Citation for this page in APA citation style.           Close


Philosophers

Mortimer Adler
Rogers Albritton
Alexander of Aphrodisias
Samuel Alexander
William Alston
Anaximander
G.E.M.Anscombe
Anselm
Louise Antony
Thomas Aquinas
Aristotle
David Armstrong
Harald Atmanspacher
Robert Audi
Augustine
J.L.Austin
A.J.Ayer
Alexander Bain
Mark Balaguer
Jeffrey Barrett
William Barrett
William Belsham
Henri Bergson
George Berkeley
Isaiah Berlin
Richard J. Bernstein
Bernard Berofsky
Robert Bishop
Max Black
Susanne Bobzien
Emil du Bois-Reymond
Hilary Bok
Laurence BonJour
George Boole
Émile Boutroux
Daniel Boyd
F.H.Bradley
C.D.Broad
Michael Burke
Lawrence Cahoone
C.A.Campbell
Joseph Keim Campbell
Rudolf Carnap
Carneades
Nancy Cartwright
Gregg Caruso
Ernst Cassirer
David Chalmers
Roderick Chisholm
Chrysippus
Cicero
Randolph Clarke
Samuel Clarke
Anthony Collins
Antonella Corradini
Diodorus Cronus
Jonathan Dancy
Donald Davidson
Mario De Caro
Democritus
Daniel Dennett
Jacques Derrida
René Descartes
Richard Double
Fred Dretske
John Dupré
John Earman
Laura Waddell Ekstrom
Epictetus
Epicurus
Austin Farrer
Herbert Feigl
Arthur Fine
John Martin Fischer
Frederic Fitch
Owen Flanagan
Luciano Floridi
Philippa Foot
Alfred Fouilleé
Harry Frankfurt
Richard L. Franklin
Bas van Fraassen
Michael Frede
Gottlob Frege
Peter Geach
Edmund Gettier
Carl Ginet
Alvin Goldman
Gorgias
Nicholas St. John Green
H.Paul Grice
Ian Hacking
Ishtiyaque Haji
Stuart Hampshire
W.F.R.Hardie
Sam Harris
William Hasker
R.M.Hare
Georg W.F. Hegel
Martin Heidegger
Heraclitus
R.E.Hobart
Thomas Hobbes
David Hodgson
Shadsworth Hodgson
Baron d'Holbach
Ted Honderich
Pamela Huby
David Hume
Ferenc Huoranszki
Frank Jackson
William James
Lord Kames
Robert Kane
Immanuel Kant
Tomis Kapitan
Walter Kaufmann
Jaegwon Kim
William King
Hilary Kornblith
Christine Korsgaard
Saul Kripke
Thomas Kuhn
Andrea Lavazza
Christoph Lehner
Keith Lehrer
Gottfried Leibniz
Jules Lequyer
Leucippus
Michael Levin
Joseph Levine
George Henry Lewes
C.I.Lewis
David Lewis
Peter Lipton
C. Lloyd Morgan
John Locke
Michael Lockwood
Arthur O. Lovejoy
E. Jonathan Lowe
John R. Lucas
Lucretius
Alasdair MacIntyre
Ruth Barcan Marcus
Tim Maudlin
James Martineau
Nicholas Maxwell
Storrs McCall
Hugh McCann
Colin McGinn
Michael McKenna
Brian McLaughlin
John McTaggart
Paul E. Meehl
Uwe Meixner
Alfred Mele
Trenton Merricks
John Stuart Mill
Dickinson Miller
G.E.Moore
Thomas Nagel
Otto Neurath
Friedrich Nietzsche
John Norton
P.H.Nowell-Smith
Robert Nozick
William of Ockham
Timothy O'Connor
Parmenides
David F. Pears
Charles Sanders Peirce
Derk Pereboom
Steven Pinker
Plato
Karl Popper
Porphyry
Huw Price
H.A.Prichard
Protagoras
Hilary Putnam
Willard van Orman Quine
Frank Ramsey
Ayn Rand
Michael Rea
Thomas Reid
Charles Renouvier
Nicholas Rescher
C.W.Rietdijk
Richard Rorty
Josiah Royce
Bertrand Russell
Paul Russell
Gilbert Ryle
Jean-Paul Sartre
Kenneth Sayre
T.M.Scanlon
Moritz Schlick
Arthur Schopenhauer
John Searle
Wilfrid Sellars
Alan Sidelle
Ted Sider
Henry Sidgwick
Walter Sinnott-Armstrong
J.J.C.Smart
Saul Smilansky
Michael Smith
Baruch Spinoza
L. Susan Stebbing
Isabelle Stengers
George F. Stout
Galen Strawson
Peter Strawson
Eleonore Stump
Francisco Suárez
Richard Taylor
Kevin Timpe
Mark Twain
Peter Unger
Peter van Inwagen
Manuel Vargas
John Venn
Kadri Vihvelin
Voltaire
G.H. von Wright
David Foster Wallace
R. Jay Wallace
W.G.Ward
Ted Warfield
Roy Weatherford
C.F. von Weizsäcker
William Whewell
Alfred North Whitehead
David Widerker
David Wiggins
Bernard Williams
Timothy Williamson
Ludwig Wittgenstein
Susan Wolf

Scientists

David Albert
Michael Arbib
Walter Baade
Bernard Baars
Jeffrey Bada
Leslie Ballentine
Marcello Barbieri
Gregory Bateson
Horace Barlow
John S. Bell
Mara Beller
Charles Bennett
Ludwig von Bertalanffy
Susan Blackmore
Margaret Boden
David Bohm
Niels Bohr
Ludwig Boltzmann
Emile Borel
Max Born
Satyendra Nath Bose
Walther Bothe
Jean Bricmont
Hans Briegel
Leon Brillouin
Stephen Brush
Henry Thomas Buckle
S. H. Burbury
Melvin Calvin
Donald Campbell
Sadi Carnot
Anthony Cashmore
Eric Chaisson
Gregory Chaitin
Jean-Pierre Changeux
Rudolf Clausius
Arthur Holly Compton
John Conway
Jerry Coyne
John Cramer
Francis Crick
E. P. Culverwell
Antonio Damasio
Olivier Darrigol
Charles Darwin
Richard Dawkins
Terrence Deacon
Lüder Deecke
Richard Dedekind
Louis de Broglie
Stanislas Dehaene
Max Delbrück
Abraham de Moivre
Bernard d'Espagnat
Paul Dirac
Hans Driesch
John Eccles
Arthur Stanley Eddington
Gerald Edelman
Paul Ehrenfest
Manfred Eigen
Albert Einstein
George F. R. Ellis
Hugh Everett, III
Franz Exner
Richard Feynman
R. A. Fisher
David Foster
Joseph Fourier
Philipp Frank
Steven Frautschi
Edward Fredkin
Benjamin Gal-Or
Howard Gardner
Lila Gatlin
Michael Gazzaniga
Nicholas Georgescu-Roegen
GianCarlo Ghirardi
J. Willard Gibbs
James J. Gibson
Nicolas Gisin
Paul Glimcher
Thomas Gold
A. O. Gomes
Brian Goodwin
Joshua Greene
Dirk ter Haar
Jacques Hadamard
Mark Hadley
Patrick Haggard
J. B. S. Haldane
Stuart Hameroff
Augustin Hamon
Sam Harris
Ralph Hartley
Hyman Hartman
Jeff Hawkins
John-Dylan Haynes
Donald Hebb
Martin Heisenberg
Werner Heisenberg
John Herschel
Basil Hiley
Art Hobson
Jesper Hoffmeyer
Don Howard
John H. Jackson
William Stanley Jevons
Roman Jakobson
E. T. Jaynes
Pascual Jordan
Eric Kandel
Ruth E. Kastner
Stuart Kauffman
Martin J. Klein
William R. Klemm
Christof Koch
Simon Kochen
Hans Kornhuber
Stephen Kosslyn
Daniel Koshland
Ladislav Kovàč
Leopold Kronecker
Rolf Landauer
Alfred Landé
Pierre-Simon Laplace
Karl Lashley
David Layzer
Joseph LeDoux
Gerald Lettvin
Gilbert Lewis
Benjamin Libet
David Lindley
Seth Lloyd
Hendrik Lorentz
Werner Loewenstein
Josef Loschmidt
Ernst Mach
Donald MacKay
Henry Margenau
Owen Maroney
David Marr
Humberto Maturana
James Clerk Maxwell
Ernst Mayr
John McCarthy
Warren McCulloch
N. David Mermin
George Miller
Stanley Miller
Ulrich Mohrhoff
Jacques Monod
Vernon Mountcastle
Emmy Noether
Donald Norman
Alexander Oparin
Abraham Pais
Howard Pattee
Wolfgang Pauli
Massimo Pauri
Wilder Penfield
Roger Penrose
Steven Pinker
Colin Pittendrigh
Walter Pitts
Max Planck
Susan Pockett
Henri Poincaré
Daniel Pollen
Ilya Prigogine
Hans Primas
Zenon Pylyshyn
Henry Quastler
Adolphe Quételet
Pasco Rakic
Nicolas Rashevsky
Lord Rayleigh
Frederick Reif
Jürgen Renn
Giacomo Rizzolati
Emil Roduner
Juan Roederer
Jerome Rothstein
David Ruelle
David Rumelhart
Tilman Sauer
Ferdinand de Saussure
Jürgen Schmidhuber
Erwin Schrödinger
Aaron Schurger
Sebastian Seung
Thomas Sebeok
Franco Selleri
Claude Shannon
Charles Sherrington
David Shiang
Abner Shimony
Herbert Simon
Dean Keith Simonton
Edmund Sinnott
B. F. Skinner
Lee Smolin
Ray Solomonoff
Roger Sperry
John Stachel
Henry Stapp
Tom Stonier
Antoine Suarez
Leo Szilard
Max Tegmark
Teilhard de Chardin
Libb Thims
William Thomson (Kelvin)
Richard Tolman
Giulio Tononi
Peter Tse
Alan Turing
Francisco Varela
Vlatko Vedral
Mikhail Volkenstein
Heinz von Foerster
Richard von Mises
John von Neumann
Jakob von Uexküll
C. S. Unnikrishnan
C. H. Waddington
John B. Watson
Daniel Wegner
Steven Weinberg
Paul A. Weiss
Herman Weyl
John Wheeler
Wilhelm Wien
Norbert Wiener
Eugene Wigner
E. O. Wilson
Günther Witzany
Stephen Wolfram
H. Dieter Zeh
Semir Zeki
Ernst Zermelo
Wojciech Zurek
Konrad Zuse
Fritz Zwicky

Presentations

Biosemiotics
Free Will
Mental Causation
James Symposium
 
Laura Waddell Ekstrom

Laura Ekstrom has surveyed the arguments for libertarian incompatibilist free will in her 2000 book Free Will: A Philosophical Study. She has also proposed her own model, based on the idea that our preferences are developed in a causal but not determined way. Our character is an aggregation of these freely developed preferences, which we have the power to alter at any time.
She is concerned that incompatibilist accounts have thus far been incoherent.
"I first elaborate the central difficulties for presenting an incompatibilist free will theory, giving sense to critics' judgments that some sort of compatibilist analysis of free will must be right, since all incompatibilist accounts are doomed to failure through incoherency. Our first question, then, is whether or not an indeterminist model of free agency can be made coherent. I believe the answer to this question is affirmative." (p.82)
Ekstrom pays a lot of attention to where exactly the indeterminism is located in the process. She says:
"Libertarians, then, share the problem of providing a positive thesis concerning the metaphysical conditions enabling free will. Without a solution to the where's the indeterminism? problem, libertarian accounts of the nature of free will cannot get off the ground. This problem is the central, most significant difficulty for the endeavor of presenting a plausible incompatibilist account of the nature of free action: locating the requisite indeterminism in the causal history of the act, together with explaining why this precise location is appropriate and important." (p.85)
She says the libertarian requires multiple possible futures:
"Indeterminism says that it is not the case that, given the actual past and the actual laws of nature, there is at every moment exactly one possible future. Alternately (and, I take it, equivalently), indeterminism holds that not every event is causally necessitated by prior events. Libertarians in common endorse this negative thesis. (Some incompatibilists concerning free will and determinism are, then, compatibilists concerning free will and indeterminism.)" (p.85)
And she says the agent must have been able to do otherwise:
"Arguments for incompatibilism...rest upon the requirement for freedom that an agent be able at the time of acting to do otherwise than she does, where this ability is understood in a categorical, and not a conditional, sense. Given the actual reigning laws and conditions, the free agent is taken to be able to perform more than one action, and her possession of this ability means not only that she has the appropriate faculties for performing a multiple number of actions, but further that she can in the circumstances exercise them. The incompatibilist argument is grounded upon, in other words, an understanding of freedom according to which it must be undetermined by natural laws and events of the past what the free agent will do next.)" (p.81)
Ekstrom is also concerned about chance as the direct cause of an action. She says:
"A further problem for libertarians is readily apparent: If what the free agent does at a given moment is not determined by the natural laws and the past, including even the agent's immediately previous preference concerning what to do, then what is it determined by? Anything at all? If nothing, then what makes the act any different from a fluke or an accident? Incompatibilists intend in speaking of agential freedom to refer to something other than haphazard events that occur willy-nilly, do they not?" (p.86)
Her proposed account of free will is an "event-causal" model, not agent causation in which the agent (mind) is a new substance, or what Campbell called "non-occurent causation." She says:
"Despite recent attempts to revive the sort of libertarian theory appealing to agent-causation and arguments to the effect that any hope for providing a defensible incompatibilist account of freedom lies in this direction, I want nonetheless to explore here a fuller variety of indeterminist options and ultimately to endorse an account not reliant upon the existence of an irreducible form of nonoccurent- or agent-causation." (p.83)
Where does Ekstrom locate the indeterminism? She wants the deliberative process itself to become probabilistic. She says:
"The indeterminism involved in free agency is thus held by this view to be located between the factors that influence an agent's decision concerning what preference to form and the decisively formed preference itself...What is involved is indeterministic but probabilistic causation — the inputs to deliberation do not determine a particular decision outcome, but they make more probable the formation of certain preferences." (p.110)

"what the libertarian affirms is precisely that there is a third option between determined and random. The third option, according to this view, is that the decision is explicable by reference to the deliberative events that caused the decision to be what it was. Why did the free agent decide in that way? Because of reasons x, y, z, and so on. Why did those reasons lead him to decide as he did? The determinist would answer: Because of a deterministic causal law linking such reasons to such a decision. But the proposed account answers: Because the agent exercised his evaluative faculty in a particular way. Why? For reasons that inclined but did not necessitate a particular outcome to his deliberation process." (p.111)

Why does Ekstrom not locate indeterminism in prior alternative possibilities? She only wants indeterminism "during deliberation," where "there are genuinely available alternatives, multiple branching paths before us." She actually argues that the Principle of Alternative Possibilities is false, although she thinks "Frankfurt-style cases" do not establish this, whereas she does. (pp.181-214)

She considers and rejects the Dennett and Mele suggestions to locate indeterminism in the early stages (random alternative possibilities) of the process. This is also our two-stage Cogito mind model.

52. In the free will literature, there are two accounts of a sort similar to the one I have proposed (besides those of Kane and Nozick) worthy of mention. Daniel Dennett sketches — without himself endorsing — what he contends is the best model of free agency available to a libertarian ("On Giving Libertarians What They Say They Want," in Brainstorms, Montgomery, VT. Bradford Books, 1978, pp. 286-299). Dennett describes a scenario in which an agent must make a decision about what to do — namely, to accept academic job offer A or to accept academic job offer B. According to the model, "It just might be the case that exactly which considerations occur to one in such circumstances is to some degree strictly undetermined" (p. 294). In Dennett's example, considerations A through F occur to Jones, and on the basis of them, she decides to take a job at Swarthmore. But had consideration G also occurred to her, she would have taken a job at the University of Chicago instead. As it happened, G did not occur to her, although it might have, given the past and the laws. The proposal is that the indeterminism required for an act to be free is quite far back in the causal history of the act in question: just prior to the events describable as certain considerations occurring to the agent as inputs to the process of deliberation over what to do.

Dennett's proposal — set out only on the way toward making the ultimate point that incompatibilist freedom is not really the sort of freedom we want — is closely related to the model of libertarian autonomy proposed (but, again, not endorsed) by Alfred Mele (see Mele, Autonomous Agents). Mele's proposal is this: What should be held by the libertarian to be causally undetermined is "which members of a shifting subset of Jones's relevant nonoccurrent beliefs will become occurrent and function in his deliberation" (Mele, Autonomous Agents, p. 214). So, some beliefs will come to mind and some will not, and of those that come to mind, a subset of them are undetermined to occur by the past and the laws — they just happen to occur when they do in the deliberation process, but they might not have. Mele takes judgments concerning what to do to be the outcome of practical deliberation. So, what an agent decisively judges best in the case of deliberation leading to free action remains causally open until deliberation ends, and when deliberation ends remains causally open, since it is causally open whether a certain belief will come to mind and prolong deliberation. Mele claims as virtues of this account of modest libertarianism the following: (1.) compatibilists have no good reason to insist on determinism at this point in the deliberative process as a requirement for freedom; (2.) this sort of internal indeterminism is, for all we know, a reality; (3.) such indeterminism does not diminish the agent's control over his deliberations (Autonomous Agents, chap. 12).

The views set out by Dennett and Mele offer another interesting T-3 libertarian alternative. But, in my view, they locate the indeterminism in the wrong place. Specifically, the views are too weak, in virtue of the indeterminism location, to secure agential freedom. On these views, free agents are subject to luck in what thoughts come into their minds as they are deliberating about what to do. But once the thoughts occur and the last of them has occurred during deliberation, there is a deterministic causal connection between the particular pattern of beliefs that has happened to occur and the subsequent decision outcome. But this is problematic. For I might be a free agent, on Dennett's or Mele's account, while being a victim, with regard to what I judge best and what I consequently intend and do, of what thoughts happen to occur to me at the time. Granted, there are "forks in the road" of some sort on this picture of free agency (alternate physically possible futures). But it is not up to me, the free agent, which one I take. Which one I take is decided by which considerations happen to come to mind, where this is indeterministically caused by some previous events. On both Dennett's and Mele's views, once a certain pattern of considerations has happened to occur to the agent, a particular action may follow of physical necessity and yet count as free. Since neither of the views includes an account of the nature of the self, they leave unanswered the question of why an act that is the causally necessary outcome of whatever considerations have happened to occur is plausibly claimed to be originated by the agent. (pp.136-7)

Ekstrom's own model attempts to move the indeterminism partly into the decision itself (as does Robert Kane). Although she is aware that to the extent the decision itself involves randomness, the agent loses control.

She begins by developing the notion of an agent's preference and extending the moment of decision to a process over time.

I wish here to broaden the notion of preference: We might form preferences for having or acting on certain desires; but more ordinarily, I suppose, we simply form preferences for acting in certain ways. I will think of a preference, then, as a specially processed desire — a desire (for instance, for performing some act) formed by a process of critical evaluation with respect to one's conception of the good. (p.106)

The term 'decision' is variably and sometimes ambiguously used in free will literature. In developing the view to be proposed, I will use the term 'decision' to refer to a process: the sequence of events describable as the consideration of various factors by the agent's deliberative faculty, leading up to the formation of a state of mind settling the agent's uncertainty. I view the momentary mental act ending the decision process as part of the process, the end event of the process. The decision process might be settled either by the formation of an intention to act or by the formation of a preference. That is to say, if the decision concerns simply what to commit oneself to doing, then the outcome of the decision process — its settlement state — is an intention to act (now or in the future). If the decision in question concerns what to desire (so that in deciding one asks oneself, What do I want to do or to desire?) and if, in the process of making the decision, the agent considers centrally her acceptances concerning what is good and as the outcome of that consideration reflectively endorses a certain desire or course of action, then the outcome of the decision — its settlement state — is a preference. The event of preference formation thus settles the agent's indecision concerning what she wants to do. When we say of an agent S that "S's decision is made," I will take this to mean that S has ceased deliberating over a given matter by forming a particular preference (or intention, depending upon the type of decision) for action. (p.106-7)

But decisions may be causally determined by the agent's character.
One might, then, propose that action causally derivative from an agent's uncoerced preference is freely performed. But here is a problem with taking action on uncoerced preference as free. Consider an agent making an important decision...further suppose that, as the agent reasoned about the situation, weighing various factors, consulting other people, considering what he accepts and prefers, he was causally necessitated to reason in just the way he did. His deliberative or evaluative faculty was the source of the preference he formed, but, in fact, he was causally determined by factors such as his genetic makeup, environment, past experiences, and physiological constitution to evaluate potential courses of action in just the way in which he did evaluate them. (If someone had known all of the past events and the laws of nature, she could have accurately predicted exactly what he would decide.) In fact, though his preferences are his in that they were formed as a result of his own reflection on what he takes to be true and good, they are not, we might say, ultimately his, since the past, the current conditions, and the laws of nature conspire to make him prefer just what he does. This is because causal necessitation implies a unique, determinate outcome, an outcome that could not have been otherwise. In such a case, the agent could accept and prefer nothing other than what he did at the time, given the conditions and natural laws as they were. (p.108)
These (adequately) determined decisions, based on the agent's character, itself the consequence of earlier free decisions, seem to be exactly what Dennett and Mele were suggesting. Ekstrom claims that the fact that the decision was caused and predictable does not "defeat" freedom because it is strongly self-determined. She calls it a "preference with undefeated authorization." Undefeated authorization is defined as follows:
Undefeated Authorization: S is authorized in preferring A in a way that is undefeated if and only if S's evaluative faculty was neither coerced nor causally determined by anything to form the preference for A, but rather the preference for A was indeterminately caused by S's considerations. (p.108)
And she goes on to claim this as the basis for her model of free agency...
Freedom of action may then be defined in terms of preferences with undefeated authorization. Many of us conceive of ourselves as an agent possessing a certain kind of power: the power to form and reform one's own character. This power, we suppose, is often an uncoerced and undetermined determinant of our actions. It is the correctness of this very conception, I suggest, that allows (or would allow) one to act freely. An agent enjoys freedom of action in performing an act at a particular time, on this proposal, if and only if it is true that the agent's act results by a normal causal process from a preference for the act whose authorization is undefeated. (p.109)
This seems on the surface similar to our two-stage model, but Ekstrom is moving the location of indeterminism.
The proposed view claims not that preferences with undefeated authorization are uncaused, but that they are undetermined — they could have been otherwise. The indeterminism involved in free agency is thus held by this view to be located between the factors that influence an agent's decision concerning what preference to form and the decisively formed preference itself. The action resulting from the preference is free only if the formation of the preference for acting is caused, but not determined, by events involving the individual's evaluative faculty or intellect. What is involved is indeterministic but probabilistic causation — the inputs to deliberation do not determine a particular decision outcome, but they make more probable the formation of certain preferences. (p.110)
It is hard to see exactly how far Ekstrom's proposal has moved away from the idea of alternative possibilities generated indeterministically as additional "factors that influence an agent's decision." Then the decisively formed preference is adequately self determined. She says
the act resulting from the preference, in the case of free action, is controlled by the agent in being causally determined by the agent, since the agent on my view, is to be identified with a set of preferences and acceptances, together with the power to fashion and refashion the character. The decision concerning what preference to form — which in part, given this analysis of the agent, is a decision about what sort of person to become — is caused, but not fully determined, by the past. (p.115)
Indeed, she sums up the central merits of her proposed libertarian view...
(1.) The account appropriately captures intuitions concerning what we want in valuing freedom: It proposes some causal openness in the construction of the self, so that who we are is not the necessary unfolding of the past and the natural laws. Yet, what preferences we form is explicable, both causally and rationally.

(2.) The account is a phenomenologically accurate one: It seems that the ending point of deliberation is the outcome of our considerations yet that during deliberation there are genuinely available alternatives, multiple branching paths before us.

(3.) The view is consistent with a wholly naturalistic metaphysics. We need not believe in immaterial souls in order to believe in free will, and we need not believe in an uncaused substance-event causal relation. The view of free action as action on undefeated authorized preference leaves free acts within the realm of what is scientifically explicable, and there is no need on this account to deny the antecedently plausible principle that every event has a cause. (p. 116)

For Teachers
For Scholars
Excerpts from Free Will: A Philosophical Study.
From the Preface

Among the other great philosophical issues — the nature of truth, the foundation of ethics, the existence of God, the source of political authority, the possibility and scope of human knowledge, and the nature of meaning, causation, and consciousness — the problem of free will takes its place prominently as one both vigorously debated through philosophical history and to this day largely unsolved. Philosophers continue to debate not only the empirical question of whether or not we have free will, but also certain deeper and more fundamental theoretical questions: What is the essence or nature of a free act? Is our having the ability to act freely consistent or inconsistent with certain other propositions about the nature of the supernatural and the natural world? Would our having free will be, in fact, a good? And if so, what kind?

This book is designed both to highlight central issues in recent discussions of free will and to propose a particular indeterminist account of its nature. Various developments of the consequence argument for incompatibilism and the consequent debates over the properties of power necessity have been for many philosophers the centerpiece of free will theory during the past two decades. Discussion of the issues relevant to assessing the incompatibilist arguments thus takes a prominent place in this book. But it is also important to set out and examine directly the differing conceptions of free action itself underlying the arguments for incompatibilism and the compatibilist responses to those arguments. Hence I spend some time, following discussion of the arguments, making explicit and critically evaluating not just one compatibilist and one incompatibilist understanding of the nature of freedom, but rather a number of particular versions of each type of account, culminating in the proposal of a particular libertarian theory. This theory makes appeal, as libertarians traditionally have, to agent causation but proposes, contrary to tradition, that the notion is reducible to purely event-causal terms.

From chapter 4, Varieties of Libertarianism

Bolstered by the sophisticated work of the past two decades designed to demonstrate the incompatibility of free will and determinism — including arguments by van Inwagen, Fischer, Ginet, and others, discussed in Chapter 2 — many incompatibilists now rest confident in their incompatibilist position. But that confidence has a way of turning to unease when incompatibilists are asked to articulate their own accounts of the nature of free action and free will. In what way precisely is the assumption of indeterminism friendly to the cause of freedom? Insofar as indeterminacy is associated with chaos and randomness, the truth of the thesis of indeterminism would seem not only to be unrequired for the right account of freedom, but further to be positively antithetical to it. For an act done of one's own free will is an act under one's own control in excelsis, yet an act undetermined by a chain of previous events, and hence subject to some indeterminacy, seems an act produced from randomness and not under control.

Arguments for incompatibilism such as van Inwagen's First Argument rest upon the requirement for freedom that an agent be able at the time of acting to do otherwise than she does, where this ability is understood in a categorical, and not a conditional, sense. Given the actual reigning laws and conditions, the free agent is taken to be able to perform more than one action, and her possession of this ability means not only that she has the appropriate faculties for performing a multiple number of actions, but further that she can in the circumstances exercise them. The incompatibilist argument is grounded upon, in other words, an understanding of freedom according to which it must be undetermined by natural laws and events of the past what the free agent will do next.

But holding fixed everything about the past up until the moment of action, including even the agent's preference or judgment about what to do next, and maintaining that, crucial to the agent's freedom is her ability to do otherwise than what she, at an immediately prior moment, prefers or judges best to do, presents an extremely puzzling picture of free agency. Is this where the incompatibilist proposes to locate the requisite indeterminism— between the agent's preference for acting and her subsequent action? If not, then where else instead?

This chapter takes up these and other fundamental problems for setting out a precise account of the nature of free action encapsulated in the intuitions driving the incompatibilist argument. I first elaborate the central difficulties for presenting an incompatibilist free will theory, giving sense to critics' judgments that some sort of compatibilist analysis of free will must be right, since all incompatibilist accounts are doomed to failure through incoherency. Our first question, then, is whether or not an indeterminist model of free agency can be made coherent. I believe the answer to this question is affirmative and that many of the prominent critical worries about incompatibilist free will accounts can be assuaged. If so, then, second, what is the best sort of indeterminist account that can be offered, the one with the fewest problems and the one best answering to the deep-seated and pervasive conception of ourselves as agents with incompatibilist freedom?

The present chapter thus addresses head-on the perplexing but crucial task of giving a positive incompatibilist theory of the nature of free agency. The chapter sets out various sorts of incompatibilist accounts of freedom, some of which have, in my judgment, more strength and plausibility than others. This discussion leads to the development of a new indeterminist model of free action. I defend this model of free action as superior to competing accounts.

Some philosophers have taken commitment to incompatibilism to imply commitment to a theory of free will invoking a notion of the causation of an event by an agent as substance, where this is taken to be fundamentally irreducible to the causation of that event by other events. For instance, in speaking of "the fundamental watershed, in philosophical anthropology, between those who think of human beings as free" in an incompatibilist sense and "adherents of determinism," Alvin Plantings writes that "what is really at stake in this discussion is the notion of agent causation: the notion of a person as an ultimate source of action." We will look at accounts of freedom appealing to an agent as cause as a primitive notion. But incompatibilists have other options for constructing a free will theory. In fact, despite recent attempts to revive the sort of libertarian theory appealing to agent-causation and arguments to the effect that any hope for providing a defensible incompatibilist account of freedom lies in this direction, I want nonetheless to explore here a fuller variety of indeterminist options and ultimately to endorse an account not reliant upon the existence of an irreducible form of nonoccurent- or agent-causation.

A large part of the work of this chapter may thus be seen by some as an in-house matter for incompatibilists: that of determining which sort of incompatibilist conception of freedom ought to be endorsed. Since, however, I take the truth of incompatibilism to be demonstrated by successful argument, in my view, this task holds significance for anyone interested in free will, for the matter is, ultimately, the discernment of the right or best account of free will. Even those unconvinced of the truth of incompatibilism nonetheless might take an interest in seeing whether or not a successful incompatibilist account of freedom can be worked out. As a benefit to the recalcitrant compatibilist, our discussion here will provide a number of libertarian accounts of freedom for critique.

Libertarianism and Indeterminism

Typically the term 'libertarian' has been used to refer to those incompatibilists who maintain that persons in fact possess free will and hence that the thesis of determinism must be false, whereas an incompatibilist, speaking more generally, might remain neutral concerning the question of freedom's actual instantiation (or even take a negative stance, as does what is sometimes called the 'hard determinist'). But in describing various accounts of free will as libertarian, I do not mean to imply that a defender of such an account necessarily takes free will to be actually possessed by anyone. In using the term, I mean only to group together different accounts of the nature of free will by their common starting point — namely, the contention that a theoretical requirement for free will is that determinism be false. I will describe theories of free will sharing this assumption as 'indeterminist', 'incompatibilist', and 'libertarian', using the terms here interchangeably.

Libertarian theories have been famously charged with being "obscure," "incoherent," and, on some versions, guilty of "panicky metaphysics." According to Arthur Schopenhauer, for instance, the libertarian makes every free action "an inexplicable miracle, an effect without a cause." Libertarians themselves sometimes admit to a certain degree of theoretical impenetrability. Richard Taylor, in speaking of his own version of libertarian theory, comments that "one can hardly affirm such a theory of agency with complete comfort and wholly without embarrassment, for the conception of men and their powers which is involved in it is strange indeed, if not positively mysterious." Especially in light of such admissions, one can understand the suspicion cast in the direction of incompatibilist free will theorists.

But mere bad press from critics (and even some proponents) is not the fundamental problem for libertarian theory. Rather, certain ways of dealing with what are the fundamental problems give rise to accounts that are vulnerable to such charges. Alternate ways of dealing with the central difficulties, however, are not so vulnerable. That is to say, while objections such as the above of incoherency and "panicked" theorizing are perhaps accurately lodged against certain libertarian accounts of the past, I believe that they can be rebuffed by a number of different sorts of contemporary indeterminist account. Let me lead into those accounts by first explaining what I see to be the fundamental problems for incompatibilist free will theory.

The First Problem: Indeterminism Where?

The construction of an intelligible incompatibilist theory of freedom must begin with the libertarian assumption: If we ever act and choose freely, then it is the case that, given the past and the laws of nature, at some moment t (or at some moments t1, t2, t3,. . . ), there is more than one physically possible future.

But to assert as a metaphysical precondition of free will that, at some moment(s), given the actual laws and past, multiple physically possible futures exist, does not yet locate agents at the crucial junctures. In other words, the denial of determinism does not itself provide an interpretation of the alternative-possibilities condition of freedom, since the denial does not itself say anything directly about agents. In order to present a complete account of freedom, the incompatibilist needs to explain in what ways determinism must be false, to say when and where the indeterminism that allows for agents' production of free actions and free choices is located. Where precisely are the gaps that must exist in nature in the chain of deterministic causal links between events in order to allow for human free will? The mere truth of indeterminism does not ensure that persons can act and choose freely.

This problem for all libertarian theories may be elaborated as follows. Indeterminism is a negative thesis: It claims what is not the case. Specifically, indeterminism is the denial of the thesis of determinism. Indeterminism says that it is not the case that, given the actual past and the actual laws of nature, there is at every moment exactly one possible future. Alternately (and, I take it, equivalently), indeterminism holds that not every event is causally necessitated by prior events. Libertarians in common endorse this negative thesis. (Some incompatibilists concerning free will and determinism are, then, compatibilists concerning free will and indeterminism.)

But here is the problem: There are various ways in which it can fail to be the case that every event is causally necessitated by prior events. That is, there are various propositions about the actual world that, if true, would entail the truth of indeterminism. Since this is so, commitment to the compatibility of indeterminism and free will (indeed, further, the requirement of the former for the latter) is where unanimity ends among incompatibilist believers in free will concerning the metaphysical preconditions of free will. Libertarians divide over the matter of precisely in what way the falsity of determinism allows for the agential possession of free will.

Libertarians, then, share the problem of providing a positive thesis concerning the metaphysical conditions enabling free will. Without a solution to the where's the indeterminism? problem, libertarian accounts of the nature of free will cannot get off the ground. This problem is the central, most significant difficulty for the endeavor of presenting a plausible incompatibilist account of the nature of free action: locating the requisite indeterminism in the causal history of the act, together with explaining why this precise location is appropriate and important.

The Second Problem:
Distinguishing Free Acts from Mere Accidents

A further problem for libertarians is readily apparent: If what the free agent does at a given moment is not determined by the natural laws and the past, including even the agent's immediately previous preference concerning what to do, then what is it determined by? Anything at all? If nothing, then what makes the act any different from a fluke or an accident? Incompatibilists intend in speaking of agential freedom to refer to something other than haphazard events that occur willy-nilly, do they not?

Of course, they do. The challenge, once having located the causal indeterminism required for the account, is to explain how it is that the event constituting the agent's action counts as sufficiently his own, sufficiently under his control, to be rightly called freely done by him. Some theorists refer to this as the problem of control. The problem gains force from the common philosophical assumption that a certain event is under control insofar as its causally antecedent events determine it to occur. If causal determination by prior events does, in fact, exhaust the scope of the notion of control, then incompatibilists indeed have a very serious problem: Any event that is undetermined by previous events is out of control and hence, it would appear, merely accidental. Incompatibilists, then, must examine the notion of control and explain how, on their own particular accounts, the free agent's act is grounded in his agency and is not a haphazard or accidental happening.

Undefeated Authorization of Preference

In previous work, I defined the notion of 'preference' as an evaluatively formed desire (formed by a process of critical reflection with respect to one's conception of the good) for a certain first-level desire to be effective in action. (p.106)

I wish here to broaden the notion of preference: We might form preferences for having or acting on certain desires; but more ordinarily, I suppose, we simply form preferences for acting in certain ways. I will think of a preference, then, as a specially processed desire — a desire (for instance, for performing some act) formed by a process of critical evaluation with respect to one's conception of the good.

The term 'decision' is variably and sometimes ambiguously used in free will literature. In developing the view to be proposed, I will use the term 'decision' to refer to a process: the sequence of events describable as the consideration of various factors by the agent's deliberative faculty, leading up to the formation of a state of mind settling the agent's uncertainty. I view the momentary mental act ending the decision process as part of the process, the end event of the process. The decision process might be settled either by the formation of an intention to act or by the formation of a preference. That is to say, if the decision concerns simply what to commit oneself to doing, then the outcome of the decision process — its settlement state — is an intention to act (now or in the future). If the decision in question concerns what to desire (so that in deciding one asks oneself, What do I want to do or to desire?) and if, in the process of making the decision, the agent considers centrally her acceptances concerning what is good and as the outcome of that consideration reflectively endorses a certain desire or course of action, then the outcome of the decision — its settlement state — is a preference. The event of preference formation thus settles the agent's indecision concerning what she wants to do. When we say of an agent S that "S's decision is made," I will take this to mean that S has ceased deliberating over a given matter by forming a particular preference (or intention, depending upon the type of decision) for action.

Action upon desire is not necessarily free. I might desire and intend to do X because I am addicted to doing X. Then, although I desire to do X, when I do X, I do not act freely but, rather, addictively. In Frankfurt's terms, although I want to X, I do not want to want to X. Then in doing X, I frustrate rather than express my self (which Frankfurt would explicate in terms of higher-order desires, but which is better characterized, I believe, in terms of preferences and acceptances, as I will explain). In acting addictively, I frustrate the goals and values (otherwise expressible as the preferences and acceptances) intimate to — that is, in part constitutive of — me.

The introduction of the notion of preference is intended to solve this problem (the internal problem) with the basic account of free action as action on desire. Since a 'preference' represents what an agent wants as the outcome of her reflection on what is good (and not what she wants simply as an instinctive drive or as an addictive desire), when an agent acts on a decisively formed preference, barring external manipulation of the reflective process, she is involved in the action — she is its source. In other words, action on preference is authored by the agent.

One might, then, propose that action causally derivative from an agent's uncoerced preference is freely performed. But here is a problem with taking action on uncoerced preference as free. Consider an agent making an important decision, such as whether or not to quit his full-time job in order to make a go at it as a musician. Suppose that he decides to quit his job and undertake the new endeavor, and this outcome of the decision appears to have been "up to him." But further suppose that, as the agent reasoned about the situation, weighing various factors, consulting other people, considering what he accepts and prefers, he was causally necessitated to reason in just the way he did. His deliberative or evaluative faculty was the source of the preference he formed, but, in fact, he was causally determined by factors such as his genetic makeup, environment, past experiences, and physiological constitution to evaluate potential courses of action in just the way in which he did evaluate them. (If someone had known all of the past events and the laws of nature, she could have accurately predicted exactly what he would decide.) In fact, though his preferences are his in that they were formed as a result of his own reflection on what he takes to be true and good, they are not, we might say, ultimately his, since the past, the current conditions, and the laws of nature conspire to make him prefer just what he does. This is because causal necessitation implies a unique, determinate outcome, an outcome that could not have been otherwise. In such a case, the agent could accept and prefer nothing other than what he did at the time, given the conditions and natural laws as they were.

Normally, we think of what an agent accepts and prefers as elements of his mental life that are up to him, under his full authority. But the arguments for incompatibilism are designed to show that, if those states are the deterministic outcome of previous events together with the laws of nature, then those states are not in fact under the agent's own full authority after all.

Consider the notion of undefeated authorization of preference. A preference with undefeated authorization is one fully attributable to the self. Its claim to being authentic to the agent is not defeated by the claim that it is merely the causally deterministic outcome — the sole physically possible outcome at the time — of the past and the laws. An act that one performs as the normal causal outcome of a preference with undefeated authorization then is strongly self-determined in a way that is sufficient for that act to be free. Undefeated authorization is defined as follows:

Undefeated Authorization: S is authorized in preferring A in a way that is undefeated if and only if S's evaluative faculty was neither coerced nor causally determined by anything to form the preference for A, but rather the preference for A was indeterminately caused by S's considerations.
Freedom of action may then be defined in terms of preferences with undefeated authorization. Many of us conceive of ourselves as an agent possessing a certain kind of power: the power to form and reform one's own character. This power, we suppose, is often an uncoerced and undetermined determinant of our actions. It is the correctness of this very conception, I suggest, that allows (or would allow) one to act freely. An agent enjoys freedom of action in performing an act at a particular time, on this proposal, if and only if it is true that the agent's act results by a normal causal process from a preference for the act whose authorization is undefeated.

To say that an action is free just in case it results, by a nondeviant causal process, from an event-causally undetermined and uncoerced decisively formed preference is to say that the agent could have, prior to forming the preference, decided in a number of ways. The outcome of her decision to prefer to act in a particular manner was neither coerced nor necessitated; there was a range of options for preference formation within the agent's power at the time. During the decision process, we experience thoughts, desires, insights, preferences, and the like. On the view being proposed, the libertarian maintains that a preference-forming decision process can lead ultimately to a free action only if the outcome of that process (the event of preference formation) is caused, but not determined, by other deliberative events—by the occurrence to the agent of particular thoughts.

Merits of the Proposed Approach

The proposed view has the merit of not making whether or not a free act follows an agent's decisively formed preference to perform it a matter of chance. Notice that the preference-forming decisions at the center of the account need not immediately precede the act in time: One might form a preference for committing a certain act at a particular time in the future, leading to the formation of a future-directed intention that, together with the belief that the time has become appropriate, generates an immediately executive intention, which, in turn, if all goes well and the agent does not change his mind, leads normally to the act. Alternately, one might deliberately form a preference to act in a certain way in the future, not at a particular time, but whenever certain circumstances arise, such that one then has a standing preference from which free acts can arise without immediately following a deliberative process. For instance, the first time I chose a caregiver for my child, I may have had to put in a great deal of consideration (concerning such matters as my values, my child's needs, and the qualifications of various candidates) before forming a preference concerning whom to hire. But subsequently, when the need arises again, I may simply phone the regular baby-sitter and do so freely, without needing to elaborately reevaluate the matter, but instead acting from a standing preference to use the person in question when he or she is available. In counting my view as a T-3 account, then, I am taking the preference leading to the act, whether standing or just formed, to be sufficiently crucial to count as a crucial precursor to the act.

The proposed view claims not that preferences with undefeated authorization are uncaused, but that they are undetermined — they could have been otherwise. The indeterminism involved in free agency is thus held by this view to be located between the factors that influence an agent's decision concerning what preference to form and the decisively formed preference itself. The action resulting from the preference is free only if the formation of the preference for acting is caused, but not determined, by events involving the individual's evaluative faculty or intellect. What is involved is indeterministic but probabilistic causation — the inputs to deliberation do not determine a particular decision outcome, but they make more probable the formation of certain preferences.

Suppose, for the purpose of illustration, that Kim is deciding whether or not Jason is the one to marry. The possible outcomes to her deliberation process are: the preference to marry Jason, the preference to marry someone else instead now or in the future, and the preference to never get married at all. Various considerations occur to her, including: the thought that Jason is intelligent and funny; a desire for security; a preference for lifelong companionship with someone who shares her values; a desire for freedom; the arguments of some feminists she has read that marriage is an outdated patriarchal institution; a slight attraction toward Steve; an overwhelming attraction toward Jason; the suggestion of her father that she marry someone financially secure; the acceptance that Jason is not at present financially secure; the suggestion of her mother that she not marry at all; a preference for having children one day soon; and her own conviction that children are best off being born to married parents. Each of these deliberative inputs points rationally toward one or another of the differing conclusive preferences; and each raises the probability of one of the three potential deliberative outcomes.

Suppose Kim decides to marry Jason. If she acts freely in marrying him, then her preference for marrying him must have been neither coerced nor causally necessitated by the past. That is, during the decision process, she was able to have formed any of the three potential outcome preferences. Given the inputs to her deliberation and the natural laws, in other words, it was physically possible for any of the alternate preferences to result.

Note that the compatibilist cannot require strictly sufficient conditions for an event in order for that event to be nonrandom. To do so is question begging, for what the libertarian affirms is precisely that there is a third option between determined and random. The third option, according to this view, is that the decision is explicable by reference to the deliberative events that caused the decision to be what it was. Why did the free agent decide in that way? Because of reasons x, y, z, and so on. Why did those reasons lead him to decide as he did? The determinist would answer: Because of a deterministic causal law linking such reasons to such a decision. But the proposed account answers: Because the agent exercised his evaluative faculty in a particular way. Why? For reasons that inclined but did not necessitate a particular outcome to his deliberation process. These causal statements report necessary, but not strictly sufficient, conditions for the decisively formed preference. In order to be explicable, the decisively formed preference need not be necessitated. But in order to be free, the decisively formed preference must not be necessitated. It may be tempting to suppose that a preference cannot be explained if it might have been otherwise, given all of the considerations of the agent and holding fixed the natural laws. But this supposition is only what Christopher Hitchcock calls a "demon of determinism." Not all explanation is deterministic.

Certainly there are a number of factors that influence us as we deliberate. These include the media, friendly advice, the attitudes of one's colleagues, one's own acceptances and preferences, the wishes of one's parents, subconscious needs and desires, dreams, the political climate, concern for the environment, the interests of one's children or siblings, regard for what is true and right, and so on. What we desire when focusing on freedom is that none of these factors necessitate particular outcomes of our decisions. We welcome the influence of any such factors if we are reasonable, but what we abhor, when concentrating on the value of freedom, is the causal determination of our deliberative process by any of these influences. We do not want the deliberative process to be driven down a single path by previous factors. Freedom is opposed to constraint, and having to form a preference to act in any particular way — no matter how appealing that way may be, or how objectively right it is, or how much one's parent approves of it — is constraining.

So the view under consideration has intuitive plausibility. Furthermore, this sort of libertarian theory has the benefit of not positing mysterious agents in the form of "unmoved movers." It is a naturalistic libertarian view, not in that it requires the truth of naturalism, but in that it is consistent with a naturalistic conception of the world. The proposed account appeals only to causal relations between events (or states of affairs). There are no entities posited that we must believe in when we think of ourselves as having libertarian freedom, such as immaterial souls, and we need not ourselves, as substances, stand somehow in causal relations to events.

Initial Objection

But wait a minute, the reader might object. True enough, the proposed T-3 [uncaused action] libertarian view does not share the problems had by an A-C [agent causal] account, and its naturalistic orientation is a benefit. But does not the proposed account face a certain difficulty, that is, providing a satisfactory answer to the problem of control? We considered earlier the view (T-3a) that an agent is free with respect to a particular action only if that action is event-causally undetermined in the sense that the action is indeterministically caused by the agent's intention. But the reason for rejecting this view is that, to require that the link between an agent's intention to act and the act itself not be necessary, but only probabilistic, is to make the act accidental, not in the agent's control. A problem for the present view is that it suffers from a similar weakness. just as indeterminism between the agent's preference and her intention, or between her intention and her action, seemed to leave the action out of the agent's control, so too indeterminism between the factors that influence the agent's formation of preference and her decisively formed preference itself seems to leave that preference, and hence the resulting act, out of the agent's control.

Surely an agent does reflect on a variety of considerations prior to making a choice, including reasons for deciding in one way and reasons for deciding in a different way. Given the influences upon the agent, including these considerations, we are asked by the proposed view to imagine that which decision outcome results from these deliberations is undetermined. Given the deliberations, a certain potential decision outcome may or may not follow. But does this not make the particular way in which the indecision is resolved arbitrary? A decisively formed preference that may or may not follow upon the particular considerations that strike an agent during the deliberation process seems to be a preference that is not under the agent's authority. If there is some random element between the considerations serving as input to the decision and the decision outcome itself (so that the decisively formed preference has no strictly sufficient conditions), then the preference and the act to which it leads are uncontrolled by the agent.

The Proposal Further Elaborated

This is a natural concern and a fair one. However, the specification of the theory is not completed. In particular, I have not yet fully filled in a conception of the agent. Suppose we take an agent to be constituted by a character, together with the power to fashion and refashion that character. A character, or character system, is an aggregate of preferences and acceptances. Then an agent is an evaluating and choosing faculty (by which she creates preferences and acceptances), along with the character system, made up of those preferences and acceptances.

In previous work, I defended this notion of the self, or agent. Surely, I urged, the faculty for shaping the character — for evaluating desires, beliefs, and courses of action with respect to standards — ought fairly uncontroversially to be taken as a constituent of the self. We. clearly have some faculty for deciding what becomes a component of our character, what remains a component of our character, and what gets discarded as no longer an element of our character (although whether or not the operation of this faculty is causally necessitated by prior factors remains less clear). And this conception of the character is plausible, I think, for several reasons. For one, unlike hierarchical conceptions of the self as structured solely by desire, the proposed conception of the self incorporates some beliefs. Our convictions concerning the truth are central to who we are. But, appropriately, not just any beliefs and desires count as parts of the character. Opinions that one holds unreflectively and passions that overtake one are too common, as well as too blindly had, to be part of the character. For a character is the complex of attributes or features that mark and distinguish the individual. Elements of the character— preferences and acceptances — are the outputs of the individual's own evaluative activity.

I may thus now alternately state the proposed T-3 [actions caused but not determined] libertarian theory: An act is free just in case it results by way of a normal causal process from a pertinent intention (e.g., an intention to perform the act here and now) that is agent caused. The agent causation of an intention is, in my view, an ontologically and conceptually reducible notion. Take an agent to consist of evaluated reasons (preferences and acceptances, as a group constituting the character), together with an evaluating and choosing faculty (in other words, a power to fashion and refashion character). Then an intention is agent caused just in case it results by a normal causal process from a preference for acting as specified in the content of the intention, where the preference itself is the output of an uncoerced exercise of the agent's evaluative faculty, the inputs into which (various considerations) cause but do not determine the decisive formation of the preference.

Why did the occurrence of considerations r1, r2, and r3 raise the probability of decision outcomes preference1, preference2, and preference3 occurring to 0.51, 0.90, and 0.75, respectively? The reply "Because r1, r2, and r3 are good reasons for the decision outcomes preference1, preference2, and preference3, respectively, and r2 is the best" can be given no substance without reference to a perspective from which the reasons are judged to be positively relevant to the particular decision outcomes and a point of view from which the reasons are estimated to be good or worthwhile, with r2 judged to be the best. This point of view is built into the proposed account: The agent's evaluative faculty is the source of the evaluation of reasons and of the subsequently produced decision outcomes. Thus I endorse an indeterminist account of the following sort:

Theory Type 3d (T-3d Theories): An action is free only if it results, by a normal causal process, from a pertinent intention (e.g., an intention to perform that act here and now) that is caused by the agent, where this latter term ('caused by the agent') is reducible to event-causal terms.
Consider again the problem of control. In response, although the notion of control does require elucidation in some causal terms, it does not require deterministic explanation. In the case of deliberation preceding a free act, which decision-settling event it will be that occurs is not determined until it occurs. The formation of preference x or preference y or preference z may happen at t-1. At t, suppose that preference x is formed. Its formation has a causal (and rational) explanation in terms of the agent's considerations.

Furthermore, the act resulting from the preference, in the case of free action, is controlled by the agent in being causally determined by the agent, since the agent on my view, is to be identified with a set of preferences and acceptances, together with the power to fashion and refashion the character. The decision concerning what preference to form — which in part, given this analysis of the agent, is a decision about what sort of person to become — is caused, but not fully determined, by the past. One may object that this means that it is partly or wholly out of one's control what sort of person to become, and that this is problematic. However, the location for the indeterminism specified by the proposed account is precisely where we want the indeterminism to be in a free will model. What is especially objectionable about determinism is the thought that our character is fully determined by the past, that we could not have become different sorts of individuals, given the causal determination of our traits by the events of the past and the natural laws.

Thus, although an agent may be unable at t to do otherwise than a, she is free in performing a at t only if she is genuinely self-determined with respect to a — that is, only if she performs a for reasons and those reasons are genuinely her own, in that she was not coercively caused to have them and in that the reasons were not causally determined to be her preferences by previous deliberative events.

In sum, the central merits of the proposed libertarian view of free action as action on undefeated authorized preference I take to be these: (1.) The account appropriately captures intuitions concerning what we want in valuing freedom: It proposes some causal openness in the construction of the self, so that who we are is not the necessary unfolding of the past and the natural laws. Yet, what preferences we form is explicable, both causally and rationally. (2.) The account is a phenomenologically accurate one: It seems that the ending point of deliberation is the outcome of our considerations yet that during deliberation there are genuinely available alternatives, multiple branching paths before us. (3.) The view is consistent with a wholly naturalistic metaphysics. We need not believe in immaterial souls in order to believe in free will, and we need not believe in an uncaused substance-event causal relation. The view of free action as action on undefeated authorized preference leaves free acts within the realm of what is scientifically explicable, and there is no need on this account to deny the antecedently plausible principle that every event has a cause.

In short, given the power of the arguments for incompatibilism, some indeterminist model of free action must have a hold on us, and the proposed account, I have argued, is the best of the available options. When I act freely, the preference leading to the act is generated by the evaluative faculty that in part constitutes me, and the preference is not coerced by an external source. Hence, on the proposed view, when the appropriately formed preference leads to my act, what I do is "up to me," since I, quite literally, causally determine it.

Excerpts from Challenges to the Proposed Indeterminist Model of Free Agency (p.120)

Any incompatibilist model of freedom is going to have to locate the event-causal indeterminism somewhere in the history of the free act, and the specified place — between the considerations and the decisive formation of preference — seems to me the most reasonable place to locate it. One might, instead, locate the indeterminism prior to some of the considerations that occur to the agent during the deliberative process, so that what is undetermined is which (and perhaps at what points during the process) particular considerations come into the agent's mind. But if this were the only place specified for required indeterminism, then an act might be the purely causally deterministic outcome of the considerations that happen to occur to the agent and yet, on the proposal, count as free. Such an account is too weak to ground agent freedom and deep responsibility, since, given the occurrence of the particular considerations in any case, one particular act follows of physical necessity, as the completely deterministic unfolding of previous events.

Consider an agent whose act is, in such a sense, "libertarian free." Now a duplicate agent in exactly similar circumstances governed by the same natural laws and subject to the same occurrence of considerations at the same points in the deliberative process will form exactly the same judgment concerning the best thing to do and will act accordingly. But then, given the consideration pattern that occurs (but might not have), there is no "wiggle room" for the agent in forming an evaluative judgment — it simply falls out, of necessity, from the consideration pattern. Hence such an account does not leave sufficient room for free agency.

Where causal indeterminism is best located, instead, is after the considerations (which in the usual case have been determined to occur by previous events, such as how much rest one has had, what one has recently eaten, with whom one has recently spoken, what one has read) yet prior to the decisive formation of preference, such that given the exact consideration pattern, the agent may decide to prefer a and may decide to prefer otherwise. The considerations themselves are indeterministic causes of the preference.

Again, every libertarian model of free action has to locate the indeterminism somewhere in the history of the act. Where else would one put it? One might, out of dissatisfaction with the proposed model, opt to be a compatibilist. But then one must face the fact that the arguments for incompatibilism considered in Chapter 2 are powerful and that no response to them is, as I have argued, particularly persuasive. Thus some incompatibilist model of free action is needed. Furthermore, the intuitions underlying our deep-rooted notion of human dignity, our commonsensical conception of the future as we engage in practical deliberation, and commonly held judgments concerning deserved attributions of moral responsibility are incompatibilist in nature. It is central to our self-image as practical deliberators that there are forks in the path in front of us representing causally open future alternatives. Where are the junctures that are important to us for acting freely?

The proposed model, I have urged, appropriately locates them. Moreover, I have argued that the other available models of incompatibilist free action are problematic. I have aimed to avoid an appeal to the agent as a primitive sort of cause, for reasons discussed above.

6. On the proposed account, it seems that what is partially not "up to me" is who I am, what sort of person I become. But this is odd as an account of free agency grounding human dignity and moral responsibility.

But it is up to me in a significant sense or to a significant degree who I am, because it is my deliberative faculty doing the reasoning in light of considerations and forming the preferences and acceptances. Also, on any indeterminist model, there is something that is not 'up to me', if by this phrase one means something that is causally undetermined by previous events. Control cannot be equated to event-causal determinism without begging the question against all sorts of libertarian theory. Also, the act done, on the proposed account of free action, is completely up to the agent. In fact, it is caused by the agent, since the agent is taken to be the evaluative faculty together with the preferences and acceptances. So one's free act is fully under one's control in being causally determined by one's self.

7. How does the proposed account meet the two initial problems posed in the previous chapter for the basic account of free action as action on desire: the manipulation problem and the internal problem?

It meets the internal problem by requiring preference for action. Action on a desire that is not a preference is not free. Hence "the unwilling addict," as Frankfurt calls him, does not act freely on my account. Does "the willing addict" (that is, an addict who prefers to take the drug) act freely on my account? No, because his preference is causally determined to be what it is by physiological factors responsible for the addiction. His preference, given the addiction, could not have been otherwise. So his evaluative faculty, as he decides what to do, is determined to reason just as it does and thus is causally determined to form precisely the preference formed. The account of undefeated authorized preference is designed to address the manipulation problem posed by threats, posthypnotic suggestions, and the like. An agent has undefeated authorization for her preference just in case her evaluative faculty is neither coerced nor causally determined by anything to form that particular preference, but rather the preference is indeterministically caused by her considerations. The preference leading to a free act must be uncoerced and caused by an indeterministic deliberative process. The proposed account has need, then, of the same list that the compatibilist needs concerning the act's (or preference's) not being the product of coercion by threat, direct brain stimulation by an evil neuroscientist, posthypnotic suggestion, and so on, in order to rule out foul play in the formation of the preference.

8. How is it, again, that the free act on your account is controlled by the agent, when there is some indeterminism in the production of it the act? Insofar as an act is not causally determined by prior causes, it is out of control, a mere random event. Your model of free agency re- places causal determinism, in the process leading to the act, with randomness. But then your account does not provide a right account of action that is under the control of the agent, or free action. Why should anyone want to be able to act freely as characterized by the proposed account, since that account only gives to an agent the status of being prone to a random process in the generation of action?

First, the incompatibilist not appealing to agent-causation is distinctive among proponents of free will models in defending a third option other than an event's being causally determined or its being random. To assert that those two alternatives are the only alternatives begs the question against all T-3 views. Every incompatibilist who thinks that the notion of free action is coherent alleges precisely that there is a third alternative between an event's being causally determined by previous events, on the one hand, and an event's being random, on the other.

Furthermore, on the particular incompatibilist account I have proposed, the act is causally determined: It is causally determined by the agent. Recall the proposed account of the self as the evaluative faculty together with the character. The free act is agent caused on the proposed account, where this notion is reducible to event-causal terms.

The problem of control is, I admit, the most serious difficulty facing the type of view I have defended. But consider the alternatives. We could, in accounting for free action, appeal to agent-causation as a primitive notion, irreducible to causation among events. But I find this approach unsatisfactory, as have many philosophers. The agent-causal theory, indeed, seems merely to label rather than to illuminate the problem of free agency. In light of problems for the A-C approach, as Robert Kane has urged (1996b), why not see how far we can get in constructing an indeterminist model of free action without relying on primitive agent-causation before giving up and making such an appeal?

Alternatively, we could adopt a T-1 approach, relying on the posit of an uncaused volition or choice. But to adopt a T-1 theory requires that we believe in something — uncaused events — with which we have no familiarity in the physical world. Uncaused events of volition or choice can be explained by reference to the agent's purposes, but they cannot be causally explained; they derive from absolutely nothing. In other words, with regard to any uncaused volition or uncaused choice, we can answer the question, Why did it occur? (by citing the agent's goals), but we cannot answer — it is impossible to answer — the question, What made it occur? This is unsatisfying. On the T-3 view I have defended, free actions result from preferences, and those preferences are formed noncoercively for reasons that are probabilistic causes of them.

9. How might the proposed indeterministic model of free action be actually realized? What evidence is there for thinking that we in fact have free will in this sense?

Several recent theorists have done interesting exploratory work on the matter of how indeterminism at the quantum level might be magnified if the brain is a chaotic system. Additionally, if we have evidence for thinking that we are morally responsible (in what I will call in the subsequent chapter the 'metaphysical' sense) and if incompatibilist freedom is required for such moral responsibility, then we have reason to think that we have incompatibilist freedom. Further, if we have reason to think that we have dignity, that our most intimate relationships are genuine, and that our commonsensical conception of the future as we practically deliberate is accurate, then we have reason to think that we have incompatibilist freedom.

10. You have proposed that, between the considerations serving as input to the agent's evaluative process and its outcome, there is non-deterministic or probabilistic causation. Suppose that one particular consideration is so powerful that it raises the probability of a certain decision outcome to 0.99999. Then suppose that that outcome in fact occurs and leads appropriately to the agent's act. Why is such an act free, whereas it would not have been free had the probability been equal to 1 that the preference for that act would occur, given the previous considerations?

We sometimes speak of a range of freedom of action, some acts being fully free and others less so. The probabilistic model gives one way of making sense of degrees of freedom. Perhaps the most free acts derive from preferences whose probability of occurring was raised by the occurrence of certain previous considerations to values within a range of, say, 0.2-0.8, whereas the act would be less free when resulting from a preference at either end of the spectrum, that is, in cases where the considerations made the probability of the preference's occurrence near 0.9 or 0.1.

Consider the case of a cancer patient's agreeing to a round of chemotherapy treatment by signing a particular document. It seems to me that in such a case the agent retains her freedom in signing the document, provided that she does so as the result of a normal causal process from an evaluatively formed preference for signing it that was neither coercively formed nor causally necessitated. The patient must have been able to form a different preference instead (namely, the preference for refusing the treatment) in order for her act of signing to be freely done.

But although the signing act is free, it seems to me less so than a different act done such that considerations did not make it so highly probable that there would be one particular outcome of the decision process. That is, given the agent's acceptance that if she refuses chemotherapy treatment, she will very likely die, her preference for signing the form has a high probability value. Compare a case in which I am deciding whether to give a certain sum of money to a charitable cause or instead to use it for a family vacation. I accept the great worthiness of the cause and prefer to use my money to benefit others as well as my own family; yet, I am convinced that my family members would enjoy the trip and relax during a vacation, and I prefer that we occasionally spend focused time together. Suppose that the matter is a "toss-up," given all of my prior considerations. Acting in one way or another then manifests my full freedom.

11. You have appealed to something unverifiable in characterizing the agent, namely, his "evaluative and choosing faculty." Why is your account not just as vulnerable as is the A-C account to a charge of mysterious agency?

In characterizing free action, I have appealed to a feature of persons that has been described as an agent's evaluative faculty for forming preferences and acceptances in light of his conception of the true and the good. But the critic may protest that this feature of persons is not empirically observable and, in this sense, is mysterious. However, this charge puts at risk of being dismissed as "mysterious" any posited ability or faculty of persons. How do we become aware of the existence of any ability? By observing its exercise or its effects. Normally, of course, we cannot discern simply by looking at a person at any time whether or not he has the ability to solve complex mathematical problems, but we can observe him in the process of exercising the ability and, from this evidence, justifiedly infer the ability's existence. Similarly, though it is true that we cannot directly observe another's evaluative faculty from the external standpoint, we can, from that standpoint, observe his exercise of the power as he fashions and refashions his character — in other words, as he deliberates among alternatives and decides upon forming one preference or acceptance rather than another. We justifiedly infer the existence of the ability from our observation of its exercise. We can and do witness others considering various factors and making up their minds, settling on a decision outcome. Others report verbally about their deliberations and their decisions; their facial expressions change as they deliberate; they hesitate, stop what they are doing, and so on.

12. But the evaluative faculty you have proposed is not just the ability to form one's character; it is further the ability to do so without being determined by anything to decide as one does. It is a particularly unverifiable position to assert that we have the ability to decide in a way that is not necessitated by anything. How could we ever know whether or not this is the case?

To claim that persons retain the power to fashion and refashion their character, without being necessitated by anything to do so in one way rather than another, is not to assert something incomprehensible or utterly opaque. We know from the inside what it is like to deliberate among competing desires or courses of action and to exercise our faculty to decide to form one particular preference rather than another, without being necessitated by any prior factor to decide as we do. It is commonplace, at any rate, to operate on the assumption that we have such an ability, and there are various considerations that justify this assumption on our part.

The account I have presented does not make free action wholly mysterious. It does not place agents entirely outside the causal nexus, and it does not require for the existence of freedom the falsity of the Principle of Universal Causation (the principle that every event has a cause). Further, free actions, on the view, are not random or arbitrary; they are directly under the agent's control in being determined by and only by the agent, and the notion of an agent is explicated rather than left as a primitive. Free actions are explicable in terms of an individual's character.

Consider, again, the internal evidence. We often introspectively observe ourselves or directly experience choosing without being necessitated to choose in any particular manner. We may, of course, be wrong. Our evidence of how it feels to make decisions about what sorts of people to be and about what to do, without being either coerced or determined by the past, may be deceiving. However, in the absence of any good reasons for thinking so, and in the presence of reasons for thinking that our internal observations of the process are accurate, it is legitimate to account for this evidence by positing an undetermined evaluative faculty, which is not to be conceived as an immaterial substance, but as an ability to form preferences.

13. A decision to form a particular preference should be reasonable or rational for the agent, rather than arbitrary, in order for the act resulting from that decision to be free. But if the agent's deliberative process comes up with the answer that, all things considered, A is the best choice, then it seems that she ought to be necessitated to choose A. To require for freedom that her choice outcome remain undetermined by the reasons seems to require an odd ability: the ability to decide against the best reasons. But no one wants this ability.

To require for freedom that an agent be able to choose otherwise in the categorical sense after the occurrence in her mind of various considerations — that she be able to choose A and that she be able to choose B, given the laws of nature and all of the facts about her and her past, including the deliberations of her evaluative faculty — seems to require for freedom the ability to act irrationally. As Robert Kane expresses the problem, what is difficult to understand and in need of explanation is "how I could have reasonably chosen to do otherwise, how I could have reasonably chosen B, given exactly the same prior deliberation that led me to choose A, the same information deployed, the same consequences considered, the same assessments made, and so on."

Once A has been judged to be the right decision outcome, a choice of B would appear to be irrational. In cases in which the deliberative faculty reveals one best answer, the proposed account of freedom, the critic charges, requires something nondesirable: the ability to choose what is not the most rational alternative. To want the ability to act freely as I have characterized it might then seem to be, in the words of Susan Wolf:

not only to want the ability to make choices even when there is no basis for choice, but to want the ability to make choices on no basis even when there is a basis. But the latter ability would seem to be an ability no one could ever have reason to want to exercise. (Freedom Within Reason, 1990, p.55)
But on the proposed account, it is not true of choice outcomes that they have no basis whatsoever; they are, rather, caused indeterministically by prior considerations. Suppose that my decision outcome is preference A. Now, what does it mean to say that I could have reasonably formed preference B, given exactly the same prior deliberation that led me, in fact, to choose A? We know what it means to say that I could have chosen otherwise than A: I had the skill required for forming some choice outcome other than A, and, given the past and the laws, it was undetermined what choice outcome would follow. In other words, the physical conditions and laws left room for the exercise of my skill. As to the question of how I could have reasonably chosen otherwise, suppose that I had decisively formed preference B (rather than preference A), that preference B was uncoerced, and that preference B was indeterminately caused by my prior considerations. Then preference B was reasonable. That is, if formed, the preference for B would have been caused and justified by reasons. In fact, so long as the choice outcome follows on reasons considered by the agent in the deliberative process, it is reasonable in some sense: in the sense that it results for a reason, although it is not necessarily the most rational option, given the total set of the agent's reasons.

If the objector claims to prefer at every juncture being determined to decide as he does by the best considerations rather than by just any considerations, then this merely shows something about the objector's values: that he prizes being right over being free. Being pushed into deciding in a certain way by anything — whether one's grandmother, one's genetic blueprint, or overwhelmingly powerful considerations — is antithetical to free agency. Contrary to Wolf's claim about what we want, in deciding what to do and what sort of person to be, we do not want to be determined by anything. Rational and causal influence is one thing; determination of choice is another. Again, freedom is opposed to constraint, and having to choose in any particular way, no matter how rational or objectively right that way may be, is constraining.

On the proposed account, then, prior to the working of the agent's evaluative faculty, the agent can form a preference for A and can form a preference for B. Both preference A and preference B could, at that point, be rational for the agent if she has reasons supporting each preference. A preference probabilistically caused by considerations of the agent's is not a preference formed "on no basis." The objection might be construed as a request for an explanation of how either of two alternate decisively formed preferences that the agent could potentially form could be rational for that if formed. The answer, then, is that whichever one is made is rationally explicable by reference to the agent's evaluative process and its inputs. Suppose she decided to form preference A. Why did she do so? Because of prior considerations that reasonably supported that particular decision outcome. The decisions from which free acts result are those that the agent forms on some basis, yet no basis is causally determining of what she decides.


Chapter 1.4 - The Philosophy Chapter 1.6 - The Scientists
Home Part Two - Knowledge
Normal | Teacher | Scholar