70 Classical Compatibilism and Incompatibilism
8 Some might object to the expression, “doing what one wants,” thinking it better
instead to speak in terms of doing what one chooses, decides, tries, or wills. But
most classical compatibilists seemed to share a principle of motivation according to
which an agent always chooses, decides, tries, wills, or does what she has the
strongest desire to do (when informed by beliefs about what desires can be satis-
fied). Hence, on the classical compatibilist approach, doing what one chooses,
decides, tries, or wills, translates into doing what one wants. Furthermore, the
formulation “doing what one wants” also has expository advantages when discuss-
ing Harry Frankfurt’s successor to the classical compatibilist account, a successor
put explicitly in terms of wants (or desires), and clearly intended as an improve-
ment building upon the old classical compatibilist account of moral freedom. (We’ll
attend to Frankfurt’s views in Chapter 9.)
9 Harry Frankfurt’s (1971) hierarchical compatibilist theory (to be discussed in a later
chapter), while not a classical compatibilist theory, can be used to address this
problem while preserving the basic classical compatibilist idea that the freedom of
will is deeply connected with an agent’s doing what she wants.
10 David Hume, 1748, An Enquiry Concerning Human Understanding, Section VIII: 71.
11 Hume, A Treatise of Human Nature, Book II, Part III, Section II.
12 Hume, A Treatise of Human Nature, Book II, Part III, Section II.
13 Recall that in Section 2.3 we considered formulations of free will problems, F7 and
F8, arising from the worry that indeterminism undermines free will. Here one can see
how the classical compatibilists exploit the worry to advance a positive compatibilist
thesis that freedom requires determination.
14 David Hume, An Enquiry Concerning Human Understanding, Section VIII, Part I: 73.
15 What is meant by an analysis? This is a subject of great subtlety and controversy. For
classical compatibilists, such as Ayer or Hobart, what they had in mind is best under-
stood in terms of meaning. As they understood it, an analysis of an expression x gives
a set of conditions y that exhaustively specifies the meaning of x. Let us say that y is a
correct analysis of x if and only if the following is true:
To say that x just is to say that y.
To illustrate, if the term “bachelor” is correctly analyzed as “an unmarried
adult male,” then to say “Casper is a bachelor” just is to say “Casper is an unmar-
ried adult male.”
While characterizing analysis in terms of sameness of meaning probably best
describes how the majority of classical compatibilists understood analysis, it is
too restrictive for how many philosophers prefer to think about it. Many philoso-
phers prefer instead to think in terms of logical equivalence; an analysis of x in
terms of y is successful just so long as the ingredients in y nontrivially provide
conditions that are logically equivalent to x. (This is consistent with y not actu-
ally fully capturing the meaning of x.)
16 For a recent insightful defense of the classical compatibilist analysis in response to
this criticism, see Vihvelin (2013: 196–208).
17 In using the notion of causation in this passage, Chisholm is assuming that the causa-
tion is necessitating or deterministic.
18 Readers will note the close similarity between these two formulations and the two
ways we considered for formulating free will problems in Section 2.3, in particular,
formulations F3 and F4.