SOURCE: "Descartes' Philosophy of Science and the Scientific Revolution," in The Cambridge Companion to Descartes, edited by John Cottingham, Cambridge University Press, 1992, pp. 258–85.

[In the following essay, Clarke examines the epistemological and metaphysical underpinnings of Descartes' philosophy of science, contrasting it with scholasticism.]

Descartes' concept of science can be understood only by paying careful attention to the historical context in which it was constructed. The scientific revolution of the seventeenth century involved two related developments: a change in scientific practice (or, more accurately, a whole series of such changes) which is reflected in the founding of new scientific societies such as the Royal Society and the Académie royale des sciences, and a complementary change in how natural philosophers described the kind of knowledge that resulted from the new scientific practices. Descartes contributed to both developments. He shared this distinction with such eminent figures as Galileo Galilei, Francis Bacon, William Harvey, Robert Boyle, Christian Huygens, and Isaac Newton, all of whom were concerned both with improving our knowledge of nature and with clarifying the status of that knowledge.

It would be an obvious oversimplification to classify all the natural philosophers of the seventeenth century as, in some fundamental sense, proposing the same scientific theories. It is equally unsatisfactory to suggest that they all accepted the same theory of science or the same model of scientific knowledge. Yet, despite the pitfalls involved, it may be helpful—at least prior to examining Descartes' texts—to think of many of the most famous natural philosophers of the scientific revolution as sharing a number of new insights about the nature of scientific knowledge and, more importantly, as repudiating certain features of the model of science that was generally accepted in colleges and universities at that time. In fact, there was more agreement about what was being rejected than about what was being proposed in its place. Descartes occupies a pivotal role in the history of this development, in the transition from a widely accepted scholastic concept of science to its complete rejection by practising scientists and the endorsement of some kind of hypothetical, empirically based knowledge of nature. The historical context in which Descartes worked should lead us to expect, therefore, that he struggled with the epistemological and methodological issues involved in this transition. It should also lead us to expect that the transition was neither quick nor clear-cut. In other words, there is a strong likelihood that seventeenth-century natural philosophers continued to accept various features of precisely the model of science which they claimed explicitly to reject, while at the same time adopting elements of the newly developing concept of science that were incompatible with their traditional allegiance.

The traditional concept of science that was almost universally taught in colleges and universities included a number of key features; one was the certainty or necessity of genuine knowledge claims, and their universality. Aristotle says in the Posterior Analytics:

We suppose ourselves to possess unqualified scientific knowledge of a thing, as opposed to knowing it in the accidental way in which the sophist knows, when we think that we know the cause on which the fact depends, as the cause of that fact and of no other, and, further, that the fact could not be other than it is…. Since the object of pure scientific knowledge cannot be other than it is, the truth obtaiped by demonstrative knowledge will be necessary.

The paradigm of this type of knowledge was pure mathematics. One begins with definitions or first principles which are known with absolute certainty, one proceeds "demonstratively" by deducing other propositions from those already known as certain, and the logical validity of our inferences guarantees the same degree of certainty for our conclusions as was available for the initial premisses. The mathematical model of demonstrated knowledge inspired one of the dominant features of the scholastic concept of science that was widely accepted in the early seventeenth century.

Another feature of this concept of science was the claim that our knowledge of physical nature depends ultimately on the reliability of our everyday observations and judgments.² This involved two elements. One was the assumption that all our knowledge ultimately depends on sensory evidence and that it includes nothing that was not learned through sensory experience.³ Secondly, the cognitive faculties with which God has equipped us are completely reliable as long as they are used within the scope of their Creator's design. Thus we know the way the world is, and we can know it with certainty, by consulting the ways in which the world appears to us in sensation.

A further element of the scholastic tradition was the assumption that, if we wish to explain the natural phenomena which appear to us in sensations, we must use the distinction between "matter" and "form."⁴ This was a very widely used distinction which varied in meaning from one context to another. It was designed to reflect our common experience of the same type of thing being instantiated in a variety of different ways; for example, dogs may be small or large, their colors can vary, as may many other inessential features, without their ceasing to be dogs. The common, essential features of a dog could be described as the form of a dog, while the nonessential, variable features could be described (metaphysically) as the matter. What appears in sensation, therefore, is the appearance of an underlying reality (form) which, in turn, is the more fundamental dimension of any reality. This underlying reality, or form, is what explains whatever is necessary or essential in anything. Because the traditional concept of scientific knowledge was limited to knowledge of what is necessarily true, it follows that scholastic scientia was directed to acquiring knowledge of forms. Thus a scholastic explanation of a natural phenomenon is a discovery of the forms that underlie the appearances manifest to the human perceiver in reliable sensations.

This very brief summary is almost a caricature of what scholastics claimed about scientific understanding. However, many of Descartes' contemporaries argued that it was precisely this philosophy that obstructed the consideration of alternative ways of investigating nature. It was this simple-minded model of knowledge that was invoked by those who objected to the new sciences, and that was used as a foil by proponents of the new sciences to show in relief the distinctive features of their own philosophy of science.

HYPOTHESES

Descartes began his account of the natural world in Le Monde (c.1632) by discussing the unreliability of our sensations as a basis for scientific knowledge.

In proposing to treat here of light, the first thing I want to make clear to you is that there can be a difference between our sensation of light … and what is in the objects that produces that sensation in us … For, even though everyone is commonly persuaded that the ideas that are the objects of our thought are wholly like the objects from which they proceed, nevertheless I can see no reasoning that assures us that this is the case…. You well know that words bear no resemblance to the things they signify, and yet they do not cease for that reason to cause us to conceive of those things … Now if words, which signify nothing except by human convention, suffice to cause us to conceive of things to which they bear no resemblance, why could not nature also have extablished a certain sign that would cause us to have the sensation of light, even though that sign in itself bore no similarity to that sensation?

(AT [Œuvres de Descartes, ed. C. Adam and P. Tannery, revised ed., 12 vols., Paris: Vrin/CNRS, 1964–76] XI 3–4)⁵

Descartes goes on to use the same example as Galileo, to argue that a tickling sensation caused by a feather does not resemble anything in the feather. "One passes a feather lightly over the lips of a child who is falling asleep, and he perceives that someone is tickling him. Do you think the idea of tickling that he conceives resembles anything in this feather?" (AT XI 6)⁶ In a similar way, there is no reason to believe "that what is in the objects from which the sensation of light comes to us is any more like that sensation than the actions of a feather … are like tickling" (AT XI 6).⁷ If we cannot argue validly from a description of our sensation of light to the claim that the light that causes this sensation resembles our experience, then we have a fundamental problem in attempting to base scientific knowledge on our sensations of the world around us. The distinction between our subjective experiences or sensations and their objective causes, between primary and secondary qualities, opens up an epistemic gap that can only be bridged by some other strategy apart from assumptions of resemblance. This strategy is hypothesis, or guesswork. Our guesses may turn out to be very secure, and there may eventually be many reasons for thinking that they are as certain as one can hope for in the circumstances; but that does not change the fact that we come to have these ideas, in the first place, by guesswork. What should a natural philosopher assume about the physical causes of our perceptions? There are a few reasons why Descartes opts for one assumption rather than another at this crucial juncture, some of which rely on his concept of explanation (which is discussed below). Apart from those reasons, he also presupposes a radical distinction between matter and mind for which he argues in the Meditations and the Principles. It follows from this that the objective causes of our sensations are material, in some sense. In order to fill in some of the relevant detail, Descartes must engage in elementary physical theory.

The speculations about matter on which Descartes' theory of matter and, subsequently, his concept of science depend include the assumption that the size, shape and motion of small particles of matter would be adequate to explain all their physical effects, including the physical effects on our sensory faculties which stimulate sensations. Some of the reasons for this degree of parsimony in theory construction are mentioned below. In postulating three types of matter in Le Monde, Descartes is not very convincing about why he assumes three (rather than more or fewer); however, once they have been introduced, he is quick to take refuge in the construction of a hypothetical world which allows his imagination complete freedom, without having to explain the rationale for each hypothesis as it is made.

Many other things remain for me to explain here, and I would myself be happy to add here several arguments to make my opinions more plausible. In order, however, to make the length of this discourse less boring for you, I want to wrap part of it in the cloak of a fable, in the course of which I hope that the truth will not fail to come out sufficiently …

(AT XI 31)⁸

By the time Descartes wrote the Principles twelve years later, he had become more self-conscious about the hypothetical character of his assumptions concerning the size, shape, etc. of particles of matter.

From what has already been said we have established that all the bodies in the universe are composed of one and the same matter, which is divisible into indefinitely many parts, … However, we cannot determine by reason alone how big these pieces of matter are, or how fast they move, or what kinds of circle they describe. Since there are countless different configurations which God might have instituted here, experience alone must teach us which configurations he actually selected in preference to the rest. We are thus free to make any assumption on these matters with the sole proviso that all the consequences of our assumption must agree with our experience.

(AT VIIIA 100–1: CSM [The Philosophical Writings of Descartes; ed. J. G. Cottingham, R. Stoothoff, and D. Murdoch; two vols.; Cambridge University Press, 1985] I 256–7)

Descartes does not claim that we are completely free to assume anything we wish about matter. He argues at great length about the fundamental properties of matter, i.e. their primary qualities, and discusses in detail the need to include or exclude certain primary qualities in a viable theory of nature. He also argues in some detail about the laws of motion or, as he calls them, the laws of nature, which determine the motions of material bodies and the ways in which they may transfer motion from one to another by contact action. However, the relevant point here is that, having decided which variables to attribute to matter, we cannot determine by similar arguments the values of these variables; we cannot decide a priori the number, size, or speed of the various small parts of matter which underpin the whole edifice of Cartesian physics. Nor could we hope to discover by observation which particles there are, what shapes they have or with what speed they move; they are much too small to be perceived directly, even with the use of a microscope. We can do no better than hypothesize answers to these questions, and then subsequently check the plausibility of our guesswork.

Thus the logic of Descartes' theory of sensation and the implications of his theory of matter both suggest that he would have to acknowledge a central place for hypotheses in any coherent account of physical phenomena. The extent to which he recognized this varied from his earlier reflections in the Regulae (c.1628), in which there was only a minimal recognition of the role of hypotheses in natural science, to his more mature considerations in the Discourse (1637), where the significance of hypotheses and experiments is explicitly acknowledged. The Discourse is of paramount importance in this context, because it was composed over a number of years while Descartes was preparing for publication the three major scientific essays for which it serves as a preface. In the "Discourse on the method of rightly conducting one's reason and seeking the truth in the sciences," Part VI, Descartes writes:

Should anyone be shocked at first by some of the statements I make at the beginning of the Optics and the Meteorology because I call them 'suppositions' and do not seem to care about proving them, let him have the patience to read the whole book attentively, and I trust that he will be satisfied. For I take my reasonings to be so closely interconnected that just as the last are proved by the first, which are their causes, so the first are proved by the last, which are their effects…. For as experience makes most of these effects quite certain, the causes from which I deduce them serve not so much to prove them as to explain them; indeed, quite to the contrary, it is the causes which are proved by the effects.

(AT VI 76: CSM I 150)

This passage raised a number of queries from readers, one of whom was Father Morin. Descartes replied to his concerns in 1638 and answered the objection that hypothetical essays should not be described as demonstrated: "there is a big difference between proving and explaining. To this I add that one can use the word 'demonstrate' to mean one or the other, at least if one understands it according to common usage and not according to the special meaning which philosophers give it" (13 July 1638: AT II 198: CSMK [The Philosophical Writings of Descartes; ed. Cottingham, Stoothoff, Murdoch, and Anthony Kenny; Cambridge University Press, 1991] 106). This shows Descartes explicitly breaking with the scholastic tradition, for which the term "demonstrate" had special connotations of deducing a conclusion rigorously from first principles. Instead he invites his readers to understand "demonstration" in a less strict sense in which it can include the reasoning process by which one argues from effects to hypothetical causes or, in the opposite direction, from assumed causes to observed effects.

The relative novelty of this type of demonstration is underlined in a letter to Mersenne in 1638, in which Descartes explains that the types of demonstration available in physics are very different from those which one expects in mathematics:

You ask if I think that what I wrote about refraction is a demonstration; and I think it is, at least insofar as it is possible to give one in this matter, without having first demonstrated the principles of physics by means of metaphysics … and to the extent that any other question of mechanics, optics or astronomy, or any other matter which is not purely geometrical or arithmetical, has ever been demonstrated. But to demand that I give geometrical demonstrations in a matter which depends on physics is to demand the impossible. And if one wishes to call demonstrations only the proofs of geometers, one must then say that Archimedes never demonstrated anything in mechanics, nor Vitello in optics, nor Ptolemy in astronomy, and so on; this, however, is not what is said. For one is satisfied, in these matters, if the authors—having assumed certain things which are not manifestly contrary to experience—write consistently and without making logical mistakes, even if their assumptions are not exactly true…. But as regards those who wish to say that they do not believe what I wrote, because I deduced it from a number of assumptions which I did not prove, they do not know what they are asking for, nor what they ought to ask for.⁹

One implication is clear. We cannot expect the same kind of demonstrations in physics as in pure mathematics, and we will have to settle for something else. However, it is not yet clear what this alternative is. Whatever its precise structure and the kind of results which it can deliver, it involves making assumptions about the causes of physical phenomena and then "demonstrating" the plausibility of these assumptions by examining their explanatory role in some comprehensive natural philosophy, a project to which Descartes repeatedly refers in his claim that he could (at least in principle) demonstrate those assumptions from some kind of metaphysical foundation.

THE CONCEPT OF EXPLANATION

Descartes shared with many of his contemporaries the insight that the forms and qualities of the scholastic tradition were, in some fundamental sense, nonexplanatory. If we notice some natural phenomenon such as the effect of a magnet on small pieces of iron, the scholastic tradition tended to explain this by saying that the magnetic stone attracts (or repels) certain bodies because it has a "magnetic form" or a "magnetic quality." There is an obvious sense in which this is true. If any natural object does something, then it must have the capacity to do so! As long as we do not understand what that capacity is or what it consists in, we might name the inscrutable property in question in terms of the effect it produces. Then sleeping pills have a dormitive power, magnets have magnetic powers, and human beings have thinking powers. So far, there is nothing wrong with this; it merely labels what needs to be explained.

However, if one follows the natural tendency of scholastic philosophy and reifies these newly named powers as if they were properties distinct from the natural objects which have them, then two problems emerge. One is a metaphysical one; namely, the multiplication of entities beyond demonstrated necessity. By applying Occam's principle, one would stop short of introducing hundreds of new forms or qualities which overpopulate one's metaphysical space.¹⁰ Descartes adverts to this question about the redundancy of forms in Chapter 2 of Le Monde, where he explains how a piece of wood burns and, as it burns, emits light and heat:

someone else may, if he wishes, imagine the form of 'fire', the quality of 'heat', and the action that 'burns' it to be completely different things in this wood. For my part, afraid of misleading myself if I suppose anything more than what I see must of necessity be there, I am content to conceive there the motion of its parts…. provided only that you grant me that there is some power that violently removes the subtler of its parts and separates them from the grosser, I find that that alone will be able to cause in the wood all the same changes that one experiences when it burns.

(AT XI 7–8)¹¹

Secondly, the introduction of scholastic forms in this context gave the impression that one had made progress in explaining natural phenomena, and that little else remained to be done. However, the very forms which are assumed as explanatory entities are themselves in need of explanation: "If you find it strange that, in setting out these elements, I do not use the qualities called 'heat', 'cold', 'moistness', and 'dryness', as do the philosophers, I shall say to you that these qualities appear to me to be themselves in need of explanation" (AT XI 25–6).¹²

Thus, for Descartes, scholastic forms are both redundant and pseudo-explanatory. The alternative suggested was to find the material and efficient causes of natural phenomena. Descartes argued that these causes must be described mechanically; in fact, he notoriously argued in a reductionist way that most of the properties that natural phenomena exhibit can be explained ultimately in terms of the size, shape, and motions of the small parts of matter into which, he assumed, physical objects can be analyzed. Therefore to explain any natural phenomenon, in this sense, is equivalent to constructing a model of how small, imperceptible parts of matter can combine to form perceptible bodies, how the properties of bodies result from the properties of their constituent parts, and why we perceive them as we do as a result of the interaction of these bodies with our sensory organs.

It has already been indicated above that Cartesian scientific explanations must be hypothetical, and that one of the reasons for this admission was the unobservability of the particles of matter in terms of which the explanation of natural phenomena must be constructed. But how are we supposed to describe and measure the properties of unobservable particles of matter? Father Morin had this type of objection in mind when, having read the scientific essays of 1637, it seemed to him that Descartes might be attempting to explain what we can readily observe by reference to what we neither observe nor understand: " … problems in physics can rarely be resolved by analogies [comparaisons]; there is almost always some difference [between the model and reality], or some ambiguity, or some element of the obscure being explained by the more obscure" (12 August 1638: AT II 291). Part of Descartes' reply to this objection includes the claim that there is no way of proceeding in physics except by constructing large-scale models of what is happening at the microscopic level. Thus, for example, we might think of imperceptible particles of light by analogy with wooden spheres the size of billiard balls.

I claim that they [i.e. models and analogies] are the most appropriate way available to the human mind for explaining the truth about questions in physics; to such an extent that, if one assumes something about nature which cannot be explained by some analogy, I think that I have conclusively shown that it is false.

(12 September 1638: AT II 368: CSMK 122)

This point had already been made in correspondence with Plempius the previous year: "There is nothing more in keeping with reason than that we judge about those things which we do not perceive, because of their small size, by comparison and contrast with those which we see" (3 October 1637: AT I 421: CSMK 65). Descartes' reply to Father Morin also included the claim that the only relevant features of the model were the size and shape of the spheres, and the direction and speed of their motions, so that the disparity in size could be ignored in constructing an explanation.

in the analogies I use, I only compare some movements with others, or some shapes with others, etc.; that is to say, I compare those things which because of their small size are not accessible to our senses with those which are, and which do not differ from the former more than a large circle differs from a small one.

(12 September 1638: AT II 367–8: CSMK 122)

Apart from the interesting assumptions about which features of a model are relevant to constructing an explanation, Descartes' comments also raise a question about the extent to which hypotheses must be true in order to be explanatory. In other words, would it help in explaining a physical phenomenon if one constructed a mechanical model of its efficient cause which, in fact, is not true to the reality? Descartes thought so, or at least he argued that a plausible though incorrect model is better than none at all. Besides, it may be the case that we can never discover the values of the variables with which we describe microscopic particles of matter, so that we will have to settle for something less than the ideal understanding which is available to God.

The first concession about false hypotheses is made in a number of places where Descartes wonders about the evolution of the universe from its initial chaos to the highly structured world we see today. Theologians commonly believed in his day, based on a nonmetaphorical reading of Genesis, that the world as we see it had been created by God. Descartes comments:

even if in the beginning God had given the world only the form of a chaos, provided that he established the laws of nature and then lent his concurrence to enable nature to operate as it normally does, we may believe without impugning the miracle of creation that by this means alone all purely material things could in the course of time have come to be just as we now see them. And their nature is much easier to conceive if we see them develop gradually in this way than if we consider them only in their completed form.

(AT VI 45: CSM I 133–4)

This suggests that an explanation of the natural world is better if we imagine the world as gradually evolving from an initial chaos under the control of the laws of nature, than if we concede to the theologians' belief that God simply made it as it is. The same idea is expressed in the Principles:

There is no doubt that all the world was created with all of its perfection from the very beginning … Nevertheless, to understand the nature of plants or of man, it is much better to consider how they can gradually develop from seeds, than to consider how they were created by God at the beginning of the Universe. Thus if we can think of a few very simple and easily known principles from which we can show that the stars and the earth, and everything else we can observe on earth, could have developed as if from seeds—although we know they did not in fact develop in this way—we could explain their nature much better in this way than if we simply described them as they are now, or how we believe they were created.

(AT VIIIA 99–100: CSM I 256)

Thus, Descartes believed for theological reasons that his evolutionary account of the development of natural phenomena was false; he also claimed that, despite being false, it was explanatory.

The second reason for accepting hypotheses which are possibly false was Descartes' pessimism about the feasibility of identifying and accurately measuring relevant variables at the microlevel. There were a number of reasons for this which, in retrospect, would seem to have been well justified and would strike the modern reader as a realistic appraisal of the experimental techniques of the early seventeenth century. If one insisted on withholding hypotheses until all the complexity of the natural world is taken into account, one would make no progress whatsoever. Descartes argued along these lines in response to Mersenne's objections, in 1629, about the interference of the air in measuring the speed of falling bodies.

However, as regards the interference from the air which you wish me to take into consideration, I claim that it is impossible to cope with it and it does not fall within the scope of science; for if it is warm, or cold, or dry, or humid, or clear, or cloudy, or a thousand other circumstances, they can all change the air resistance.¹³

The same justification was offered, almost eighteen years later, for the apparent failure of the impact rules to coincide with our experience of colliding bodies. A number of correspondents objected that the rules proposed by Descartes in the Principles (Book II, arts. 46 ff) were contradicted by our experience. Descartes' response was:

Indeed, it often happens that experience can seem initially to be incompatible with the rules which I have just explained, but the reason for this is obvious. For the rules presuppose that the two bodies B and C are perfectly hard and are so separated from all other bodies that there is none other in their vicinity which could either help or hinder their movement. And we see no such situation in this world.

(AT IXB 93)

This was a standard reply to objections about a lack of fit between theory and reality. Cartesian explanations were constructed by analogy with the interactions of macroscopic physical bodies in motion. The underlying reality they purported to explain is microscopic, is inaccessible to human observation, and may involve so many interfering factors that our model is far short of adequately representing it.¹⁴

Thus a Cartesian explanation is a hypothesis that may be acknowledged to be either false or significantly inadequate to the reality it purports to explain. When we lack the evidence required to identify the actual cause of some phenomenon, "it suffices to imagine a cause which could produce the effect in question, even if it could have been produced by other causes and we do not know which is the true cause" (letter of 5 October 1646: AT IV 516). The suggestion that we settle for the best hypothesis available is reflected in the epistemic status claimed for various explanations in the Principles. For example, different astronomical hypotheses are examined, not to decide which one is true, but rather to find out which is more successful as an explanation: "Three different hypotheses, that is suggestions, have been discovered by astronomers, which are considered not as if they were true, but merely as suitable for explaining the phenomena" (AT VIIIA 85: CSM I 250). Descartes' preferred hypothesis is chosen "merely as a hypothesis and not as the truth of the matter" (AT VIIIA 86: CSM I 251).

Evidently it would be better if we could discover the true causes of natural phenomena; but if we cannot, it is still worth while to settle for a possible or plausible cause:

As far as particular effects are concerned, whenever we lack sufficient experiments to determine their true causes, we should be content to know some causes by which they could have been produced …

I believe that I have done enough if the causes which I have explained are such that all the effects which they could produce are found to be similar to those we see in the world, without inquiring whether they were in fact produced by those or by some other causes.

(AT IXB 185, 322)

The methodology suggested here, of constructing mechanical models as best we can, coincides with Cartesian scientific practice. Descartes and his followers in France in the seventeenth century were almost profligate in imagining hypothetical models to explain natural phenomena and, in some cases, to explain what could only be called alleged phenomena; they even constructed explanations of nonevents. It was this widespread and notorious dedication to unrestrained hypothesis construction that helps explain Newton's famous disclaimer: "I do not construct hypotheses."¹⁵

Yet, despite the fact that the logic of Descartes' philosophy implied that explanations of natural phenomena had to be hypothetical, there are equally clear intimations in his work of a very different methodology. Descartes often referred to the possibility of constructing a natural philosophy based on a metaphysical foundation that would realize the kind of certainty and unrevisability which is apparently at issue in the Meditations. This feature of his methodology needs some clarification before inquiring if it is compatible with the story told thus far.

FOUNDATIONS OF SCIENCE

In the Preface to the French edition of the Principles, Descartes introduces a metaphor that accurately expresses his views about the relationship of physics to metaphysics. "Thus the whole of philosophy is like a tree. The roots are metaphysics, the trunk is physics, and the branches emerging from the trunk are all the other sciences, which may be reduced to three principal ones, namely medicine, mechanics and morals" (AT IXB 14: CSM I 187). There was nothing unusual in this suggestion. Descartes had maintained for about twenty-five years prior to this that physics, as he understood it, is based on or depends on metaphysics and that any natural philosopher worth his salt had better get his metaphysics in order first, before tackling the explanation of specific natural phenomena. For example, he wrote to Mersenne in 1630 about a short essay on metaphysics he himself had begun to write: "It is there that I have tried to begin my studies; and I can tell you that I would not have been able to discover the foundations of physics if I had not looked for them in this direction" (15 April 1630: AT I 144). This helps explain why he objected to Galileo's methodology. According to Descartes, the Italian natural philosopher had ignored questions about foundations and had applied himself instead directly to explaining particular physical phenomena: "without having considered the first causes of nature, he [Galileo] has merely looked for the explanations of a few particular effects, and he has thereby built without foundations" (to Mersenne, 11 October 1638: AT II 380: CSMK 124). The question arises, therefore, about the kinds of foundations Descartes envisaged for physics, and the connection between those foundations and the various sciences that depend on them. One way of focusing on this issue is to contrast Descartes' approach with what is standard practice in modern science. Physicists or physiologists of the twentieth century do not begin their research with a study of metaphysics, although they may well make metaphysical assumptions in the course of constructing their theories. Instead, they first develop scientific theories which are tested for viability, and the metaphysical implications of the theories are subsequently read off from the finished scientific product. In this approach there is no independent criterion for the acceptability of ontological commitments, apart from the success or otherwise of a given theory. Descartes held the opposite view. He assumed that we can, and ought, to construct our metaphysics first, and that we should subsequently consider physical theories which are consistent with our metaphysical foundation. Thus there must be available independent criteria for deciding which metaphysics to adopt.

On this issue Descartes is very close to scholastic philosophy. The epistemic foundation of Cartesian metaphysics is reflection on "common sense" or on our everyday experience of the natural world. Rule II of the method proposed in the Discourse, which reflects Rule IX of the Regulae, was "to begin with the simplest and most easily known objects in order to ascend little by little, … to knowledge of the most complex" (AT VI 19: CSM I 120).¹⁶ Where metaphysics is concerned, we begin with such everyday experiences as the experience of thinking, of feeling, of moving, etc. Among these experiences, Descartes favors the most simple, accessible and widely available experiences because he hopes thereby to find indubitable foundations. This strategy was outlined in Part VI of the Discourse:

I also noticed, regarding observations, that the further we advance in our knowledge, the more necessary they become. At the beginning, rather than seeking those which are more unusual and highly contrived, it is better to resort only to those which, presenting themselves spontaneously to our senses, cannot be unknown to us if we reflect even a little. The reason for this is that the more unusual observations are apt to mislead us when we do not yet know the causes of the more common ones, and the factors on which they depend are almost always so special and so minute that it is very difficult to discern them.

(AT VI 63: CSM I 143)

The privileged position of everyday experience coincides with a complementary distrust of sophisticated experiments; the latter are likely to mislead us because they may be poorly executed, their results may be incorrectly interpreted, or they may be compromised by various interfering factors of which we are unaware.¹⁷ Therefore, experimental evidence is too unreliable to provide metaphysical foundations for scientific theories; that can only be done by reflection on ordinary experience.

The central claims of Cartesian metaphysics are summarized in the Meditations and in Part I of the Principles. While they are discussed elsewhere in this volume, the relevant feature here is the extent to which Descartes relies on a scholastic set of concepts to interpret metaphysically the personal experiences for which he claims indubitability. For example, the distinction between a substance and its modes is central to the Cartesian argument in favor of a radical distinction between things that can think and those that cannot.¹⁸ The same distinction is put to work in defining the essence of matter and in denuding matter of many of the primary qualities other natural philosophers were willing to attribute to it, such as gravity or elasticity. In summary, Descartes' metaphysics is a subtle combination of scholastic categories, metaphysical axioms (e.g., ex nihilo nihil fit), and apparently incontrovertible common experience.¹⁹

Once this foundation is in place, the second stage of theory construction is the formulation of the so-called "laws of nature." Despite the fact that these are said to be "deduced" from a metaphysical foundation, the evidence adduced in favor of the laws, both in Le Monde and the Principles, is a mixture of metaphysical axioms and everyday observation. For example the first law, to the effect that a material object continues in its condition of rest or motion unless some cause intervenes to change its condition, is partly justified by reference to the general axiom that every event or change requires a cause, and partly by reference to our everyday experience: "our everyday experience of projectiles completely confirms this first rule of ours" (AT VIIIA 63: CSM I 241).²⁰ The other two laws of nature are confirmed in the same manner, by appealing to metaphysical axioms and to our everyday experience of physical objects that move about in the world (AT VIIIA 64–5: CSM I 242).

Thus the metaphysical foundations Descartes claimed to establish for scientific knowledge included a number of related elements, which relied on the kind of the evidence just discussed: (a) a radical distinction between matter and spirit, and a preliminary identification of the primary qualities of matter. This included an equally confident dismissal of various properties which Descartes claimed matter does not have; (b) a rejection of the scholastic understanding of explanation and, in its place, the substitution of an uncompromising model of mechanical explanation; (c) a sketch of three fundamental laws of nature according to which material particles interact and exchange various quantities of motion.

Once these were in place, the question arose of how Descartes might make progress in constructing the type of mechanical models required by his method. What kind of inference was available to move from general principles to the explanation of specific natural phenomena?

Descartes' actual scientific practice coincided with his description of theory construction in Part VI of the Discourse. As he moved further away from general principles and closer to particular phenomena, he found he needed hypotheses and experimental tests:

First I tried to discover in general the principles or first causes of everything that exists or can exist in the world…. Next I examined the first and most ordinary effects deducible from these causes. In this way, it seems to me, I discovered the heavens, the stars, and an earth … and other such things which, being the most common of all and the simplest, are consequently the easiest to know. Then, when I sought to descend to more particular things, I encountered such a variety that I did not think the human mind could possibly distinguish the forms or species of bodies that are on the earth from an infinity of others that might be there if it had been God's will to put them there. Consequently I thought the only way … was to progress to the causes by way of the effects and to make use of many special observations…. I must also admit that the power of nature is so ample and so vast, and these principles so simple and so general, that I notice hardly any particular effect of which I do not know at once that it can be deduced from the principles in many different ways; and my greatest difficulty is usually to discover in which of these ways it depends on them. I know no other means to discover this than by seeking further observations whose outcomes vary according to which of these ways provides the correct explanation.

(AT VI 63–4: CSM I 143–4)

This text is clear in admitting that it is not possible to deduce, in an a priori manner, an explanation of particular natural phenomena from the very general laws of nature Descartes defended, because there is an almost infinite number of alternative paths—all consistent with the laws of nature—by which God might have caused particular natural phenomena. To discover which path he chose, i.e. to discover the mechanism by which natural phenomena are caused by the interaction of particles of matter, one has to have recourse to crucial experiments. And, as has been already acknowledged above, the results which can be gleaned by this method are still hypothetical.

However, Descartes is not consistent in acknowledging that hypothetical initiatives must remain hypothetical, and that they cannot be converted subsequently into something more like the purely formal deductions of mathematics. And, despite the need for experiments to help decide how a natural phenomenon occurs, he sometimes described the results of his scientific method in language which could almost have been taken directly from the section of Aristotle's Posterior Analytics quoted above: "As far as physics is concerned, I believed that I knew nothing at all if I could only say how things may be, without being able to prove that they could not be otherwise" (letter of 11 March 1640: AT III 39: CSMK 145). This raises a question about the kind of certainty Descartes claimed for the results of his scientific method when applied to natural phenomena.

CERTAINTY AND PROBABILITY

Descartes' claims about the relative certainty of scientific explanations are appropriately ambivalent. The ambivalence reflects the comparatively unsophisticated concepts of certainty and uncertainty available to the early seventeenth century. The scholastic tradition was committed to a sharp dichotomy between two kinds of knowledge-claim; one was certain and demonstrated, and the other was dialectical and uncertain. As far as scholastics were concerned, therefore, one had to choose between claiming to have demonstrated, certain knowledge—which was the only kind worth having—or the type of uncertain opinion which hardly deserved further discussion, since it was completely uncorroborated. Descartes' efforts to describe the degree of certainty that resulted from his scientific practice are best understood as a doomed attempt to classify the probability produced by the new scientific method in the language of the scholastics. Thus he sometimes claims that his explanations are certain; he cannot concede that they are uncertain without automatically excluding them as genuine alternatives to the established explanations of the schools. At the same time he recognizes that they are not absolutely certain, that they do not enjoy the type of certainty that can be realized in mathematics, that they are only morally certain or as certain as one could hope to be in this type of enterprise.²¹ Another compromise, consistent with the claims about a metaphysical foundation, is the argument that the first principles are certain whereas the explanations of particular natural phenomena are more or less uncertain.

Descartes consistently claims that his first principles, or the more general claims about matter and the laws of nature, are very certain.

as regards the other things I assumed which cannot be perceived by any sense, they are all so simple and so familiar, and even so few in number, that if you compare them with the diversity and marvellous artifice which is apparent in the structure of visible organs, you will have far more reason to suspect that, rather than include some which are not genuine, I have omitted some which are in fact at work in us. And knowing that nature always operates in the most simple and easy way possible, you will perhaps agree that it is impossible to find more plausible explanations of how it operates than those which are proposed here.

(AT XI 201)

This point was reiterated on a number of occasions; the basic hypotheses of the Cartesian system were said to be simple and relatively few, and at the same time they explained a great variety of disparate natural phenomena. "Simple" had connotations of being easily understood, possibly by analogy with some natural phenomenon with which we are ordinarily familiar. It also implied that a hypothesis was consistent with the limited categories available in Cartesian natural philosophy, such as size, speed, and quantity of motion. In other words, it was possible to imagine or construct a mechanical model of a so-called "simple" hypothesis, whereas the kinds of explanations proposed by others were allegedly difficult to understand, not amenable to simple modeling, and probably expressed in the metaphysical language of the schools. Thus he wrote in Part III of the Principles: "I do not think that it is possible to think up any alternative principles for explaining the real world that are simpler, or easier to understand, or even more probable" (AT VIIIA 102: CSM I 257).

Descartes was aware of the objection that one could construct a hypothesis to explain any conceivable phenomenon and that, as a result, hypotheses could be accused of being ad hoc. His answer to this objection included a number of elements. One was that he used only a few hypotheses to explain many different phenomena: "it seems to me that my explanations should be all the more accepted, in proportion as I make them depend on fewer things" (AT VI 239). Given the few principles from which he begins, the variety of phenomena which are explained provides an extra degree of confirmation.

In order to come to know the true nature of this visible world, it is not enough to find causes which provide an explanation of what we see far off in the heavens; the selfsame causes must also allow everything which we see right here on earth to be deduced from them. There is, however, no need for us to consider all these terrestrial phenomena in order to determine the causes of more general things. But we shall know that we have determined such causes correctly afterwards, when we notice that they serve to explain not only the effects which we were originally looking at, but all these other phenomena, which we were not thinking of beforehand.

(AT VIIIA 98–9: CSM I 255)

Apart from the points just mentioned, Descartes also argued that the new natural philosophy should be compared, not with some abstract criterion of what counts as a good theory, but with other theories available in the 1630s to explain the same range of phenomena. In that context, Cartesian science was claimed to be the best available. This is clear from a letter to Father Morin of 13 July 1638:

Finally, you say that there is nothing easier than to fit some cause to any given effect. But although there are indeed many effects to which it is easy to fit different causes, one to one, it is not so easy to fit a single cause to many different effects, unless it is the true cause which produces them. There are often effects where, in order to prove which is their true cause, it is enough to suggest a cause from which they can all be clearly deduced. And I claim that all the causes which I have discussed are of this type … If one compares the assumptions of others with my own, that is, all their real qualities, their substantial forms, their elements and similar things which are almost infinite in number, with this one assumption that all bodies are composed of parts—something which can be observed with the naked eye in some cases and can be proved by an unlimited number of reasons in others … and finally, if one compares what I have deduced about vision, salt, winds, clouds, snow, thunder, the rainbow, and so on from my assumptions, with what they have deduced from theirs … I hope that would suffice to convince those with an open mind that the effects which I explain have no other causes apart from those from which I deduce them.

(AT II 199–200: CSMK 107)

The conclusion of the Principles repeats the same claim; if a few assumptions can explain a wide variety of disparate phenomena, then that argues well for their plausibility:

Now if people look at all the many properties relating to magnetism, fire and the fabric of the entire world, which I have deduced in this book from just a few principles, then, even if they think that my assumption of these principles was arbitrary and groundless, they will still perhaps acknowledge that it would hardly have been possible for so many items to fit into a coherent pattern if the original principles had been false.

(AT VIIIA 328: CSM I 290)

If we accept the point being made, that a few basic hypotheses are put to work in explaining all the natural phenomena mentioned, what degree of certainty should Descartes claim for his first principles? Not surprisingly, one finds two rather different claims in this context: one of them concedes that the confirmed principles are only more or less probable, whereas the other assumes that they are certain and demonstrated. The more modest claim is found in a letter to an unknown correspondent, written about 1646: "I would not dare claim that those [principles] are the true principles of nature. All I claim is that, by assuming them as principles, I have satisfied myself in all the many things which depend on them. And I see nothing which prevents me from making some progress in the knowledge of the truth" (AT IV 690). The more confident claim about moral and metaphysical certainty comes in the penultimate article of the Principles:

there are some matters, even in relation to the things in nature, which we regard as absolutely, and more than just morally, certain.… This certainty is based on a metaphysical foundation … Mathematical demonstrations have this kind of certainty, as does the knowledge that material things exist; and the same goes for all evident reasoning about material things. And perhaps even these results of mine will be allowed into the class of absolute certainties, if people consider how they have been deduced in an unbroken chain from the first and simplest principles of human knowledge…. it seems that all the other phenomena, or at least the general features of the universe and the earth which I have described, can hardly be intelligibly explained except in the way I have suggested.

(AT VIIIA 328–9: CSM I 290–1)

The French version of this text is even more explicit on the demonstrative character of the explanations found in Cartesian physics:

I think that one should also recognise that I proved, by a mathematical demonstration, all those things which I wrote, at least the more general things concerning the structure of the heavens and the earth, and in the way in which I wrote them. For I took care to propose as doubtful all those things which I thought were such.

(AT IXB 325)

The problem of classifying the type of certainty Descartes might reasonably have claimed for his principles and hypotheses is best understood historically, by taking account of the lack of a concept of probability in the early part of the seventeenth century and of the assumption of the scholastic tradition that anything less than demonstrated truths was as unreliable as mere opinion or guesswork. In this context, Descartes claimed that his natural philosophy was certain and demonstrated; at the same time, realizing that it could hardly be as certain as the formal proofs of mathematics, he conceded that only the more general assumptions of his system were certain, whereas the explanations of particular natural phenomena were more or less certain.

This point reopens the question about the kind of evidence Descartes thought was appropriate to supporting scientific claims, and the relative importance of metaphysical arguments vis-à-vis experiential evidence. There is no suggestion that Descartes ever reneged on the conviction, so clear in the Meditations, that one can realize a degree of certainty which is equivalent to indubitability by reasoning about concepts and axioms. This kind of metaphysical certainty is appropriate to the foundations of our knowledge, whether that knowledge is mathematical, physical, or otherwise.

However, if we wish to make judgments about the physical world, then we cannot assume naively that our sensations reflect the way the world is. Nor can we discover in any detail what kind of natural phenomena occur, nor what mechanisms explain their occurrence, by introspecting our ideas. There has to be some provision, therefore, for beginning with clear and distinct metaphysical concepts and axioms and somehow making the crucial transition to describing and explaining the natural world around us. This can be done only by consulting our experience of the natural world, and this implies that we use our senses in order to gain scientific knowledge.

At the same time, Descartes can be correctly described as a critic of the reliability of empirical evidence. His critique was carefully developed to identify a number of ways in which we might draw erroneous conclusions from our sensory experience. Two of these have already been identified: (a) We might ignore the distinction between primary and secondary qualities and, as a result, assume that our sensations resemble the causes of our sensations; and (b) we might argue too hastily from an experiment to some conclusion without taking account of the many ways in which an experiment can mislead. In general, we are in danger of spontaneously making naive, uncritical judgments about the physical world without questioning the reliability of our sensations or the logic of conclusions drawn from reliable observations. Such spontaneous judgments should be distinguished from other judgments, equally based on sensation, which we make after due deliberation and reflection. Unfortunately for the modern reader, Descartes expressed this distinction in terms of a contrast between experience and reason; what he meant was a contrast between two types of judgment, both equally based on experience. This is made explicit in the following text:

It is clear from this that when we say 'The reliability of the intellect is much greater than that of the senses,' this means merely that when we are grown up the judgments which we make as a result of various new observations are more reliable than those which we formed without any reflection in our early childhood; and this is undoubtedly true.

(Sixth Replies: AT VII 438: CSM II 295)

For this reason, a true philosopher "should never rely on the senses, that is, on the ill-considered judgments of his childhood, in preference to his mature powers of reason" (AT VIIIA 39: CSM I 232).

It is obvious, then, that one cannot avoid the necessity of relying on experientially based evidence. Descartes acknowledges the need for this kind of evidence in natural philosophy and uses it extensively in the scientific experiments which he describes. He says openly, in Part VI of the Discourse, "regarding observations, that the further we advance in our knowledge, the more necessary they become" (AT VI 63: CSM I 143). On this point, his scientific practice corresponded with his methodological rule, for he spent much more time doing experiments or reading about those done by others than he ever spent in mere thinking. However, for reasons already mentioned, he had little confidence in experiments he had not checked himself.²² Hence there were serious limits to the extent to which he could hope to complete a comprehensive explanation of nature; he was likely to be frustrated "by the brevity of life or the lack of observations" (AT VI 62: CSM I 143). For this reason, Descartes decided to devote his life to the pursuit of what he called a "practical philosophy which might replace the speculative philosophy taught in the schools" (AT VI 61: CSM I 142). "I will say only that I have resolved to devote the rest of my life to nothing other than trying to acquire some knowledge of nature from which we may derive rules in medicine which are more reliable than those we have had up till now" (AT VI 78: CSM I 151). This is equivalent to a commitment to doing experiments, the cost of which he often complained of. To attempt to gain this practical knowledge in any other way, apart from experimentally, would be to join those "philosophers who neglect experience and think that the truth will emerge from their own heads as Minerva did from that of Jupiter" (Regulae Rule V: AT X 380).

A full account of the contribution of Descartes to the history of philosophies of science would involve examining his work in the light of his successors in the seventeenth century. Without examining this supplementary evidence here—which would include the ways in which Descartes was understood by, for example, La Forge, Malebranche, Rohault, Poisson, Cordemoy and Régis—there is reason to believe that his successors shared a common interpretation of the main features of Descartes' philosophy of science.²³ These common features are best understood in contrast with the scholastic philosophy for which they were proposed as a substitute. For Descartes, the contrast was between the practical and the speculative, the explanatory and the nonexplanatory, the critical and the naively uncritical, the mechanistic and the formal, the mathematical and quantitative versus the qualitative. Despite the favorable contrast with the natural philosophy of the schools, however, Descartes continued to accept the scholastic assumption that we should construct our metaphysics first, on the epistemic basis of reflection on ordinary experience, and that any subsequent explanations of natural phenomena must be consistent with the foundational metaphysics.

Once the foundations were in place, it was accepted that we could never know the way the world is by consulting our sensations and inferring from them that the causes of our sensations must resemble our subjective experiences. Besides, if we assume that physical phenomena are constituted by the interactions of very small particles of matter, then the sheer size of such particles of infinitely divisible matter would put their observation beyond our reach. For these two reasons, we can only come to know how the physical world is by hypothesis.

For Descartes, to explain a natural phenomenon is not to redescribe it in the language of forms and qualities, as was done in the schools. To explain, in this context, is to construct a mechanical model of how the phenomenon in question is caused. This model construction is necessarily hypothetical. So, beginning with the basic laws of nature and the metaphysical foundations established in the Meditations or in Book of the Principles, Descartes set out to construct the kind of models his concept of explanation demanded. Although he continued to claim absolute certainty for the foundations, it was clear that he could not be as confident about the more detailed explanations of natural phenomena. These explanations depended on observations, and on performing complex experiments the interpretation of which introduced new reasons for doubt. There was also another reason for caution which emerged at this stage, namely Descartes' skepticism about the possibility of ever identifying the multiplicity of variables involved in any complex natural phenomenon. What begins on "indubitable" foundations, therefore, quickly gets mired in the almost immeasurably complex detail of unobservable particles of matter interacting at unobservable speeds. The crucial experiments which we perform to help choose the most plausible explanation are open to various interpretations. Hence the birth of the well-known Cartesian tradition of simply imagining some mechanism by which small parts of matter in motion might have caused some natural phenomenon which we observe.

To those who objected: this does not result in the kind of demonstrated knowledge prized by the scholastic tradition, Descartes replied that those who demand such demonstrations do not know what they are looking for, nor what they ought to look for. It is not possible to realize the same kind of certainty in physics as in mathematics or metaphysics. We have to settle for less.

This suggests that Descartes' philosophy of science was very much a product of the time in which it was developed. The 1630s and 1640s were a time of transition from the science of forms and qualities to what we describe now as modern science. One finds features of both of these philosophies of science in Descartes. What was significantly new was the commitment to mechanical explanation rather than the "occult powers" of the scholastic tradition, and the recognition that this type of explanation must be hypothetical. But for Descartes, lacking a theory of probability, this seemed compatible with the continued claim that his natural philosophy was not only superior in explanatory power to that of the schools, but that it was just as certain; or at least, that its more fundamental principles were demonstrated.

Notes

¹Posterior Analytics, 71b 8–12, 73a 21–2.

² The extent to which scholastic philosophy influenced the curriculum of colleges and universities in France in the seventeenth century is comprehensively documented in Brockliss, French Higher Education in the Seventeenth and Eighteenth Centuries.

³ This was summarized in the axiom: "nihil est in intellectu quod prius non fuit in sensu." French Cartesians in the period immediately after Descartes understood his theory of innate ideas as, in part, a response to what they considered to be a generally accepted scholastic doctrine, that all ideas derive originally from sensation. See, for example, Poisson, Commentaire ou remarques sur la méthode de M. Descartes, unpaginated preface, which discusses the "famous principle on which depends some of the dogmas of scholasticism, that nothing enters the mind which does not pass first through the senses." The same doctrine is discussed at some length on pp. 124–38. Cf. Le Grand, An Entire Body of Philosophy, p. 4. Among scholastic defenders of the thesis, even after Descartes, see Huet, Censura Philosophiae Cartesianae, pp. 51–3.

⁴ Even dedicated Cartesians, such as Jacques Rohault, continued the tradition of explaining natural phenomena in terms of matter and form. See Rohault, A System of Natural Philosophy, translated by J. Clarke, pp. 21–2. The original French text was published in 1671.

⁵ Mahoney (trans.), The World, pp. 1–3.

⁶ Mahoney, The World, p. 5.

⁷ Mahoney, The World, p. 7.

⁸ Mahoney, The World, p. 49.

⁹ Letter to Mersenne, 27 May 1638 (AT II 141–2, 143–4:CSMK 103). The same use of the word "demonstration" is found in Descartes' letter to Plempius, 3 October 1637 (AT I 420:CSMK 64).

¹⁰ The principle of parsimony in metaphysics, that one should not postulate the existence of more distinct entities or types of entity than is necessary, is usually attributed to William of Occam (1280?-1349?). See for example his Quodlibeta V, Q.1

¹¹ Mahoney, The World, p. 9.

¹² Mahoney, The World, p. 39.

¹³ Although the letter was written in French, the italicized phrase was in Latin: sub scientiam non cadit. Descartes to Mersenne, 13 November 1629 (AT I 73). See also Descartes to Mersenne, 11 June 1640 (AT III 80); Descartes to Cavendish, 15 May 1646 (AT IV 416–17).

¹⁴ Cf. similar responses to Mersenne, 23 February 1643 (AT III 634) and 26 April 1643 (AT III 652).

¹⁵ In the original Latin text, "hypotheses non fingo." Isaac Newton, Mathematical Principles of Natural Philosophy and His System of the World, ed. Cajori, p. 547.

¹⁶ Cf. Rule Nine of the Regulae: AT X 400: CSM I 33.

¹⁷ Descartes frequently pointed to problems in interpreting experimental results, especially when they seemed to disconfirm his own theories. However, the objections he raised were, in principle, legitimate. See, for example, Descartes to Mersenne, 9 February 1639 (AT II 497–8), 29 January 1640 (AT III 7), 11 June 1640 (AT III 80), 4 January 1643 (AT III 609).

¹⁸ Cf. Principles Part I, arts. 51–7: AT VIIIA 24–7: CSM I 210–12.

¹⁹ In the Third Meditation, Descartes argues that "something cannot arise from nothing" (nec posse aliquid a nihilo fieri) (AT VII 40: CSM II 28). In the Second Replies to Objections, he says that the causal principle on which he relied in the Third Meditation was equivalent to "nothing comes from nothing" (a nihilo nihil fit) (AT VII 135: CSM II 97).

²⁰ Cf. Mahoney, The World pp. 61–76: AT XI 38–47.

²¹ There was a tradition in scholastic philosophy and theology of distinguishing various degrees of certainty in terms of the kind of evidence required to achieve them and the relative importance of acting on our beliefs in different contexts. "Moral certainty" referred to the certainty required for important human actions, such as marrying one's partner or defending oneself against an aggressor. In this type of case, one does not usually have mathematical certainty about various relevant features of the context, but one is sufficiently certain to act and to be excused of responsibility if, despite taking normal precautions, one is mistaken. Cf. French version of Principles, Part IV, art. 205: "moral certainty is certainty which is sufficient to regulate our behaviour, or which measures up to the certainty we have on matters relating to the conduct of life which we never normally doubt, though we know that it is possible, absolutely speaking, that they may be false" (CSM I 289).

²² "I have little trust in experiments which I have not performed myself (letter to Huygens of 1643: AT III 617).

²³ For an analysis of how these authors understood Descartes' philosophy of science, see Clarke Occult Powers and Hypotheses.