Both Descartes and Leibniz are on record as maintaining that acting freely requires that the agent 'could have done otherwise.' However, it is not clear how they could maintain this, given their other metaphysical commitments. In Leibniz's case, the arguments connected with this are well-rehearsed: it is argued, for example, that Leibnizian doctrines such as the Principle of Sufficient Reason and the thesis that God must will the best possible world preclude that the human could ever do other than she did.¹ The question of whether Descartes can maintain that the agent is able to do otherwise in the face of his wider metaphysical commitments has received comparatively little attention. However, Chappell has recently noted that Descartes's thesis that God is the 'total cause' of everything seems to preclude the possibility of human freedom (where this includes the ability to do otherwise).² [End Page 387] 


This paper compares a particular attempt by Leibniz to reconcile his wider metaphysics with the human agent's ability to do otherwise, with an attempt by Descartes to do the same which involves similar moves. It begins with a brief exploration of Leibniz's and Descartes's respective accounts of free-will, and how their claims that the free agent is 'able to do otherwise' apparently conflict with their other metaphysical commitments. I then examine a fairly prominent attempt by Leibniz to defend this ability to do otherwise, centering on the claim that the agent is able to do otherwise insofar as she has alternatives that are 'possible-in-themselves.' An apparently insignificant Cartesian defense of the ability to do otherwise is then considered, and I show that this defense follows a similar strategy. Finally, it is argued that this particular kind of defense is ultimately unsuccessful for Leibniz, but may be more successful for Descartes.


I Leibniz on the ability to do otherwise


In the Theodicy, Leibniz lists three conditions that must obtain for there to be human freedom: intelligence, which involves 'knowledge of the object of deliberation'; spontaneity, which 'belongs to us insofar as we have within us the source of our actions'; and contingency, where this involves that a claim that 'agent A wills that X' is not a metaphysically or logically necessary truth.³ 


We may look more closely at the third requirement of contingency. For Leibniz, a proposition is necessary⁴ iff its denial is a contradiction. Thus, the proposition [End Page 388] 


1. Agent A wills that X


is necessary if there are no conditions under which it is true that


2. ∼ (Agent A wills that X).

Now (1) would not be necessary if agent A need not have existed. If agent A need not have existed, then (2) would be possibly true — in which case (1) would be contingent. However, the agent's possible non-existence does not by itself satisfy Leibniz's contingency requirement in respect of human freedom. In the Discourse on Metaphysics, Leibniz declares that freedom involves that 'the will has the power to act otherwise or to suspend its action completely.'⁵ 


There are thus two senses in which Agent A's willing of X is a contingent act. In the broader sense, the act is contingent insofar as a denial of (1) is not a self-contradiction. In this sense, contingency is satisfied either by the agent's possible non-existence or the agent's being able to do otherwise. In the narrower sense, the act is contingent insofar as the agent is able to do otherwise than she did. The Leibnizian contingency requirement for freedom is satisfied only if the act is contingent in this narrower sense. For Leibniz, then, human freedom requires intelligence, spontaneity and the ability to do otherwise. (When I refer henceforth to the contingency requirement for freedom, I mean specifically the narrower requirement that...