Cartesian skepticism and empirical beliefs

In the Dictionary of Philosophy of Mind, Cartesian skepticism is defined as follows:  Any of a class of skeptical views against empirical knowledge based on the claim that claims to empirical knowledge are defeated by the possibility that we might be deceived insofar as we might be, for example, dreaming, hallucinating, deceived by demons, or brains in vats.

Pete Mandik , explains that the gist of Cartesian-style skeptical arguments is that some empirical proposition (e.g. that there are trees) cannot be known because we might be deceived (e.g. we might be brains in vats hallucinating that there are trees). In addition, he observes that there are related forms of these arguments, which attack our justification for believing some empirical proposition on grounds of possible deception. These 'justification' versions are intended to undermine claims to knowledge on the base that justification is a necessary condition on knowledge . Under the assumption that the arguments of the 'knowledge' variety and the arguments of the 'justification' variety all have the same general form in common, Pete Mandik gives an example to be taken as a representative of the whole class of examples which can be engendered by this general form. Though it is of the 'knowledge' variety, it can easily be transformed into an argument of the 'justification' variety by simply replacing all occurrences of 'knowledge' with 'justification', he says.

The general form is

A. I know that p only if I know that I am not deceived that p.

B. I do not know that I am not deceived that p.

C. Therefore, I do not know that p.

Where p has to be replaced by an empirical proposition.

The Cartesian-style skeptical argument Mandik  presents as an example, is the following:

A’. I know that there are trees only if I know that I am not deceived that there are trees.

B’. I do not know that I am not deceived that there are trees.

C’. Therefore, I do not know that there are trees.

It seems pretty clear that this is a valid argument. But is it sound? Mandik argues in defense of the inference inquiring into the truth of the premises.

What I find it questionable here is that a discussion about such a model-argument, when its soundness is at issue (note: not its validity), can neglect any consideration concerning the surrounding contexts of its actual use. If we distinguish the case of our ordinary language from the case of theoretical or scientific discourse, for instance, will the answer we are seeking remain the same ? In the following lines, I’m going to assume this distinction and inquiry into the premises from this point of view (for my purposes will be sufficient to examine the above premise A’). In particular, I’m going to focus on the first case and show that when the range of the inference is our ordinary language, the inference is unsound. The conclusion has implications regarding Cartesian skepticism and in the last section I’ll try to sketch them.

Let’s start, now, with fixing the first step : the range of the inference is our natural language.

Since if premise A’ doesn’t hold, the argument won’t hold, we’ll turn to premise A’, as said, concentrating our attention on it.

Assumption *: a) knowledge of a proposition p entails the true of p; b) deception about p entails the falsity of p. [Assumption * gives for granted what Mandik says about the relation holding between knowledge or deception of a proposition and the true or the falsity of it].

To discuss premise A’, I’m going to present three examples out of our ordinary language, each one featuring:  a proposition p; the two schemas: “I know that …”, “I’m deceived that … “ – where the symbol ‘ … ‘ has to be replaced by the proposition p; and a short suggestion about what the true or the falsity of p presupposes.

1.     p = “There are red apples on the table”. “I know that there are red apples on the table” means: ‘p’ is true i.e.: there are apples on the table and they are red. “I’m deceived that there are red apples on the table” means: there are apples on the table, but they are not red (maybe they are green). Note what the true or the falsity of p presupposes: ‘There are apples’; i.e.: the set of apples is not empty; there exist apples.

2.     p = “There are apples on the table”. The deception of p means: there are no apples on the table (maybe there are oranges or no fruits at all). Here again there must be apples, for p to be true or false.

3.     p = “There are apples”. The true of p means: the set of apples has at least one element. So the falsity of p means: the set of apples is empty. The (3) presupposes that a set, though named, could have no elements. Therefore, when I say: “I know that there are apples”, I’m accepting the idea that that part of the world I’m referring to could be empty.

If in the above general form we replace p with “There are red apples on the table” or “There are apples on the table”, the argument will turn sound. But to attain such a success will surely not satisfy the Skeptic. Since both the two examples presuppose that “There are apples” is true, the whole argument will sum up to say that there are apples, but we are not able to state a true proposition about apples. Skepticism is much more general and radical a view than this. For the above general form to yield a Skeptic argument, p has to be replaced by the proposition of example 3 or the like. As a matter of fact, Mandik goes this way and it is at this point that the decisive objection arises. To be properly construed, the Skeptic argument has to attribute to the ordinary speaker the idea that that part of the world he’s referring to could be empty.

That a reference to an empty set can positively play a role in our everyday language is highly questionable. The subject is well known and in philosophical literature one can find many contributions about it. Strawson[1952] – see below, for example, discusses the problem when he takes under consideration some inferences allowed by Aristotelian logic, which seem to need the hypothesis of an “existential assumption” to be valid. The same “existential assumption” seems presupposed by our everyday language, he says. If someone tells you that “ All the apples on the table are John’s” and you find out that there are not apples on the table, what will you think about ? In all likelihood, you won’t say that the proposition was false, or that it was true, you will instead notice that there are no apples at all. Maybe the speaker was mistaken or he was joking. The fact that we answer this way means it is natural for us to give for granted that what we are talking about there is.

Since premise A’ implies the idea of an empty set, it can’t belong to our everyday language. The whole argument can’t be formulated within our everyday language, where it would be nonsense.

Criticism against the theory of “existential assumption” confirms what I’m trying to say here. This criticism consists in observing that if we admitted the theory we would impose dangerous restrictions to scientific work. Newton’s first law of motion, for instance, states that some things are true about objects which don’t depend on any external force: if they are at rest they tend to stay at rest and if in uniform motion they tend to remain in that state of motion unless an external force is applied to them. The law can be true and a physicist may want to express it and defend it without being committed to presupposing any actual existence of objects which an external force is not applied to. Theoretical abstraction and scientific generalisation need a very special flexibility of mind as far as to take under consideration possibilities which are excluded by our ordinary view of the world.

In the well known wax example of Meditation II, Descartes explains this difference as a difference between imagining and conceiving. Let’s read the passage:

‹‹ [ … ] But, while I am speaking, let it [the piece of wax ] be placed near the fire--what remained of the taste exhales, the smell evaporates, the color changes, its figure is destroyed, its size increases, it becomes liquid, it grows hot, it can hardly be handled, and, although struck upon, it emits no sound. Does the same wax still remain after this change ? It must be admitted that it does remain; no one doubts it, or judges otherwise. What, then, was it I knew with so much distinctness in the piece of wax? Assuredly, it could be nothing of all that I observed by means of the senses, since all the things that fell under taste, smell, sight, touch, and hearing are changed, and yet the same wax remains.
It was perhaps what I now think, viz, that this wax was neither the sweetness of honey, the pleasant odor of flowers, the whiteness, the figure, nor the sound, but only a body that a little before appeared to me conspicuous under these forms, and which is now perceived under others.
But, to speak precisely, what is it that I imagine when I think of it in this way? Let it be attentively considered, and, retrenching all that does not belong to the wax, let us see what remains. There certainly remains nothing, except something extended, flexible, and movable. But what is meant by flexible and movable ? Is it not that I imagine that the piece of wax, being round, is capable of becoming square, or of passing from a square into a triangular figure ? Assuredly such is not the case, because I conceive that it admits of an infinity of similar changes; and I am, moreover, unable to compass this infinity by imagination, and consequently this conception which I have of the wax is not the product of the faculty of imagination. But what now is this extension ? Is it not also unknown ? for it becomes greater when the wax is melted, greater when it is boiled, and greater still when the heat increases; and I should not conceive [clearly and] according to truth, the wax as it is, if I did not suppose that the piece we are considering admitted even of a wider variety of extension than I ever imagined, I must, therefore, admit that I cannot even comprehend by imagination what the piece of wax is, and that it is the mind alone ( mens, Lat., entendement, F.) which perceives it. ››

Cartesian skepticism does not aim to question empirical beliefs in general as far as our everyday life and conduct are concerned. But are they also an adequate basis for scientific purposes ?

Antonio Calonego


Strawson, F. H. (1952). Introduction to Logical Theory, Metheuen & Co., London.

Descartes, R. (1641)  Meditationes de Prima Philosophia.


| Dialegesthai |