One objection in this book acquired a name of its own. It appears as a charge of circular reasoning in the Fourth Objections. Mersenne’s group had already raised the underlying difficulty. Both exchanges are in the volume Descartes sent to press.

What is supposed to go in a circle?

Here is the route as Arnauld lays it out. Whatever I perceive clearly and distinctly is true. On that basis I establish that God exists. Since God does not deceive, what I perceive clearly and distinctly must indeed be true.

Arnauld wonders how Descartes avoids a circle. We are certain that clear and distinct perceptions are true only because God exists; yet we are certain of God’s existence only because we perceive it clearly and distinctly. Must we not already trust that kind of perception before establishing God?

The problem came to be known as the Cartesian Circle. It has outlived both men and remains a subject of philosophical dispute.

The person who objected, and the person who printed it

What matters here is where the objection appears. It was not an embarrassment dug up centuries afterwards. It arrived in a solicited report before publication, and the author placed it after his own main text.

Arnauld’s manner is courteous. He does not announce that he has caught a fatal error. He says one doubt remains. He has divided his report between philosophy and theology; this doubt closes the philosophical portion.

Then he adds something else he had nearly forgotten: Descartes’ claim that the mind is aware of everything within it seems wrong. Consider a child in the womb.

A challenge to the foundations of the argument is followed by an additional item almost as an afterthought. It is the voice of someone doing the job he was asked to do.

The distinction in the reply

Descartes answers within the same volume. He separates two conditions: attending to something clearly now, and remembering that one attended to it clearly before.

On his account, the first does not require an appeal to God. While the mind attends to the reasons establishing God’s existence, their evident force is present. The difficulty concerns what happens afterwards, when one is no longer considering the reasons and remembers having understood them.

That later confidence requires a guarantee. God is invoked for it, not to supply the immediate force of the original insight. In Descartes’ view, distinguishing the two prevents the supposed circle from closing.

Whether the answer succeeds is not decided here. There is something else to notice about how it is delivered.

A book that sends you back through itself

When he replies to Arnauld, Descartes says he has already explained the point sufficiently in the third and fourth sections of his response to the Second Objections. Rather than start over, he sends the reader to an earlier reply.

That reply answers a difficulty raised by Mersenne’s group. If nothing can be known with certainty before God’s existence is known, how can Descartes already be certain that he is a thinking thing? And what about an atheist geometer who clearly understands that a triangle’s angles equal two right angles?

Replying to them, Descartes again says he has already made the point clear. He means the certainty of conclusions remembered after their grounds are no longer before the mind. The earlier place he invokes is the Fifth Meditation.

The path therefore runs backwards. A distinction appears in the Fifth Meditation. Mersenne’s group challenges the position, and the reply points back to that distinction. Arnauld raises the difficulty again as a circle; Descartes points back to the answer he gave Mersenne.

These are not merely detachable appendices. Objections, replies and meditations refer into one another. The author has already asked us to read the exchanges before passing judgment on the main text. Skipping them means missing part of the argument the book asks us to assess.

A reference, and a gas tap

The alleged circle has an everyday counterpart. You ask whether someone is trustworthy and consult a reference they themselves supplied. That alone need not be circular: the reference may genuinely know them. The circle begins if you trust the referee only because this person recommends them, and trust the person only because the referee recommends them. Neither judgment has an independent support.

The distinction in Descartes’ answer also has an everyday counterpart. Two streets from home, you cannot remember whether you turned off the gas. You did turn it off, and watched yourself do so. But what you have now is a memory.

The difference between seeing the closed tap then and recalling it now is more than elapsed time. One person goes back because the memory does not settle the matter; another keeps walking because it does.

Descartes wants the later confidence to have a secure footing. Underneath it, he places God. The household example makes the distinction visible; it does not vindicate his guarantee.

At the end

Later readers gave the difficulty a name. Two groups asked it, and Descartes printed both. Arnauld’s courteous opening and his next, almost casual objection remain there too.

To answer Arnauld he directs us to Mersenne’s exchange; that exchange directs us to a meditation. Before judging the final answer, we have to follow the route by which it arrives.