I.

Reason has been a central value in the sciences, mathematics, natural history, and various other knowledge enterprises for many centuries; at the same time, it would obviously be ahistorical to imagine that “reason” has always referred to the same thing. The eighteenth century is the canonical “age of reason,” but what did that name denote, and how had the theme developed? In the preceding century, reason as a category or a discourse was sometimes subjected to uses and even explicit discussions that exposed it to different kinds of scrutiny than became the norm in the eighteenth; some of those discussions indicate important issues that subsequently became hidden beneath the overlay of shifting questions, but which never fully went away.

For two or three centuries it has been usual to speak of philosophical approaches in the seventeenth century as falling into two main camps, the “empiricist” and the “rationalist.” But it has also become quite common to notice the lack of clarity in that dichotomy, and the ways in which the two alternatives tend to bleed into one another; so Locke, an exemplar of empiricism, can be represented as being much like Descartes, the exemplar of rationalism, when it comes to the innate endowments of the mind; whether these endowments are called “innate ideas” or not becomes the only real point at issue (e.g. Aarsleff Reference Aarsleff and Aarsleff1982, esp. 172–175). But attempts in the seventeenth century to draw sharp battle-lines were often quite vigorous nonetheless, and the principal movers in doing that drawing tended to be the people we think of as empiricists, classically the English empiricists Locke and Boyle. Locke's famous diatribe against “innate ideas” in the Essay, usually taken to be directed against Cartesians, may also be directed, more locally, at the Cambridge Platonist Ralph Cudworth (Locke [1690] Reference Locke and Nidditch1975, Book I; Cudworth Reference Cudworth1678). These disputes nonetheless avoided contesting “reason” itself as a central component of knowledge-generation, even in attacking “rationalism.” For an “empiricist” philosopher in the seventeenth century, “experience” meant going and looking, as well as thinking in some appropriate way about what you saw; but in addition, “reason” was an essential part of that appropriate way of thinking. Empiricism could not dispense with reason any more than rationalism ignored the lessons of the senses. Most practical uses of the term “reason,” or else “rational,” failed to interrogate it very much, any more then than now; usually people talked as though everyone knew perfectly well what such terms meant, and as though these self-evident terms always represented something good.

Usefully pedantic explications of “reason” and the “rational,” according to different disciplinary contexts, appear in philosophical lexicons of the seventeenth century, and they largely confirm what one might guess from noticing the miscellany of contemporary uses (e.g. Goclenius [1613] Reference Goclenius1980; Micraelius [1662] Reference Micraelius1966; cf. the Cartesian Chauvin Reference Chauvin1713; also Locke [1690] Reference Locke and Nidditch1975). The Latin word ratio needed to be defined as it was used in mathematics (more or less the same as the modern English mathematical word “ratio”); in metaphysical and physical contexts it was often used as a synonym for the Aristotelian concept of “form,” or forma, and more generally as “cause.” In the sense in which we tend to think of “reason” in modern English, the appropriate context was that of logic, where reason could be explicated in terms of deductive, inductive, analogical, probable, or demonstrative kinds, in various combinations; ratiocinatio meant something like “ratio in action,” where ratio was in this case an intellectual capability. However, in regard to actual uses of these terms by philosophers, writers tended in general not to tell the reader to which relevant sense of the term “reason” they were appealing; it was presumably supposed to be evident from context, or else functionally ambiguous, as with most language.

The particular aspect of reason in this period that seems the most crucial, however, is one that seems on the face of it the most uninteresting. No mention is made of it, for example, in the chapter on “Reason” in Geoffrey Lloyd's recent considerations in comparative anthropology, Cognitive Variations (Lloyd Reference Lloyd2007, chap. 8; see also Prakash Reference Prakash1999). For Lloyd, cultural variations in “reason” are about the differential plausibility of various sorts of informal inference: persuasive metaphors, contextualized versus decontextualized inferences about specific subject-matters, and so on. As with most examinations of active “reasoning,” Lloyd (ironically, for him) ignores questions concerning the sorts of inferences that Aristotle was concerned to formalize: the domain of deductive logic and its mathematical analogues. For many philosophers in the seventeenth century, including Descartes and Newton, these were the inferences that mattered.Footnote ¹

In seventeenth-century mathematics, deductive proofs were still the gold standard. They required, in good Euclidean style, the step-wise movement from one set of statements to a new statement that the previous ones were held to imply. But there were some people who found those movements worthy of comment. There were no difficulties, in principle, in actually performing them, but understanding how they were done was another matter. Rather than “reason,” the relevant category is perhaps better rendered as “reasoning.”

Descartes addresses the issue of human reasoning head-on in his early text from the 1620s, the Rules for the Direction of the Mind. Here he, in effect, analyzes reasoning down to its fundamentals. In Rule 10, Descartes addresses the procedures of syllogistic logic as a model for disciplined reasoning, and criticizes them for being too rote; so automatic that the mind needs to pay but little attention. The logicians prescribe precepts that supposedly govern human reason, “in which,” he says, “the conclusions follow with such irresistible necessity that if our reason relies on them, even though it takes, as it were, a rest from considering a particular inference clearly and attentively, it can nevertheless draw a conclusion which is certain simply in virtue of the form” (Descartes Reference Descartes, Adam and Tannery1996, 10: 405–406; trans. Descartes Reference Descartes, Cottingham, Stoothoff and Murdoch1985–91, 1: 36, my emphasis).Footnote ² However, a great danger lies in the use of these procedures: one can easily become trapped in the formal structure, and be misled by sophisms. By contrast, “the cleverest sophisms hardly ever deceive anyone who makes use of his untrammeled reason” (Descartes Reference Descartes, Adam and Tannery1996, 10: 406; trans. Descartes Reference Descartes, Cottingham, Stoothoff and Murdoch1985–91, 1: 36). So natural reason is something that escapes from, or underlies, formal argumentation; the precepts of formal argumentation do not themselves capture it or sum it up. Instead, the operations of reason depend on something that Descartes calls “intuition.”

Thus in Rule 3: “The self-evidence and certainty of intuition is required not only for apprehending single propositions, but also for any discourse whatever.” So, Descartes continues: “Take, for example, the inference that 2 plus 2 equals 3 plus one: not only must we intuitively perceive that 2 plus two make four, and that 3 plus one make four, but also that the original proposition follows necessarily from the other two” (Descartes Reference Descartes, Adam and Tannery1996, 10: 369; trans. modified from Descartes Reference Descartes, Cottingham, Stoothoff and Murdoch1985–91, 1: 14–15).Footnote ³ The certainty of lengthy deductions is a more difficult problem, however, because, as Descartes explains, a long string of intuitively-linked propositions only confirms the eventual conclusion with the additional support of the memory. Nonetheless, the fundamental units from which the string or chain is constructed are still those of intuitively-recognized true propositions and their intuitively-perceived individual relations each to the next.

In effect, Descartes represents this kind of intuition, whereby one makes an inference from one set of statements to the next, as a kind of seeing: you just see that the conclusion follows, in the much the way that Aristotle assumed when he noted that you cannot prove foundational statements; axioms are simply to be accepted, but they are also supposed to be self-evident (e.g. Aristotle, Posterior Analytics I, 10). In Rule 9, Descartes explicitly analogizes mental intuition to ordinary vision (cf. Locke [1690] Reference Locke and Nidditch1975, 31).

Both here and elsewhere, Descartes also describes how the best kind of deductive demonstration, of the sort that consists of a large number of primitive intuitive steps, will be one where you have trained yourself to run over the stages of the proof so smoothly that your mind can move from the beginning to the end without any longer needing to break down the deduction into its discrete steps. Thus in Rule 7, referring to such separate inferential steps, he says:

Thus deductive reasoning reduces, for Descartes, to brute (or, perhaps, angelic) intuition – in its most primitive form, to the irreducible leaps from one statement to the next that you just have to see; and in a more capacious sense, to a trained capacity to encompass an entire deductive sequence as a single intuitive perception. In that sense, the core of deductive reasoning is anything but mechanical: even deduction was not simple deduction.

A particular feature of Descartes's ideas is worth noting: historians have sometimes remarked that Descartes's notion of res cogitans, the immortal thinking stuff that constitutes the mind, comes dangerously close to medieval Averroïst doctrine (e.g. Gaukroger Reference Gaukroger1995, 346–348). Averroës made a distinction between the active and the passive intellect: the passive intellect registers an individual's experiences, and dies with that individual, and the active intellect is shared by everyone and is, in that shared sense, immortal. The basic idea behind the identity of this immortal active intellect seems to be that reason is taken to be the same for everyone; there is no individuality, in effect, in the correct use of reason, in the sense that two plus two equals four for me in exactly the same way as it supposedly does for everyone else. Extending this point to Descartes's analysis – for which that commonality is also clearly assumed – we have the intuitions that knit together each step of a deductive sequence themselves being necessarily shared and identical – in turn, again, suggesting a common origin for the validity of those intuitions (cf. Normore Reference Normore1993 on the issue of whether or not, for Descartes, the certainty of deductive conclusions was a matter of their necessity).

This point in turn relates to a major area of discussion in modern Cartesian scholarship to do with the status for Descartes of the eternal truths: axiomatic logical and mathematical statements of the sort that provide the foundations for solid deductions. Descartes famously argued in writing to Mersenne that the eternal truths – things like the Euclidean axioms – are themselves radically dependent on God's will; their truth is not separate or autonomous from God, and these statements are only true because God wills them to be true (Descartes Reference Descartes, Adam and Tannery1996, 1: 145–146; 15th April 1630). This assertion alarmed Mersenne, not surprisingly, as a more extreme position than Thomas Aquinas’ view that these true statements are “co-eternal” with God; God is not Himself the cause of their truth.Footnote ⁴ But more immediately important than these metaphysical finesses is the issue of how we ourselves know these eternal truths, regardless of their exact relationship to God's will. This brings us into territory more familiar to seventeenth-century theological orthodoxy, albeit territory that was itself something of a battleground.

The widely-accepted Thomistic position, adopted by most Jesuit theologians and in general a conventional Catholic position in this period, was that we recognize the truth of such statements through what Aquinas calls the “natural light” of our minds. This is an inborn human endowment that God has put there, not just another feature of the human intellect, and this “natural light” has considerable theological significance: Aquinas calls it the “image of God” in man, regarding it as what was meant by the assertion that man was made in God's image (Harrison Reference Harrison2008, 43–46).

There was also another Latin theological tradition that represented such issues in a rather different, and in practice less optimistic, way. St. Augustine's neo-Platonic Christian theology, which was quite central to Lutheran and especially Calvinist theologies, assigned a role for divine assistance in the acquisition of knowledge and the operations of reason that was analogous to the more Aristotelian one put together by Aquinas: instead of God giving man a natural light by which to perceive truth on his own, Augustine has God illuminate the human mind whenever it directly perceives truth – this is the operation of the so-called divine light rather than natural light, and in that sense it makes man more directly dependent on God (Dear Reference Dear1988, 85–89). It is therefore a more desperate resort, in a sense, than Thomistic natural light, because on Augustine's view “original sin” (which Augustine, after all, invented), attendant on the expulsion from Eden, had severely compromised human intellectual capacities (Harrison Reference Harrison2008, 38–40, chap. 2).

Peter Harrison has recently argued for the importance of the role of the Fall in seventeenth-century philosophical and natural-philosophical projects (Harrison Reference Harrison2002 and Reference Harrison2008). Views on the Fall often correlated with confessional allegiances: within Catholic, typically Thomistic, theology, the Fall tended not to be seen as quite so catastrophic as in Protestant, typically Augustinian, theology: for Augustine, the Fall had resulted in the profound darkening of human intellectual perceptions, and made certainty considerably harder to attain, whereas the Thomistic, and typical Catholic, view restricted the damage done by the Fall largely to the loss of guaranteed salvation rather than to the loss of intellectual capacities – the image of God in man is still alive and well from that perspective, and hence the natural light of reason remains intact (cf. Ashworth Reference Ashworth1989).Footnote ⁵ Thus Harrison shows that this theological question played a significant role in seventeenth-century English Puritan critiques of university learning, as in the Webster-Ward debate, with the pretensions of Aristotelian philosophy to access the essences of things being equated with Papist superstition (Debus Reference Debus1970).

From this perspective, Descartes had no real difficulty: his attempts to square himself with Catholic orthodoxy were no kind of stretch. The crucial role of God in his metaphysical system accommodated itself pretty well to a Thomistic notion of “natural light”; thus Descartes says in the Discourse that “all our ideas or notions must have some foundation of truth, for otherwise it would not be possible that God, who is all-perfect and all-truthful, should have placed them in us” (Descartes Reference Descartes, Adam and Tannery1996, 6: 40; Reference Descartes, Cottingham, Stoothoff and Murdoch1985–91, 1: 131; see also Morris Reference Morris1973). And in Principles of Philosophy, in talking about knowledge of efficient causes, he says, “we shall see, with the aid of our [God-given] natural light, what conclusions should be drawn concerning those effects which are apparent to our senses.” Nonetheless, he then adds a safety-clause: “we should remember . . . that the natural light is to be trusted only to the extent that it is compatible with divine revelation” (Descartes Reference Descartes, Adam and Tannery1996, 8: 16; Reference Descartes, Cottingham, Stoothoff and Murdoch1985–91, 1: 202–203).

Unlike the enthusiastic overtones of intellectual perception triggered by the immediate gift of divine light, natural light gave everyone at all times the potential for true perception through reason – but “only to the extent that it is compatible with divine revelation.” Only strong, unequivocal social authority could establish the uncontrolled variable “divine revelation”; at the same time, possession of natural light united all human beings. Its correct use, however, which Descartes wanted to teach, was not so natural: honor, or its specific seventeenth-century French version, honnêteté, underlay such intellectual facility, just as it underlay social order.Footnote ⁶

II.

One immediate association with deductive reasoning that we might tend to make nowadays is computation carried out by a machine – artificial intelligence, as it were. Questions of causation in machines set up a useful counterpoint to early-modern ideas of “natural light,” the piquancy of which has in effect been observed in the 1982 essay “Relativism, Rationalism and the Sociology of Knowledge,” by Barry Barnes and David Bloor. Barnes and Bloor at one point criticize philosophers, including Gilbert Ryle, for saying that, in psychology, causal accounts are only appropriate for pathological phenomena, not for normal or “correct” ones. They then ridicule that view:

Barnes and Bloor make a crucial assumption here: in accepting that what's at stake is indeed a “causal” account of an outcome they are, like their philosophical opponents, imprecise about what “causal” means. A seventeenth-century Aristotelian philosopher – or a Galenic physiologist – could have replied that a normal perception would be economically explained by making reference to the proper functioning of the sensory apparatus: what would amount to a formal cause explanation. Why bother with providing an efficient cause for something normal? In the case of properly functioning adding machines, they are correctly designed according to sound principles – they work properly because they were built in such a way as to do so. A machine that does not work properly is then anomalous; something must have gone wrong, or a mistake made in the design, and that error would be the feature in need of special explanation.

But what, in fact, did seventeenth-century philosophers have to say about arithmetical machines? Blaise Pascal invented, and had built, just such a machine in the 1640s. Its sets of wheels displayed base-ten digits, and were designed to add and subtract, the chief trick being the technical means of carrying numbers. Pascal described the machine as allowing the performance of arithmetical operations “without any work of the mind”; he makes no claim that there is “artificial intelligence” embodied in it. Further, Pascal does not attempt to compare its operations to a person doing mental arithmetic; instead, the only things to which he compares it are other means of performing calculations. He speaks of the

Pascal understands his arithmetical machine as an alternative technique for performing computations, on a par with working on paper or using the common system of counters (a kind of abacus on a table).Footnote ⁷ His machine chiefly serves to relieve the memory, as well as being quick. He also mentions that the machine performs the desired operations without its operator even having to think about it (sans même qu'il y pense).

Descartes had written of the intellectual virtues possessed by the ability to pass in a single movement of thought from premises to conclusion in deduction: he spoke of learning “to pass from the first to the last so swiftly that memory is left with practically no role to play, and I seem to intuit the whole thing at once” (cf. Jones Reference Jones2006, 76–7). Pascal here attributes a similar virtue to his machine: there is no need to rely on one's own memory, because the machine manages those operations by itself. But there is a central and crucial difference between the two: the intended outcome of the process in the case of Pascal's machine is the result; whereas the virtue of Descartes's procedure was that it created an “intuition” of the correctness of that result. Indeed, Pascal's version, as he stresses, dispenses with the very need for thought.

How, then, can Pascal establish that the machine, as a technique for performing calculations, does indeed yield results that are correct? Certainly, the machine does not intuit their validity, in the manner of an ideal Cartesian deduction. Instead, Pascal discusses two different foundations for regarding his machine as reliable. One is the possession of a full understanding by the builder of the principles on the basis of which the machine is designed and constructed; the reliability of the calculations follows from those principles, in what one might see as a mechanically deductive fashion: if you know how the wheels work, you'll know that they must give correct results. In mentioning this issue, however, Pascal is only concerned with having the artisan manufacture the machine properly. This process, he says,

The manufacturing of the machine, which guarantees its reliability, is itself, on this view, reduced to a kind of mechanical procedure, albeit a human one.

The other basis for believing the machine's continued reliability is that Pascal has tested it. The only way that he could gain a complete assurance for himself, which could then be passed on to others, was, he says, to “make the experience, by the transport of the instrument over more than two hundred and fifty places on the road without any alteration” in the performance (ibid., col. I) – rather like Périer's famous stops during the ascent of the Puy-de-Dôme to measure the height of the mercury (Dear Reference Dear1990, 676–79).

The examples underscore the point that Pascal's understanding of what computation was all about had little to do with the mind, or with reasoning. The mechanical behaviors of the machine as its wheels were turned by the operator could perhaps be seen as analogous to the stages in a deductive argument as Descartes had described them, but for Pascal they had nothing to do with understanding a deductive demonstration. The results of calculations made with the machine were just that – results brought about by appropriately mindless manufacture on the part of habituated workmen, and by the grinding of gears. No intuition was needed, any more than in performing sums on paper or with counters, or in using such gadgets as Napier's bones; Pascal promoted as its chief selling-point the fact that the machine dispensed with the need to use the memory. In fact, the same thing seems to apply to later elaborations of Pascal's type of machine too, including Leibniz's from the 1670s, as well as other kinds of mechanized calculators such as Samuel Morland's in England (Ratcliff Reference Ratcliff2007). In his Pensées, Pascal's preferred distinction concerned the will: he observed that “the arithmetical machine creates effects that approach thought more nearly than anything animals do; but it does nothing that might make us say that it possesses will as animals do” (Pascal Reference Pascal and Lafuma1963, 596; L741).Footnote ⁸

The key difference between Descartes on deduction and Pascal on his arithmetical machine can be illustrated by a simple point, reminiscent of Barnes’ and Bloor's remarks: it would have been possible for Pascal to have made an arithmetical machine that reliably produced wrong answers, whereas Descartes's view of deduction presupposed that his codified mental procedure will always produce correct answers. Pascal's machine is just a gadget, like an abacus, not an example of Artificial Intelligence, and it was never intended to be anything else (cf. esp. Collins Reference Collins1990).

But the central meaning of the term “reason” for Pascal is still fairly close to what it was for Descartes. Recall again Descartes's remarks in the Regulae concerning intuition, intuitus. Intuition was an irreducible sense of conviction, like a “clear and distinct” mental perception, and it applied both to simple, single deductive steps and to awareness of the truth of one's fundamental, unprovable axiomatic principles. Pascal understood matters in much the same way, which is unsurprising given the role of these themes in Aristotle's logic. What makes Descartes, and Pascal, different from Aristotle is what they add to the Aristotelian analysis: divine authority.

One of the most famous of Pascalian aphorisms, from the Pensées, itself concerns reason: usually, in English, “the heart has its reasons that reason knows not of” (Pascal Reference Pascal and Lafuma1963, 552, col. I [L423]). Although this seems at first sight to be a typically paradoxical theological aphorism, it can be read as fairly straightforward when understood in terms of the variety of associated meanings that the word “reason” – here, la raison – had in seventeenth-century philosophical and literary parlance. Pascal is referring here to fundamental, unprovable principles at the foundations of formal “scientific” deduction. “The heart has its reasons” in the sense of “reason” that meant underlying causal principles, as with the ratio of an Aristotelian philosopher. Such reasons are those that “reason knows not of” because this second sense of reason concerns the act of “reasoning,” which is unable in itself to establish the very principles on the basis of which it builds. In effect, Pascal seems to be saying that “the heart has its cherished convictions that the reasoning intellect cannot touch.” And part of his point is that this is just as true of religious faith as it is of Euclidean geometry.

Unsurprisingly, Pascal, like Descartes, sometimes discusses formal reasoning with reference to the “natural light,” which enables us to obtain certainty in geometry despite the obvious impossibility of defining and proving every element of our starting principles; our “natural light” sees us through. In another passage from the Pensées, Pascal puts the matter this way:

The irreducibility of this kind of intuition (or, as he sometimes called it, sentiment Footnote ⁹) is also expressed, albeit in an indirect way, in Pascal's account of his arithmetical machine. Referring to the operation of the machine – its practical employment – he says that his reader will forgive Pascal's decision not to give a detailed account of the working of the machine:

Pascal subsequently recommends the mathematician Gilles Personne de Roberval as the person to consult if his reader happens to be in Paris and wants instruction on how to use the machine (Pascal Reference Pascal and Lafuma1963, 191, col. II). Furthermore, this recommendation is not casual; there is for Pascal more involved in teaching the use of the machine than a formalized conveyance of information. In effect, Pascal invokes the importance of tacit knowledge: not everything can be made explicit, but there are ways around that problem; sometimes God, but in this case, Roberval.

III.

Perhaps the most uncompromising, and, for modernist sensibilities, the clearest position on the nature of reason was presented by Thomas Hobbes in a number of his writings, most familiarly in Leviathan: he tells us what, in his view, “is meant by this word Reason, when wee reckon it among the Faculties of the mind. For REASON, in this sense, is nothing but Reckoning (that is, Adding and Subtracting) of the Consequences of generall names agreed upon, for the marking and signifying of our thoughts” (Hobbes [1651] Reference Hobbes and Tuck1996, 32). Presumably, then, Hobbes would have regarded Pascal's machine as in some sense a “reasoning engine.” But even for Hobbes, this does not imply that reason was entirely straightforward, entirely “mechanical.” Besides so-called “natural reason,” there was also the category of “right reason.”

The locus classicus for “right reason” was Aristotle's Nicomachean Ethics: in Greek, orthos logos. The Stoic adoption of the term is then rendered by Cicero as recta ratio; and the overall sense of the term is something akin to “sound reason,” “proper reason,” “solid practical reasoning.” The term stresses reasoning as a process rather than reason as a capacity. One seventeenth-century discussion of these issues is by Mersenne, in his first publication, from 1623, a devotional work called L'Usage de la raison. Mersenne is interested in showing how reason can aid in salvation; he explains that proper reason, in the sense of active reasoning, is the result of combining understanding with the will. Although this rather Molinist account, concerning “hypothetical necessity,” makes no use of the term “right reason,” its theme is the pursuit of salvation resulting from the will's decision to embrace what the understanding recognizes as the good; it fits very well various uses of the term in this period (Mersenne [1623] Reference Mersenne2002, esp. 26–33; also Dear Reference Dear1988, 90–91).

Lotte Mulligan has shown in several articles how “right reason” was widely used in Anglican apologetics to designate reasoning in religion that was correctly aligned with accepted assumptions or premises derived from orthodox interpretations of Christian revelation, or in general that led to theologically desirable conclusions (Mulligan Reference Mulligan and Vickers1984, Reference Mulligan1994, and Reference Mulligan and Crocker2001). For most users of the term, the hallmark of right reason was that it supported, and never contradicted, orthodox religious teaching (cf. Boyle Reference Boyle, Hunter and Davis1999–2000, 14: 180–181, contrasting “right reason” to “hypothetical reason”). In effect, it was a form of reasoning that was uncorrupted by the Fall, or else that had been restored with the help of God's grace and the sweat of man's brow (Schaffer Reference Schaffer and Holland1985 also addresses English uses of “right reason”). But for Hobbes this was clearly a tricky issue, because of his determination that civil authority not be threatened by an autonomous church. “Right reason” for Hobbes was still reason that had been morally purified, but for him that meant a form of reason that promoted the interests of the civil power. When Hobbes discussed even so basic a subject as arithmetic, he required that this kind of authority be central.

Hobbes held that “Reason” itself could always be reduced to a matter of adding and subtracting. Arithmetic was therefore the most fundamental of sciences, because it represented formalized “reckoning,” as he put it, which was all that reason truly was. But even reckoning was not a self-sufficient, self-justifying procedure. As Hobbes says in Leviathan:

Reason, in other words, needed to be policed correctly; its virtue, represented above all in the strict procedures of mathematics, could not stand by itself. Otherwise, exactly as in the case of religious enthusiasm, the social fabric would fall apart, as Hobbes goes on to explain:

Hobbes had expressed the point even more bluntly some years before, in his De corpore politico (published 1650):

So, he concludes, “this is certain, seeing right reason is not existent, the reason of some man or men must supply the place thereof,” and those men must be those “that have the sovereign power” (Hobbes Reference Hobbes and Molesworth1839–45, iv, 225).Footnote ¹⁰

Notice, here, what the wicked Hobbes is doing. For his contemporaries, “right reason” was generally contrasted with so-called “natural” reason, which included no truths of revelation with which it had to accord. For Hobbes, on the other hand, sound reason, natural reason, is the same as right reason; there is no distinction between the two. The supernatural has been quietly replaced with civil authority, which provides the absolute criterion for reason. For Hobbes, notions of God could not by themselves be cashed; the brute force of civil authority necessarily played that role for Him. The king was both God's banker and God's trustee.

Perhaps there should be nothing surprising in finding the idea of “reason” and the “rational” being grounded in some kind of transcendental supra-human power. In a sense, seventeenth-century views about God's role in reason may be easier to take, historiographically, than Hobbes’ equally seventeenth-century totalitarian vision. Hobbes sometimes seems to speak directly to secular modernist sensibilities, but when he asserts that our reasoning capacities depend on political power we tend not to like it, even in our more Foucauldian moments; Hobbes seems to go too far for comfort. Nonetheless, God may be our most perspicuous focus in trying to understand even Hobbes’ arguments: the functioning of “reason” in the seventeenth century, and perhaps, by implication, for a much longer period, depended on some transcendental guarantor. Hobbes saw it most clearly: someone or something is needed to tell you when you've got it right: reason does not shine by its own light. That was why, for most commentators on the subject, natural or divine light, both owing to God, was essential for completing the circuit – unless you were willing to allow the king to do it for you.

IV.

If such were the issues circulating in the seventeenth century, clearly much had changed by the time of the famous eighteenth-century question “What is Enlightenment?” (Schmidt Reference Schmidt1996). The Age of Reason had by then established a new idol, neither illumination, exactly, nor civil authority. What the seventeenth-century discussions had shown is how “reason” itself was not free-standing, and never could be without becoming completely circular: even formal logic failed to show a way out, because, despite the conflations, there was a lot more to reason than mere logic. No doubt this is how reason, like nature, could take on such strong moral connotations in the eighteenth century (Daston and Vidal Reference Daston and Vidal2004).

Historically, “reason” as an epistemic theme has always been a flexible resource.Footnote ¹¹ In the seventeenth century, its persuasiveness usually rested on a transcendence that underlay and guaranteed it. This transcendence routinely passed through an established Christian church, or through some other theological authority. If it did not so pass, it would be a threat to stability; and that is precisely how some in the eighteenth century were to use it. Hobbes, of course, disagreed.