most relentless opponent of the antiphlogistic theory’, did not venture to offer a distinct concept, let alone a ‘particular theory’, of phlogiston.Footnote ²⁴ In his explicit (and very late) attempt rekindle the phlogistic debate in 1796, he willingly conceded that the ‘phlogistic theory is not without its difficulties’.Footnote ²⁵ But this theoretical frailty did not shake his conviction that phlogiston, as a substance, exists:

‘Something’ was lost, the precise nature of which Priestley no longer ventured to define. He was willing to accept the antiphlogistians’ crucial claim that metals absorb, ‘gain something’, during their calcination. What he was unwilling to eschew, however, was that they ‘likewise part with something’. Some physical, material substance must be admitted, even if its particular import in the chemical process had changed radically. Priestley went on to explain the necessity of this assumption:

The smell emanating from the metal undergoing calcination cannot and should not be ignored. It signals ‘something’ – ‘to which we give the name of phlogiston’ – before and independently of what this thing is. The undeniable sensual experience implied for Priestley an indubitable material existence. It was as important, if not more so, than the acknowledged fact that weight was gained. Priestley's reasoning was not an isolated instance. Johan Christian Wiegleb, for instance, German pharmacist and chemical educator, advanced a similar point in asking, rhetorically, ‘would it be reasonable to question the existence of [an inflammable] principle, because one cannot pick it up immediately?’Footnote ²⁸

The surprising concurrence between the likes of St John, Wiegleb and Fourcroy indicates that there was more to Priestley's point than an old man's desperate clinging to a defunct, pre-theoretical perception of what the essence of inflammability or combustibility might be.Footnote ³⁰ It suggests that Priestley's insistence that sensual experience is a definite sign of ‘something’ being released reflects a common underlying assumption concerning the nature of the chemical substance and its place and function within the chemical process. Kirwan, the most prominent spokesman of the late phlogistic camp,Footnote ³¹ reasoned along similar lines, explaining in his Essay,

The ‘red vapours’ are a sign of phlogiston. Like the ‘strong smell’ noticed by Priestley and Wiegleb, they cannot be ignored, for they signal the presence of a material entity. This presence, so vivid and unquestionable to the traditional chemist, is absent in the new chemistry. Responding to Kirwan's analysis of the marine (muriatic) acid, Berthollet ignored the ‘red vapours’ altogether, asserting that ‘if hydrogene [Kirwan's phlogiston] exists in the muriatic acid, there is no fact which shows its existence’. Once the vapours and their particular colour had been rendered ‘no fact’, Berthollet was justified in asking, ‘is not the phlogiston therefore that [Kirwan] supposes, an useless being, which has no influence in any of the phenomena we have endeavoured to explain, relating to the properties of the oxygenated muriatic acid?’Footnote ³³

Phlogiston became ‘an useless being’ precisely because ‘red vapours’ became a ‘no fact’. In the new chemistry facts were of a different order than vapours, colours and smells. The ‘explanation of what happens in calcination’, Berthollet emphasized, together with the other authors of the Method, ‘is not an hypothesis, but the result of absolute facts’, which are established in a particular and distinct fashion:

In the new chemistry, ‘absolute facts’ were ‘certain quantities’, the results of precise weighing of the outcomes of processes in carefully enclosed systems.Footnote ³⁵ The competing analyses of aqua regia (a mixture of concentrated nitric and hydrochloric acids) exhibit a similar disagreement. Advancing a causal explanation, Kirwan referred to qualities picked up by the senses, suggesting that ‘part of the nitrous acid is converted into nitrous air, which immediately unites to the undecomposed part of the nitrous acid, and forms phlogisticated nitrous acid, and hence the red colour of the liquor’.Footnote ³⁶ Berthollet's response, in contrast, consisted of a strictly quantitative analysis. ‘The part of the muriatic acid, which combines with the oxigene’, he reasoned, ‘does not remain in the liquid, but is disengaged, and maybe received in proportion as it is formed, at the expense of part of the oxigene’.Footnote ³⁷ Attention to vapours, colours and smell, Berthollet stresses, with less respect to Kirwan than usual, is not only spurious; it is positively misleading:

For Berthollet the cause of Kirwan's mistake was obvious: the great Irish chemist was confused by smell and vapours.

This is not to suggest that chemists before Lavoisier had any qualms about the chemical practice of weighing or the implementation of precise techniques of measurement. Nicholson emphasized that ‘the beginning and end of every exact chemical process consists in weighing’.Footnote ³⁹ Cavendish, who in 1784 was still convinced that ‘the commonly received principle of phlogiston explains all phenomena, at least as well as Mr Lavoisier's’,Footnote ⁴⁰ was particularly skilled in precise quantitative experimentation and ‘his measurements [were] made with “superhuman care”’.Footnote ⁴¹ Furthermore, the defenders of phlogiston acknowledged the great advance in weighing procedures introduced by the antiphlogistians. Lavoisier, Kirwan commended, is ‘a philosopher of great eminence, who was the first that introduced an almost mathematical precision into experimental philosophy’. Nicholson, however, in his introduction to Kirwan's Essay, suggested treating some of these claims to accuracy with caution:Footnote ⁴²

Nicholson's ironic attitude towards the rhetoric of precision does not reflect a rejection of ‘exact chemical … weighing’ as a crucial empirical practice. Rather, it conveys the phlogistians’ discomfort with the subjection of all chemical experimentation to weighing and with the conceptions of matter and of chemical knowledge implied by this reduction (see Figure 1).

‘The determination of weights’ was always crucial for the chemist, but it became the ultimate empirical tool only with the assumption of a homogeneous material infrastructure to all chemical phenomena. The observation that bodies gain weight in combustion was never contested. Yet this observation was a decisive argument against the existence of phlogiston – and against its theoretical and practical framework – only for those who accepted that matter is homogeneously heavy. Only with the assumption of such homogeneity does every increase in weight indicate the addition of matter and any loss of matter correspond directly to loss of weight. The assumption of homogeneity, however, was in contrast with the chemistry of someone like Priestley, whose professional pride as a chemist rested on an intimate familiarity with differences and particularities, which for him comprised the chemical realm. Substances, in the chemistry defended under the phlogistic banner, were first and foremost particular entities, and since their chemical functions were irreducible to the manifestations of a homogeneous material substratum, neither their presence nor their absence could be inferred solely from the detection of weight changes and exchanges.

Figure 1.

Weighing before Lavoisier. Note the prominence of accurate balances (on the right, against the window) alongside traditional chemical alembics, cucurbits, furnaces, distillation retorts and so on in this figure from William Lewis's Commercium Philosophico-Technicum of 1765.

For Priestley, weight computations could not comprise chemistry's ultimate analytic tool. Weight change was an important indicator, of course, but its causes were rarely transparent and it required as much skilled interpretation as any other chemical effect. Overemphasizing weight calculations, Priestley complained, the antiphlogistians forgot the basic chemical lore. ‘Because the calx of mercury derives its additional weight from dephlogisticated air’, he explained,

The continuous avowal of the existence of phlogiston, even by chemists less committed to it and more impressed than Priestley by Lavoisier's arguments, expresses the refusal to ignore smells and colours and the insistence on interpreting those as a sign of ‘something’, a chemical entity. For the chemistry they practised, every substance comprised unique properties and every chemical phenomenon was an effect of causal processes involving particular substances endowed with such unique properties. Chemical knowledge was predicated upon the study of these properties through an array of empirical procedures, consisting of various phenomenological expressions, as Bergman explained in words Priestley and Scheele would have readily accepted: ‘the knowledge of the form, taste, solubility, tendency to effloresce, and other properties … of the substances, is of great use in enabling us to judge … whether any, and what decomposition has taken place’.Footnote ⁴⁵

Like other phlogistians, Bergman was not opposed to chemical accuracy or careful mathematical representation, and he noted that ‘a more accurate measure of [attraction], which might be expressed in numbers, is as yet a desideratum’.Footnote ⁴⁶ Yet aspiring to greater accuracy in chemical practice and theory could not justify the reduction of matter to sheer bulk, of chemical phenomena to the aggregation of mass, or of the wide variety of chemical experimental practices to weighing. Echoing Priestley's lamentations, Bergman observed how ‘smell also often indicates what is taking place’, and how ‘the taste likewise often informs an experienced tongue’.Footnote ⁴⁷

Priestley's arguments concerning the interpretation of the famous experimentum crucis with the calx of mercury demonstrate how fundamental this reasoning was for the defence of phlogiston and, through it, of traditional chemistry. ‘That mercury revived either by inflammable air or in close vessels has the same properties will not be denied’, Priestley conceded, ‘and if so, it must consist of the same principles, and in the same proportion, or nearly so’. According to the traditional chemistry, as Nicholson summarized, ‘metals, like all other inflammable bodies, contain phlogiston united to a base’,Footnote ⁴⁸ so Priestley (like Scheele two decades earlier) assumed that ‘phlogiston passes … thro’ the glass when the calx is revived’.Footnote ⁴⁹ Yet Priestley was willing to assume, for the sake of argument, that the antiphlogistians were correct and that the ‘difference between the calx [of mercury] and the metal [mercury], is that the latter has parted with the air which it had imbibed’. Yet for this to be a ‘proof that metals are simple substances’, and that calcination does not involve the release of phlogiston, one had to conclude that this is true ‘in all other cases of calcination, as well as this’. But this was patently false, for ‘this is the case of only this particular calx of this metal’. Homogeneity of substances cannot be inferred from similarity in appearances, Priestley stressed, for ‘though with the same external appearance’,

What the antiphlogistians considered a cornerstone of their reinterpretation of pneumatic processes – inferred directly from rigorous computations in close accord with the principle of weight conservation – was for Priestley a mere instance of a ‘particular case’. Mercury's specific constitution, in this particular chemical reaction, allowed it to absorb air while parting with almost no phlogiston; such were the ‘proportions’ of its ‘particular principle[s]’. Even ‘different proportions’ of constituents did not necessarily entail a change in a substance's essence. After all, Priestley concluded by way of question, ‘what is the evidence of a change in the nature of any thing, but a change in its properties?’Footnote ⁵¹

The insistence on particularities as the foundation for chemical knowledge was what led Priestley to reject the Lavoisians’ interpretation of another crucial experiment: the synthesis and decomposition of water. This, the phlogistians had to admit, was a strong argument against their stance, as Kirwan remarked, concerning the synthesis of water:

Like Kirwan, Priestley did not deny that water was a compound that could be decomposed and recomposed, analysed and synthesized. He even seems to have accepted the Lavoisians’ account of what takes place in the experiment itself:

Priestley, however, could not consider this experiment as an instance of oxidation; that would equate it with other processes that according to the antiphlogistians were similar chemical instances – of the same oxidation – such as rusting. Such identification flew in the face of fundamental chemical knowledge, as Priestley understood it: ‘common rust of iron, has a very different appearance from this finery cinder, being red, and not black’. For Priestley, appearances stood for properties and properties were carried by substances (we explain below how his concept of ‘degrees of phlogistication’ fits into this close relation between substances and properties). Consequently, even Fourcroy's explanation that ‘iron is partially oxygenated’ was unacceptable to Priestley: it implied that ‘common rust’ and ‘finery cinder’Footnote ⁵⁴ were essentially the same substance, differing only in that they exhibited varying degrees of oxidation. But the different properties of rust and cinder, according to Priestley, demonstrated their essentially different ‘nature’: ‘iron in this state is saturated with some very different principle, which even excludes that which would have converted it into rust’.Footnote ⁵⁵ As late as 1800 Priestley was still complaining of ‘the experiments adduced in support of [oxidation] being not only ambiguous, or explicable on either hypothesis, but exceedingly few’.Footnote ⁵⁶ But just as his defence of phlogiston was aimed at rescuing the irreducible particularity and specificity of the chemical account of phenomena, his arguments against the explanatory power of oxidation and its empirical support were aimed at fending off the threat of reduction that the overarching concept of oxidation presented.

‘Certain quantities’ versus ‘proportions’

Lavoisier famously succeeded in portraying ‘Dr. Priestley's work’ as but ‘a train of experiments, not much interrupted by any reasoning, an assemblage of facts’.Footnote ⁵⁷ It is therefore important to stress that Priestley's underscoring of differences and particularities was not just a stubborn anti-theoretical or naive empiricist stance. His rejection of Fourcroy's analysis was based on a distinctly theoretical consideration. For Priestley, as well as for other pneumatic chemists, phlogistication and dephlogistication of pneumatic entities such as airs (for Scheele also heat and light) were processes of transmutation. Aerial entities transmuted one into another upon the loss and gain of phlogiston (as we shall explain below); they changed some of their properties but retained their common pneumatic essence. Reacting to phlogiston they changed their ‘nature’ – they became different substances. From the phlogistians’ perspective therefore, although gradual addition (or subtraction) of phlogiston to airs made sense, Fourcroy's ‘partially oxygenated’ metals did not: ‘nature’ could not be changed ‘partially’.

In the new French chemistry air and different vapours represented particular substances as much as they expressed states of matter (solid, liquid, gaseous). For Lavoisier each gas was a different chemical species whose involvement in chemical processes was as different as that of various solids and liquids. Solids could combine with solids, liquids, and airs in the same fashion as airs combined with airs. The state of the body was irrelevant in this context. This implied an elimination of the distinction between air(s) and metals that the phlogistians could not accept. Defending phlogiston against Lavoisier's assault meant, among other things, defending a concept of a chemical substance as a particular species of matter, endowed with substantial qualities irreducible to quantities of homogeneous bulk.Footnote ⁵⁸ This did not preclude thinking about chemical substances in terms of relative proportions and combinations, for ‘the same substances in different combinations, and in different states, have different properties’ and ‘substances possessed of very different properties may be composed of the same elements, in different proportions, and different modes of combination’.Footnote ⁵⁹ Indeed, the basic pneumatic phlogistic hypothesis, going all the way back to ‘the phlogiston of Stahl’, was that combustibility and inflammability of bodies are functions of the degree of phlogiston they contain:

The ‘kinds of air’, Priestley summarizes, ‘differ chiefly in the quantity of phlogiston’, but it will be a mistake to read him as suggesting a truly quantitative analysis of phlogistication. The ‘gradation’ of phlogiston did not stand for its different relative amounts but implied the level of an air's purity or impurity, namely its ability to support respiration and combustion; there was no real sense in which the ‘quantity of phlogiston’ could be isolated, measured or perceived independently of the different level of impurity it was linked with. Phlogistication or dephlogistication did not consist of an arithmetic addition or subtraction of quantities of phlogiston but entailed a transmutation, by which one kind of air turned into another, following regular and constant sequences such as dephlogisticated, common, phlogisticated and nitrous airs, according to Priestley's basic sequence. The different ‘proportion[s] … of that principle’ produced differences in pneumatic ‘kinds’. The various steps of this sequence implied various ‘proportions’ of phlogiston, yet these proportions represented the distinct properties of each of the distinct airs. As such, they were expressions of these steps as much as they were their causes.Footnote ⁶¹

To wit, although expressed in terms of quantities of phlogiston, it was a process – phlogistication – that Priestley described and analysed, a process he had interpreted as a transformation of aerial qualities. Priestley expressed this understanding in his account of the nitrous air test, designed to evaluate what he referred to as the ‘goodness’ of a given air.Footnote ⁶² After explaining how to perform the test by mixing different airs with nitrous air, then measuring the diminution in aerial bulk, Priestley added that ‘any other process by which air is diminished and made noxious answers the same purpose … In fact, it is phlogiston that is the test’.Footnote ⁶³ This status accorded to phlogiston shows why the demand to isolate it, which became a major crux philosophiae,Footnote ⁶⁴ could be rendered at times as a categorical misinterpretation of chemical knowledge, as Richard Watson claimed in 1787. Answering his own question, ‘what is phlogiston?’, Watson presented the demand to isolate it as a complete fallacy:

The difference between phlogiston and other imponderable entities, such as light and caloric, expounded by the new chemistry, was not that the latter were less hypothetical or more immediately quantifiable. In this context, Lavoisier's use of caloric may seem symmetrical to Priestley's use of phlogiston:

But whereas Lavoisier's quantities of caloric put ‘the same body’ into different states, Priestley's ‘gradations’ of phlogiston produced different aerial substances, which transmuted into one another. It is the transmutational sequence that other phlogistians like Scheele shared with Priestley, even though, as Fourcroy complained, they could not agree on the specific products:

Whereas Priestley's sequence consisted of four kinds of air (dephlogisticated, common, phlogisticated and nitrous), Scheele's sequence included empyreal air (analogous to dephlogisticated air), light, heat and inflammable air. Later on Scheele adds that ‘with a small addition of phlogiston it produces light; and with still more, the well known inflammable air’.Footnote ⁶⁸ Inflammable air, Scheele's most phlogisticated entity, was later advanced by Kirwan as phlogiston in an aerial state (or phlogiston gas).Footnote ⁶⁹ However different the sequences they stipulated, pneumatic chemists like Priestley, Scheele and Kirwan all subscribed to the notion of transmutational sequences.

Bergman, Nicholson, Priestley and their phlogistian colleagues shared Lavoisier's aspiration to quantitative accuracy. But the explanatory role Lavoisier had assigned to quantification militated against their basic concept of matter. For them, the various ‘kinds of airs’, endowed with different properties, were ipso facto different substances. They could not be merely different states, distinguished only quantitatively. Put differently, from the ontological perspective of the traditional chemistry that they were struggling to preserve, substance was a chemical entity whereas a material state was merely physical. ‘Dr. Priestley’, Nicholson observed, discovered ‘a considerable number of aerial fluids.’ These airs, he explained, reflecting on Priestley's sequence, were distinguished by degree of phlogistication: ‘common or atmospheric air’, ‘fixed air’, ‘nitrous air’. They were, nevertheless, distinct substances; particulars endowed with essential properties and not states of the same substance. According to Lavoisier, ‘air is a fluid naturally existing in a state of vapour’, but Nicholson viewed ‘vapours’ as mere ‘elastic fluids or subtle invisible matters which fly from bodies subjected to chemical operation or otherwise’. ‘Accurate chemical writers’, Nicholson stressed, ‘confine this appellation [of vapour] to such exhalations only as may be condensed into the fluid state by cold’.Footnote ⁷⁰ A substance's ephemeral existence in a particular state was evidently ‘in contradistinction to the aerial fluids, of which scarcely any are so convertible by any means in our power’.Footnote ⁷¹ Airs of varying degrees of phlogistication were no vapours:

The force of affinity and affinity as a force

This distinction between airs and vapours reveals the fundamental ontological units that people like Nicholson and Bergman were defending when they were clinging to the concept of phlogiston. ‘Condensable’ vapours were mere physical states of matter; airs, pneumatic entities charged with various ‘gradations’ of phlogiston, were chemical substances. To eschew phlogiston was not merely to accept the superiority of a new hypothesis over an old one, which most of these chemists were (grudgingly) willing to admit, but to depart from a peculiar and distinct understanding of what chemical substances were and of what constituted the relations between them.

If the fundamental chemical units of the phlogistians were substances with their irreducible properties, the fundamental relations between these substances were affinities. Eighteenth-century chemists developed a growing mastery of the intricate web of these relations and their roles in various chemical processes, which they represented in the affinity table, their emblematic ‘paper tool’.Footnote ⁷³ Phlogiston was an essential component of the affinity table from its first introduction by Geoffroy in 1718, and was still a crucial element in Bergman's 1785 version, the most extensive affinity table of his time (see Figure 2).Footnote ⁷⁴ The defence of phlogiston was thus also a defence of a particular understanding of what affinities were – again, over and above phlogiston's own declining theoretical merit. As Bergman's example demonstrates, his table reflected and resulted from a particular type of chemical order, which entailed particular types of chemical transformation:

Figure 2.

Affinities in the traditional and the new chemistry. On the left: ‘Single Elective Attractions’ from Bergman's 1785 Dissertation (phlogiston in column 36); on the right: Lavoisier's ‘table of oxygenous principle’ from Kirwan's 1789 Essay. Comparing the two tables demonstrates the visual dimension of the reduction from the ‘richness’ of phlogistic traditional chemistry to the ‘bareness’ of the French system.

The affinity table was perhaps the most conspicuous tool of traditional chemistry to be retained by the Lavoisians. Their conception of chemical affinities and the respective manner of their application, however, failed to impress Kirwan:

This may seem a harsh assessment of Lavoisier's work on affinities. Lavoisier was aware that ‘the table of affinities of the oxigenous principle with the different substance’ is far from perfect; ‘Mr. Kirwan’ he averred, ‘does not judge me with more severity than I do myself.’Footnote ⁷⁷ Yet Kirwan's complaint, as always, did not concern the antiphlogistians’ competence but the chemical principles – ontological and epistemological – underlying their reasoning and practice. Even more so, Kirwan was defending the principles the antiphlogistians rejected.

Lavoisier conceded that affinities were complex and difficult to pin down, admitting that ‘the force of affinity, which unites two principles’, for example, ‘is not the same in … two degrees of saturation’.Footnote ⁷⁸ Yet it is a ‘force’ that he is searching for; it is not a quality of any particular substance, but a unifying principle of nature, subject to mathematical laws which are simple and precise even if hard to discover: ‘affinity is a variable force, which decreases according to certain laws, not yet determined’.Footnote ⁷⁹

This idea of homogeneous force captures exactly what Kirwan and the late defenders of phlogiston could not accept in the new chemistry's perception and use of affinities. The term ‘force’ did have a natural place in the old system of affinities as well. ‘In this dissertation’, Bergman announced at the beginning of his work on affinities, ‘I shall endeavour to determine the order of attractions according to their respective force’.Footnote ⁸⁰ Moreover, in the homage to Newton with which he commenced his work, Bergman made it clear that he understood well the concept of force as a universal presence governed by simple mathematical laws.Footnote ⁸¹ For Lavoisier, the two uses of ‘force’ – the chemical and the physical – had to coincide, at least in principle; attractions and elective affinities were to be understood along the lines of Newtonian gravity: ‘two forces, both of which are variable; the first, according to a certain law dependant of temperature; and the second, according to the distance’.Footnote ⁸² Bergman, however, would never confuse this universal simplicity of gravity governed by mathematical law with the complexity of the attraction of affinity. For Lavoisier, ‘force’ was a physical, unifying principle, common to all matter. For Bergman, attractions were particular properties of substances; force was but a second-order property, the measure of their difference.

What the insistence on phlogiston represented, then, was resistance neither to the replacement of one hypothesis with another, nor to the introduction of ‘mathematical precision into experimental philosophy’, as Kirwan framed Lavoisier's contributions. Kirwan and Bergman were defending the autonomy of chemistry from physics.Footnote ⁸³ What they refused to accept was the idea that the unifying principle of nature consisted, in essence, of a simple material infrastructure governed by simple mathematical laws. What they tried to maintain in their concept of affinity was regularity of cause and effect, rooted in the qualities of individual substances, which were, at the same time, wholly embedded in continuous chemical processes: ‘By chemical operations or processes, we mean the application of the proper means to affect the decomposition or composition. Every one of them is grounded on the various degrees of affinities of heterogeneous substances amongst each other.’Footnote ⁸⁴ Affinities, a chemist like Friedrich Gren would still assert in 1800,Footnote ⁸⁵ could not be perceived outside ‘chemical operations or processes’. Their ‘various degrees’, moreover, did not follow independent mathematical laws but the relations ‘amongst heterogeneous substances’. Such substances, Bergman explained, were ‘heterogeneous’ by virtue of possessing different qualities:

Lavoisier's interpretation of the reasons for the complexity of affinities is revealing. Justifying his own ‘table of affinities’, in reaction to Kirwan's critique, he argued,

And in order ‘to form accurate ideas respecting these phenomena’, Lavoisier added,

There was no place for the sharp distinction between ‘heterogeneous’ and ‘homogeneous’ substances in Lavoisier's system. Since all matter was homogeneous and consequently homogeneously heavy, all unions had ultimately to abide by the rules of ‘aggregation’. Both Bergman and Lavoisier, then, had recourse to the concept of aggregated matter. Yet even the difference in their particular phrasings is instructive. Bergman paired ‘aggregation’ with ‘attraction’, distinguishing it from ‘attraction of composition’, as the other type of possible combination: the chemical. This distinction was altogether ignored by Lavoisier, who intentionally blurred the division between the chemical and the physical, referring to the ‘affinity of aggregation’ as the only existing type of combination. For him, only one kind of chemical ‘affinity’ existed – that of ‘aggregation’ – which was physical in the sense that it was universal to ‘all the bodies in nature’ and abided by the (Newtonian) laws of ‘mutual attraction’. This was how the chemical complexity of affinities and their particularities, as entailed by their relational character, were replaced by ‘accurate ideas’. This accuracy hardly served to settle Kirwan's worries. He complained that the ‘simplicity’ of the ‘antiphlogistic hypothesis … though seducing in some cases, becomes insufficiency in many others. I pass over many other embarrassing objections’ (see below).Footnote ⁸⁹

Bergman's table is at once a tour de force of chemical knowledge and an admission of its limits. It is an attempt at the highest level of formalization attainable, given the limited regularity of chemical processes and their irreducible complexity. In fact, this complexity is the single most distinctive characteristic of affinities. Justifying ‘the necessity for a new Table of Attractions’, Bergman wryly admitted,

Uniformity versus simplicity

From the point of view of the late defenders of phlogiston, therefore, phlogiston and oxygen were not symmetrical. To say that all inflammables contain phlogiston was to suggest a unifying hypothesis. To say that they all have affinity to oxygen was only to repeat that they are inflammable. Moreover, the Lavoisians used the term ‘affinity’ in a way traditional chemists could not accept. Affinity as understood by the latter was not a universal, physical force but an expression of the interrelation between various substances. Affinity tables described patterns of such relations as displayed by the elective chemical behaviours of recognizable substances.

Yet it is not because affinities were lost on the new chemistry that the adherents of ‘the old system’ complained.Footnote ⁹¹ Quite the contrary: Kirwan, still in his phlogistic mode,Footnote ⁹² had to admit that some accounts by chemical affinities are much more suited to ‘the antiphlogistic hypothesis’. For example, ‘why iron precipitates copper from the vitriolic acid, may be explained in the antiphlogistic hypothesis, since iron is said to have a greater affinity to the oxigenous principle than copper has, and also to take up more of it’.Footnote ⁹³ Yet there is a fundamental tenet to Lavoisier's analysis by affinities that Kirwan could not permit. In one of his responses to Kirwan's Essay, Lavoisier claimed to have

Unlike Priestley's ‘quantities of phlogiston’, Lavoisier's ‘quantities of oxigene’ are gauged strictly by weighing; characteristically, Lavoisier applies here a quantitative analysis to a distinctly chemical pattern: the recovery of substances by precipitation. Kirwan found this type of analysis – the attempt to infer the chemical (substances, nature of reactions, etc.) directly from the physical (ultimately represented by quantities) – severely lacking, ignoring many of the crucial phenomena to be accounted for:

The difference in the understanding of the explanatory role and ontological status of affinity highlights the dissatisfaction with the new chemistry's account of combustion and calcination processes which clinging to phlogiston expressed, a dissatisfaction emerging from a fundamental conviction of what comprises a natural law and its use in chemical explanations. This difference is best perceived by a closer look at the use of the term ‘force’ in the context of affinities.

Lavoisier's Newtonian allusion in this context was not haphazard. Bergman was using attractions to mean both chemical affinities and physical forces, but was always careful to distinguish between the chemical ‘attractions of composition’ and the physical ‘attractions of aggregation’ which were expressions of Newtonian universal gravitation. Lavoisier, as we have seen above, used both notions interchangeably, blurring the difference, crucial to traditional chemists, between distinctly chemical properties and the universal physical force:

Interpreting affinity as an underlying force, identical in all the substances exhibiting it, Lavoisier could utilize it to replace phlogiston as the general principle of inflammability: where inflammable bodies were the ones containing phlogiston, they were now those possessed of a high affinity to oxygen. ‘Several conditions are requisite to enable a body to become oxygenated’, Lavoisier explained, ‘and primarily, that the particles of the body … shall have less reciprocal attraction with each other than they have for the Oxygen’.Footnote ⁹⁷ This, from the perspective of traditional chemistry, was a hopelessly simplistic and reductive account. Even one type of combustion, like calcination, was a complex process, differing from one substance to another, as Kirwan tirelessly described:

It is precisely this emphasis on ‘differences’, the emphasis on the intrinsic complexity of nature, that the antiphlogistians challenged. ‘We may justly admire the simplicity of the means employed by nature to multiply qualities and forms’, declared Lavoisier,

Nature, according to the new chemistry, was in essence simple, and its diversity was hence produced by equally simple means; chemistry should follow suit. ‘The method we have adopted’, declared Lavoisier, ‘following nature in the simplicity of her operations, gives natural and easy nomenclature applicable to every possible neutral salt’.Footnote ¹⁰⁰

The traditional chemists never denied this feature of the new theory. Kirwan, for instance, readily admitted that it was ‘recommendable by its simplicity’. Yet it was a ‘false shew of simplicity’.Footnote ¹⁰¹ ‘The more we succeed in simplifying the principles of bodies, the more difficult it is to determine truly what passes in chemical operations’, was the way Nicholson formulated a similar concern.Footnote ¹⁰² The conclusion from the irreducible complexity of chemical processes, however, was not a call for phenomenological scepticism (or for self-imposed modesty of an art versus a science). Kirwan's complaint was not that the antiphlogistians tried to explain too much but that they explained too little. The ‘simplicity’ of their ‘doctrine’ was ‘insufficiency’ not because it offered more order than nature allowed, but because, as he pointed out in the preface to his Essay, ‘the ancient doctrine [is] the more uniform of the two’.Footnote ¹⁰³ It is more uniform, Bergman explained, because it pays full attention to the richness of causal relations that bring about natural phenomena:

The chemical subject matter is essentially complex, hence chemical knowledge cannot be reductive. Chemistry requires immediate recognition of the different substances in their different states and stages. Not a ‘false shew of simplicity’ but a close acquaintance with substances that allows the discernment of uniformities and regularities within the complexity. This was the kind of interpretation of principles and properties that the defence of phlogiston strived to preserve. Priestley, the last major chemist to engage actively in this defence, was perhaps the most eloquent in characterizing the type of uniformity that traditional chemistry offered, and how it was phlogiston that embodied the unique balance between uniformity and diversity, between ‘differences’ and ‘known laws’, which formed the foundation of chemical knowledge:

Just as the emanating ‘smell’ signalled the indubitable presence of ‘something’ – a substantial chemical agent of sorts, for which ‘the term phlogiston’ was as fitting as any other – so the very same ‘thing’, Priestley claimed, cannot be ‘a mere quality, or mode of bodies’. Its presence and nature are inferred from relations between qualities or chemical ‘circumstance[s]’, which derive from experiments and experiences. It must be a ‘substance’, and as such an integral part of the material realm, the ultimate subject of chemical knowledge.