Newton and scholastic pedagogy: the Trinity Notebook

Newton's most obvious engagement with scholastic philosophy occurred during his years as a BA student at Trinity College, Cambridge, for which the key evidence is the so-called Trinity Notebook, containing notes taken c.1661–1664. The document is well known: it begins with passages in Greek from Aristotle's Organon and Porphyry's Isagoge, and goes on with notes from two near-contemporary scholastic textbooks: Johannes Magirus's Physiologiae peripateticae libri sex (Cambridge, 1642, 1st edn 1600) and Daniel Stahl's Axiomata philosophica (Cambridge, 1645). To what extent do these notes – which were not published in the modern edition of the notebook, and so must still be consulted in the original manuscript in Cambridge University LibraryFootnote ⁵ – reveal much about Newton's debt to scholastic natural philosophy?

In the 1980s, William Wallace – who has examined the notes more fully than anyone else – argued that they reveal a great deal on this score, to the extent that Newton's debt to the scholastic tradition was comparable to the much more obvious debts incurred by Galileo. To take only a few examples, he speculated that Magirus's positing of a vis insita in the stars ‘could not but have had an influence on the young Newton’ and even that some of the notes from Stahl pre-empt Newton's discovery of universal gravitation.Footnote ⁶

Such claims do not convince. Newton's notes from Magirus and Stahl show little independent reflection, and we may safely agree with Jed Buchwald and Mordechai Feingold's recent comment that ‘Newton's absorption of scholastic material bears every sign of having been superficial’.Footnote ⁷ To this we might add some comparative contextual evidence. The two scholastic textbooks from which Newton took the most notes (as far as we know, barring further manuscript discoveries) were, relatively speaking, quite primitive compared to some of the textbooks being read by Newton's close contemporaries. Take, for example, Francis Willughby, who like Newton was at Trinity, albeit slightly earlier, between 1652 and 1660, and who would also become an important member of the Royal Society. During his university days Willughby read some rather more advanced scholastic texts, such as Marcus Wendelin's Contemplationes Physicae (Cambridge, 1648, 1st edn 1625).Footnote ⁸ Indeed, there is considerable evidence of Wendelin's work being used by Cambridge students: one surviving copy, for example, belonged to John Patrick, having been used by him in 1650, either at the very end of his studies for a BA, or at the very start of those for an MA.Footnote ⁹

If these students were reading texts like Wendelin's much more advanced textbook at quite late stages of their student careers, Newton's case seems very different: as far as the evidence we have goes, he was simply taking basic notes from much more introductory texts. That is not to say that the notes are of no significance at all for his mature writings: as we shall see below, he would later attack scholastic concepts that he could have encountered in both Magirus and Stahl. But even those attacks will be revealed to have been haphazard and commonplace.

Analysis, synthesis and the regressus tradition

A much more subtle, significant and philosophically precise argument for the putative influence of neo-scholastic philosophy on Newton's mature thought than that of Wallace has recently been made in the pioneering study by Steffen Ducheyne. He argues that Newton's natural-philosophical methodology – specifically his self-proclaimed use of the method of analysis and synthesis – was deeply indebted to the scholastic tradition of regressus, developed by Aristotelians in sixteenth-century Padua, most famously by Jacopo Zabarella, as that tradition was manifested in some seventeenth-century English philosophy textbooks. Most broadly, this was a tradition of thinking about the nature of demonstration in which one first derives an inference from an observed effect to its cause (demonstratio quia), and then reverses the procedure, and infers from the cause to the effect (demonstratio propter quid), leading to truly ‘scientific’ (that is to say, causal) knowledge of the effect.Footnote ¹⁰ If Ducheyne is correct, it would have a very major impact indeed not only on our understanding of Newton's intellectual formation, but also on the importance of the regressus tradition to early modern naturalphilosophical change more generally. It thus seems worth treating the issue at some length.

From around the 1690s onwards, Newton tended to describe his natural-philosophical methodology as proceeding by a dual process of analysis and synthesis. A particularly famous summative quotation can be found in Query 31 of the Opticks:

Newton here, as elsewhere,Footnote ¹² states that as in mathematics, so in natural philosophy, investigation should proceed first by analysis, which consists of making experiments and observations, generating general conclusions by induction; in synthesis (or composition), those general conclusions are taken as principles and new phenomena are derived and proved by them.

The long-standing view in the literature has been that Newton was here applying the terminology of classical and contemporary mathematics to natural philosophy more generally.Footnote ¹³ But Ducheyne seeks to refute this, through an argument which – while rich and complex – can be fairly summarized as consisting of three main parts:

1. 1 Newton's version of analysis/synthesis was incompatible with the mathematical tradition, primarily because ‘Newton's conception of natural philosophy is explicitly causal … a feature that is obviously lacking in the mathematical tradition of analysis–synthesis’.Footnote ¹⁴

2. 2 Newton thus drew on Aristotelian scholastic regressus theory: ‘Newton conceptualised natural-philosophical analysis and synthesis in line with the tenets of the Aristotelian textbook tradition on the matter’.Footnote ¹⁵

3. 3 A specific text – Samuel Smith's Aditus ad logicam (1st edn 1613) – ‘can quite safely be considered an actual source for Newton's ideas on natural philosophical analysis and synthesis’.Footnote ¹⁶

For reasons that will shortly become clear, I would like to proceed through these three parts in reverse order. We begin with the final claim: that Newton relied on Samuel Smith's Aditus ad logicam as the key source for his regressus arguments. Newton owned a copy of the sixth edition of 1649, and it survives among his books in Trinity College, Cambridge: Ducheyne claims that ‘there is serious evidence that Newton thoroughly studied this work’,Footnote ¹⁷ leading him to the very assertive conclusion quoted above. But as always with evidence of early modern reading, we have to be extremely careful; on this occasion, slightly more caution might have yielded a different conclusion.

For example, Ducheyne asserts that the book ‘shows numerous signs of dog-earring’ (Newton's favoured method of bookmarking), and lists forty-four pages where he claims to discover such signs.Footnote ¹⁸ But the book in fact contains no dog-earing at all, only the typical signs of wear and tear that one finds on almost any early modern book, as is clear from a comparison of the book with another that Newton certainly did later dog-ear (Figure 1). Indeed, the most reliable evidence points to the fact that Newton only started folding pages to specific passages somewhat later than the 1660s.Footnote ¹⁹

Figure 1.

Left: Newton's copy of Smith's Aditus ad logicam (Trinity, NQ 9.166), showing p. 12, one of the supposedly dog-eared pages (according to Ducheyne). Right: Newton's actual practice from much later, in his copy of G.J. Vossius's Theologia gentilis (NQ 8.46), p. 35.

Some of Ducheyne's other claims also fail to stand up to detailed scrutiny. He claims, for example, that the drawing of the ‘Arbor Porphyriana’ – that is to say, the Tree of Porphyry traditionally used to classify the Aristotelian categories – that we find at the beginning of the Trinity Notebook is ‘based on Samuel Smith's division’, as found in the version of the ‘Arbor’ in the Aditus ad logicam. But if we compare carefully Newton's tree with that in Smith, we find that they are in fact quite different. To take only one example, Newton's contains divisions for simple and mixed non-living bodies, whereas Smith's does not (Figure 2). But such divisions are present in the version of the ‘Arbor’ in other contemporary pedagogical texts, such as Melchior Cornaeus's Curriculum philosophiae Peripateticae (1657), among many others.Footnote ²⁰ This is not to argue that Cornaeus was Newton's source, but simply to point out that it seems much more likely that something else like his work (rather than Smith's) was a direct source for Newton.

Figure 2.

Left: Smith's ‘Arbor Porphyriana’ (between pp. 32 and 33 in the Aditus). Right: A close-up of the Arbor from Newton's Trinity notebook (CUL MS Add. 3996, fol. 2r). Note the differences. But Newton's divisions were present in other near-contemporary versions of the Arbor.

The reader may consider these bibliographical points to be relatively minor. But another deserves our serious attention. Ducheyne claims that the markings and notes on the Trinity copy of Smith's textbook – which are indeed clustered around the chapter on demonstratio that contains the discussion of regressus – are Newtonian. But here we encounter a major stumbling block, for it is quite clear that the handwriting is not Newton's (Figure 3).Footnote ²¹ In short, we must conclude that there is no evidence that Newton was particularly inspired by Smith's book, or even that he read it with any care, if he read it at all.Footnote ²²

Figure 3.

Left: Newton's handwriting in the 1660s (CUL MS Add. 3996, fol. 16r). Right: two examples of the very different handwriting in NQ 9.166.

We see, then, that there is no bibliographical, material or intellectual biographical evidence for Newton's deep engagement with the regressus methodology of textbook scholasticism. This leaves us with Ducheyne's first two arguments, which are purely philosophical–intellectual in nature. His second point is that Newton must have been drawing on Aristotelian regressus because his methodological position, like that of the Paduans and their followers, was a causal one. As he summarizes, ‘Newton conceptualised natural-philosophical analysis and synthesis in line with the tenets of the “Aristotelian” textbook tradition on the matter’.Footnote ²³ But if we compare Newton's words with those of the regressus theorists we find far more differences than similarities. First of all, where Newton foregrounds the language of analysis and synthesis, the Aristotelian regressus theorists placed emphasis not on this (Smith, for example, does not use the terms at all) but on the classically Aristotelian language of explanation τοῦ ὅτι and τοῦ διότι, generating demonstratio propter quod and demonstratio propter quid in Latin. Even more importantly, all this is part of the process of constructing a syllogism. If we take Smith as a typical example, at the beginning of the chapter on demonstratio that contains his discussion of regressus, he summarizes that logical demonstration involves forming a syllogism, beginning with true premises and deducing necessarily true conclusions.Footnote ²⁴ Newton, of course, never says anything like this: he never talks about syllogisms or about these different types of demonstration. Indeed, in the quotation from Query 31 that we met earlier, Newton explicitly says that his method ‘be no Demonstration’, whereas the whole point of regressus was that it offered demonstratio – for example, this is the chapter title of the section in which Smith's discussion is to be found. All this is not to say that it was impossible for natural philosophers to adopt regressus theory while dropping the syllogistic aspect – a good example is William Harvey, in the methodological preface to his Exercitationes de generatione animalium (1651).Footnote ²⁵ But even he (imbued with the Paduan methodological ideas during his time there) retained the references to Aristotle, the language of τοῦ ὅτι and τοῦ διότι demonstration, and the whole complex mechanism of mental abstraction that the Aristotelians believed this to involve. Newton simply never does any of this.

To this it might be replied that Newton's vision of analysis and synthesis has to be derived from the Aristotelian textbook tradition because it simply is not compatible with the other suggested source: the mathematical tradition as we understand it (this is Ducheyne's first point).Footnote ²⁶ This may be true – Newton certainly does not use the terms in precisely the sense they were used by Pappus (who, it should be said, is very ambiguous on the issue) and others whom he claimed to be emulating (see below) – but it is surely beside the point, precisely because Newton did not have our understanding of ancient mathematics, but an early modern one. As is now well established, there had been, since the mid-sixteenth century, a concerted attempt to increase the status of mathematics by presenting it as a causal scientia both equal in status to natural philosophy and capable of offering explanations within natural philosophy. This attempt had been made first by a group of Jesuits like Christopher Clavius, then by Galileo, then by many others, including in England Thomas Hobbes and Newton's mentor Isaac Barrow.Footnote ²⁷ As Alan Shapiro, Mary Domski and Niccolò Guicciardini (among others) have shown, Newton almost certainly inherited a stake in this debate from Barrow, his early mentor at Trinity.Footnote ²⁸ Guicciardini above all has charted how Newton's 1690s mathematical manuscripts built on the preface to the Principia, which presented the genre of the work as ‘rational mechanics’ supposedly in the vein of ancient geometers like Pappus, and that it was in these manuscripts that Newton developed his vision of natural philosophy being founded on the ancient method of analysis and synthesis.Footnote ²⁹ As far as Newton was concerned, mathematics was directly connected to physical reality, and ‘mathematical representations should be closely aligned with the properties of physical bodies and their motions’,Footnote ³⁰ geometrical objects having been generated mechanically, as they were for Barrow too.

This much is well established. But it still leaves us with an important unresolved question. Barrow, after all, did not use the language of analysis or synthesis, and his history of mathematics was very different from Newton's, focusing not on Pappus and the putative tradition of rational mechanics, but rather on Aristotle and some other classical Greek mathematicians and natural philosophers. Why, then, did Newton turn away from Barrow's formulation to the language of analysis and synthesis and his own historical justification for it? Here I offer an explanation that no previous scholar (as far as I know) has broached: that while Newton would have agreed with the broad principles behind Barrow's vision of the history of mathematics, he may well have been deeply unconvinced by the specifics of Barrow's history, and that Newton's consequent shift away from those specifics explains his turn to the language of analysis and synthesis.

In his Lucasian lectures, Barrow had argued, ‘Mathematics is co-extended with physics’, and the division of mathematics into pure and mixed was an artificial historical development introduced first by Aristotle and then by Geminus, when in reality all mathematics always abstracted from physical bodies.Footnote ³¹ While this is structurally very similar to Newton's argument about mechanics in the preface to the Principia and the 1690s manuscripts, had Newton investigated Barrow's actual history in any detail – as he is likely to have done – he may well have found it very unconvincing.

Take, for example, Barrow's argument that all the ancients believed that mathematics offered certainty in natural philosophy. To justify this, Barrow claimed that Aristotle had inherited this belief from his predecessors, citing passages from De caelo and the Metaphysics. But in reality Aristotle had in those loci argued something completely different, insisting only on the accuracy of mathematics. Indeed, he quite clearly opined, ‘The minute accuracy of mathematics is not to be demanded in all cases, but only in the case of things which have no matter’:Footnote ³² exactly the opposite of what Barrow was claiming for him.

Such misreadings are even more apparent in Barrow's treatment of number theory. Barrow believed so strongly in the pre-eminence of geometry as the science of real magnitude that he even subsumed arithmetic under geometry and claimed that this had been the case in antiquity. This claim caused immediate headaches, for all the ancient mathematicians had quite clearly differentiated arithmetic from geometry: most importantly both Aristotle and Euclid supported a strict division between number and magnitude, a division reflected in the separation of the geometric and arithmetic material in Euclid's Elements.Footnote ³³ Barrow was aware of this problem, and sought to alleviate it by manipulating some of the evidence. Instead of citing the mathematical tradition as it was commonly understood, Barrow cited a very ambiguous fragment from the obscure Pythagorean Archytas, which stated that arithmetic was only the ‘sister’ to geometry. Yet had Barrow been honest, he would have admitted that the same source which contained this fragment – Nicomachus’ Introduction to Arithmetic – also stated that Archytas believed that ‘arithmetic seems to be far superior to the other arts’, geometry in particular.Footnote ³⁴ Barrow was effectively twisting the truth on its head, and this kind of scholarship is impressive neither by modern nor by early modern standards: certainly, it was an idiosyncratic reading in the seventeenth century.Footnote ³⁵

In short, Newton had very good reasons to think that he needed a rather different history of mathematics to that offered by Barrow, even if they were making broadly the same methodological point. Moreover, Newton also had very good reasons to settle on the language of analysis and synthesis, because plenty of others before him had already used it to make precisely the same point. We might here quote only one author whose works Newton certainly read as a student,Footnote ³⁶ Walter Charleton:

Like Newton, Charleton believed that the ancient mathematicians possessed an esoteric method of analysis but that they generally used synthesis in the writings;Footnote ³⁸ like Newton, he also believed that the analysis–synthesis method could be ‘accommodated’ to the study of a natural-philosophical subject: the soul.Footnote ³⁹

Charleton, in turn, may well have been influenced by his acquaintance Thomas Hobbes. On the surface, there are very strong similarities between Newton and Hobbes: not just the use of analysis–synthesis, but also the mathematical physics it is used to defend, including a dislike for the unnecessary abstractions supposedly introduced into analysis by the modern algebra (of course, very ironic in Newton's case), a subsequent belief that analytic geometry added little to ancient methods, and the belief that arithmetic depended on geometry.Footnote ⁴⁰ As with Newton, there have been some ambitious attempts to connect Hobbes to the Paduan regressus theorists.Footnote ⁴¹ And just as with Newton, these do not stand up to detailed scrutiny.Footnote ⁴² As Noel Malcolm has summarized,

Since the initial reception of the Greek mathematical writings in the sixteenth century was partially shaped ‘through their real or perceived compatibility with the Aristotelian ideal of science’,Footnote ⁴⁴ it is not surprising that the terms were used rather promiscuously. But in the absence of direct evidence for Newton's debt to the regressus tradition – an absence that we have confirmed – the safest option seems to be simply to believe what Newton himself said: that he saw himself to be operating in a mathematical tradition stemming from Pappus.

There is, of course, a larger philosophical significance to all this. Ducheyne's attempt to find a scholastic Aristotelian lineage for Newton's methodology is part of a broader project to present Newton as a firm causal realist, in opposition to the phenomenalist reading of Newtonian physical explanation. Relying in particular on the previously quoted passage from Query 31 to the Opticks, he claims that when Newton spoke of ‘forces’ in the Principia he was really speaking of ‘proximate causes’ as the concept had been used by Smith (i.e. that inverse-square centripetal forces are the proximate causes of universal gravitation). With Smith – and thus with the concept of ‘proximate causes’ (never used by Newton in printFootnote ⁴⁵) – we have already dispensed. We can consequently more generally question the causal reading of Newton on force. The language of analysis and synthesis that Newton began using in the 1690s manuscripts related most directly to the few methodological parts of the Principia: the preface – where we will remember that Newton links his methodology to the putative rational mechanics of the ancient geometers – and the Scholium at the end of Book I, Section 11. Nowhere here does Newton intimate that ‘he had unveiled gravity as primary or proximate cause for the heavenly and terrestrial motions'.Footnote ⁴⁶ Rather, he informs us, he had ‘found propositions’; that he had not ‘inquir[ed] into the causes or species of the force, but consider[ed] only their quantities, directions, and effects’; that he had ‘not yet learned the cause of gravity from the phenomena’; that he had ‘discover[ed] the forces of nature from the phenomena of motions’; and that he spoke of attraction and impulse ‘considering … not the species of forces and their physical qualities but their quantities and mathematical proportions’.Footnote ⁴⁷ Analysis, Newton confirmed in a draft preface to the first edition of the Opticks composed c.1703–1704, leads only to ‘the general properties of things’, which one then only ‘assum[es]’ as causes during the process of synthesis.Footnote ⁴⁸

Indeed, Newton only really emphasized the language of ‘causes’ in the passage from Query 31, and in the many manuscript drafts for it which are repeatedly quoted by Ducheyne, because at this point he was also making the rhetorical point that natural philosophy leads to knowledge of the operation of God, the ‘first cause’.Footnote ⁴⁹ Even nearby in the Opticks, Newton, when not required to make these rhetorical concessions to theological language, clearly differentiated between gravity as a ‘principle’ or ‘law of nature’ whose ‘Truth appear[ed] to us by Phaenomena’, and its ‘Causes’, which are ‘not yet discover'd’.Footnote ⁵⁰ By the process of analysis he had discovered ‘properties’, but not causes.Footnote ⁵¹ And it was for this, after all, that he was chastised again and again by Leibniz and by others, who claimed that Newton's explanation of gravity was not philosophical at all. Had Newton been influenced by Aristotelian scholastic regressus theory, he would surely have simply made this clear in his responses. Instead, he continued to insist on the new, mathematical nature of his natural philosophy: after all, it was precisely this new type of explanation that his ‘mathematical principles of natural philosophy’ entailed.

Indeed, Newton seems to have been concerned from an early stage in his natural-philosophical career to avoid – or at least treat with extreme caution – the language of causes. In his earliest defences of his theory of colour from the early 1670s, Newton was already using the rather idiosyncratic concept of ‘properties’ that we have just encountered him using on several occasions in his post-Principia phase. His discovery, he announced, ‘consists only in some properties of light’, a claim that he repeated again and again.Footnote ⁵² Even in the drafts to the Opticks passage on which Ducheyne relies, we see Newton agonizingly equivocating over whether to use ‘properties/principles’ rather than ‘causes’.Footnote ⁵³

Given all the evidence presented above, we can say with some certainty that Newton was not a causal realist in the Aristotelian sense, and – with much more caution – we might also suggest that the phenomenalist reading of Newtonian method is not so misguided. It should be made clear that there can be no doubt that Newton was rather indiscriminate in his terminology. But equally there can be no doubt that he did not think of the immediate results of his mathematical physics as causes (proximate or otherwise) in an Aristotelian scholastic sense.Footnote ⁵⁴ As much is clear from the famous part of the General Scholium where Newton announced that although he had not found the cause of gravity, he had still shown that it exists – or more accurately that ‘these properties of gravity’ do.Footnote ⁵⁵ As we shall have further cause to point out, Newtonian methodological statements were often geared more towards polemic than towards internal precision or consistency. As the manuscript evidence shows, the polemical force of the passages on analysis and synthesis was directed more towards the debate on the certainty of mathematics than towards the status of causation (although Newton would surely have also believed that his physicalist conception of mathematics mitigated the ‘non-philosophical’ implications of this). Piercing through the polemic to the reality of Newton's practice is not simple (and often impossible), but the one thing that seems certain is that there is, for the time being, no hard bibliographical or biographical evidence of Newton engaging deeply with the scholastic Aristotelian regressus tradition.

The rejection of metaphysics

The foregoing argument has been a negative one: Newton did not use Aristotelian scholastic regressus, he did not hold to an Aristotelian scholastic conception of ‘cause’, and the claim that he did depends on a bibliographical and palaeographical misreading. The following argument is also negative, but in a different sense, for it will be shown that famous passages in the General Scholium and elsewhere that have been adduced to claim that Newton possessed a sophisticated ‘metaphysics’ were in fact designed not to make any positive claims about God's relationship with the world (or any positive claims at all), but – unbeknown to the many modern scholars who have attempted to analyse them – to attack very specific aspects of scholastic philosophy.

How we have reached the stage where so many scholars now assume that Newton had an elaborate metaphysics is unclear.Footnote ⁵⁶ Their interpretations vary, but what they tend to mean by this is that Newton held some kind of non-empirical vision of God's presence and action in the world that shaped the rest of his natural philosophy, and perhaps explains such issues as physical causation, because God may in one way or another act as the medium through which matter interacts. But as I have shown elsewhere, Newton quite explicitly rejected the discipline of ‘metaphysics’, considering it to lie at the origins of both theological and natural-philosophical error; because of this rejection, he was deliberately agnostic about the precise mechanisms of God's operation in the world (to step beyond that agnosticism would be to descend into the dreaded metaphysics).Footnote ⁵⁷ Here we might quote only one of the myriad of passages in which Newton rejected the discipline:

Newton could hardly be clearer that he had no time for metaphysics as a discipline.

In opposition to such a reading, it might be claimed that these were just throwaway polemical commonplaces (for example, the anti-Cartesian implications of the above passage are clear), and that in reality Newton did actively practise metaphysics. Before we turn to a rebuttal of such a claim, we need to ask what the concept actually meant to Newton and his contemporaries. Here the scholastic context is crucial, for contrary to Howard Stein's much-cited article on ‘Newton's metaphysics’, we cannot simply use the definition of ‘metaphysics’ from the 1911 edition of the Encyclopaedia Brittanica, but we must identify what early modern philosophers themselves meant by the term.Footnote ⁵⁹

Aristotle had been famously ambiguous about what metaphysics constituted. Sometimes he equated it with ‘first philosophy’, understood as the study of being qua being or of universal principles, but at other times he equated it with ‘theology’ – the Greek θεολογικα – understood as a science of eternal causes and immaterial entities.Footnote ⁶⁰ This ambiguity generated much dispute in the medieval and Renaissance Aristotelian tradition, and continued through to seventeenth-century philosophical pedagogy. To offer a very rough taxonomy (which makes allowances for many important exceptions), the Scotist tradition in which figures like Francisco Suárez operated tended to see metaphysics as the science of being in general, a move that was also followed by notable Lutheran metaphysicians such as Christoph Scheibler, whose Metaphysica (1617) became probably the most popular metaphysics textbook in seventeenth-century England.Footnote ⁶¹ On the other hand, a line of thinkers stemming primarily from one of Suárez's Jesuit counterparts, Benito Pereira, postulated that ‘metaphysics proper’ had as its subject only God and immaterial being, and thus equated it with theologia naturalis. This definition was taken up by some Calvinist metaphysicians, such as Rudolph Goclenius, and in England it was embraced perhaps most passionately by none other than Henry More, who spent the first three chapters of his Enchiridion metaphysicum (1671) arguing that metaphysics was not a philosophia prima or a science of being qua being, but directly equivalent to theologia naturalis as the study of immaterial substances.Footnote ⁶² Descartes, as always, wanted to have his cake and eat it: as the full title of his Meditations (1641) made clear, his ‘metaphysics’ both was primarily concerned with immaterial substance, and was also a philosophia prima – as we know, it was indeed a propaedeutic to the rest of his philosophy.Footnote ⁶³

Newton made it so abundantly clear that he was not doing ‘metaphysics’ in a Cartesian sense that we need not consider the possibility at much length.Footnote ⁶⁴ But what about the possibility that Newton had a ‘metaphysical’ vision of how God interacted with the world, and that his metaphysics was thus broadly equivalent to natural theology? (It could still thus function as a propaedeutic to the rest of his philosophy.) Once again, Newton said quite clearly that he was not doing this, and that it was the attempts to do this, and specifically to explain the nature of God's omnipresence, that had led both to idolatry and to natural-philosophical error. Instead, he was deliberately agnostic about precisely how God was omnipresent in the world. This is clear enough from this draft to the Queries to the Opticks:

As we see, according to Newton the rejection of metaphysics means that we simply cannot know how several natural phenomena occur, including the interaction between soul and body and the action at a distance that causes gravitational attraction. From phenomena the world may appear ‘alive’, but to speculate further and to try to explain this would be to return to the dreaded and misleading discipline of metaphysics. Newton, on his own terms, was not a metaphysician, but deliberately and self-consciously agnostic about issues of the sort he considered metaphysical.Footnote ⁶⁶

If the above reading is correct, however, it might be thought to leave us with a major dilemma. For there are several passages, much adduced and analysed in the secondary literature, in which Newton seems to offer some much more positive and non-agnostic statements about the relationship of God and the world. The first is from the General Scholium: ‘God is omnipresent not only virtually but also substantially’.Footnote ⁶⁷ (A better translation, the philosophical significance of which will appear imminently, is ‘God is omnipresent not only in power [virtus] but also in substance’). The second is from the much-discussed ‘De gravitatione’ manuscript, where Newton says variously that space is an ‘emanative effect of God’, that ‘space is an emanative effect of the first existent being’, and that ‘space is … the emanative effect of an eternal and immutable being’.Footnote ⁶⁸

At first glance (and scholars have directed many more glances than that at these passages) these seem to be much more positive statements than the agnosticism that I am attributing to Newton. But is this really the case? Here the scholastic context proves crucial, for it emerges that without it these statements are quite simply incomprehensible.

No one, to my knowledge, has ever asked what, contextually speaking, Newton might have meant when he attacked the idea that God was only present virtually, or ‘in power’. But if we turn to the tomes of the late scholastics, the answer becomes quite clear. Newton was dismissing the commonplace scholastic argument that substances – especially immaterial substances – may have a ‘virtual extension’; that is to say, that they operate through their power – virtus – without having the properties of normal extended substances like parts and divisibility. To give three almost random but very popular examples, the Jesuit Pedro Hurtado de Mendoza devoted a whole section of his 1624 philosophy textbook to ‘Judging the doctrine of the virtual extension of points’, concluding that he ‘admits virtual extension in spiritual substances, but rejects it in matter’, and then suggesting that virtual extension could explain the classic problem of how incorporeal objects could act on corporeal ones.Footnote ⁶⁹ One finds very similar arguments in a myriad of other texts, such as the popular textbooks of Pedro de Fonseca (1599–1604) and Luis de Ribas (1643).Footnote ⁷⁰ Nor was it only used by scholastics; Descartes, for example, used it in his correspondence with Henry More (a text known to Newton).Footnote ⁷¹

Unsurprisingly, the ubiquity of this concept led to it being attacked by many of the anti-scholastic novatores. The vast majority of them offered variants of the same criticism: that virtual extension was meaningless, for extension was either real or no extension at all. This was proclaimed by Descartes's friend and editor Louis de la Forge,Footnote ⁷² but also by Descartes's experimentalist critic Pierre Petit, who not only suggested that extensio virtualis is scholastic nonsense, but also even went on to claim that the concept threatened rather than preserved God's omnipresence, which it had been used to defend by scholastic philosophers.Footnote ⁷³ Petit was elected FRS in 1667, and it is no surprise that we also find the concept of virtual extension attacked by English natural philosophers of the second half of the seventeenth century. To take only one example, the physician Samuel Haworth argued that if the soul is not properly extended it cannot exert any action upon the body: the idea that the soul ‘operated upon the Body’ virtually was simply a typical scholastic absurdity.Footnote ⁷⁴

Placed in this context, Newton's dismissal of virtual extension/extension in power thus now appears only as a rather commonplace anti-scholastic salvo. This seems to be unequivocally confirmed by a passage from the ‘Tempus et locus’ manuscript from the 1690s, where, as part of a long condemnation of scholastic philosophy, Newton attacks the idea that God is only omnipresent as regards power: ‘Let them [i.e. the scholastics] therefore consider whether it is more agreeable to reason … that God is everywhere as regards power [virtus] and nowhere as regards substance, or that God's power should subsist everywhere in the divine substance … and exist nowhere separately.’Footnote ⁷⁵

Far from engaging in an elaborate metaphysical enterprise, then, Newton was simply pointing out that his defence of God's omnipresence was qualitatively different from that of his scholastic predecessors. This is no concession to occasionalism, materialism, Neoplatonism, Spinozism or any of the other positions that Newton has been asserted to have come close to in the recent literature.Footnote ⁷⁶ It is a polemical self-defence against those who might accuse him of not properly accepting God's omnipresence, as one could accuse the scholastics and Descartes.

That still leaves us with ‘De gravitatione’, and the idea that we find there that space is an ‘emanative effect’ of God. The only precursor to Newton in using the term that half a century of scholarship has examined is Henry More, who in his Immortality of the soul (1659) states that an ‘emanative effect’ is the product of a cause that acts simply by existing:

Unfortunately, this has once again generated the rather misleading impression that Newton's discussion of emanation must have been informed by idiosyncratic Neoplatonic traditions, or even by Spinoza.Footnote ⁷⁸

Such an impression is misleading because if we delve into the scholastic textbooks we find that More was not at all unique or idiosyncratic, for this was again a very common concept. In scholastic texts the language of ‘emanative cause and effect’ had a very precise meaning: an ‘emanative cause’ was, alongside its counterpart the ‘active cause’, one of the first subset of efficient causes. It was a cause that functioned simply though the existence of an object, from the action of which was generated an ‘emanative effect’. The standard example almost always offered to explain this was that the heat produced by fire was an emanative effect, whereas the heat of any object heated by the same fire was an active effect. As with virtual extension, the real philosophical purpose was usually to explain the action of non-extended substances like the soul.

It is precisely this use that we find in some of the most popular scholastic manuals and textbooks of the seventeenth century, such as Rudolph Goclenius's Lexicon philosophicum (1613),Footnote ⁷⁹ Franco Burgersdijck's Institutionum logicarum libri duo (1626),Footnote ⁸⁰ or Adriaan Heereboord's ΕΡΜΗΝΕΙΑ logica (1651),Footnote ⁸¹ all of which were republished in England and used in English philosophical pedagogy throughout the century.Footnote ⁸² Nor was the use of the concept limited to pedagogical texts, for it appeared in otherwise very different philosophical works by the likes of Zabarella and Johann Heinrich Alsted.Footnote ⁸³ Most interestingly for us, the undergraduate Newton made a note on the scholastic concept of emanation when reading Stahl's Axiomata.Footnote ⁸⁴ This may thus be one of the very rare cases when we can directly connect the Trinity Notebook to the later Newton, although, as we have seen, the concept was so commonplace that that does not really reveal very much.

We have seen, then, that in ‘De gravitatione’ Newton was using a standard scholastic concept. Moreover, it was a concept which, like extensio virtualis, had been mocked by plenty of non-scholastic philosophers, whether on the Continent or in England.Footnote ⁸⁵ This conclusion leaves us with two obvious questions. First, can Newton's use of the concept in ‘De gravitatione’ be used as a guide for his later statements about the relationship between God and the world (such as those in the General Scholium), as it has been by a host of interpreters? Here the answer is surely ‘no’: everything we know about Newton's intellectual biography suggests that he would have balked at using concepts that he recognized as scholastic in his mature writings. In support of this, it is important to recognize that ‘emanative causation’ had a direct relationship with the other concept that we have just found Newton rejecting: virtual extension. Fonseca, for example, when explaining how the soul interacts with the body, claimed that it is ‘“eminently” or “virtually” capable of all the effects found in lesser forms’.Footnote ⁸⁶ Closer to home, Henry More, when responding to Petit, claimed something very similar, even asserting that this fact meant that God's extension had to be understood not in a physical but in a ‘metaphysical’ sense.Footnote ⁸⁷ But as we have seen time and time again, Newton rejected metaphysical approaches to these problems. It seems just as unlikely that he would have seriously considered the use of the scholastic concept of ‘emanative effect’ in his mature works (of course, this is why he never did use the words).

But this leaves us with the obvious second question: why did the young Newton use the scholastic language of ‘emanative causation’? One very simple answer is that ‘De gravitatione’ was written at a relatively early stage of Newton's philosophical development, around 1670, when he had not yet fully arrived at his anti-metaphysical and anti-scholastic dogmatism. This conclusion can be somewhat contextually refined. Mordechai Feingold has discovered a manuscript that shows that ‘De gravitatione’ was very likely part of a course of lectures that Newton delivered in Cambridge in 1671.Footnote ⁸⁸ It would be far more understandable for Newton to use scholastic metaphysical terminology that he did not personally agree with – terminology that his students would have imbibed repeatedly through the textbooks of the arts course – in such a pedagogical setting. No less a philosopher (and friend of Newton) than John Locke would later recommend precisely this practice. Like Newton, he thought metaphysics a nonsensical discipline that should be abandoned by all serious modern thinkers. Nonetheless, he recommended its incorporation into a course of philosophy study for young students.Footnote ⁸⁹

If I am right, then, here we finally have Newton being a scholastic philosopher, but only when he was teaching. In his own mature natural philosophy there was no place for metaphysics, scholastic or otherwise.

Broader conclusions

It is time to stop attempting to use ‘De gravitatione’ as some kind of metaphysical key to Newton's mature natural philosophy. Some of our colleagues who approach Newton from a philosophical perspective have said that the dating of the work is irrelevant, because it still represents Newton's central philosophical concerns. But if Feingold and I are correct, the issue is not simply that it is an early text; rather, it is that it may well be a text that is not at all representative of Newton's true opinion, but rather a pedagogical exercise – modern academics hardly need reminding that one's actual beliefs are not always represented in, and are often contradicted by, one's pedagogical pronouncements. If we are to reconstruct the historical Newton, the genre of various texts matters as much as their chronological context.

This leads to a second, connected point (one that I am by no means the first to make).Footnote ⁹⁰ This is that we should treat Newton's methodological statements with extreme care. The more research is conducted, the more it appears that Newton used methodological statements not really to explain precisely what he was doing, but to score polemical points. To the anti-Cartesian and anti-Leibnizian points that other scholars have identified, I hope that I have now been able to add the anti-scholastic ones.

Finally, I would like to come back briefly to the broader issue raised in the Introduction, and to place Newton in the wider context of the history of scholastic philosophy in seventeenth-century England. Since the work of Charles Schmitt, it has been common to emphasize the continuing relevance of the Aristotelian tradition(s) to the novatores of the seventeenth century.Footnote ⁹¹ As was stated at the outset, this has borne very fruitful results. Nonetheless, it should not obscure the fact that anti-scholasticism was a very real, and a very powerful, force, which – however gradually – genuinely eroded the relevance of Aristotelian philosophy.Footnote ⁹² In the case under consideration here, it does seem that Newton was rather detached from scholastic philosophy, however much he was prepared to attack sporadically some of its more awkward concepts. Already in his earlier works, Newton saw Aristotelian scholastic hylomorphism not as something to be refuted, but only to be mocked.Footnote ⁹³ The vitality of early modern Aristotelianism notwithstanding, late seventeenth-century natural philosophers, at least in England, were not particularly concerned with relating their research to the textbooks that they had learnt from in their university days. Pace Wallace, there seems little hope of relating Newton's undergraduate notes to his natural philosophy.

This late seventeenth-century distance between school philosophy and natural philosophy as it was actually being practised emerged for various reasons. Partially – and this is especially the case in England – it emerged because Cartesianism had overtaken scholastic Aristotelianism as the chief ‘systematic’ bête noire to be refuted.Footnote ⁹⁴ But more significantly, it also emerged because there had been a genuine anti-scholastic movement, which, crucially, had been institutionalized in the universities. The heroes of that story are neither Newton, who was a second-generation inheritor of this shift, nor Bacon, who for all his anti-Aristotelianism never demonstrated that a rival system or set of results was available. The real heroes of such a story may be the still rather understudied and underappreciated figures like Ralph Bathurst or Nathaniel Highmore: English physicians and naturalists of the 1650s who absorbed the virulent anti-Aristotelianism of Continental natural philosophers such as Sébastien Basson and Pierre Gassendi. They used it not to serve a destructive radicalism but to continue to promote the actual practice of natural philosophy within the learned institutions: the universities and the College of Physicians.Footnote ⁹⁵ It was they who created the institutionalized intellectual culture in which Newton only had to develop a superficial familiarity with scholastic philosophy, and it was they and their colleagues who were the driving force behind the other institution which facilitated the practice of (experimental) natural philosophy, and of which Newton would eventually become president: the Royal Society of London.