107 10 697KB
English Pages 122 [124] Year 2010
Free Will in Criminal Law and Procedure
ARSP BEIHEFT 120
Archiv für Rechts- und Sozialphilosophie Archives for Philosophy of Law and Social Philosophy Archives de Philosophie du Droit et de Philosophie Sociale Archivo de Filosofía Jurídica y Social
Friedrich Toepel (ed.)
Free Will in Criminal Law and Procedure Proceedings of the 23rd and 24th IVR World Congress Kraków 2007 and Beijing 2009
Franz Steiner Verlag Stuttgart 2010
Bibliografische Information der Deutschen Nationalbibliothek: Die Deutsche Nationalbibliothek verzeichnet diese Publikation in der Deutschen Nationalbibliografie; detaillierte bibliografische Daten sind im Internet über abrufbar. ISBN 978-3-515-09320-0 Zugleich: 978-3-8329-5499-4 Jede Verwertung des Werkes außerhalb der Grenzen des Urheberrechtsgesetzes ist unzulässig und strafbar. Dies gilt insbesondere für Übersetzung, Nachdruck, Mikroverfilmung oder vergleichbare Verfahren sowie für die Speicherung in Datenverarbeitungsanlagen. © 2010 Franz Steiner Verlag, Stuttgart Gedruckt auf alterungsbeständigem Papier Druck: Druckhaus Nomos, Sinzheim Printed in Germany
CONTENT I. Is Free Will an Incoherent Notion? (Proceedings of the 23rd IVR World Congress Kraków 2007) Friedrich Toepel, Bonn Preface ......................................................................................................................
9
Ronald Allen, Chicago Introduction ............................................................................................................ 11 David Hodgson, Sydney In Defence of Voluntariness.................................................................................... 15 John Montgomery, Strasbourg The Freewill Issue in Theological Perspective ........................................................ 23 Friedrich Toepel, Bonn Free Will and Infinite Regress ................................................................................. 29 Allan McCay, Sydney Bad Luck and Compatibilism; Smilansky‘s Concerns about the „Shallowness of Compatibilism“ ............................................................................ 39 Juan Pablo Mañalich, Santiago (Chile) Determinism, Free Will and Criminal Responsibility ........................................... 49 Ronald Allen, Chicago Replies ...................................................................................................................... 63 II. The Connection between Punishment, Retribution and Free Will (Proceedings of the 24th IVR World Congress Beijing 2009) Friedrich Toepel, Bonn Preface ...................................................................................................................... 71 Angus Menuge, Mequon (Wisconsin) Does Neuroscience Undermine Retributive Justice? ............................................. 73 John Montgomery, Strasbourg Some Remarks on Punishment and Free Will in Legal Theory and Classical Christian Theology.................................................................................................. 95 Friedrich Toepel, Bonn Free Will and Alternative Possibilities .................................................................... 101 Mateusz Klinowski, Kraków What can Logic of Action Tell Us About Responsibility and Free Will? ............. 115
I. IS FREE WILL AN INCOHERENT NOTION? (PROCEEDINGS
OF THE
23RD IVR WORLD CONGRESS KRAKÓW 2007)
PREFACE Perhaps as a result of my background in the Continental legal tradition, for a long time I avoided writing about free will and in this respect shared the attitude of many German legal scholars. In the discourse that I was most familiar with, there seemed to be two groups of opinions separated by an impenetrable wall, the practically thinking die-hards who always presuppose free will no matter how strong the evidence is against them and the allegedly more enlightened thinkers who reject free will on account of the psychological evidence or at least allow for the possibility that there is no free will. If members of the two groups meet, their discussion typically takes the following form: Those who call free will into doubt focus on statements of psychologists and results of brain-imaging studies which support their opinion, while their opponents question the reliability of the studies and point to the devastating practical effect of abandoning the presumption of free will. Such discussions are notorious for never leading to a conclusion. Nevertheless, the idea to organize a workshop on free will emerged as I listened to Professor Ronald Allen’s presentation of a paper at Northwestern University School of Law on the occasion of the 100th anniversary of the Northwestern Law Review.1 Discussing the notion of free will in the context of Miranda rights, he presented an argument that free will is not only a concept to be avoided because of insufficient scientific evidence in favour of its support, but that we actually have no clear idea what we are talking about when using the term – that free will is an incoherent notion. Professor Allen is not alone with this opinion. Renowned philosophers like Galen Strawson2 have developed a similar argument and his arguments have been favourably received by other philosophers including Saul Smilansky.3 Professor Allen’s remarks completely altered my view of the free will problem because here at last was a question that could and should be answered, namely whether the term free will can be explicated in a coherent manner. Incoherent terms should not be tolerated in law, much less serve as an important criterion in making legal decisions because there is always the possibility that the public will one day discover this incoherence, and that will substantially undermine trust in the legal system. Further, the problem of incoherence is not much discussed by legal scholars. I am not aware of any Continental works on this topic and therefore I feel that here is a substantial amount of work to do in this area. At the same time, examining the coherence of the term would bring in much of the complicated debate related to the classical problem of whether it is reasonable to accept free will given the empirical evidence at our disposal. It would not be sufficient to prove that there is a coherent shallow notion of free will which would be totally irrelevant for the law. Rather, it is vital to examine whether a coherent notion can be found which is deep enough to meet the needs of the legal system. If this is not possible, we have to look for a substitute, which in turn could require structural alterations to the legal system.
1 2 3
Ronald J. Allen, Miranda’s Hollow Core, Northwestern University Law Review 100 (1) (2006), 71–85 Galen Strawson, Freedom and Belief, 1998, 58 Saul Smilansky, Free Will and Illusion, 2000, 65–67
Friedrich Toepel
10
Thus the same options have to be taken into consideration as when discussing the classical problem of whether free will should be accepted given the empirical evidence at our disposal. We can accept the possibility of a classical libertarian variant of free will, we can reject such a notion as unattainable because of its incoherence and rest content with a more shallow notion which is compatible with determinism (compatibilism), or we can reject free will altogether and accept a hard determinist view. A fruitless empirical discussion in this way could be turned into a fruitful conceptual discussion. After I had talked over the matter with Professor Allen, he encouraged me to organize a workshop on the conceptual problem of free will at the IVR World Congress in Kraków in 2007. I felt it would be good to let a historian participate in the discussion because he could enlighten us on how free will was used at crucial points in history, but I wondered where I could find the right person. When I visited Professor John Warwick Montgomery in Strasbourg, it occurred to me that what was actually needed for such a debate was a theologian rather than a historian because the free will problem received its most intense discussion during the Period of the Reformation. Professor Montgomery is not only a professor of law but also a wellknown theologian – a rare combination. I need only add that after I had talked to him about the incoherence question and he, independently from Professor Allen, suggested that we should have a workshop on the topic, I proceeded to put my plan into practice. We were fortunate to be joined by Justice David Hodgson from the Supreme Court of New South Wales who in a book4 and numerous distinctive articles (among them a contribution to the Oxford Handbook of Free Will)5 has become a prominent contributor to the free will debate by developing a libertarian view of free will which he also defends in the present volume. We further received two interesting contributions by representatives of the younger generation, Allan McCay from Sydney University who is writing his Ph.D. thesis on the subject and Juan Pablo Mañalich from the University of Santiago de Chile who is staying at Bonn University as a Humboldt grantee at the moment. McCay explored a crucial point made by Saul Smilansky, luck’s undermining effect on moral responsibility which becomes relevant when considering the views of those who see free will as compatible with determinism. Mañalich undertook to show that a compatibilist background is sufficient as a basis of modern criminal responsibility. My impression is that the diversity of standpoints among the participants of our workshop ensured lively presentations of the problem so that no important aspect of the conceptual discussion has been left out. Bonn, August 2008
4 5
Friedrich Toepel
David Hodgson, The Mind Matters, 1991 David Hodgson, Quantum Physics, Consciousness and Free Will’, in The Oxford Handbook of Free Will, ed. R. Kane, 2002, 85–110
RONALD J. ALLEN, CHICAGO1 INTRODUCTION It is both a great honor and a deep pleasure to be introducing this symposium on free will and the criminal law. My papers on the intersection of free will and certain aspects of American criminal procedure seem to have been the rough grain of sand that helped precipitate the pearls that are to follow, and it is a great honor indeed to have the members of such a distinguished and insightful panel be motivated at least in some small way by them. It is also a very deep pleasure because it allows me to express my thanks to and admiration for Friedrich Toepel, who organized this event. I and my colleagues were blessed with his presence at Northwestern as a Fulbright scholar. It was conversations during that time about my and his work that led in part to his organizing this symposium. The enthusiasm and energy represented by this effort gives one a small flavor of the enormously positive influence that he is on those around him. We cherished his time at Northwestern and very much look forward to welcoming him back, hopefully in the not-too-distance future. It is my task here briefly to lay on the problem that brought the commentators together, and in a phrase that problem is free will. Although my work may have been the occasion for the papers that follow, in one sense that is ironic. Free will is normally thought of as philosophical problem, which of course it is, but it also has enormous, if oft-neglected, practical consequences. The articles below are largely philosophical in their approach, but I am not even an amateur philosopher. I am a scientist interested in understanding legal phenomena, and thus I am more interested in the practical consequences of the concept or entity than I am in the rich philosophical discourse it has generated. I mention this point because I was writing about free will only in passing. The object of my inquiry was instead the famous American case of Miranda v. Arizona.2 Miranda is not only a famous American case; it is also one of the most controversial cases ever decided by the United States Supreme Court, ranking with such cases as Dred Scott v. Sandford,3 Plessy v. Ferguson,4 and Roe v. Wade.5 Indeed, it is not much of an overstatement to say that a substantial portion of Presidential politics in the U.S. over the last thirty years has been driven by the effort of one part of the political spectrum to overrule and another to preserve Miranda and Roe v. Wade. That is also, however, not what interests me particularly. What interests me, and what I set out to explain, is why the case has largely been failure. When Miranda was first decided, there was a sense of crisis in American law enforcement because of the fear that the case augured the end of confessions, and confessions, like it or not, are a critically important component in efficient and effective law enforcement. A surprising thing happened on the way to the Forum, however, which was – not very 1 2 3 4 5
John Henry Wigmore Professor, Northwestern University School of Law. 384 U.S. 486 (1966) 60 U.S. 393 (1856) 163 U.S. 537 (1896) 410 U.S. 113 (1973)
Ronald Allen
12
much. In what must pass as one of the high ironies of modern American jurisprudential work, the defenders of Miranda today largely argue for its irrelevance,6 and it does appear as though the criminal justice system seems to have accommodated (and in some people’s eyes, subverted) its commands. To me, the interesting question is why – what explains this remarkable phenomenon. I will give the briefest summary of my explanation here; the interested reader can consult my previous work for elaboration. The Fifth Amendment to the U.S. Constitution provides a right not to be compelled to incriminate oneself. That is all well and good, but of course the question is what does it mean. It originally meant that courts could not order a person to testify over the threat of contempt, but for many convoluted reasons this historical purpose was merged in a general concern over not compelling a person to speak. But, when does that occur? The obvious answer is that a statement is compelled when it is not voluntarily given or alternatively is coerced – it is not voluntary, in other words, when it is not made as an exercise of free will. This set the courts down the path of trying to distinguish between those statements that were exercises of free will and those that weren’t, although the vocabulary employed was that of voluntariness. In retrospect, it is obvious why this would be a disaster. If free will exists, it will still be enormously difficult to sort out when one is exercising it and when one is not. Indeed, I would go further and argue that, if free will exists, one always exercises it when one confesses. Unlike physical actions that can be forced – I can take your finger, put it on the trigger of a gun, and pull it back – vocalizations cannot. At the point at which I speak, I am obviously willing my body to act in a way that can only be understand as an exercise of free will. Even if my motive is to stop the pain that is being inflicted on me, I still have that choice. Suppose free will does not exist. Well, then, what exactly does “voluntary” now mean? It sounds like it should still refer to something in the mental realm, but there seems to be nothing left to which it can refer.7 The conjoining of these two points surely explains in large part the discontentment in the courts that led to the Miranda experiment. The Miranda court responded to these problems in an astonishingly creative way, but one that in retrospect merely recapitulated the very problem it was trying to resolve. The Court did not see that the problem lies in the very concepts at the heart of compulsion, coercion, and free will. Rather, it thought the problem was that a person’s will was too readily overcome by the “inherently compelling atmosphere of the jail house.” Thus, the solution was to bolster the exercise of free will. This was to be done through the four interrelated warnings: – –
6 7
You have a right to remain silent, which informs the person of a right that they may not know they possess. Everything you say can and will be used against you. In other words, you need to know and understand–comprehend–the significance of this right.
For a discussion of this literature, see Ronald J. Allen, J. L. Hoffmann, D. A. Livingston, W. J. Stuntz, 2005. Comprehensive Criminal Procedure. 2nd edition. 2005, 887–890. As I point out in my papers on this problem, the solution to this is precisely what the American courts were moving toward–a catalogue of acceptable and unacceptable police practices.
Introduction
–
–
13
You have the right to a lawyer. It is not just you against the police where you are helpless and be held indefinitely until you cooperate. Rather, you have the right to have the assistance of a lawyer, which opens a window to the outside world from the isolation of the interrogation room. If you cannot afford a lawyer, one will be appointed for you. What use is a lawyer for the typical indigent suspect? Not much, and thus they need to be told that their wealth or ability to secure counsel does not matter.
As I say, this was a tremendously creative effort, yet it fell completely flat. Why? The answer it turns out is pretty simple. Regardless whether free will exists, if the Court is correct about the inherently compelling atmosphere of the jailhouse – so compelling that innocent people will confess falsely to serious crimes – how could warnings make a difference? The very same compelling atmosphere will compel waivers and confessions at just about the same rate as before – which is more or less what happened. Moreover, moving from a voluntariness to a warnings and waiver regime replaced at least a somewhat sensible effort – examine what happened and try to figure out why the suspect confessed – with an extremely peculiar surrogate–were the warnings given and waived? In articulating this analysis, I went a step further. The deepest failure of Miranda was its blindness to the nonexistence of free will. That was its deepest failure because that is what drove it to offer an alternative that merely reflected the very problem at hand. Thus, the real irony of Miranda is that it replaced one search for free will with another, which is why it has failed. This, then, led me to defend with a straight forward argument the nonexistence of free will. It is a standard but powerful argument. Choices are caused by reasons or not. If they are not caused by reasons, whatever causes action cannot do the work of free will that the law requires, which involves identity, subjectivity, and agency. If they are caused by reasons, the reasons determine the actions, and therefore the actions are not free. Reasons themselves are held for (caused by) other reasons or not. Again, if not, there is no scope for free will. If reasons are held for prior reasons, those prior reasons cause us to hold the primary ones, and so on into an infinite regress leading back to the beginning of the universe. So either we live in a causal or some weird random universe, and there is no way out of that box. But in any case, the box does not possess free will. This argument on the surface appears to have shattering implications. Criminal law, for example, is based on blame, as are many other aspects of the law. If this argument is true, it suggests the end of responsibility, and so on. This is, I think, what caught Friedrich’s attention and led to his organizing this symposium. In my opinion, not much turns on the matter. For example, it is not shattering for confession law, for we can still regulate state force. More generally, regardless of whether free exists, we live in an “as if ” world and will continue to do so – even if you believe as I do; we will continue to operate as if free will exists, which is all that is needed. This comforting fiction does not deal with the philosophical problem, where perhaps the conclusion of no free will is even more shattering because of its denial of self, autonomy, agency, subjectivity, and so on. And thus one of the holy grails of philosophical thought has long been and remains putting free will on a firm footing. One sees this in the papers that follow, in which the authors strain against the burden of the apparent nonexistence of free will. Some try as philosophers have for
14
Ronald Allen
centuries to reestablish its existence; some accept determinism but try to preserve our way of life nonetheless; others simply accept the loss of innocence and try to reconcile to it. Let us hear from them, and then I will have a few words to say, not so much in rebuttal but in the spirit of continuing the conversation.
DAVID HODGSON, SYDNEY IN DEFENCE
OF
VOLUNTARINESS
There are four main points I wish to make in response to Ronald J. Allen’s stimulating article ‘Miranda’s hollow core’:1 (1) the idea of free will is defensible, and not ‘hokum’; (2) if free will exists, its exercise can be affected by the physical world; (3) it is reasonable to make voluntariness of confessions a condition of admissibility; (4) similar questions of voluntariness arise in other areas of law. FREE WILL DEFENSIBLE In a book2 and several published articles,3 I have been developing a version of free will that I believe is not merely reasonable but actually more likely to be true than contrary views. Rather than repeat or summarise what is in these publications, I will take a number of assertions in Professor Allen’s article and set out briefly why I disagree with them. The first is his assertion that free will requires ‘uncaused choices in the mental realm’.4 This assertion covertly assumes that causation can only occur by way of developments that exhibit either determinism (things occurring as uniquely determined by pre-development circumstances and laws of nature) or some combination of determinism and randomness (things occurring randomly within probability parameters determined by pre-development circumstances and applicable laws of nature). This is an assumption widely made by persons who criticise free will, generally without being either made explicit or supported by reasoned argument. And while there are good arguments that can support this assumption, there are also good arguments against it, which I believe are stronger.5 The view I advocate is that decisions, such as a person’s deciding what to do when there are conflicting reasons, are instances of causation in which the transition 1 2 3
4 5
Ronald J. Allen, Miranda’s hollow core, Northwestern University Law Review 100(1) (2006), 71– 85 David Hodgson, The Mind Matters, 1991 Particularly David Hodgson, Hume’s mistake, in The Volitional Brain, ed. B. Libet, A. Freeman and K. Sutherland, 1999, 201–224; Constraint, empowerment, and guidance: a conjectural classification of laws of nature, Philosophy 76 (2001), 341–370; Three tricks of consciousness, Journal of Consciousness Studies 9 (2002), 65–88; Free will, in MacMillan Encyclopedia of Cognitive Science, ed. L. Nadel (2003),Vol. 2, 150–156; A plain person’s free will, Journal of Consciousness Studies 12(1) (2005), 1–19; Response to commentators, Journal of Consciousness Studies 12(1) (2005), 76– 95; Partly free, Times Literary Supplement 5440 (6 July 2007), 15–16; Making our own luck, Ratio 20 (2007), 278–292; many of which articles are available on my ‘Reasonable free will’ website (http://users.tpg.com.au/raeda) Allen (note 1), 76–77 Hodgson (note 2), Hume’s mistake
David Hodgson
16
from the pre-decision circumstances to the post-decision circumstances is determined, within a spectrum of possibilities constrained by pre-decision circumstances and laws of nature, by the person’s decision itself, made for the reasons that have weighed with the person. The decisions are not uncaused: the pre-decision circumstances cause the decision-making process to occur; and the decision-making process is itself a natural form of causation. This leads to the next assertion I contest, namely that ‘if choices are made for reasons, those reasons determine the choices’.6 I think it is clear, on the contrary, that consciously-held reasons for action are characteristically inconclusive, and that there is a corresponding gap between reasons on the one hand and decisions and actions on the other.7 Hume said we always act in accordance with the preponderance of our desires, but that falsely assumes that desires, like forces in Newtonian physics, are commensurable, so that there is always a single ‘resultant’ desire that can direct our actions; whereas in truth there is no common scale on which (say) a feeling of hunger can weigh against a feeling of obligation to carry out a promised task. If ‘desires’ such as these conflict, the outcome is not determined by a preponderance of one over the other (because there can be no preponderance of incommensurables), but (I contend) by a choice between them that takes account of their different characters by means of a global assessment to which laws cannot apply. The third assertion I contest is that ‘modern science seems to place us in a world where all states of affairs are determined by previous states of affairs’.8 Of course, this assertion is true if it is taken as meaning merely that this is how it seems to some people; but I take Professor Allen to be asserting that this is the reasonable conclusion. Now not all states of affairs are the same as previous states of affairs, so it is clear that they cannot be determined by previous states of affairs alone. It is a possible view that science suggests that all states of affairs are determined by previous states of affairs and laws of nature, although I believe the majority of physicists would prefer the view that what previous states of affairs and laws of nature determine are probability parameters within which things occur at random (albeit that things at the scale we can observe generally look as if they were determined by previous states of affairs and laws of nature). I do accept that randomness as such is not conducive to free will; and the enormous success of science in explaining things in terms of laws of nature and randomness is one good argument in support of the assumption I mentioned earlier. However, science has not yet explained consciousness on this basis; and I believe it is in fact impossible to explain consciousness, or its role in behaviour, in terms of laws of nature and randomness. If it was, then anything that can be done with consciousness could be just as well done without it. Consciousness would be a superfluity, just as it would be in a computer: even if it were possible to make a computer that felt pain, it would be absurd to think it necessary or appropriate to use pain to motivate a computer to carry out its program.
6 7 8
Allen (note 1), 77 Hodgson (note 2), Hume’s mistake; John R. Searle, Rationality in Action, 2001 Allen (note 1), 77
17
In Defence of Voluntariness
I believe consciousness enables decisions to be made that are based on reasons that do not compel the conclusions on the basis of rules of any kind, but rather can support them as a matter of reasonable judgment. I will say more about this later. The fourth assertion I contest is that ‘if we are not responsible for our genetic makeup and past experiences, how can we be responsible for our actions? … We cannot.’9 In this assertion, Professor Allen is summarising a powerful argument made by Galen Strawson10 to the effect that we can never be truly responsible for anything we do, because we do what we do because of the way we are, so we cannot be responsible for what we do unless we are responsible for the way we are; and we cannot be responsible for the way we are when we first make decisions in life (that must be all down to genes and environment), so we can never become responsible (through earlier decisions) for the way we are later in life. I have answered this argument11 by arguments to the effect that the sense in which it is true that we do what we do because of the way we are is that (1) the way we are plus our circumstances plus laws of nature provide alternatives, inconclusive reasons, and tendencies, and also the capacity to choose between the alternatives on the basis of the reasons, and (2) what we do is what we choose in exercise of that capacity, the choice not being influenced by any differentiating features of the way we are otherwise than through the alternatives, reasons, and tendencies; that this leaves us with a degree of ultimate responsibility for what we do even if we are not responsible for the way we are; and that this means in turn that we can become partly responsible for the way we are, as our choices, for which we are partly responsible, come to supplement the effects of genes and environment on the way we are. For the details of my argument, I refer to those articles, which I contend do give the ‘cogent and robust account of free will’ that Professor Allen says is required.12 AFFECTATION
OF
EXERCISES
OF
FREE WILL
The next of Professor Allen’s assertions I contest is that ‘if free will exists, its exercise is unaffected by the physical world’.13 As noted above, I contend that the way we are plus our circumstances plus laws of nature provide alternatives, inconclusive reasons, and tendencies, and also the capacity to choose between the alternatives on the basis of the reasons; so that the exercise of free will is affected by the physical world in a number of ways. We have no alternatives outside the spectrum of possibilities left open by physical circumstances and physical laws. We have no experiences that can give us consciously-held reasons for choosing within this spectrum apart from experiences that arise from pre-choice circumstances and are correlated with physical brain processes. The way these reasons feel and appeal to us, and the tendencies to act that these and other 9 Allen (note 1), 77–78 10 Galen Strawson, Luck swallows everything, Times Literary Supplement (26 June 1998), 8–10; Galen Strawson, The bounds of freedom, in The Oxford Handbook of Free Will, ed. R. Kane (2002), 441–460 11 Notably in Hodgson (note 2), A plain person’s free will; Partly free; Making our own luck 12 Allen (note 1), 77 13 Allen (note 1), 77
David Hodgson
18
brain processes produce, also arise from pre-choice circumstances and are correlated with physical brain processes. Thus, while I do say that we have and cannot help having the capacity to make choices that are not pre-determined by pre-choice circumstances, I also say that this capacity is subject to the constraints and influences I have indicated, provided by the physical world. VOLUNTARINESS
AS A
REASONABLE CRITERION
One reason Professor Allen gives for contending that the test of voluntariness is useless is that it requires ‘an algorithm for distinguishing freely willed actions from determined ones’.14 I contend this is mistaken, for two reasons. First, the distinction between voluntary confessions, on the one hand, and involuntary or coerced confessions on the other hand, is not a distinction between freely willed actions and determined ones. Rather, the distinction being made is one between actions the motivation for which does not include unreasonable coercion or other unreasonable inducement from other persons, and actions the motivation for which does include such coercion or inducement. I agree with Professor Allen that, if there is free will, it is operating in both classes of case; so that it is not free will as such that provides the distinction. But I contend that the test of voluntariness does not suggest otherwise; that the test can operate even if free will is an illusion; but that at least part of its justification is an assumption that free will is not an illusion. I will return to these contentions shortly. Second, in requiring an algorithm, Professor Allen is requiring something that is not required or even possible in most legal reasoning. Most legal reasoning, indeed most human reasoning, is not algorithmic; that is, it does not proceed in accordance with rules of logic and/or mathematics and/or probability, or any other rules that could be incorporated into a computer program.15 Rather, it is informal plausible reasoning, in which the premises or data do not entail the conclusion by virtue of applicable rules, but rather support it as a matter of reasonable judgment. Arguments of Hume, Popper and others, particularly as developed by Hilary Putnam,16 strongly suggest that reasoning of this kind cannot be fully explained in terms of rules for good reasoning, whether they be rules of logic or mathematics or probability or whatever. I suggested earlier that it is consciousness which enables decisions to be made that are based on reasons that do not compel the conclusions, but rather can support them as a matter of reasonable judgment; and I contend that plausible reasoning depends in part on experiences, ideas and feelings consciously grasped as gestalt wholes, which enable judgments to be made that have regard to incommensurable reasons and to analogies that do not depend either on identity or on quantitative assessment of common features. This is part of my argument in favour of our having 14 Allen (note 1), 76 15 David Hodgson, Probability: the logic of the law – a response, Oxford Journal of Legal Studies 14 (1995), 51–68; Three tricks of consciousness, Journal of Consciousness Studies 9 (2002), 65–88; A plain person’s free will, Journal of Consciousness Studies 12(1) (2005), 1–19 16 Hilary Putnam, Reason, Truth and History, 1981, 174–200
In Defence of Voluntariness
19
the capacity to make decisions that are not determined by pre-decision circumstances and laws or rules of any kind. I believe the only alternative to this account of plausible reasoning is to seek to explain it wholly in terms of unconscious computational processes that do not have any validity on the basis of logical rules or other rules for good reasoning, but which work because they have been selected in evolution for their effectiveness in promoting survival and reproduction. But this introduces a vicious circle into justification of plausible reasoning. If we cannot rely on our plausible reasoning as the conscious non-algorithmic process it seems to be, and on associated feelings of assurance, then any confidence we could have in it would have to depend on the belief that plausible reasoning is supported by computational processes whose reliability is assured by the evolutionary tests they have passed; yet this belief would itself have to depend on extensive plausible reasoning, giving rise to a vicious circle.17 Disagreements in matters of plausible reasoning could not be addressed rationally: so long as identifiable fallacies were avoided, there would be no basis on which one process of plausible reasoning could be preferable to another. Returning to the test of voluntariness, although this test is often associated with language that refers to involuntariness or the overbearing of the will, I think it is plain, when the context is considered, that what is being referred to is circumstances where coercion or other improper influence has induced the person to confess; so that the confession is not voluntary in the sense that it is not one the person made unaffected by any improper influence of this kind. That is, the concept of involuntariness operating here is one of actions induced by improper influence. Although, on my account of free will, it may not be a matter of absolute certainty what would have happened had the improper influence not occurred, the constraints and influences on choice discussed above generally make it possible for judgments to be made as to what would (more or most) probably have happened; and this is all that is required in this and other areas of law in order to decide whether or not some conduct by a person was induced by wrongful conduct of others. On this analysis, the test of voluntariness does not require that there be free will in the sense I support: it is sufficient that judgments can be made about whether or not coercion or other improper influence has been applied, and if so, whether or not it induced the confession. There is no doubt that the objectives of discouraging abusive police tactics and excluding confessions that are likely to be false are important reasons for requiring that confessions satisfy the test of voluntariness, reasons that apply whatever one’s view is about free will; and Professor Allen raises the question whether these objectives could be pursued more effectively by using some different tests. But I say that there are further reasons that justify using the test of voluntariness, and that there are, in Australia at least, already additional tests more directly related to these objectives. My view is that considerations of justice or fairness justify the test of voluntariness. It is considered fair that a confession that a person makes, which is not induced by improper influences, should be admissible against him or her, because it is the 17 Thomas Nagel, The Last Word, 1997, Ch. 7; Alvin Plantinga, Warrant and Proper Function, 1993, Ch 12
David Hodgson
20
person’s own choice, free of improper influences, that subjects him or her to the risk of being wrongly convicted on the confession if it is untrue or misinterpreted. It is considered unfair that a confession induced by improper influences should be admissible, because it is not the person’s ‘free’ choice, but something induced by improper conduct, that would, if the confession were admitted, subject the person to the risk of wrongful conviction if the confession is untrue or misinterpreted. I believe that most of us, if considering under a Rawlsian veil of ignorance what rules should apply to confessions, would be prepared to accept the risk of wrongful conviction based on confessions freely given in this sense, but not prepared to accept the risk of wrongful conviction based on confessions induced by improper means. Indeed, I think that this consideration of fairness is the fundamental reason for excluding improperly obtained confessions, and that the considerations that such confessions are more likely to be false and that exclusion of such confessions tends to discourage police abuses are secondary and consequential, albeit still highly important. I note also that in Australia at least, there is in criminal cases a requirement that evidence whose prejudicial effect outweighs its probative value should be excluded, and also a judicial discretion to exclude evidence that has been improperly obtained; so the two objectives identified by Professor Allen can be pursued directly quite independently of rules about the voluntariness of confessions. If, as I believe, this idea of fairness is an important part of the justification for the exclusion of confessions that are involuntary in the sense I have explained, then ideas about free will may have a part to play in this justification. Whether a person’s choice, free of improper influences, should be given such significance may depend on whether or not one takes the view that the choice is, in any event, pre-determined by physical circumstances outside the person’s control. In a similar way, ideas about free will play a part in justifying the idea that it is fair to punish people for voluntary conduct in breach of a public law, to an extent that is somehow proportionate to their degree of responsibility, but unfair to punish them in other circumstances or to an extent greater than they ‘deserve’. I believe this idea is an important aspect of human rights, defended in most Western countries and abused in some other countries; and in a paper presented at the IVR conference in Lund in 2003 I argued that this aspect of human rights could be damaged by dismissal of free will on the basis of arguments that were at best inconclusive and premature.18 VOLUNTARINESS
IN
OTHER AREAS
OF
LAW
The last main point I want to make is that similar considerations arise in other areas of law where the voluntariness of conduct is considered important. One such area is the question of consent of a woman to sexual intercourse in a case where a man is charged with rape. One element of the crime of rape is that the sexual intercourse be without the consent of the woman, and purported consent has to be voluntary, in the sense that 18 David Hodgson, Responsibility and good reasons, Ohio State Journal of Criminal Law 2.2 (2005), 471–483
In Defence of Voluntariness
21
it does not count as consent if it was coerced or obtained by threats or deception. In cases where consent was obtained by coercion or threats or deception, the consent would still generally be an exercise of free will, if free will exists; but it would not make the sexual intercourse permissible because it would not be considered voluntary, in that it was induced by influences that are considered improper. The question whether a woman’s consent to sexual intercourse was voluntary in this sense is a question that has to be decided, beyond reasonable doubt, in some criminal cases. Such decisions do not require resolution of subtle questions about free will, but rather common sense judgments about whether what happened amounted to coercion or threats or deception, and if so whether the purported consent was induced by this coercion or threats or deception. There is no algorithm to determine these matters, only plausible common sense reasoning. However, questions of free will do underlie this area of law as well. Sexual intercourse between a man and a woman that is not consented to by the woman is a very serious assault indeed, and may affect her significantly, albeit that there will be no crime if the man believes that the sexual intercourse is consented to. The essential thing that distinguishes between conduct that is a very serious assault and conduct that is lawful is consent, or the man’s belief that there is consent. Underlying this is the idea that the woman herself can choose what happens in this serious way to her body, so that sexual intercourse is generally permissible if she consents and impermissible if she does not. That idea has less force if the woman’s choice is seen as something other than one she is free to make, in that it is determined by physical circumstances outside her control. One other area where questions of voluntariness of conduct can arise is the law concerning the setting aside of transactions for undue influence or unconscionable conduct. If property is transferred by one person to another, for example, for no consideration or for less than it is worth, generally the transferor cannot recover the property unless it is proved that the transfer was not voluntary in the sense that it was induced by undue influence or fraud or other unconscionable conduct. Questions are raised that do not require resolution of questions about free will, but rather require consideration of common sense issues about motivation. However, underlying and justifying this law is the idea that it is fair that a person should be bound by a transaction that was ‘freely’ chosen, but not bound by a transaction which was induced by improper conduct. CONCLUSION So my contention is that the criterion of voluntariness, understood in the way I have explained, is an appropriate criterion for admissibility of confessions in criminal cases, as it is for various issues in other areas of law; that its application does not require consideration or resolution of questions concerning free will; but that there are ideas of free will underlying principles of justice and fairness in all these areas of law; and that dismissal of free will on the basis of inconclusive arguments could prejudice these principles.
JOHN WARWICK MONTGOMERY, STRASBOURG1 THE FREEWILL ISSUE
IN
THEOLOGICAL PERSPECTIVE
The freewill/determinism issue has not been a concern solely of secular metaphysicians and philosophers of law. Theologians also have wrestled with this intractable problem. The present paper considers the three major approaches to the issue as presented in Western theology: that of Roman Catholicism/Protestant Arminianism, Lutheranism, and Calvinism. A sound theological approach is seen to have distinct advantages over against secular treatments, both in general terms and in the sphere of legal philosophy. Part One of the The Oxford Handbook of Freewill is devoted to “Theology and Fatalism.”2 The Handbook quite properly recognises the place of freewill discussions in the history of Christian theology and their potential value to the analysis of that crucial issue in other domains such as legal theory. The purpose of this essay is to outline the positions classically taken on the freewill issue in Christian dogmatics and to see whether they can shed light on the freewill/determinism controversy in general. Before presenting the theological alternatives, however, it may be worthwhile to observe the state of the question in secular thought. On the one hand, it seems logical to assume that the genetic makeup of the individual covers all aspects of his or her actions; were we to have a complete map of that genetic situation in the case of any given person, we could presumably predict all of that individual’s life decisions. However, such a deterministic conclusion flies in the face of our need to establish responsibility for human action—particularly in the case of antisocial behaviour, where one can hardly be allowed to push responsibility back upon one’s progenitors and thereby avoid the consequences of one’s acts. To take but one legal example, the French Cour de cassation in the important Laboube case declared: “Encore faut-il, conformément aux principes généraux du droit, que le mineur dont la participation à l’acte matériel à lui reproché est établi, ait compris et voulu cet acte; toute infraction, même non intentionnelle, suppose en effet que son auteur ait agi avec intelligence et volonté”.3
In short, in spite of Herculean efforts to arrive at rational compatibility between genetic determinism and freely chosen human actions,4 the paradox remains: in 1
2 3
4
Ph. D. (Chicago), Th. D. (Strasbourg, France), LL.D. (Cardiff, Wales). Distingnished Research Professor of Philosophy and Christian Thought, Patrick Henry College; barrister-at-law, England and Wales; avocat à la cour, barreau de Paris. The Oxford Handbook of Freewill, ed. R. Kane, 2002 Crim. 13 déc. 1956, Recueil Dalloz, 1957.349, note Patin (italics ours). “It is still necessary, in conformity with the general principles of law, that the minor whose participation in the actus reus has been established, should have understood and willed this act; every offence, even nonintentional ones, suppose in effect that its author has acted with intelligence and will.” One of the most striking is the argument for psychological dualism/interactionism by Nobel prize winning neurophysiologist Sir John C. Eccles: “If my uniqueness of self is tied to the genetic uniqueness that built my brain, then the odds against myself existing in my experienced uniqueness are 1010,000 against”, Popper, K. R. and Eccles, J. C., The Self and Its Brain, 1985, 559.
John Montgomery
24
theory, our acts are predetermined, yet in practice we must take personal responsibility for them in order to maintain a functioning civilised society. Einstein put it succinctly: “I am a determinist, compelled to act as if free will existed, because if I wish to live in a civilized society, I must act responsibly. I know philosophically a murderer is not responsible for his crimes, but I prefer not to take tea with him.”5 THE THREE CLASSIC THEOLOGICAL APPROACHES The history of Christian thought has provided three major understandings of the relationship between divine providence and human freedom: the Roman Catholic/ Arminian view; the Calvinist view; and the Lutheran view.6 Notably, since the point de départ of Christian theology is divine revelation rather than human speculation, these approaches are not general attempts to resolve the destiny/freewill issue, but focus (as does Holy Scripture) on the matter of personal salvation. The question for the theologians has been “What, in the final analysis, accounts for the saved person being saved, and what accounts for the unbeliever remaining in his or her unbelief?” The following diagram sets forth the three classic positions: Election/Predestination and Freewill in Human Salvation
The Saved
The Unsaved
RC/Arminian
Lutheran
Calvinist
Election
Election
Election
Freewill
Freewill
Freewill
Election
Election
Election
Freewill
Freewill
Freewill
Several clarifications are immediately necessary to make this conceptualisation understandable. These are best presented by way of the three confessional positions represented in the chart. The Roman Catholic view has always emphasised the controlling place of the human will in salvation. God’s grace alone provided the means of human salvation through the gift of His Son Jesus Christ, but to benefit from this gift one must exer5
6
Albert Einstein, quoted in Denis Brian, Einstein: A Life, 1996, 185: “Quantum theory, to be sure, does not support such determinism. “As in Newton’s world, the actors in Einstein’s world parrot their lines from a script that was written beforehand. But in a quantum play, the actors suddenly throw away the script and act on their own. The puppets cut their strings. Free will has been established”, Kaku, M., 2006, 149. Einstein would of course reply that without consistent physical laws, one could not establish the soundness of quantum theory in the first place. A valuable historical survey of western Christian approaches to the freewill issue (though lacking in sympathy for Protestant viewpoints and appreciation for the logical irreconcilability of the Lutheran approach with the classic Roman Catholic position) is Quilliet, B, L’Acharnement théologique: Histoire de la grâce en Occident IIIe–XXIe siècle, 2007.
The Freewill Issue in Theological Perspective
25
cise his or her freewill—by personally accepting the Church’s sacramental provisions whereby the “treasury of Christ’s merits” becomes available to those who repent and agreeing to the penitential ministries of the Church as the extension of Christ’s body in history. An interesting recent illustration of this viewpoint is seen in a comment by Monseignor Ravasi, prefect of the Ambrosian Library, Milan, and member of the Pontifical Biblical Commission, when asked why Judas’ betrayal of Jesus has been included in the recent revision of the Stations of the Cross; said he: “The episode shows that we all have been given free will and a conscience.”7 The freewill explanation of both salvation and damnation has also been maintained by the followers of Dutch Protestant theologian Jacob Arminius (1560–1609), who developed his position over against that of the strict Calvinism of his time. It is also represented by the so-called “Freewill Methodists,” “Freewill Baptists,” and Arminian Evangelicals who assert a direct causal relationship between making a “decision for Christ” and salvation. Polar opposite to the Roman Catholic/Arminian view is that of classic Calvinism, which holds that the efficient cause of both salvation and of damnation is the “election” (i.e., decision) of God in eternity. Predestinarian Calvinism comes in two varieties: “supralapsarian” (God’s decision preceded even the Fall of man) and “infralapsarian” (that decision was not made until after our first parents sinned). Not all Calvinists by any means take these positions today. A number of Presbyterian church bodies have removed the predestination article from the text of the Westminster Confession of Faith (a prime Calvinist doctrinal statement). Historically, the French Calvinist theologian Moïse Amyraut or Amyraldus (1596–1664) formulated a theology of freewill virtually indistinguishable from the Lutheran view. The Lutheran position, which is also that of mainline Anglican theology,8 endeavours to take into account the full range of biblical teaching on the election/ freewill issue. On the one hand, Scripture is definitive in its teaching that no one can save himself or herself by any good work, including any act of human will (John 1:12–13, Ephesians 2:8–9). The believer must not therefore attribute his or her salvation to any other source than God Himself, working through His Holy Spirit. On the other hand, unbelief is never presented as the result of God’s decision to damn; damnation is the product of the misuse of the creature’s freewill (Matthew 23:37). Efforts have been made to assimilate the Lutheran view to that of Calvinism. Thus James Packer, in his edition of Luther’s De servo arbitrio, presents Luther as holding that “the cause of salvation and damnation alike is the sovereign will of God.”9 However, this interpretation of Luther simply does not wash. True, Luther fought tooth and nail against the Roman Catholic and Renaissance humanist position of Erasmus that freewill is the effective cause of salvation; and, so important to Luther was the doctrine of salvation by grace alone through faith, that he sometimes expressed that teaching in extreme terms. But if we compare Luther with Calvin, 7 8
9
Gianfranco Ravasi, Way of Sorrows To Call at New Stations, The Times London: 6 April 2007 I.e., the Anglican mainstream which is neither “low church” (essentially Calvinistic) nor “high church” (essentially Roman Catholic in its theology – without, to be sure, accepting the authority of the Pope) J. I. Packer and O. R. Johnston, eds. and trans., The Bondage of the Will by Martin Luther, 1957, 55
John Montgomery
26
Robert Will is surely correct in seeing “une différence de tempérament très nette (…) entre la liberté [de Luther] (…) et la détermination avec laquelle Calvin, dans l’intransigeance de sa raison française, mit en pratique ses principes de liberté.”10 As Robert Kolb has emphasised, for Luther, “God is not responsible for evil. No explanation of the existence of evil and its continuation in the lives of believers is possible.” We are to recognise how “unsearchable are God’s ways” (Romans 11) and be driven “to reliance on the goodness of God and to trust in Jesus Christ.”11 If further evidence of Luther’s true position were needed, one could simply go to his Theses for the Heidelberg Disputation of April 1518, where he states in Thesis 14 that “’Freewill’ after the fall has the potentiality toward good as an unrealisable capacity only [subiectiva potentia]; towards evil, however, always a realisable one [activa potentia].”12 The Lutheran position—often referred to as “single predestination,” since divine election applies only to the saved—lacks the consistency of the Roman Catholic/ Arminian viewpoint (freewill across the board) and that of the Calvinist “double predestination” (divine election across the board, affecting both the saved and the lost). But it has the great merit of taking into account all the biblical data. An interesting illustration in this regard is the passage in the Acts of the Apostles (16:30–31) recounting the Apostle Paul’s encounter with his Philippian jailer. The jailer asks, “What must I do to be saved?” Paul (who wrote Ephesians, declaring that one is saved solely by God’s grace and not by what one does or wills and that faith itself is God’s gift) replied: “Believe in the Lord Jesus Christ and you shall be saved.” An act of will is required; but once that act has taken place, it must be attributed to God the Holy Spirit and not to the individual—“lest anyone should boast.”13 FURTHER ANALYSIS AND A CONCLUSION Several questions are worth raising at this point. (1) Could one not eliminate the paradox in the Lutheran viewpoint (and vindicate the Roman Catholic/Arminian approach) by observing that in Scripture predestination is made conditional upon divine foreknowledge (Romans 8:29)? The problem here is that the biblical understanding of “foreknowledge” entails the notion of divinely created knowledge—and is thus simply another way of expressing divine sovereignty. God’s foreknowledge “is an election or foreordination of His people (R. 8:29; 11:2) or Christ (1 Pt. 1:20).”14 The idea is not that God looks forward in time to see who will 10 R. Will, La liberté chrétienne. Etude sur le principe de la piété chez Luther, 1922, x–xi; cf. G. O. Forde, The Captivation of the Will: Luther vs. Erasmus on Freedom and Bondage “Lutheran Quarterly Books”, 2005, especially 32 ff. 11 R. Kolb, Bound Choice, Election, and Wittenberg Theological Method “Lutheran Quarterly Books”, 2005 12 Martin LutherWA [the standard, authoritative Weimarer Ausgabe of Luther’s Works], vol. I, 1883, 353–54 13 See Montgomery, J. M. 1997, The Holy Spirit and the Defense of the Faith, Bibliotheca Sacra 154 (October–December 1997), 387–95. 14 Theological Dictionary of the New Testament, ed. G. Kittel, trans. Geoffrey W. Bromiley (10 vols.), vol. I, 1964–1976, 715.
The Freewill Issue in Theological Perspective
27
believe and who will not and then ratifies what the human creature decides, but rather that divine election/predestination takes place as a result of the action of the divine mind. In short, one cannot solve the paradox by pitting foreknowledge against election. (2) Can one not get around the problem by Ockhamist thinking or by Molinist “middle knowledge”? (After all, Luther himself had Ockhamist instructors early in his theological and philosophical education!) Zagzebski points up the great difficulties with both of these approaches. “In my opinion a serious problem with Ockhamist solutions is that even if they can produce an account of temporal asymmetry that has the consequence that God’s past beliefs do not have the necessity of the past, it is unlikely that this can be done in a way that is independently plausible.”15 On Molinism, Zagzebski cites Walls who “argues that since Molina maintained that God chooses to put people in situations in which he knows they will choose damnation, Molinism is as morally abhorrent as the Calvinist doctrine of predestination.”16 (3) Should not the Lutheran approach be rejected simply on the ground that it embraces a formal contradiction? If one subtracts the saved elect from the totality of the human population, must not the lost be regarded as in that category because they are the non-elect? Or if one subtracts the lost—who misused their freewill—from the sum total of humanity, must not the saved be seen as having arrived there through a different but equally real act of will (they did not reject the grace of God)? Luther’s answer is simply that since “God’s thoughts are higher than our thoughts” (Isaiah 55:9), one must stick with the Word of God in Holy Scripture no matter what, and must never draw inferences from one passage of Scripture which would contradict the clear teaching of other biblical passages. Putting it another way, Luther places fact (here biblical fact) above formal questions of contradiction. Life, for him, is bigger than logic. This may seem initially irrational, but at the frontiers of science, the same approach operates. Thus, though the properties of particles are not those of waves (and the two are in various respects logically incompatible), where two sets of equally good experiments lead to the conclusion that light is both particulate and undulatory, one works with the “photon” (a “wave-particle”) regardless of the logical difficulties present in such a solution. The alternative is clearly unacceptable, for it would involve refusing to recognise one set of sound experiments or the other.17 Of course, Luther’s reasoning is founded on a confidence that the Bible is indeed God’s Word and therefore that none of its asseverations can rightly be ignored. Support for that claim would take us well beyond the bounds of this paper.18 15 L. T. Zagzebski, Recent Work on Divine Foreknowledge and Free Will, in: The Oxford Handbook of Freewill, ed. R. Kane, 2002, 54–55 16 Zagzebski (note 14), 57; J. Walls, Is Molinism as Bad as Calvinism? Faith and Philosophy 7 (1990), 85–98 17 Cf. K. S. Thorne, Black Holes and Time Warps, 1994 (with a Foreword by Stephen Hawking), especially 147. 18 For a full-scale argument to this effect, see J. W. Montgomery, Tractatus Logico-Theologicus, 2003;
28
John Montgomery
(4) What can a theological approach—and, specifically, the Lutheran—offer to the general and the legal discussion of the freewill issue? We shall make four suggestions in conclusion. First, freewill is established on a transcendent foundation—on the basis of clear revelatory teaching, and each individual is held responsible morally and legally for his or her acts (Galatians 6:7). Secondly, because the most important possible decision in life, that of entering into a saving relationship with God, does not have its ultimate explanation in man’s freewill but rather in God’s sovereign love, humans are given every reason not to exercise hubris in thinking that they can build towers of Babel so as to climb up to God by their own self-centred efforts. Thirdly, the promise of a Last Judgment means that where judicial error has occurred or for any other reason human beings have escaped the consequences of the misuse of their freewill on earth, they will not escape those consequences in eternity. Finally, though the theological answer does not resolve the paradox of determinism/freewill as it exists in secular thought, it places it in the context of a loving God who sent His only Son to die for an undeserving race and “who will have all people to be saved and to come to the knowledge of the truth” (1 Timothy 2:4).19 How much more satisfactory is the biblical gospel than the conclusions to which secular theorising leads in an attempt to resolve the destiny/freewill issue. Consider again Einstein: “Human beings, vegetables, or cosmic dust, we all dance to a mysterious tune, intoned in the distance by an invisible player.”20 Or playwright Glen Berger (Underneath the Lintel): “A magician tells you to choose any card in the deck, (Increasingly bitter) and so with free will you do choose…but you don’t realize the magician has already subtly forced you to pick the exact card he wanted you to pick. Magicians call that a ‘Hobson’s Choice.’ And in life we think we make choices…but they’re Hobson’s Choices. So who is this Hobson? Who is this magician gulling us? That’s the question. Simply something named Chance? Or Fate? (Looking up) Or Something Else?”21 “Something Else,” indeed. Rather, Someone Else. And Someone who says that not a sparrow falls from a tree without the knowledge of our Heavenly Father and that we are of more value than many sparrows. The sovereign decisions of this “magician” are saving acts, and if we insist on employing our genuine freewill to thwart His love and grace, we have only ourselves to blame for the results, both in time and in eternity.
also John W. Montgomery, History, Law and Christianity, 2002. 19 This essay is not the place to enter into casuistical areas such as that represented by a recent Roman Catholic work written by Simon Francis Gaine, O.P, ‘Will There Be Free Will in Heaven?’ Freedom, Impeccability and Beatitude, 2003. The answer to that question, by the way, is Yes: owing to a radical character change in the saved individual (2 Corinthians 5:17), he/she in eternity will no longer seek to use freewill negatively so as to reach sinful decisions. Augustine properly described this as the state of non posse peccari. 20 Albert Einstein in Brian (note 4) 21 Glen Berger, Underneath the Lintel 2003, 28–29. Currently (April 2007) Richard Schiff is starring in the play at the Duchess Theatre in London’s West End.
FRIEDRICH TOEPEL, BONN FREE WILL I. REGRESS
AND
OR
INFINITE REGRESS
SURD
OF
REASON?
Professor Ronald Allen has highlighted the oddity of the notion ‘free will’ in his paper ‘Miranda’s Hollow Core’ in the following way:1 Free will exists when free choices exist. But choices can only be made either for reasons or not for reasons. Choices which are made for reasons seem to be free only if the reasons for which the choices are made are not themselves determined. But reasons are also chosen either for prior reasons or not for prior reasons. Consequently, we are faced with a dilemma. Only two possibilities remain and neither can ever justify us to call a choice free with the connotations associated traditionally with the term. If choices are always made for reasons which are justified by choices of prior reasons, we are never able to call a choice free, because it will not be possible for us to see the whole chain of reasons, and if we do not see the whole chain before us we have no justification to pronounce it free. The regress into which the choices lead, may be infinite, but it needs not even to be infinite to worry us. It is enough that the likelihood increases with the length of the chain that it at some point completely depends on our desires which in turn are at the mercy of external influences and insofar definitely unfree. If on the other hand the chain leads back to some choice which is made for no reason at all, then we may be sure that the chain of reasoning is not determined, yet it also seems not to be appropriate to call it free in the sense of such choices which need protection from the legal point of view. The exercise of autonomy traditionally associated with free will consists in the possibility of a person to follow one’s own reasoning and to decide accordingly. A choice for no reason at all therefore is no exercise of autonomy but comparable to random fluctuations in the universe. If this would be a correct account of the situation, then ascription of free will to a particular person is always incoherent talk. The two prongs of Professor Allen’s dilemma are an infinite regress in case of choices for reasons and a surd of reason in case of choices for no reasons at all. It seems to be plausible that free will matters because persons should be able to follow their own reasoning, not because persons should be able to make decisions for no reason at all.2 Consequently, I do not begin with an attack of the surd-of-reason prong of the dilemma. I rather concentrate on the remaining prong, the danger of a regress:
1 2
Ronald J. Allen, Miranda’s Hollow Core, Northwestern University Law Review 100 (1) (2006), 77 Though – as we shall see in section VI – I do not think that following one’s reasoning must imply complete determination of the choices by the reasons with no arbitrariness involved at all.
Friedrich Toepel
30 II. CONDITIONAL ANALYSIS: FREE CHOICES WEAK SENSE
AS
SUFFICIENT CONDITIONS
IN A
If we ask why choices for reasons and prior choices of these reasons for prior reasons can be considered a chain which leads to a regress, a general answer would be that prior reasons and prior choices are conditions for reasons and choices. However, conditions are either necessary or sufficient conditions. A chain in Professor Allen’s sense consists of two kinds of members. There is a choice to have a reason and this reason in turn becomes the reason for another choice etc. until the last reason becomes a reason for a choice to act. The link between the choice to have a particular reason and having this reason presents no problem. The relationship is analytic. Whoever chooses a reason, by definition also has this reason. Choosing the reason is analytically sufficient for having the reason – at least in the moment of the choice, though I may change or give up my reasons later. Not so with our ordinary understanding of the way in which a particular reason influences a particular choice. Having the reason does not logically necessitate the choice. It would perhaps be possible to call the relationship between having the reason and making the choice a causal one,3 but with an important difference as to regular causal sequences. When I kick a stone over a rocky edge into a pond below, the stone causes the splash in the moment it drops into the water in accordance with the laws of nature. The dropping of the stone is a causally sufficient condition for the splash. But when a person has a reason to decide and decides in accordance with the reason, then we do not view that person’s reason as a causally sufficient condition in a strong counterfactual sense for the person’s choice.4 At best we say that having the reason only becomes a sufficient condition of the choice in a strong sense when it is supplemented by a special acceptance of the reason as relevant in the particular case. We say this precisely because we associate decision-making for reasons with free persons. I shall call such a condition a sufficient condition in the weak sense:5 A reason normally is only a condition sufficient in the circumstances if we presuppose that it has already been accepted in a free choice. We impute a nephew’s killing his uncle for 3
4 5
Insofar I agree with Justice David Hodgson, In Defence of Voluntariness, in this volume on page 16, where he calls the relationship between pre-decision circumstances and the decision-making process a form of causation. To my mind however, nothing important should depend on the choice of words whether to call the pre-decision circumstances causes or mere conditions. Philosophers have proposed to abolish the notion of causation completely in favour of the more precise talk about conditions (Willard Van Orman Quine, Pursuit of Truth 1990, 76; Skyrms, Choice and Chance – Introduction to Inductive Logic, 1985, 84 ff.). John L. Mackie, The Cement of the Universe, 1974, 39–40; cf. also John R. Searle, Freedom and Neurobiology: Reflections on Free Will, Language and Political Power, 2007, 58–66. Mackie has likened persons whose reaction is mediated psychologically to indeterministic slotmachines. If Tom insulted Bill, we can say that the insult caused Bill to hit Tom, even if we presuppose that Bill had free will. Bill is behaving like Mackie’s indeterministic slot-machine L: “I put a shilling into L and receive a bar of chocolate. Putting in the shilling was, in the circumstances, necessary for the result. It was also sufficient in the circumstances in a weak, not in a strong counterfactual sense. A possible world, with the same laws of working as the actual world, can lack the result, and yet still contain the inserting of the shilling.” Mackie (note 4) 39 – 41; see also Richard W. Wright Causation, Responsibility, Naked Statistics and Proof: Pruning the Bramble Bush by Clarifying the Concepts. Iowa Law Review 73 (1988), 1001.
31
Free Will and Infinite Regress
the reason that the uncle is rich and the nephew his sole heir to the nephew because we expect the nephew to resist his impulse to kill the uncle even after he has discovered the reason to kill the uncle having become aware of the uncle’s richness and the fact of being the uncle’s sole heir. Our concept of reasoning and making choices implies this. If the nephew would be drawn with an irresistible impulse to killing his uncle after having become aware of the uncle’s richness and the fact of being the uncle’s sole heir, this reason would indeed be a strongly sufficient condition for killing the uncle, but the reason would not work in a way in which we normally think that reasons influence choices. It would rather work akin to the reasoning of a person under the influence of a psychosis where we exclude freedom of choice. III. TESTING FREEDOM ALL EARLIER CHOICES
OF
CHOICE
WITHOUT
HAVING
TO
TEST
THE
FREEDOM
OF
Is it relevant to the impact of the regress whether causes are sufficient only in the weak sense just described or in a strong counterfactual sense? Yes, for the strong counterfactual sense does not leave room for a gap which can be filled by an indeterministic decision to complete the sufficient condition while the sufficient condition in a weak sense does exactly that. A sufficient condition in a weak sense does not determine events in the way of sufficient conditions in a strong counterfactual sense. Each choice rather demands a new acceptance of the reasons relevant for the choice. A prior choice never triggers a later choice without the possibility that the person choosing changes his mind in the mean time. Therefore, in a regress consisting of choices the prior unfree choices can but do not have to infect the later choices. Later choices may still be free, even if the earlier choices are not. Whether a choice is defective in the sense of lacking freedom then depends on whether the defect is still present in the choice at issue. A choice can be defective in two ways: Either the freedom of the choice may be impaired by a false cognitive basis or it may be impaired by coercion on account of threats or violence which have influenced the decider. Presence of the defect means that the error or the coercion must be of relevance for the choice at issue. However, if only the relevance of the defect for the choice is of interest, then we need not take account of the freedom of former choices to determine whether a choice is free. We rather have to look only for defects like error and coercion and decide whether someone can be called the author of a choice only in the absence of such error or such coercion. The relevance of the defect may depend on the object of the choice. Difficult questions may arise when trying to separate relevant from irrelevant defects, but the difficulties have nothing to do with an infinite regress anymore. To demonstrate this, I discuss the following examples: 1. Error Let us imagine a situation before a police interrogation. The suspect deliberates whether to confess or not. He knows that his mother has strict moral principles and would always want him to confess, while his father is a criminal who would never
32
Friedrich Toepel
wish that he would confess. He also knows that either his father or his mother must have reported his crime to the police, but he does not know who of both has done that. He could not accept the strict principles of his mother in case that she would have reported his crime to the police. Therefore, he asks his brother who knows the truth. The brother gives him the false information that their father had reported the suspect to the police, although in reality their mother has been the one who did that. Consequently, the suspect reasons: My mother has not given me away. So, I have more respect for her moral principles than for those of my father. My mother thinks that it is always best to make a clean breast of what you have done. This principle I accept. Therefore, I confess. Whether the suspect freely chose to confess does not depend on the fact that he fell into an error when making a former choice which was part of a chain of choices leading up to the one at issue. The question rather depends on the object of the choice whose freedom is legally guaranteed. If a suspect’s right would exist not to follow the moral principles of someone who has reported the suspect’s deed to the police, then the object of the choice at issue would be the acceptance of the moral principles of another person. Yet, the legislator has not equipped the suspect with a right to silence so that the suspect might freely follow his own motives. The object of a decision whether to confess or not rather is the waiver of the suspect’s privilege against self-incrimination. False information about the person who betrayed the suspect does not affect a choice to confess because it does not in any way conceal to the suspect how far he may influence the further trial with his decision. Therefore, such misinformation cannot remove the freedom of the choice to confess. However, the wrong information that the suspect’s accomplice has already confessed would, for example, be relevant for the suspect’s freedom of the choice to confess because the accomplice’s confession relativises the importance of the suspect’s own choice. It may make it appear to the suspect as if the truth would be known without the suspect’s confession. Therefore, this information is able to make the impact of this choice appear less than it really is. 2. Coercion Gaps between reasons and choices also influence the understanding concerning the relevance of coercion like threats or violence. If a boy has accepted moral principles because of having been caned quite often, it does not follow necessarily that he cannot make free choices anymore when he later applies the principles he learnt in that way in a situation where no caning is threatening. In the same way, if a suspect has been induced to make the decision to confess under threats, it is not excluded that he may make a free confession if he still maintains his opinion which he acquired under threats when he makes the confession at a time at which he no longer fears that the threats might be carried out. We therefore have to exclude the continuing influence of the coercion to call a choice free, not to determine whether the coercion first influenced a choice in a chain leading up to the choice at issue.
Free Will and Infinite Regress
33
IV. INCOMPATIBILISM At the beginning of section II I referred to our ordinary understanding of the way in which a particular reason influences a particular choice. Therefore, my solution does only explicate ordinary language which favours indeterminism. It does not prove that the use of ordinary language is correct. To the contrary, Professor Allen correctly describes the dilemma, if he already presupposes that determinism is an option. He who presupposes determinism must indeed assume that searching for a free choice leads to an infinite regress because for him reasons are sufficient conditions in a strong sense for the subsequent choices. Thus the dilemma becomes another argument against compatibilism: If determinism is taken serious, it is not possible to speak of free will without falling a prey to the dilemma. Yet, the dilemma does not show anything regarding a position which is indeterminist and incompatibilist at the same time, for such a position excludes by definition the chain which is necessary to support the infinite regress argument. V. THE PRINCIPLE
OF
SOLE ATTRIBUTION
At present, all who object to an infinite regress argument by pointing to its questionbegging character must be prepared for the rejoinder to be found for example in Smilansky’s defence6 of Galen Strawson’s regress argument7: Someone just referring to the question-begging character of the regress would ignore that there is ethical motivation of the deepest sort for a principle of sole attribution. Having such a sort of foundation in radical self-creation would be required for the libertarian aim of achieving moral worth. So, requiring an explainable ultimate responsibility would be nothing more than the libertarian must require from himself: An account would have to be given as to which non-arbitrary factor brought about the decision or action and why. Attributing moral worth to a person would require that this decision or action follow from what she is, morally. The decision or action could not be produced by a random occurrence and yet count morally. It should be noted that a recurrence to reasons is not sufficient to meet this rejoinder. The decider’s character resulting from past decisions and other circumstances in the past must also be taken into consideration. We must take such a rejoinder seriously. Indeed, we already conceded that the surd in Professor Allen’s dilemma ought to be avoided. Free will does not matter in law just because people should be able to make decisions for no reason at all. Yet, we did not plead for totally arbitrary decisions. We just pointed out that reasons usually are at most sufficient conditions in a weak sense for choices. The reasons in other words only partially explain the choices. This inevitably leads to the consequence that a “residual arbitrariness” remains in all of these choices.8 The same is true for other conditions which form the character of a person: They only partially 6 7 8
Saul Smilansky, Free Will and Illusion, 2000, 66 Galen Strawson, Freedom and Belief, 1986, 58 Robert Kane, Some neglected Pathways into the Free Will Labyrinth. In The Oxford Handbook of Free Will, ed. R. Kane, 2002, 425
34
Friedrich Toepel
explain the choices which flow from the character. On the one hand we need these partially explanatory factors to avoid the surd. On the other hand we need the incompleteness to retain free will. Degrees of ultimate responsibility may vary with the degrees of “how hard it is, by reason of the pre-choice state, to make the selection”,9 for example to resist acquired bad habits or tendencies developed during an unfortunate childhood. But we could only impute to someone a behaviour like a crime if a rest of undetermined behaviour remains which makes such resistance possible at all. Otherwise, imputation of behaviour makes no sense. Now our opponents reply that by leaving our explanation incomplete we have failed to exclude arbitrariness. Only if we are sure that a certain good decision flows from the good character, we are able to say that the decider is really ultimately responsible for the decision in a meritorious way. However, proof that the decision flows from the character cannot be furnished by something short of a complete explanation. An explanation is complete when no room remains for indeterminacy. Yet, if there is no indeterminacy, we have a decision completely determined by the pre-choice state. Consequently we have only either the possibility of incomplete explanations – which is not enough to demonstrate that decisions are meritorious and that therefore free will worthy of respect does exist, or the possibility of complete explanations – which presuppose determinism and therefore exclude free will. Q.E.D.? I do not want to dismiss easily Galen Strawson’s contention insofar as he demands radical self-determination for an ascription of decisions as meritorious. Martin Luther already made a similar point when he rejected an ascription of an isolated action as meritorious. He contended that to be really good an action must flow from a good character – something which only God can give and which therefore is no reason “to boast”, nothing meritorious.10 Yet, to my mind our opponents concentrate too much on the paradigm of meritorious decisions. The argument for the impossibility of ultimate self-creation does fit neither for the imputation of crimes and other bad deeds (henceforth “negative imputation“) nor for the imputation of an action like the exercise of a right (henceforth “neutral imputation”). Nobody would say for example that a bad deed could only be imputed to a person if his character would be thoroughly bad. David may have been one of the holiest men in the Old Testament before his fall, but I presume that neither Galen Strawson nor Saul Smilansky would say that this holy character hinders to impute David the adultery and murder he committed afterwards. The incompleteness of an explanation of David’s fall also would be no obstacle to the negative imputation of his deeds. To the contrary, the more complete the explanation would be the less we could impute his crimes to him. A complete explanation would mean that David would be irresistibly driven by his desires to commit his deeds, and then a basis for a serious reproach would be missing. The same point could be made for neutral imputation. No legal system would deem the exercise of a right null and void if the decision is not in line with the 9 David Hodgson, A Plain Person’s Free Will. Journal of Consciousness Studies 12(1) (2005), 18 10 Martin Luther, Letter to Spalatin dated 11 August 1522. In Martin Luther, Saemtliche Schriften, vol. XV, ed. Walch, 1880–1910, col. 2569: Luther here uses the picture of a tree. “A bad tree cannot bring forth good fruit.”
Free Will and Infinite Regress
35
known character or past decisions of the holder of the right. Conferring a right to someone implies that the holder of the right may do what he likes within the scope of the right without having to fear that his motives will not be approved. Consequently, we may concede that no ultimate responsibility for good works can be coherently argued for, yet this does not hinder a coherent concept of free will based on incomplete explanations for negative and neutral imputation. Galen Strawson’s famous argument in this way supports in an astonishing way Luther’s old model of imputation who also thought that positive imputation would be impossible because lack of an adequate concept of free will admitting of such positive imputation but who nonetheless recognized an asymmetry which made negative imputation possible.11 VI. THE SURD
OF
REASON REVISITED
If we therefore accept negative and neutral imputation on account of incompletely explicable choices, we have to qualify our concession at the beginning of this paper that free will does not matter just because persons should be able to make decisions for nor reason at all: It is true that we grant autonomy to enable persons to follow their own understanding of the relationship between reasons and choices. But it is also true that autonomy presupposes the possibility deliberately to make a choice for no reason at all. The “residual arbitrariness”12 is implied in the meaning of free choice. Consequently, it is not quite precise that (for a libertarian) a surd of reason would not be a way out of Professor Allen’s dilemma. Only if we would have no other evidence for free will than choices for no reasons at all, we would not have anything worth to protect as an autonomous choice. However, incompleteness of explanation is acceptable as involved in all free choices. Thus we accept in discussions the possibility that reasons are chosen for no prior reasons at all, except if we in turn have a special reason why we regard the prior reasons as important. Otherwise we could never remain focused on a relevant topic of conversation but would spend time with examining the other persons’ background. If someone thinks, for example, that it is right to give first aid to the enemy on the battlefield for humanitarian reasons, we typically do not ask why he has chosen to be humane. We rather try to draw his attention to other reasons or to question whether his reasons are really as humanitarian as he thinks. Speaking in terms of the picture of the infinite regress introduced at the beginning of this paper, we can say that a surd of reason is unacceptable insofar as we really attach importance to a certain choice being autonomous. Yet, with respect to the prior choices of the reasons which were relevant to the choices at issue, we do not require that they must be made for reasons. The choices at issue are still worth to be protected as autonomous choices if they have been made for a reason and only the prior choice of this reason was made for no
11 Martin Luther, Defence against the Judgment of Eck. In Martin Luther, Saemtliche Schriften, vol. XVIII, ed. Walch, 1880–1910, col. 1404–1405 12 Robert Kane (note 8), loc. cit.
Friedrich Toepel
36
reason at all. The autonomy here is mixed with arbitrariness, but that is the hallmark of all free choices. There is still more than just arbitrariness. VII. IF
WE CAN FORESEE WHAT AN INDIVIDUAL WILL DO, DOES THAT EXCLUDE
AUTONOMY?
It might also be objected that we at least sometimes take account of other people in a strategic way. The science of economics, for example, depends on the possibility of predicting other people’s behaviour. But can an observer try to predict how other persons are going to act and at the same time think of the actions of these persons as the result of free choices without holding a contradictory view? Rationally predicting other persons’ actions seems to presuppose the persons as part of a causally predetermined world. Thinking of the persons as acting freely seems to presuppose the opposite. However, to my mind it is possible to reconcile both viewpoints by taking account of probability. Free choice is only excluded if an outcome is determined by a prior state of affairs in the sense that the prior state is a sufficient condition in the strong sense for the outcome. The picture changes as soon as probability is introduced. It may be that many features of the prior state of affairs are necessary for the outcome to happen but that jointly these features are not sufficient to make the outcome happen. In such a situation the necessary conditions raise the probability of the outcome from zero to some indefinite level. It may even be the case that some features are not necessary conditions for the outcome but that they raise the probability nonetheless. In all such constellations a free choice could just be what is needed to make an event happen. An economic man, for example, may be more disposed to sell his shares if he receives news of certain unfavourable events. In case he is not confronted with convincing counter arguments, it may be quite safe for another person to predict that the man will indeed sell his shares. But this must never mean that he is internally compelled to sell his shares; and according to our conditional analysis the gap necessary for a free choice must be reflected by a lack of sufficiency of the prior state of affairs. It is well possible that men in some situations like at the stock exchange let themselves be influenced chiefly by data calculable for others, but the chance of an irrational decision is always to be taken into consideration by leaving a probabilistic gap,13 to exclude a complete sufficient condition.14 13 It must not be considered totally irrelevant how small the degree of probability for such a decision against the reasons is. We rather have to take account of the fact that in modern times all causal laws are conceived as probabilistic to a certain degree. If the probability that someone would act differently would just leave a theoretical option for the person acting, then the picture would have completely merged with the model of the modern determinist. Indeterminism requires a practically relevant possibility that the actor chooses an alternative to his actual behaviour. 14 My result matches with Mackie’s view that the concept of cause as necessary condition would explain why it may be possible to assume causation involving sequences of psychic events without prejudging the question, Mackie (note 8), 43. If the concept of causes as necessary conditions but not necessarily sufficient conditions is endorsed (but-for test), it may be stated after the completion of the chain of events that A’s insulting B caused A to be hit by B without disputing
Free Will and Infinite Regress
37
Thus, our conclusion must be that it is not possible to talk of a physical event as the result of determination by a prior physical event and simultaneously as the result of free choice if determination by a prior physical event is understood as the prior physical event being a sufficient condition of the result. Nonetheless it is possible to conceive of an event as the result of a free choice and at the same time as being fostered by certain characteristics which raise the probability of that result. These characteristics can make it possible in some situations to predict behaviour even if the agent has freely chosen to behave in such a way.15 VIII. SUMMARY Choices are made for reasons which in turn may need justification by other reasons. In this way, an infinite regress seems to be built into the notion of free choice: If a reason determines a choice, then the choice is only free if the reason is not itself determined. If the reason is chosen for a prior reason, then the prior reason also must not be determined. And so on in an infinite regress. Such an argument would show the incoherence of free will if the relation between reasons and choices could only be reconstructed as a relation consisting of reasons and choices determined by these reasons. The reasons would have to be sufficient conditions for the choices in a strong counterfactual sense. Yet, such a view already presupposes determinism. An indeterminist does not need to follow the argument because for him reasons would be only necessary or at most sufficient conditions in a weak sense for the choices. The chain supporting the infinite regress argument breaks down if its members are merely connected by sufficient conditions in a weak sense. In our everyday life we need to look at other people as predictable beings as well as responsible persons. This becomes possible if we concede that a person may be probabilistically influenced by factors in his environment, but that such reasons are not sufficient in a strong sense to determine the outcome of the person’s decisions.
that B had acted on account of a free choice. Similarly, John Searle in his new book on Freedom and Neurobiology, Searle (note 4), 42, finds four kinds of gaps between deliberating and acting concerning each of which the antecedent situation is not sensed as setting the causally sufficient conditions for the next conscious state: the gap between the reasons for the decision and the making of the decision, between the decision and the onset of the action, and for any extended action as well between the onset of the action and its continuation to completion. 15 It may be that this possibility of simultaneous predictability and freedom of behaviour is of relevance for the question how to reconcile the characteristic of a human being as a person who legislates for himself (homo noumenon) and a being subject to the appearances (homo phainomenon). Kant thought that we have to imagine these two sides of a human being as simultaneously situated in two different worlds, Immanuel Kant, Metaphysics of Morals, Koenigsberg, 2nd ed. 1798, p.28 (Introduction, IV).
ALLAN MCCAY, UNIVERSITY OF SYDNEY FOUNDATION PROGRAM, NEW SOUTH WALES BAD LUCK AND COMPATIBILISM SMILANSKY’S CONCERNS ABOUT THE “SHALLOWNESS OF COMPATIBILISM” Some legal philosophers have seen the compatibilist view of free will as a justification for the criminal law’s practice of holding people to be morally responsible for their criminal behaviour. Thus if offenders possess compatibilist free will and wrongfully exercise it in the commission of their offence, ascriptions of moral responsibility are justified. In this paper I will discuss the work of the philosopher Saul Smilansky to support the view that any compatibilist justification of the criminal law is at best partial. The aspect of Smilansky’s work that I will focus on is that of luck, and I will place his work in the context of others who have worried about luck’s undermining effect on moral responsibility. I will conclude that whilst Smilansky’s position, if tenable, does not refute those legal philosophers who see compatibilism as a moral platform for practices of the criminal law, it does leave legal compatibilism as it relates to the criminal law as ‘shallow’.1 Before discussing the views of Smilansky, I will briefly provide an overview of some significant positions in the free will debate. It is necessary to do this in order to understand Smilansky’s position. I. FREE WILL
AND
MORAL RESPONSIBILITY
Discussions of free will have been seen to have significance in many areas of life including theology, artistic creation, meaning in life but the area that is most relevant for the law is moral responsibility. A significant proportion of the debate has centered on the question of whether free will is compatible with determinism. If determinism is true then there is one possible future. For some this means that: “Our personal lives have already been punched out on the pianola scroll. Like the intricate clockwork dolls that amused the baroque courts of Europe, the only tune we can play is the tune we will play; there is no free will.”2
For the criminal law, the concern is that we may be using notions such as moral responsibility and retribution on people who may be seen as constituent parts of a mechanistic universe rather than agents who freely will their own and (of particular relevance to the criminal law) others’ futures. If determinism was true could even the most heinous offender still be held morally responsible and perhaps subject to ret1 2
Saul Smilansky, Compatibilism: The Argument from Shallowness. Philosophical Studies 115 (2003), 257–282 Brian Silver, The Ascent of Science, 1998, 235
Allan McCay
40
ribution if his life had already been ‘punched out on a pianola scroll’? The answer to this question depends on which view of free will is being taken. II. HARD DETERMINISM A hard determinist would say no. Hard determinists argue that free will is not compatible with determinism and that determinism is true. As determinism is true, nobody has free will and nobody is responsible for anything. The eighteenth century French philosopher Holbach expresses the incompatibility of free will and responsibility with determinism in the following: “Man’s life is a line that nature commands him to describe on the surface of the earth without his ever being able to swerve from it, even for an instant…. Nevertheless, in spite of the shackles by which he is bound, it is pretended that he is a free agent, or that independent of the causes by which he is moved, he determines his own will and regulates his own condition…”.3
Under this view free will is completely non-existent. Along with the denial of free will, hard determinists also deny moral responsibility It is easy to see why the criminal law does not appear to have been described as a hard determinist system. Notions such as retribution and so much else in the criminal law are in opposition to hard determinism that such a view would be hard to maintain. Hard determinism provides no justification for legal or any other notion of free will or moral responsibility. III. HARD INCOMPATIBILISM Holbach was writing at a time when scientists were more confident in the truth of determinism than they are now. Developments in quantum theory have caused some physicists to believe that determinism may not be true.4 Some writers have sought to use the possibility of quantum level indeterminism in the brain as perhaps allowing for choices, which are genuinely undetermined and therefore free. Derk Pereboom has advanced a view that whether or not determinism is true, it seems unlikely that free will exists.5 He argues that such randomness could not allow agents to be the ultimate source of their behaviour in order to for them be truly free. Pereboom argues that agent causal free will, a variety of libertarian free will, could conceivably exist but that there is not much evidence for it. It is unlikely that the law could be conceived of as a hard incompatibilist system, as like hard determinism, hard incompatibilism seems to rule out concepts that the law regards as important. According to the hard incompatibilist thesis, no one freely 3 4
5
P. Holbach, The Illusion of Free Will in Reason and Responsibility; Readings in Some Basic Problems of Philosophy, ed. J. Feinberg & R. Shafer-Landau, 2005, 392 Whereas Newtonian physics has often been said to be deterministic, some interpretations of quantum mechanics are indeterministic. Thus some physicists believe that there may be effects that are not determined by causal antecedents but are in fact random. If this were true it would seem that there may be many possible futures and not one. Einstein famously opposed this view when he asserted that: God does not play dice. Derk Pereboom, Living Without Free Will, 2001
Bad Luck and Compatibilism
41
wills anything nor is anyone ever morally responsible for any action. No justification for legal free will, notions of moral responsibility or retribution can be found in hard incompatibilism. IV. LIBERTARIAN FREE WILL Hard determinism and hard incompatibilism both see free will as incompatible with the truth of determinism. However there is another incompatibilist position in the free will debate that provides a prima facie justification for legal free will. That is the libertarian position. Libertarians deny the truth of determinism and assert that free will does exist. The exercise of free will therefore actually alters the course of the future in a way that was not determined. It is not merely part of a deterministic causal chain, which extends back at least to the big bang and inexorably forward, but by its exercise, a future comes into existence that was not inevitable beforehand. Thus libertarians would deny that: ‘our personal lives have already been punched out on the pianola’.6 They would deny that there is ‘only one possible future’.7 In short they would reject the truth of determinism. If the libertarian position is true perhaps notions of free will, moral responsibility and retribution in the criminal law are justified. However, Smilansky is amongst those philosophers who argue that libertarian free will is incoherent and cannot therefore provide any justification for notions of moral responsibility. For Smilansky, quantum level randomness in the brain cannot provide a basis for regarding a choice as free nor can agents create themselves from nothing in such a way as to be truly responsible for their choices. Libertarian free will cannot avoid falling ‘into being merely the unfolding of an arbitrarily given, whether determined or random’.8 If Smilansky is right about libertarian free will then the criminal law must turn to compatibilist free will as a platform for its notions of moral responsibility and retribution. V. COMPATIBILIST FREE WILL Even if it turned out that determinism was true of all human actions, a compatibilist would argue that determinism, of itself, was not enough to negate free will. The argument is that determinism and free will are compatible. Indeed some have argued that free will would be impossible without determinism. This view has achieved some popularity in legal philosophy.
6 7 8
See Silver (note 2), 235. Peter Van Inwagen, An Essay on Free Will, 2002, 3 Saul Smilansky Free Will, Fundamental Dualism and the Centrality of Illusion. In The Oxford Handbook of Free Will ed. R Kane, 2002, 491
Allan McCay
42 1. Compatibilism in Legal Philosophy
Kaye has asserted that ‘[a]ccording to the dominant view in criminal theory, we have a compatibilist criminal law’.9 This view is echoed by Norrie who states that ‘the law adopts in practice the compromise between free will and determinism, known as compatibilism’10 Whilst neither Kaye nor Norrie are compatibilists, many in legal theory are. For example Norrie points out that Hart is a compatibilist when he states that ‘the soft determinist position is compatible with the free will required to found a satisfactory concept of criminal responsibility’.11 That this free will is a prerequisite to moral responsibility is clear when Hart states that a ‘moral licence is required in the form of proof that the person punished broke the law by an action which was the outcome of his free choice’.12 Another prominent compatibilist Morse contends that ‘only compatibilism can explain and justify our legal practices”. 13 Thus for some legal philosophers it would seem that compatibilism can provide a justification for notions of free will and moral responsibility in the criminal law. 2. Some Problems for Compatibilist Justification of the Criminal Law One problem for any compatibilist justification of the criminal law is the possibility that free will may not be compatible with determinism. Peter van Inwagen has presented a more formal argument against compatibilism. He presents a short form of the argument thus: If determinism is true, then our acts are the consequences of the laws of nature and events in the remote past. But it is not up to us what went on before we were born, and neither is it up to us what the laws of nature are. Therefore, the consequences of these things (including our present acts) are not up to us.14 This argument has generated much discussion and remains a significant source of dispute amongst compatibilists and incompatibilists.15 3. Is the Compatibilist Justification Shallow? Saul Smilansky is another philosopher who has concerns about compatibilism. Unlike other incompatibilists, he does not see either free will, moral responsibility or retribution to be entirely incompatible with determinism. For Smilansky there is a sense in which compatibilism is true and another sense in which it is false. This posi9 10 11 12 13 14 15
Anders Kaye, Resurrecting the Causal Theory of the Excuses, Nebraska Law Review 83 (2005), 1158 Alan Norrie, Crime, Reason and History A Critical Introduction to Criminal Law, 2001, 51 See H.L.A. Hart in Norrie (note 10),143 See H.L.A. Hart in Norrie (note 10), 143 Stephen Morse, Reason, Results and Criminal Responsibility, University of Illinois Law Review, 2004, 431 Peter Van Inwagen, An Argument for Incompatibilism in: Free Will, 2004, 39 For an overview of the argument see Kane (2005).
43
Bad Luck and Compatibilism
tion he calls the ‘fundamental dualism’.16 This will be discussed below but for the moment I will focus on the issue that causes the concern that compatibilist moral responsibility is ‘shallow’.17 This is the problem of luck. Before examining Smilansky’s views on luck, I will briefly discuss some earlier work on the topic. This is order to put Smilansky’s work in the context of others who are troubled by the impact of luck on moral responsibility. VI. MORALITY
AND THE
POISONING EFFECT
OF
LUCK
If determinism is true, then human agency forms part of a causal chain, which extends back to times which predate the birth of the agent. Smilansky has pointed out that on this view, an unpalatable degree of luck pervades all human actions.18 It seems that when we engage in moral assessment of human actions, there is a disturbing element of luck, which enters into the assessment. Luck seems to be a poison that weakens moral responsibility and our confidence in our moral judgments. The problem of “moral luck” has its own place in the philosophical literature and one of the seminal articles on the topic is Thomas Nagel’s article “Moral Luck” in which he outlines the problem. Nagel first of all notes the Kantian view that luck should not impinge upon our moral judgments; one’s moral worth should not be affected by matters of luck. He points out that ‘prior to reflection it is intuitively plausible that people cannot be morally assessed for what is not their fault, or for what is due to factors beyond their control’.19 However there is a conflict here, as we do seem to hold people responsible for events, which are outside their control. Thus for some, a drunk driver who harms no one is considered to be morally different from an equally drunk driver who kills someone. This is the case even though the fact that one driver encountered an empty road on their return from the pub and the other was unlucky enough to encounter another car, could not be controlled. Much of the literature on moral luck is an attempt to reconcile this conflict. For Nagel, moral luck occurs ‘[w]here a significant aspect of what someone does depends on factors beyond his control, yet we continue to treat him in that respect as an object of moral judgment’.20 It is indeed a problem to explain why luck should not undermine moral responsibility.
16 17 18 19 20
Saul Smilansky Free Will and Illusion, 2000, 94. See Smilansky (note 1) above. See Smilansky (note 16), 45. T Nagel, Mortal Questions, 1979, 25 See Nagel (note 19), 26
Allan McCay
44 Nagel identifies four kinds of luck; 1
‘Constitutive Luck’21
He explains that ‘constitutive luck – the kind of person you are, where this is not just a question of what you deliberately do, but of your inclinations, capacities, and temperament’ is ‘disturbingly subject to luck’.22 2
‘Luck in one’s circumstances’23
Aside from luck in who you are there is luck in the circumstances that you find yourself. Thus a person is morally lucky if they never have the opportunity of genocide. 3
‘Luck in the way one’s actions and projects turn out’24
An example of someone who has this kind of luck could be a criminal who has attempted a crime and is morally lucky through not succeeding. The fact that the attempt failed may be due to factors outside the control of the criminal and this is morally lucky. 4
‘Luck in how one is determined by antecedent circumstances’25
One may be lucky or unlucky in ones antecedents as these factors are outsides one’s control. Some have antecedents that lead to moral behaviour and others have antecedents that lead to immoral behaviour and which kind of antecedent one has seems to be a matter of luck as it is not in one’s control. It is in the discussion of luck in one’s antecedents that Nagel notes the connection to the free will problem. If the will itself is the product of antecedent circumstances then acts which result from the will would seem to be disconcertingly subject to luck.26 Whilst Nagel does not talk about compatibilism, it is here that luck’s poisoning effect on compatibilist moral responsibility would seem to be most clearly manifested. For compatibilists one can be morally responsible even if determinism is true. But if determinism is true then all causal chains leading to an act originate prior to a person’s birth. No one has any control over things which happened prior to their birth and it seems be a matter of luck as to whether one has antecedents which lead to an immoral and criminal life or antecedents which lead to a praiseworthy and law abiding life.
21 22 23 24 25 26
See Nagel (note 19), 28 See Nagel (note 19), 28 See Nagel (note 19), 28 See Nagel (note 19), 28 See Nagel (note 19), 28 See Nagel (note 19), 35
45
Bad Luck and Compatibilism
For those who uphold notions of moral responsibility, moral luck is a problem. It seems to impinge on or even entirely account for all actions which have moral significance. VII. BAD LUCK
AND
COMPATIBILISM
IN
SMILANSKY’S WORK
Smilansky’s discussion of the compatibilist’s problems with luck appear in his book Free Will and Illusion (2000) and a number of articles. I will discuss the role of luck in the context of an overview of his unusual position in the free will debate. 1. Smilansky on Free Will: An overview Saul Smilansky holds an unorthodox position in the free will debate. His most substantial exposition of his view can be found in his book Free Will and Illusion (2000) in which he holds there to be truth in both the compatibilist and the hard determinist positions.27 He rejects the libertarian position on the grounds of incoherence. He tries to establish grounds for moral responsibility that do not rely solely on consequentialism. Smilansky challenges the ‘assumption of monism’; the view there is one conception of moral responsibility and that one can choose between compatibilism and hard determinism but cannot support both views. Smilansky argues for a ‘fundamental dualism’ in which some aspects of moral responsibility require libertarian free will (which does not exist) but that for other aspects, compatibilism suffices. This means that when one considers an action one must be at once both a compatibilist and a hard determinist. At one stage Smilansky uses the example of a fraud.28 Regarding the fraud, there are two perspectives which are both true. If determinism is true then viewed from the hard determinist level the offender is a victim if he suffers as a result of a criminal sanction as ultimately his or her causal antecedents are a matter of luck. However if the offender was not coerced, or subject to a psychological compulsion, then he is morally responsible in a compatibilist sense. The important point to note is both of the perspectives are valid and do not contradict each other because moral responsibility is not a monistic concept, it is split and so one can be both morally responsible in a sense and yet not morally responsible in another sense. This is ‘fundamental dualism’. To be ultimately responsible for an action we would need libertarian free will but for Smilansky this is impossible as the libertarian view is incoherent. Therefore some aspects of moral responsibility are impossible and so hard determinism is partially true. 27
Smilansky uses the term hard determinism in a non standard way. Hard determinism for Smilansky refers to the incompatibility of free will with either a deterministic world or a world that includes some non-libertarian indeterminism. Smilansky’s use of the term would therefore embrace Holbach’s position and Pereboom’s hard incompatibilist position on free will and moral responsibility. 28 Smilansky (note 16), 51
Allan McCay
46
However he also sees compatibilism to be partially true and thus to some extent he is a compatibilist. He uses the example of a lazy waiter to argue for compatibilist desert.29 It is Smilansky’s contention that there is no significant moral concern if the waiter’s tips relate to his effort even if determinism is true. Thus compatibilist voluntariness is morally relevant and there is a partial truth in compatibilism. Smilansky argues there is a morally significant difference between reward based on some kind of voluntariness (like in the lazy waiter example) and reward based on race. Why is he not a wholehearted compatibilist? Smilansky argues that compatibilist moral responsibility is ‘shallow’ (2003). Smilansky elaborates on the disturbing nature of luck to notions of moral responsibility in the following: “A person can change her character, but not the original ‘she’ that can choose whether or not to make an effort to change. In the end, even if we ‘freely’ do what we want in compatibilist terms, what we want, our desires and beliefs, is not ultimately something we choose: in a deterministic picture there was no real opportunity for us to be people who do otherwise. If in the end it is only our bad luck, then it is not morally our fault-anyone in ‘our’ place would (tautologically) have done the same, and so everyone’s not doing this, and the fact of our being such people as do it, is ultimately just a matter of luck. Matters of luck, by their very character, are the opposite of the moral – how can we ultimately hold someone accountable for what is, after all, a matter of luck? How can it be fair that she ‘pay’ for this?”30
It seems unfair that some are biologically and environmentally predisposed to criminal behaviour as a result of events in their childhood and the biology of their parents. The element of luck seems to undermine our moral certainty in holding them responsible and responding retributively towards them. Smilansky argues that if a person does suffer retributive punishment, then there is a sense in which they are victimized through being made to suffer for something that they were ultimately not responsible for. Compatibilism is partly true and, compatibilist distinctions are needed to maintain a ‘community of responsibility’31 but compatibilism lacks moral depth. As a result of partial validity of both the hard determinist and compatibilist claims and possibility of a dualism with regard to moral responsibility a ‘joint perspective’32 is in order. However Smilansky is concerned about the consequences should this dualistic view of moral responsibility become prevalent. He identifies a number of issues including concerns about whether people would continue to take responsibility for their actions or feel remorse for their wrongdoing. He also notes that in reality people have illusory beliefs about free will and that these illusory beliefs are for the most part positive: ‘[h]umanity is fortunately deceived on the free will issue, and this seems to be a condition of civilized morality and personal value’.33 There is an interesting connection here to the paper of Ronald J Allen in the present working group. Professor Allen asks a general question pertaining to the legal system as a whole: ‘[t]o what extent does one work with rather than explicitly reject operating assump-
29 30 31 32 33
Smilansky (note 16), 80 Smilansky (note 16), 45 Smilansky (note 16), 83 Smilansky (note 8), 495 Smilansky (note 8), 500
Bad Luck and Compatibilism
47
tions widely held but wrong’? Smilansky would say sometimes we should work with them. Smilansky considers how this illusionism can work and identifies two possibilities. One is the ‘elitist solution’ where the realisation that there is no libertarian free will and thus injustice is visited on criminals through retribution, is confined to an elite. Those outside the elite keep the unqualified belief that criminals entirely deserve their punishment. He sees this to be at best a partial solution and fraught with moral and practical problems. The more significant solution is found in the fact that people generally do have illusory beliefs about free will and that these illusory beliefs are beneficial in maintaining a ‘community of responsibility’. 2. Bad Luck Compatibilism and the Criminal Law If we think of Smilansky’s work in the context of the criminal law, what is to be done about holding responsible those who have no libertarian free will and have had bad luck in their antecedents, which have lead to the commission of a crime? Of course some of the criminal law’s practices relate to consequentialism and thus have no need for an agent with moral responsibility. However other practices relate to the moral responsibility of the offender; judges at sentencing try to craft a sentence which takes into account the moral culpability of the offender and many jurisdictions have retributivism as a sentencing aim. If Smilansky is right about the absence of libertarian free will then any moral responsibility must be compatibilist moral responsibility but this is ‘shallow’ due to the effect of luck in one’s antecedents. Judges must be compatibilists but they must, at the same time, look to hard determinism, to the ultimate perspective where seen in the context of a world of universal causation, criminal conduct is a matter of luck. But this raises the question of how can one do this? Smilansky states that ‘there are no general methods for discerning which element ought to prevail, except in the more extreme circumstance’.34 He does however say that ‘very severe punishment, even for a deed compatibilistically up to the agent, would create acute Ultimate Injustice’. 35 Since some criminal punishments are sometimes ‘very severe’, if Smilansky is right then there is significant degree of victimization of the offender. It is perhaps a criticism of the work that he does not make it entirely clear where ultimate injustice becomes too extreme but perhaps some inference can be made from his use of examples. It has already been said that when arguing for the shallowness of compatibilism and the extent of victimization of those who are punished he uses the example of a criminal case, a person who has perpetrated a fraud 36 but when arguing to partially restore compatibilism, he uses an example from distributive justice that raise no issues of criminal law; that of the lazy waiter.37 34 35 36 37
Smilansky (note 16), 134 Smilansky (note 16), 134 Smilansky (note 16), 51 Smilansky (note 16), 80
48
Allan McCay
It seems that Smilansky’s view would suggest that in cases of extreme punishment of those who are guilty in a compatibilist sense too much ultimate injustice is done. The compatibilist justification may be too shallow in the case of the extremes of the criminal law. It seems that the criminal law’s response to many serious crimes may constitute extreme circumstances in which the severity of the law’s response may result in significant ultimate injustice in the face of luck in the offender’s antecedents. VIII. CONCLUSION The idea of being morally responsible for things that are a matter of luck is a disconcerting idea. Since for Smilansky there can be no libertarian free will, some form of compatibilism is needed to maintain notions of moral responsibility. However, this kind of compatibilism is made ‘shallow’ through the impact of luck in one’s antecedents. Whilst this does not refute legal compatibilism, it leaves it weakened. The criminal law may thus do wrong if it relies on legal compatibilism and thereby converts bad moral luck in one’s antecedents into bad retributive luck through excessive suffering.
JUAN PABLO MAÑALICH, SANTIAGO (CHILE) DETERMINISM, FREE WILL
AND
CRIMINAL RESPONSIBILITY
The article presents an argument in favor of a compatibilistic account of free will as criterion for the attribution of personal responsibility, stressing that freedom of the will must be understood not in terms of indetermination but rather in terms of selfdetermination. In applying the argument to the specific structures of criminal responsibility, the article discusses how law-abidance must be seen as personal preference of an agent who is member of a democratic political community. 1. DETERMINISM,
INTENTIONALITY AND AGENCY
a) Determinism and avoidance Determinism is the thesis that “there is at any instant exactly one physical possible future”.1 Yet determinism does not imply fatalism, that is, the inevitability of events. In a fully determined world, there are things some agents can in fact avoid, inasmuch their evolutionary design allows them to foresee the occurrence of a certain event and to decide to prevent it. Neither does determinism imply the predictability of events.2 Determination is an ontological property of events, whereas the predictability of an event concerns an observer’s epistemic capacity. For the prediction of an event it is essential that generalized correlations between certain types of events and other types of events be available, which can be more or less strict. The thesis that the identity of mental and physical events entails the determination of the former stems from the so-called principle of the causal closure of the physical world. Against this claim it is often argued that quantum physics would refute it, since the behavior of electrons at subatomic level would appear to be undetermined.3 There seem to be good reasons, however, to restrain this interpretation of quantum physics to a microphysical level, so that no direct argument at the level of the philosophy of biology or of philosophy of mind be made out of it.4 But even if one accepted the transference of microphysical indetermination to a macrophysical level, no foundation of an indeterministic conception of free will would be gained, since undetermined events do not necessarily represent freely willed actions.5 The evolutionary design of agents that are capable of intentional avoidance concerns some properties that are described, for the most, by reference to biological 1 2 3 4 5
Peter Van Inwagen, An Essay on Free Will, 1983, 3; Daniel Dennett, Freedom Evolves, 2003, 25 Gerhard Roth, Fühlen, Denken, Handeln, 2003, 504–11; Michael Pauen, Illusion Freiheit?, 2004, 32–33 John Searle, Mind. A Brief Introduction, 2004, 16–17 Michael Esfeld, Holismus in der Philosophie des Geistes und in der Philosophie der Physik, 2002, 351– 60 Jürgen Schröder, Einführung in die Philosophie des Geistes, 2004, 310–11
50
Juan Pablo Mañalich
and mental features. The fact that these properties entirely supervene on properties that can be described in terms of purely (macro-)physical features does not entail the reducibility of the sort of explanation achieved through reference to biological and mental properties to a sort of explanation that exclusively refers to properties described in a physicalist vocabulary. b) Intentionality and agency Intentionality stands for a distinctive property of certain mental states, which rests on the fact that these states refer to a given object the existence or inexistence of which is independent of the subject of the intentional state.6 In this technical sense, intentionality means “aboutness”. When the given intentional object has the logical form of a proposition, the correspondent intentional state is defined as a propositional attitude. “Believe”, “think”, “wish”, “expect”, among others, are verbs which express propositional attitudes, in the sense that in ordinary speech we typically use these verbs by uttering sentences which are normally followed by a subordinate sentence which begins with the conjunction “that” and expresses a certain proposition.7 The ascription of propositional attitudes is the definitive moment of the explanation of actions.8 An event is interpreted as action when there is at least one true description under which it appears as intentional behavior of someone.9 This does not mean that all action must be intentional action. But it does mean that all action must be intentional behavior. Oedipus’s killing of Laius is an action, because there is a true description of Oedipus’s behavior under which it was intentional (which means: the object of a propositional attitude of Oedipus), for instance the killing of a traveler, which does not mean that the action of killing Laius was intentional, since Oedipus did not know he was killing Laius. What comes to attention here is the opacity or “intensionality” of the ascription of propositional attitudes. Intentional explanation may thus be defined as explanation by means of new description.10
6
Franz Brentano, Psychologie vom empirischen Standpunkt I, 1924, 124–28; Schröder (note 5), 133– 43; Ernst Tugendhat, Selbstbewusstsein und Selbstbestimmung, 1979, 18–22 7 See Tugendhat (note 6), 20–21. The description of the expression “that …” as a subordinate sentence concerns its grammar, not its semantics, since from this point of view the clause works as a singular term. Hereto Donald Davidson Inquiries into Truth and Interpretation, 1984, 93–108. 8 Donald Davidson, Essays on Actions and Events, 2001, 3–19, 43–61; Donald Davidson, Problems of Rationality, 2004, 101–16, 151–66 9 See Davidson (note 8), Essays 43–47. Simester argues, on the contrary, that intentionality cannot make a general feature of action, since that would not give account of some forms of negligent actions. See A.P. Simester, Agency, Law and Philosophy 15 (1996), 172–75. Simester does not recognize, however, that negligence may indeed be understood as a criterion of responsibility that enables imputation although the agent was not capable, at the decisive moment, to intentionally avoid wrongful conduct, whereas she could have intentionally avoided to put herself in that situation. See H.L.A. Hart, Punishment and Responsibility, 1968, 136–57; Joachim Hruschka, Imputation, Brigham Young University Law Review 1986, 686–90, 696–98. 10 Davidson (note 8), Essays, 109–10; Davidson (note 8), Problems, 105
Determinism, Free Will and Criminal Responsibility
51
Intentionality designates therefore a necessary condition of the counterfactual assumption that underlies all ascription of actions.11 “Action”, used as ascription term, expresses a relation between an actual event and a counterfactual event, namely the event that would have taken place, if the agent had omitted the action that he actually performed. This counterfactual assumption that underlies the ascription of actions entails an implicit reference to the possibility of avoidance and thus to the controllability of an event’s occurrence.12 Hence, intentionality represents a central criterion of agency. The ascription of propositional attitudes, that is, of beliefs, desires, expectations and so on, is essentially holistic, for it is not possible to have one belief without having, at the same time, a pretty large number of further beliefs.13 The ascription of propositional attitudes, just like the ascription of meaning to the utterances of a speaker, occurs under a condition of potential indeterminacy.14 Therefore, the generalizations that are relevant for the explanation of events described through reference to intentional properties always include ceteris-paribus clauses.15 The decisive feature of these generalizations is that they cannot be homonomic, as the strict laws of physics are, but only heteronomic, which means that they consist in generalizations that correlate types of events possibly described in different vocabularies.16 Some of these generalizations correspond in fact to the standards we apply when we explain action in terms of folk psychology. Explanations of this sort can be understood as rationalizations of actions, since they are based on the identification of propositional attitudes, or more precisely, of certain “pro-attitudes” combined with the proper set of belief states, which constitute the reasons that explain their performance.17 This kind of rationalization of action is made possible by the use of standards that are partially normative, in the sense that they require that the interpreter ascribe a consistent set of correct beliefs to the agent, a supposition that can be articulated in terms of the so-called principle of charity, which represents a “policy of rational accommodation”.18 c) Physicalism, multiple realization and non-reductionism The recourse to intentionality as essential criterion of agency does not presuppose a dualistic ontology, even if one rejects the reducibility of intentional to purely physical explanations. The critical consideration hereto concerns the so-called multiple realization problem. Multiple realization of mental properties means that the very 11 See Urs Kindhäuser, Intentionale Handlung, 1980, 111–12. This does not mean, however, that actions should not be seen as a special class of events. One can rather assert that actions are events, the description of which includes the mention of properties which imply such counterfactual moment. Hereto Davidson (note 8), Essays 43–61, 105–22, 293–304. 12 Kindhäuser (note 11), 172–75, 202–15 13 Donald Davidson, Subjective, Intersubjective, Objective, 2001, 123–34; Davidson (note 8), Problems, 10–12; Esfeld (note 4), 48–68 14 Davidson (note 7), 125–70. See further W.V.O. Quine, Word and Object, 1960, 26–30, 68–72. 15 Davidson (note 8), Problems, 111–12; Dennett (note 1), 40 16 Davidson (note 8), Essays, 219–23 17 Davidson (note 8), Essays, 3–8 18 Davidson (note 8), Problems, 35–36
Juan Pablo Mañalich
52
same psychological state can have more than one physical realization, such as when two different persons share the same belief.19 The important argument regarding multiple realization claims that the supervenience of the mental to the physical, which allows the assertion of identity between the realized and the realizing states without having to imply any epiphenomenalism of the mental,20 cannot be conceived as a relation of local supervenience. The supervenience of mental states on physical states means that although there can be no qualitative difference between belief-states (or other propositional attitudes) if there is no qualitative difference between the corresponding physical states on which the former are realized, the contrary is not the case, so that two sets of belief-states can indeed be identical even if the corresponding physical states are qualitatively different. Yet: […] subjective states are not supervenient on the state of the brain or nervous system: two people may be in similar physical states and yet be in dissimilar psychological states. This does not mean, of course, that mental states are not supervenient on physical states, for there must be a physical difference somewhere if psychological states are different. But the interesting physical difference may not be in the person […].21
The important point here is that multiple realization does not exclude the possibility of identity between mental and physical states (that is, between events described in mentalistic vocabulary and events described in physicalist vocabulary), but only the possibility that such identity relation be a 1:1 relation.22 The supervenience of mental states on physical states must thus be seen as a relationship of global supervenience, in the sense that the total amount of a person’s mental states supervene on physical states of her brain and the environment, without to imply that each given mental state could be identified with one corresponding physical state.23 2. Compatibilism and autonomy a) Intentionality, free will and compatibilism The ascription of propositional attitudes is the decisive moment of the explanation of actions.24 This is why the imputation rules of the criminal law refer to intentional capacities. The complex capacity to follow a legal conduct rule, which establishes 19 Esfeld (note 4), 196–201; Jaegwon Kim, Mind in a Physical World, 1998, 2–8 20 The thesis of the epiphenomenalism of mental properties, which is founded on the supervenience argument, asserts that, although mental events are causally effective, this causal effectiveness exclusively depends on their physical properties, which would entail that mental properties are purely epiphenomenal. See Kim (note 19), 38–47. This argument does not refute Davidson’s Anomalous Monism, since this position holds that there is no ontological difference between physical and mental properties, for mental properties are just physical properties under another description. Hereto Donald Davidson, Truth, Language and History, 2005, 186–200. 21 Davidson (note 13), Subjective, 61–62 22 Esfeld (note 4), 198–99 23 Davidson (note 13), Subjective, 43–45; Davidson (note 8), Problems, 130–31; Esfeld (note 4), 149–60 24 Davidson (note 8), Essays, 3–8, 72–74; Davidson (note 8), Problems, 105–09
Determinism, Free Will and Criminal Responsibility
53
the object of possible criminal responsibility, is no other than the capacity to build and fulfill intentions of first and second order, that is, the (first-order) intention to do or not do something, and the (second-order) intention to build such an intention in compliance with the norm that commands or forbids the deed in question. One can thus designate the first capacity as action-capacity and the second as motivationcapacity, so that the ascription of responsibility can then be reconstructed as a twolevel judgment.25 Michael Moore’s distinction between prima facie and definitive culpability can be interpreted exactly in this sense.26 To be an agent who is capable of being held responsible for wrongful (but also rightful) conduct is therefore to be an agent towards whom it is not only possible to adopt an intentional stance, but also a personal stance: persons are, at least, second-order intentional systems.27 Such an account of the attribution of criminal responsibility can be seen as an application of Harry Frankfurt’s well-known conception of free will,28 which has received a good amount of attention in the discussion on freedom and determinism. Frankfurt defines freedom of the will in terms of a person’s capacity of willing to have the will she wants to have.29 This capacity must be regarded, according to Frankfurt, as the decisive criterion for the application of the concept of a person. For only a person can have the kind of mental state, which can be seen as a kind of propositional attitude,30 that Frankfurt calls a second-order volition. Thus, one can differentiate the kind of entity that we acknowledge as a person, on the one hand, from the kind of entity that Frankfurt calls a “wanton”, on the other hand, who is someone that, although capable of having second-order desires, does not appear to have second-order volitions. This points to the fact that volitions are desires which show effectiveness, which means that a first-order volition is a desire “that moves a person all the way to action”,31 so that a second-order volition would be a desire that moves a person “all the way” to having a certain volition. This capacity of caring about one’s own will, which according to Frankfurt is the distinctive feature of personhood, constitutes the core of freedom of the will. Frankfurt claims that his analysis of free will is fully compatible with the truth of determinism, since the fact that a person’s second-order volitions could be causally determined places no obstacle to the assertion that she may indeed be capable of having the will she wants to have.32 This is why Frankfurt’s view on free will makes a common place in the discussion about the plausibility of compatibilism.33
25 26 27 28 29 30 31 32 33
Hruschka (note 9), 672, 682–84 Michael S. Moore, “Prima facie moral culpability”, Boston University Law Rev. 76 (1996), 319–20 Daniel Dennett, Brainstorms, 1981, 236–39, 273–85 Harry Frankfurt The Importance of What We Care About, 1988, 11–25; Harry Frankfurt, Taking Ourselves Serious & Getting It Right, 2006, 14–16 Frankfurt (note 28), Importance, 20; Dennett (note 27), 283–85 Ernst Tugendhat Einführung in die sprachanalytische Philosophie, 1976, 97–102; Tugendhat (note 6), 18–22, 200–1 Frankfurt (note 28), Importance, 14 Frankfurt (note 28), Importance, 20 See the contributions in David Widerker and Michael Mackenna (ed.), Moral Responsibility and Alternative Possibilities, 2003.
54
Juan Pablo Mañalich
b) Freedom without alternate possibilities? From the beginning one should notice that the plea for compatibilism entails, at least for the sake of argument, the assumption of the possible truth of determinism, whereas incompatibilism can be committed either to the truth of determinism or to the truth of indeterminism. It is often presupposed, however, that only determinism would face difficulties in arguing in favor of free will, the most important of which refers to the so-called principle of alternate possibilities (PAP). But this is far from meaning that no difficulty arises for indeterminism, since this position must indeed face the problem that without determination of a deed it seems hard to justify its attribution as action to someone. The difficulty comes, therefore, with the requirements of agency. For it is essential to agency that the agent can be seen as effectively conditioning the event that constitutes the object of ascription, since otherwise its occurrence would be purely random.34 The question that now arises points to the way in which the deterministic character of willed actions could be compatible with the principle of alternate possibilities (PAP). An easy way-out would consist in renouncing to PAP, in order to defend a conception of free will for which the existence of alternate possibilities were irrelevant. This is exactly the view Frankfurt maintains in his very well-known and much discussed article “Alternate possibilities and moral responsibility”,35 whereas his argument concerns cases of possible over-determination of a decision to perform a given action. This would be for instance the case of a person that has already decided to do something, but who afterwards is coerced by another person to do exactly the same thing. If one assumes that the first person has been impressed by the threat in the same way as any normal person, one could then assume that she would still have had the same will to perform the given action, had she not made the preceding decision. Yet according to Frankfurt, this would not exclude that, even when acting motivated by the threat, the coerced person has still acted in correspondence with a second-order volition of her own, since she had already decided to perform the given action. Despite its recognition, Frankfurt’s argument does not hold.36 For his argument violates a decisive condition of the application of PAP. Regarding the case just presented, Frankfurt assumes that the threatened person has acted in accordance with a decision made freely by herself, although she had no alternative to having the will that she had, since, because of the threat, she would have decided to perform the action anyway. This presupposes, however, that the coercion exercised by the second person was not relevant for the first person’s motivation to act. But by adding that still without having made such prior decision, the first person, due to the threat, would have decided to perform the same action anyway, Frankfurt introduces a modification of the circumstances in the context of which the presence of the alternative to the actual course of events is to be analyzed. This move infringes a crucial
34 Peter Bieri, Das Handwerk der Freiheit, 2003, 21–24, 81–83, 230–49; Pauen (note 2), 15–16, 48–49 35 Frankfurt (note 28), Importance, 1–10 36 Davidson (note 8), Essays, 74–75; Pauen (note 2), 115–18
Determinism, Free Will and Criminal Responsibility
55
requirement of the application of PAP, namely that the background conditions of the decision to act must remain unaltered.37 c) The conditional interpretation of PAP The argument for compatibilism cannot simply avoid the test of PAP. It is important here to notice that PAP is usually formulated in terms of the question of whether a person could have acted differently as she indeed acted. In this way, however, one does not reach yet the problem of freedom of the will, but only the question of freedom of action. In Frankfurt’s own words, the problem of free will does not concern the issue of whether someone is free to do what he wants, but rather the issue of whether he is free to have the will he wants to have.38 To ask about alternatives at this level is to ask whether the person was capable of willing to have a different will from the one she actually had. The argument is usually traced back to G.E. Moore’s analysis of the term “could” as used in the formulation of PAP.39 Moore differentiated two interpretations of “could”, whereas the traditional assumption that determinism would be incompatible with PAP presupposes an unconditional interpretation: an action is free only if the agent could have done otherwise (“no matter what”). Under this unconditional interpretation, the truth of determinism precludes freedom in terms of PAP, since there can be no alternative course of events in a world which is fully determined. Moore’s proposal, on the contrary, consisted in a conditional interpretation of “could” as used in the formulation of PAP: an action is free only if the agent could have done otherwise, in case he had chosen to do so.40 Such a conditional formulation of PAP does not yet touch, however, the problem of free will. Freedom of the will requires that the person could have willed to have a different will from the one with which she actually acted. Hence, one has to formulate the principle at the level of second order volitions. Under the proper restatement, what Moore’s conditional interpretation suggests is the possibility of regarding the agent’s own second-order volitions as part of the total causal history of the action, which then could be seen as self-determined by the agent. Under this formulation, PAP does not require that any alternative to the decision was in fact possible, but only the alternative that would have corresponded to the person’s decision if she had wanted to decide otherwise. Thus, it becomes possible to restate the application condition of PAP that was violated by Frankfurt’s intent to define free will without alternate possibilities. That condition must be conceived in the sense that the non-variation of the circumstances that build the situation of the actually performed action must cover all the conditions that are relevant for the given decision, except for the agent’s own volitional dispositions that were relevant for it. That the talk on decisions is justified in
37 38 39 40
Pauen (note 2), 108–9, 188 Frankfurt (note 28), Importance, 19–20 G.E. Moore Ethics, 1912, 84–95 Michael S. Moore, “Causation and the Excuses”, California Law Rev. 73 (1985), 1142–44
Juan Pablo Mañalich
56
this context results from the fact that decision-making can indeed be seen as a process which concludes with the production of reasons for action.41 3. SELF-DETERMINATION
AS SELF-CONSTITUTION
a) Infinite regress? Against this solution it could be argued that it is indeed no solution at all, since the reference to higher-order volitions would inevitably lead to an infinite regress: in order to establish self-determination at the level of second-order volitions, one would have to introduce a reference to third-order volitions, and so forth. Yet the answer must be construed not as a solution to the problem of infinite regress, but rather as its dissolution. What one requires is a view that enables the identification of properties of decisions that are internal to the agent’s identity in the sense that a decision which exhibits those properties can be interpreted as self-determined.42 To ask for alternate possibilities at that stage would be to ask for an alternative to the person herself, which is pointless. For the important distinction here is not between determination and indetermination, but between autonomy – that is, self-determination– and heteronomy.43 The crucial remark concerns the fact that it is not the case that behind every cause there must be a causer.44 The very idea of “ultimate agency” or “ultimate responsibility” becomes thus pointless. For at least in principle, the internal and external factors which determine the will expressed in the decision to perform the given action cannot be seen as impediments, but rather as genuine preconditions.45 An unbounded, unconditioned will would be, on the contrary, nobody’s will and therefore no will at all,46 since an undetermined will, to speak with Hegel, cannot be but the pure abstraction of the absolute possibility and hence “the freedom of emptiness”.47 Freedom of the will must rather be understood as a will that contemplates its own conditioned nature as a self-determined one.48 Among the properties that co-define a person’s self one should distinguish between personal capacities and personal preferences. Personal capacities are those which enable a person to make decisions and perform or omit the actions that she decides to perform or to omit, which means: the capacities that are necessary to act according to one’s own preferences. For the consideration of an action as freely willed it is thus critical to establish which preferences can be held as constitutive for the person’s self, so that an effective decision to perform or not to perform an action in which the agent’s personal preferences are expressed can be regarded as self-determined. 41 42 43 44 45 46 47 48
Joseph Raz, Practical Reason and Norms, 1975, 65–71 Frankfurt (note 28), Importance, 21 Pauen (note 2), 59–65 Davidson (note 8), Essays, 19 Bieri (note 34), 49–53; Moore (note 40), 1130 Bieri (note 34), 239–42 G.F.W. Hegel, Grundlinien der Philosophie des Rechts, 1821 § 5 Hegel (note 47), § 7
Determinism, Free Will and Criminal Responsibility
57
b) Personal preferences There are two possible versions for this criterion.49 According to the first version, a preference counts as personal if it is itself the object of an effective autonomous decision, which requires the person to be in position to critically evaluate and reject her own preferences. According to the second version, what is emphasized is the person’s disposition to reflexively identify herself with the given preferences, so that the decisive question points to the possibility of a positive self-ascription of such preferences. Under this last version, freedom of the will becomes a matter of selfinterpretation. The claim that, under any of both versions, the compatibilistic account of free will could not avoid the infinite regress, since the recognition of personal preferences would now depend on the recognition of further preferences, which would itself depend on further preferences, and so on, is misplaced. For the answer to the question whether a person is capable of criticizing or identifying herself with a system of preferences does not depend on an actual ascertainment of those more fundamental preferences, to which the exercise of this capacity relates. That there are such preferences does not mean that one can always establish which they are. To be sure, both versions of the criterion present difficulties. For the autonomy of a decision is, on both versions, a function of the correspondence between the preferences that determines which reasons explain the given action, and the system of preferences that define the person’s self-concept, whereas the former is to be evaluated in the light of the latter. Yet the crucial remark is that a clear delimitation of the preferences that are constitutive for the person’s self, on the one hand, from those that are to be evaluated in terms of their congruence with the first, on the other, cannot be thought of as being static.50 And this is true of each ascription of responsibility, at least if one acknowledges that the total image of an agent’s self also depends on which are the deeds that can be seen as his own.51 The concept of the self exhibits a certain plasticity, in the sense that there are no essential, irreplaceable properties of a person’s self, as well as a certain “scalarity”, in the sense that its application is not an all-or-nothing matter, but rather one of degree.52 This does not alter the fact, however, that such a self-conditioning of the self can only limitedly take place at any one time. At each moment in which I ask whether my own will is free, there is a piece of internal soil that cannot be subject of this question, remaining as its very context.53 Therefore, the articulation of the concept of the self must avoid any commitment to essentialism, namely to the idea that the self could designate a mysterious entity that would correspond to the essence of the person. For there is no such entity, but only a complex interplay of properties which, although variable over time, stand as answer to the question, at a given time, of who a person actually is. Which means: the self is essentially contingent.54 Just as 49 Pauen (note 2), 75–96 50 Bieri (note 34), 400–2 51 Meir Dan-Cohen “Responsibility and the Boundaries of the Self ”, Harvard Law Rev. 105 (1992), 970 52 Dan-Cohen (note 51), 965–68 53 Bieri (note 34), 409–10 54 Pauen (note 2), 65–68
Juan Pablo Mañalich
58
the strength of a thread does not rest on a single strand that runs trough its whole extension, but rather on the fact that many different strands overlap with each other,55 the identity of a person over time does not require the subsistence of a mysterious substance, but only the recognition of a relative continuity of interrelated components. Yet at some stage a point must be reached where the question whether we have positively decided ourselves for a given feature of our own personality becomes pointless. The image of a person being free from the features that, on the whole, constitute her own personality, has no sense. For there is no point in raising the question of whether I am free from myself, since then the pronoun “I” has lost its reference. As Dennett puts it: “If you make yourself really small, you can externalize virtually everything”.56 4. AGENCY
AS UNDETERMINED DETERMINATION OF EVENTS?
An important argument for compatibilism asserts that only through the assumption of determination of the events that are ascribed as freely willed actions one can make sense of the notion of agency regarding the person to whom the event is ascribed. The commitment to indeterminism regarding freely willed actions would then appear to conflict with the requirements of agency, since their indetermination would imply their randomness. A way-out for the indeterminist would be open if he were able to hold that the agent’s determination of her own action could itself be undetermined. Such a conception of agency can be found in the work of some German criminal law theorists, who defend a concept of freely willed actions that pursues to establish the basis for the traditional doctrine of the so-called Regressverbot or “regress prohibition”.57 The core of their position concerns the opposition between the notions of causa libera and causa naturalis, according to which a “prohibition” to trace back the former causal history of an event applies if one identifies a singular cause that itself has no cause – for strictly speaking, it could have no prior causal history.58 This strategy does not lead, however, to the expected outcome. The attempt to introduce an interruption of the causal chain that fully explains an action, in order to achieve a concept of freedom that makes sense of the agency principle, is inevitably helpless.59 For an interruption of the precedent causal chain would always imply a denial of the causal relevance of the agent’s own history (i.e. her biography) regarding her decision to act, so that she could not be seen as conditioning her action by
55 Ludwig Wittgenstein, Philosophische Untersuchungen, 1984, § 67 56 Dennett (note 1), 122 57 Joachim Hruschka “Regreßverbot, Anstiftungsbegriff und die Konsequenzen”, Zeitschrift für die gesamte Strafrechtswissenschaft 110 (1998), 581–610; Jan Joerden, Strukturen des strafrechtlichen Verantwortlichkeitsbegriffs (1988), 16–20, 30–35 58 Joerden (note 57), 30 59 Dennett (note 1), 103–26; Bieri (note 34), 239–42; Alfred Mele, Autonomous Agents, 2001, 8–13; Pauen (note 2), 162–75
Determinism, Free Will and Criminal Responsibility
59
means of a prior decision that itself were not random. The performance of the given action would then have no internal connection to the person’s identity. The problem of this indeterministic conception of freely willed action rests on the assumption that, in order to be labeled as freely willed, action must be, to some extent, undetermined. Such an account entails a purely negative understanding of autonomy as absence of determination. The picture of a new beginning of a causal chain due to the performance of freely willed action, however, does not make sense, since actions and decisions can only be characterized as autonomous if one can explain them as the product of the beliefs, desires and preferences of the agent.60 For there is no way of making sense of the idea that an agent could have control over a random course of events: “Wouldn’t it be much safer –and hence more responsible– to keep the randomizer inside you, under your watchful eye in some sense? No. Randomness is just randomness; it isn’t creeping randomness”.61 5. SELF-DETERMINATION
AND CRIMINAL RESPONSIBILITY
a) Law-abidance and democratic culpability It is the capability of intentionally acting on reasons that an agent can see as his own, because they are reasons which express personal preferences of him, what makes the core of the idea of freedom of the will understood as criterion for the attribution of definitive personal responsibility. Yet there is a particular feature of criminal responsibility that is to have critical impact on this compatibilistic account of free will. For this account assumes that there is a variable set of preferences that are constitutive for a person’s contingent identity. And although it may be particularly difficult to determine which preferences count as such, this difficulty does not alter the fact that there are such preferences.62 When it comes to the connection between free will and criminal culpability, however, it suffices to acknowledge that the commitment to the existence of personal preferences is indeed plausible. For the ascription of criminal responsibility must be conceived as taking place under a counterfactual assumption of personal compliance with the legal norm as personal preference of the citizen who may be made responsible for its violation. This counterfactual assumption can be understood as an hypothesis of loyalty to law, which means that the subject of imputation must be regarded, from a legal point a view, as law-abiding person. And exclusively on the legitimacy of this hypothesis certainly depends the justification of what can be called material – and not purely formal – culpability.63 The construction of this hypothesis concerns the problem of whether the defeasible and provisional legitimacy of a norm could be inferred from its legality, that is, from its legal validity. One way to render such an inference rests on the possibility of regarding the person from whom the compliance with the norm is legally ex60 61 62 63
Pauen (note 2), 147–52; Mele (note 59), 11–13 Dennett (note 1), 133 Bieri (note 34), 61–65; Pauen (note 2), 74–75 Urs Kindhäuser, “Rechtstreue als Schuldkategorie”, Zeitschrift für die gesamte Strafrechtswissenschaft 107 (1995), 701–2, 725–27
Juan Pablo Mañalich
60
pected as its very author, so that the norm may be seen as her norm. This presupposes the acknowledgment, essential to democratic justifications of the authority of law, that the person who is made responsible for criminal wrongdoing may be seen as playing a double role with regard to the norm, the violation of which is imputed as culpable wrongdoing. The attribution of such a double role follows from the internal connection that holds between citizenship and legal personhood, which makes the core of what Klaus Günther calls “deliberative personhood”.64 The ascription of this double role rests upon an interplay of two different dimensions of autonomy. Citizenship relates to the factual possibility of exercising political autonomy, a distinctive feature of which is the potential participation in public discourses referred to the validity of common norms. Citizenship means thus possible participation in “justification discourses”. The opposite side concerns the position of the very same person in exercise of private autonomy, where the scope for permissible action is limited by the rules which prescribe how she is to act. This double ascription of autonomy enables a symmetrical expectation of reciprocity regarding the compliance with the common norms that bind each member of the political community. Only under conditions of democratic legitimacy can one consider the binding effect of legal norm as autonomous. The material ground for criminal responsibility must thus be seen as the unilateral appropriation of the switch between those two roles. The regulation and institutionalization of the interplay between the positions of author and addressee of valid legal norms stands as critical feature of democratic legal systems.65 As citizen, the subject can question and pursue the modification or abrogation of legal norms, but she may not invoke the exercise of this political right in order to discharge herself of following the norm in a situation in which she is immediately obligated to do so. A person’s culpability within a democracy can only be traced back to a breach of the mutual expectation of reciprocity with regard to the definition of the political game as the instance in which citizens can argue for the revision and cancellation of their own norms. b) Imputation rules and legal personhood The notion of material culpability can only be settled if one is able to offer a justification for the claim that the norm, which defines the wrongdoing for which I can be held responsible, is my norm.66 For the justification of that claim makes a genuine precondition of criminal liability, upon which all further conditions of liability depend.67 In order to be responsible for wrongdoing one must be answerable before someone. And this someone can only be the correspondent political community itself: “For one account of the moral conditions of the obligation to obey the law, and of being answerable through the courts, is expressed in terms of community. The defendant is obligated to obey the law in virtue of his membership of a com64 65 66 67
Klaus Günther, Schuld und kommunikative Freiheit, 2005, 97–105, 245–58 Günther (note 64), 252–53 Michael Detmold, “Law as Practical Reason”, Cambridge Law Journal 48 (1989), 455–63 R.A. Duff, “Law, Language and Community: Some Preconditions of Criminal Liability”, Oxford Journal of Legal Studies 18 (1998), 192
Determinism, Free Will and Criminal Responsibility
61
munity whose law it is”.68 And this is only possible if the defendant can see the language of the law as his own, that is, “speak it in an authentically first personal voice”.69 It is critical to recognize, however, that in order not to collapse the judgment on legal culpability into a moral one, a critical self-identification with the legal norm as a valid reason for action is not to be expected from the defendant.70 This means that the actual motivation of the subject plays no role in establishing whether she is criminally responsible for wrongful action. The motivation to observe the norm as effective reason for action must be seen, on the contrary, as a counterfactual assumption that underlies imputation, so that the conditions of personal responsibility can be formulated as a set of specific negative conditions fixed in legal imputation rules. When these special conditions are indeed satisfied, compliance with the norm ceases to be expected, either because the actual motivation capacity is seen as failing – e.g. in case of insanity or infancy – or because in the specific given situation a lawful motivation would lead to a major personal sacrifice – e.g. in case of duress. The fact that these conditions are fixed in legal rules means that a normative concept of personhood is also fixed under these rules. This suggests that, to some extent, the criminal law’s concept of personhood becomes objectified. Yet the generalization implicated in such objective definition is, to some degree, inevitable. For the possibility of identifying necessary conditions of personhood does not imply that the identification of sufficient conditions must also be possible.71
68 Duff (note 67), 197 69 Duff (note 67), 198–99; Detmold (note 66), 468–69 70 Klaus Günther “Individuelle Zurechnung im demokratischen Verfassungsstaat”, Jahrbuch für Recht und Ethik 2 (1994), 156; Kindhäuser (note 63), 726–27 71 Dennett (note 27), 285
RONALD J. ALLEN, CHICAGO1 FREE WILL IN CRIMINAL LAW RESPONSES
AND
PROCEDURE:
I am very indebted to all of the commentators for their invigorating presentations. Now, my task is briefly to take on the entire field of Christian theology and apologetics, ably represented here by Prof. Montgomery, the new approach to the free will problem that Judge Hodgson has developed over the years, respond to the effort to locate free will in the sophisticated analysis of causation that underlies Friedrich Toepel’s paper, and of course to take on the whole new generation of epistemologists represented by our two Ph. D. Students here, Allan McCay and Juan Pablo Manalich. And again, I have about 5 pages to do all this, which should be no problem at all. It won’t be any problem at all because obviously it cannot be done! All that can be done is to give a few cursory and shallow but hopefully provocative responses that at best weakly reflect the depth and significance of each of the contributions, and by doing so further the debate. So, I have no hopes of systematically engaging the many interesting contributions, but I have great hopes of being a provocateur who incites each of them to robust debate. On that note, I turn first to: 1. PROF. JOHN WARWICK MONTGOMERY I am quite honored to be on the panel with such a distinguished academic, from whose presentation I learned much. There is a sense in which he is doing in his paper something quite different from my work, but also something different from what motivates much of his previous work. As I said, I am a scientist, and interested in testable propositions. Much of Prof. Montgomery’s work has been a careful analysis of the evidence supporting aspects of Christian belief, like the historical truth of Christ, and so on. That is not what’s going on in this paper, at least that part of this paper that intersects with my own work. Here is posited an unobservable entity to resolve an empirical puzzle. The creation of unobservable entities as place holders in theoretical systems (electrons, quarks, whatever) is about as far as I’m willing to go, and even then there must be testable consequences. So, to ground an argument on things being “established on a transcendent foundation–on the basis of clear revelatory teaching.” is to me to forgo the scientific enterprise. That is not to say that there are not mysteries in life, and there are: consciousness, infinity (including an infinite number of sentient beings), the beginning of the universe (a version of infinity), and so on. But I don’t see what good it does to posit another mysterious entity without testable implications–God–to solve these problems. All it does is relocate them. Thus, as a legal scientist, the contents of this paper are important as disclosing the views of members of the community who are to be governed by whatever laws are created, laws which must be responsive to the governed community, but the 1
John Henry Wigmore Professor, Northwestern University School of Law.
64
Ronald Allen
substantive content of those views, their truth or falsity, is another matter. One cannot rationally test the truth or falsity of revelatory teaching, and thus these views have no power rationally to persuade. But, what is interesting in another sense is that the incompatibilists who wish to defend free will end up doing the same thing – they, too, hypothesize mysterious and untestable entities to house free will. This leads me to 2. JUSTICE DAVID HODGSON Interestingly, Justice Hodgson and I seem to agree on the issue that motivated my paper that in turn motivated this panel, which is how to regulate police interrogations. We agree that the regulation should be of the police practices directly, without filtering the analysis through things like free will or voluntariness, regulation that can be done by straight forwardly mining the conventions of the community. The term “voluntary” would then mean that a confession was given in the absence of those practices, such as unacceptable physical or mental coercion/threats, that determine the inadmissibility of any statement given by a suspect during custodial interrogation. And remember, my original point is that this is the case regardless of one’s views of the free will problem. We get to this conclusion from completely opposite approaches, however, with Justice Hodgson relying on the fairness of taking into account exercises of will, but I think this is wholly unpersuasive, for all the reasons I identified in my original paper. I will not reiterate here, but simply pose the question: Why, or how would one know, that a person subjected to a polite half hour interrogation confesses as an exercise of free will but one subjected to 45 minutes, or an hour, or two hours, of similar interrogation is compelled? But more importantly, this leads us to the question of free will. In a series of books and articles over the last 15 years, Justice Hodgson has provided what he believes to be a robust defense of free will. In his own remarks today, he did not have time to develop it, and obviously, my time is even more constricted. So, I can just be the provocateur here. In essence, his argument is that free will resides in the interaction of the conscious agent, with its understanding of its unique history and experience, with choices that are available in the universe. As he puts it in his article in the J. of Consciousness Studies, “we take into account our assessments of whole particular gestalt experiences; yet these experiences cannot, as unique and particular wholes, engage with general rules.” Of course, this does not mean that those responses are not random, and he mounts no robust defense of his assertion that “it is not reasonable to think that the outcome in such cases is random.” Still, I agree that they don’t look random. I consistently rely on a limited set of variables in choosing the wine that I drink, and so on. But, what about determination? Judge Hodgson rightly points out that no computable algorithm could correctly describe or predict the complexity of the interactions human have with their environment, but he infers from this that therefore such interactions cannot be algorithmic, and since they don’t look random, ergo there is free will. But, here is the mistake, I think. There is a critical difference between a computable and a non-computable algorithm. Or more carefully put, that we lack the computational capacity to compute determinate outcomes does not
Replies
65
make those outcomes indeterminate. We are surrounded by examples of this. I do not think that Justice Hodgson thinks that water molecules or rivers or oceans have free will, and I am sure he believes that the actions of water molecules in rivers and oceans are completely determined, yet no computable algorithm can describe or predict those individual actions. And interestingly for this discussion, those actions do not at the discrete level look random. Gravity is another example. Gravitational relationships between two objects in a vacuum and not in the gravitational field of any other objects are easily computable. Add a third body, and we can no longer do it. We can approximate but we cannot compute. Add a bunch of bodies and we can no longer even approximate; all we can do is watch and wait. So too with humans and their environment. That we all have unique experiences does not mean general laws do not apply; it means at most only that we could not compute outcomes for the specific application of those laws. Justice Hodgson makes the standard argument that no computer can do what the mind can, and therefore, the mind cannot be like a computer, which of course opens up space for some mysterious entity like free will to be the explanation of what we observe. This has always struck me as a curious argument. First, our capacity to deal with computationally complex problems is increasing. DNA computation allows the solution of some problems that previously were thought not solvable; fuzzy set theory allows formalization of fuzzy relationships, and so on. More importantly, what possible significance is it that present computers are not as competent in certain respects as typical human minds? It does not necessarily follow from that that minds are somehow different. An equally plausible explanation is that computers are not as advanced in computation (of the right sort) as minds. Justice Hodgson’s argument privileges as the standard what might be a relatively crude stage in the development of a machine. Rather than accept any of this, however, Justice Hodgson hypothesizes a mysterious decision making capacity that is neither random nor determined, but he gives literally no reason to think such a thing exists other than the unavailing reliance on computational complexity. More problematic still, he gives literally no explanation of how it works. And how could he? It is plainly obvious that any particular choice will be caused by something, especially so for someone who rejects dualism, as he says he does. Thus, as Tom Clark has put it, Justice Hodgson, rather than giving us a transparent argument for free will has given us instead free will as an impenetrable black box into which are put all the interesting questions, and from which one gets no answers. There are many other aspects of Justice Hodgson’s argument that he repeats here that have met sustained and serious, and in my opinion persuasive, criticism (his claim about the vicious circle of reasoning justifying evolutionary explanations, for example), but hopefully I have met my obligations as provocateur for the moment. I should, however, say, that all of you here should know that criticism, not imitation, is the sincerest form of flattery. In any event, Judge Hodgson, like Prof. Montgomery, creates a mystical, mythical, unobservable entity and asserts it is the solution to the problem of free will. In my opinion, so, too does our host,
66
Ronald Allen
3. FRIEDRICH TOEPEL I am particularly happy to have Friedrich’s paper, for it brings to completion a process that began over a number of years ago. We were blessed with Friedrich’s presence as a Fulbright Scholar, and I was his faculty contact. He attended my Constitutional Criminal Procedure class, in which I discuss many of the issues we are discussing today. I made a simple argument for the non-existence of free will and asked if anyone had any questions. I called on Friedrich who stood up, walked to the black board and began what was obviously going to be an interesting, but long, lecture on the problem of free will. Well, I was trying to teach American Constitutional Criminal Procedure, so I fear I told him that he would have to give a radical summary of the argument and that we could talk about it more after class. Now we have that luxury. It is interesting to note, which I think is a commentary on how intractable our problem is, that he paper he presents now is not quite what he was presenting then. Friedrich, like the previous two speakers, needs to find some place to locate something that avoids the horns of the dilemma of being either determined or random. He suggests that the actual decision facing people involves a complex set of reasons, which are best thought of as weakly sufficient to explain an outcome. Since they are only weakly sufficient, free will resides in the choice over these variables. This point may be valid, but it does not dispose of the problem of free will as an uncaused attribute of a subjective agent, which is what the problem that I engaged with requires. Rather, it just blurs over, or like Justice Hodgson, puts in a black box, what is actually occurring. The argument that in the actual human decision making context, there will be a lot of reasons, and often no one in particular will be dispositive, is picking up on the complexity of causation, and there are indeed complexities here. Two bullets strike a person in the heart at the same time. Which does the killing? More to Friedrich’s point, a person bleeds to death from a thousand cuts caused by a thousand different people, or ten people each put a dose of poison sufficient to kill in the goblet: Who is responsible? Well, these are interesting hypotheticals in first year criminal law class, and they do show how simple notions of causation don’t suffice, but they certainly do not show that the results are not determined by pre-existing conditions. Same too with a complex set of weak reasons, some subset of which causes a person to choose to do X. However the person takes into account the universe of reasons, again she is doing so for reasons or randomly, and thus we are back to our main motif. It may be that in light of the ten reasons I would act one way, but you would act another, but it is still that set of reasons causing choice in me. Friedrich’s argument is not a justification of free will; rather, it is an “as if ” argument; we can behave as if we had free will because decision making is often quite complicated. He is right that it is often quite complicated. He is also right that we can behave as if we were free and autonomous. But these arguments seem to concede what is actually at stake in his discussion, that free will is a fictional account of the human condition that, like religion, may serve important psychological purposes but, like unicorns, doesn’t exist in this universe.
67
Replies
4. JUAN PABLO MANALICH
AND
ALLAN MCCAY
I am very pleased to have the next generation of philosophers represented here. Both Juan Pablo Manalich and Allan McCay have written interesting papers, with great promise. The primary task of each new generation of scholars is to correct the errors of their predecessors, and thus it is that knowledge advances. I am sure that they both think that they have observed a number of errors in need of correction here today, and I am sure they are right about that. So, I mainly want simply to encourage them both in their endeavors, although I have one piece of advice, criticism, for each of them. Allan McCay is, among other things, examining the relationship between various philosophical perspectives and the justification for the criminal law. This is a worthwhile project, but I would caution that perhaps it misevaluates the criminal law, or any other substantial body of law, to think of it as a logically closed system that could be “justified” or “explained” on the basis of any relatively simple basis. The law is an accumulation of untold numbers of responses to an almost infinite variety of problems. It is a human construct, not a logical entity. In Hayek’s famous phrase, it is grown and not made. To the extent this is true, treating it as a made system subject to top down, normative critique, may simply be making a mistake about the object of the evaluation, and thus it is unclear what follows from the inability to give a sound logical or philosophical underpinning to it. Juan Pablo Manalich seems to me to be working with and examining the German philosophical tradition, which, as I understand it, long has tried to explain how the obvious truth of determinism can be compatible with the Principle of Alternative Possibilities. I have nothing to add to that discourse, but I do have something to add from the perspective of a legal scientist. He asserts that “The image of a person being free from the features that, on the whole, constitute her own personality, is a nonsensical image. For there is no point in raising the question whether I am free from myself.” For what it is worth, in the law this point is very much worth asking, and may have dramatic consequences. Or, put another way, it is not obvious what “myself ” is. Let me finish today where I began, which is back in the doctrinal thicket of American Constitutional Criminal Procedure, and let me give you another of the puzzles I pose to my students. Remember that the search under the right to be free from compelled self-incrimination is to sort out what the state and what the person is responsible for. Promises or threats that cause a confession make it inadmissible because it is no longer the consequence of the person’s free choice. Now, consider programs that the state requires to be taught in schools that have the effect of heightening the “conscience” of individuals, or making them less able to resist their internal sense of guilt. And imagine further a person who confesses in part because of having being exposed to such programs. Has that state coerced him into speaking? Or is he the person who he is later in life, regardless of the earlier inputs? How do we know who the relevant “myself ” is?
II. THE CONNECTION BETWEEN PUNISHMENT, RETRIBUTION AND FREE WILL (PROCEEDINGS
OF THE
24TH IVR WORLD CONGRESS BEIJING 2009)
PREFACE Why this second workshop on free will? The idea of a second workshop was born in a discussion by the participants of our first workshop when we met at the outdoor seating area of a restaurant in the market square of Kraków. I tried to explain that I am not able to see how a compatibilist could create a notion of free will that would be good for anything after he conceded the possibility of a completely deterministic world. Some of my colleagues did not agree with me. They were convinced that compatibilists could solve a good deal of problems. They only seemed to hesitate as to whether a sufficiently ‘deep’ account of legal or moral responsibility could be given if we allow for the possibility that all choices are predetermined. I agreed that for me certainly something necessary to impute legal or moral responsibility to an agent would be missing if I would adhere to a form of compatibilism. However, it was difficult for me to describe what this missing part might be. I think this problem has not been much discussed in legal philosophy until now, yet it is of vital importance because our attitude towards the discussion of free will and new discoveries in the neurosciences depends on whether we take the view that compatibilism provides us with a notion of free will ‘deep’ enough for the imputation of responsibility or not. If the compatibilist account is ‘deep’ enough, we would not need to worry about neuroscientists who purport to have discovered that all choices are determined. Compatibilism is reconcilable with determinism, so that’s it. If, however, something is missing in the compatibilist account of free will and this prevents us to justify holding persons responsible, then it might well be that empirical evidence for determinism becomes the source of a problem. Therefore, a further workshop on free will in which we could examine whether there are shortcomings of compatibilism seemed to me to be very promising. I suspected that the idea of personhood underlying a retributive theory of punishment could not be accommodated within a compatibilist system. Accordingly, I chose ‘The Connection between Punishment, Retribution and Free Will’ as the title for our workshop, not as a limitation for the content of the papers, but as a working hypothesis which reflected my impression of the situation and could be corroborated or refuted by more detailed analysis. To me, the inadequacy of compatibilism seems to become apparent most clearly in the discussion of alternative possibilities. I therefore try to establish the following line of reasoning in my paper: A retributive theory of punishment demands an open future and therefore genuine alternative possibilities for the perpetrator at the time at which he commits the crime. Genuine alternative possibilities are not compatible with a completely deterministic world. If a deterministic world is ruled out as background for the imputation of responsibility, then compatibilism is ruled out. Frankfurt-type examples carry no persuasive power to the contrary. The retributive component of my topic has obvious roots in theological thinking which need to be explored. Therefore, it was natural that Professor John Warwick Montgomery and I should continue our cooperation that we had begun with the first workshop on free will. I think that particularly his interesting remarks on original sin and on the relationship between the New Testament and retribution make his paper a valuable contribution.
72
Friedrich Toepel
We thought it appropriate that our discussion should be supported by a professional philosopher who has made himself a name on the subject. Therefore we invited Professor Angus Menuge from Concordia University, Wisconsin, to our workshop. We were rewarded with a paper that not only digs deep into the philosophical controversies but also gives an insight into the modern free-will debate in the neurosciences. We were joined by Professor Mateusz Klinowski from the Jagiellonian University at Kraków who in his paper explores the logical implications of free will. His interesting contribution shows how the ‘stit’-calculus explains the ascription of responsibility in an indeterministic world and supports the view that the principle of alternative possibilities is a necessary condition for such ascription of responsibility. To my mind, this second workshop on free will was not less lively and less diversified than the first. It also neatly complements the first workshop by throwing some light on the impact of retribution and alternative possibilities. Bonn, November 2009
Friedrich Toepel
ANGUS MENUGE, MEQUON (WI), USA DOES NEUROSCIENCE UNDERMINE RETRIBUTIVE JUSTICE? I. INTRODUCTION Retributivism – the idea that there is a fitting punishment which a criminal deserves – has come under heavy fire from scientific materialists.1 According to psychologists Joshua Greene and Jonathan Cohen, the crux of the issue is that retributivism assumes a ‘common sense, libertarian conception of free will’ which, they claim, is (or soon will be) undermined by ‘new neuroscience’.2 Greene and Cohen think scientists are justified in assuming that the brain can be understood in completely materialistic terms. Materialism recognizes only undirected causes (i.e., those governed by blind chance and necessity). This excludes teleological (goal-directed) causes, making it impossible for the purely mental volitions of an agent to intervene in and redirect the physical order in novel ways. Against defenders of libertarian free will, they assert: There is not a shred of scientific evidence to support the existence of causally effective processes in the mind or brain that violate the laws of physics …. [A]ny scientifically respectable discussion of free will requires the rejection of what Strawson … famously called the ‘panicky metaphysics’ of libertarianism.3
What Greene and Cohen mean by the libertarian’s ‘panicky metaphysics’, is the ‘magical mental causation’4 of an agent’s volitions, which would exert a downward, teleological causal influence from the mind to the brain. If downward mental causation is excluded, then libertarian free will is impossible, and, they argue, there is then no sound basis for retributive justice. Greene and Cohen maintain, correctly in my view, that compatibilist accounts of free will do not ground the kind of ultimate responsibility for action which retributivism requires.5 Without libertarian free will, 1
2 3
4 5
For example, Henrik Walter, while trying to salvage a compatibilist notion of responsibility, argues that “There is no ‘ultimate responsibility’ of the kind postulated by libertarians”; see Henrik Walter, Neurophilosophy of Free Will: From Libertarian Illusions to a Concept of Natural Autonomy, Cambridge, MA: MIT Press, 2001, 290. Likewise, Daniel Wegner asserts that ‘We come to think of … prior thoughts as intentions, and we develop the sense that the intentions have causal force even though they are just previews of what we may do’; see Daniel Wegner, The Illusion of Conscious Will, Cambridge, MA: MIT Press, 2002, 96. Joshua Greene and Jonathan Cohen, For the law, neuroscience changes nothing and everything, Philosophical Transactions of the Royal Society of London, Series B (2004) 359, 1775–1785, 1776 Joshua Greene and Jonathan Cohen, For the law, neuroscience changes nothing and everything, 1777. The phrase ‘panicky metaphysics’ is found in Peter F. Strawson’s classic paper, Freedom and Resentment, Proceedings of the British Academy, xlviii (1962), 1–25. Joshua Greene and Jonathan Cohen, For the law, neuroscience changes nothing and everything, 1780 Compatibilist attempts to justify some form of retributive punishment on the basis that people can be punished for behaviors resulting from their own reasons fail to convince, because there is nothing in the mental life of the agent which is really ‘up to the agent’, as it can all be traced to prior, impersonal causes outside of his control. In this sense, as Robert Kane has argued, compatibilism fails to ground the idea that ‘genuinely free agents must have ‘ultimate’ or ‘buck-stop-
Angus Menuge
74
punishment may continue, but only for such consequentialist reasons as deterrence and containment, not because it is what the criminal deserves: The law will continue to punish misdeeds, as it must for purely practical reasons, but the idea of distinguishing the truly, deeply guilty from those who are merely victims of neuronal circumstances will … seem pointless.6
Within Greene and Cohen’s remarks, we can distinguish three main lines of argument against libertarian free will. First, they believe in the causal closure of the physical, according to which every event has a sufficient, completely physical cause: this is why the volitions of an agent cannot be ‘causally effective’. Second, they think that the downward mental causation presupposed by libertarian free will would be ‘magical’, because it would ‘violate the laws of physics’. And finally, they claim there is no empirical evidence for downward causation and strong evidence against it, citing the work of Daniel Wegner.7 To the contrary, I will argue in this paper that: we have good reasons to deny that the physical is causally closed (section 2); libertarian free will does not need to violate physical laws – the apparent conflict rests on a mistaken conception of laws (section 3); the empirical case against downward causation is weak (section 4); and there is strong evidence that conscious will can contribute to healing neurological disorders by exerting a downward causal influence on the brain (section 5). 2. AGAINST CAUSAL CLOSURE The causal closure of the physical (CCP) is the foundational principle of strict naturalism8 or physicalism. Jaegwon Kim, a major proponent, formulates CCP as follows. Pick any physical event … and trace its causal ancestry or posterity as far as you would like…this will never take you outside of the physical domain. Thus, no causal chain involving a physical event ever crosses the boundary of the physical into the nonphysical.9
6 7
8
9
ping’ responsibility for their free actions and character’; see Robert Kane, Free Will: The Elusive Ideal, Philosophical Studies 75 (1994), 25–60, 33. Notice that a suitably programmed computer can do damage to other people e.g. extortion, on the basis of a rational program (e.g. one which implements an extortion scheme), but no-one thinks that the extortion was up to the computer or that the computer was ultimately responsible for the extortion. This seems a very good explanation of why we do not punish the computer for extortion. But on compatibilist accounts, human beings are precisely biological computers, programmed by genes and environment, so apparently we should not be punished because we deserve it either. Joshua Greene and Jonathan Cohen, For the law, neuroscience changes nothing and everything, 1781 Greene and Cohen think that Daniel Wegner’s The Illusion of Conscious Will provides good evidence that ‘our first-person sense of ourselves as having free will may be a systematic illusion’; see Greene and Cohen, For the law, neuroscience changes nothing and everything, 1783. The term ‘strict naturalism’ is used by Stewart Goetz and Charles Taliaferro to denote those naturalists who deny that any teleology is real; see their Naturalism, Grand Rapids, MI: Eerdmans, 2008. By contrast, ‘broad naturalism’ accepts the reality of teleology and attempts to explain it in naturalistic terms. It is clear from their remarks that Greene and Cohen are strict naturalists or physicalists. Jaegwon Kim, Philosophy of Mind, second edition, Cambridge, MA: Westview Press, 2006, 194– 195
Does Neuroscience Undermine Retributive Justice?
75
An important implication of CCP is that an agent’s volitions or other mental acts or states have no distinctive impact on the physical world: they cannot produce any new effects which could not be explained by physical causes. Thus, if the instantiation of a physical neurological property P is causally sufficient for behavior B, CCP implies that for any mental property M (including any M which supervenes on P): (1) M has nothing to do with why B happens (M is excluded by P—the so-called exclusion problem); and (2) M could not change B to some other behavior B* instead. While Kim proposes that we preserve some sort of mental causation by identifying mental properties with physical properties, the fact remains that the mental makes no causal contribution over and above the physical. What this means, however, is that if CCP is true, then all of those characteristics of the mental realm which do not reduce to the physical can have no causal influence on the physical: they must be epiphenomenal. This includes much more than the volitions of a libertarian agent. A strong case can be made that consciousness, intentionality and rational agency are all irreducible to the physical, yet also that all do have a causal influence on the world. To be sure, some naturalistic philosophers have attempted to avoid these arguments by defending some form of ‘non-reductive physicalism’ or ‘emergentism’, according to which mental properties can have their own causal powers even though they are determined by, or the product of, a purely physical base.10 Thus non-reductive physicalists (such as Terence Horgan11) have defended the compatibility of downward causation with materialism. And so have emergentists,12 those who claim that mental properties arise from sufficiently complex physical systems, and bring new causal powers into the world. However, in my view Jaegwon Kim has argued persuasively that neither non-reductive physicalism13 nor even emergentism14 is compatible with downward causation so long as the positions are committed to supervenience, CCP, and the denial of systematic causal overdetermination. This case has recently been strengthened by the work of J. P. Moreland, who shows with considerable rigor that the kind of emergent properties capable of accounting for downward causation are incompatible with the physicalist’s underlying metaphysics.15 10 Thanks to Mateusz Klinowski for discussion on this issue. 11 See, for example, Horgan’s entry, Reduction, reductionism, in: The Encyclopedia of Philosophy: Supplement, ed. D. M. Borchert, New York: Macmillan, 1996. Horgan also defends a position he calls ‘Robust Causal Compatibilism’ in Terence Horgan, Kim on Mental Causation and Exclusion, Noûs, Vol. 31, Supplement: Philosophical Perspectives, 11, Mind, Causation, and World (1997), 165–184. 12 A useful survey of views on the mind-body problem which includes several versions of emergence is Robert Van Gulick, Reduction, emergence and other recent options on the mind-body problem: a philosophic overview, Journal of Consciousness Studies 8 (2001), 1–34. 13 Jaegwon Kim has made this case against non-reductive physicalism in many places, including his: Mind in a Physical World: An Essay on the Mind-Body Problem and Mental Causation, Cambridge, MA: MIT Press, 1998; his Philosophy of Mind, second edition Cambridge, MA: Westview Press, 2006 and his essay ‘Causation and Mental Causation’, in: Contemporary Debates in Philosophy of Mind, eds. Brian P. McLaughlin and Jonathan Cohen, Malden, MA: Blackwell, 2007 (see especially his critique of Horgan’s views in section 6). 14 Kim develops this problem for emergentism most fully in Jaegwon Kim, Emergence: Core ideas and issues, Synthese 151 (2006), 547–559. 15 See James. P. Moreland, Consciousness and the Existence of God: A Theistic Argument, New York: Routledge, 2008. These arguments do not rule out the emergent dualism, however, a position
76
Angus Menuge
First, consider consciousness. The so-called ‘hard problem’ of consciousness is that a complete, impersonal knowledge of how brains work does not entail either the existence of subjective conscious states or anything about what it is like to be in those states.16 Furthermore, as Goetz and Taliaferro point out, conscious experiences, such as the experience of pain, cannot be understood in physical terms because, while ‘physical explanations of the intrinsic natures of things / events are typically given in terms of part-whole compositional and spatial terms … an experience of pain is simple … in the sense that it is not made up of event parts’.17 But if consciousness does not reduce to the physical, then (given the failure of non-reductive physicalism and emergentism) the physicalist must claim that it is epiphenomenal, something which mysteriously emerges from the physical, but which has no effects on the physical, like the sparks thrown up from a machine. This is a consequence which even naturalistic philosopher John Searle finds absurd: If, for example, you think you ate because you were consciously hungry, or got married because you were consciously in love … or withdrew your hand from the fire because you consciously felt pain, or spoke up at the meeting because you consciously disagreed with the main speaker, you are mistaken in every case.18
Moreover, the idea that consciousness is epiphenomenal is implausible even on physicalist grounds. Physicalists assume that our cognitive faculties exist because they conferred an adaptive advantage on our ancestors (they increased the chances of survival and reproduction).19 But natural selection can only select an organism’s traits on the basis of its behavior, so we would expect mental faculties or abilities to be selected only if they have a causal influence on behavior. Granted epiphenomenalism, however, consciousness has no such influence, so it cannot be selected and this means that the existence of consciousness must be a remarkable fluke, an accidental byproduct of some feature which did confer an advantage. As many philosophers have noted, on the physicalist’s assumptions, there are possible ‘zombie’ worlds behaviorally identical to our own, but in which consciousness did not happen to appear as an accidental byproduct.20 But given the biological cost of supporting consciousness, together with its presumed uselessness, natural selection makes one of the zombie worlds much more probable than the actual world. So physicalism fails to give a plausible explanation of the existence of consciousness.21
16
17 18 19 20
21
ably defended by William Hasker in his The Emergent Self, Ithaca, NY: Cornell University Press, 1999. See, for example, David Chalmers, The Conscious Mind: In Search of a Fundamental Theory, New York: Oxford University Press, 1996; and Colin McGinn, The Mysterious Flame: Conscious Minds in a Material World, New York: Basic Books, 1999. Stewart Goetz and Charles Taliaferro, Naturalism, 45–46 John Searle, The Mystery of Consciousness, New York: New York Review of Books, 1997, 154 See, for example, Steven Pinker, How the Mind Works, New York: W.W. Norton and Company, 1997. For a penetrating analysis of the incompatibility of Darwinist epistemology and a causal role for consciousness, see William Hasker, The Emergent Self, Ithaca, NY: Cornell University Press, 1999, 75–80. In fact, as J. P. Moreland has shown with impressive rigor, theism provides a superior explanation for consciousness. See J. P. Moreland, Consciousness and the Existence of God: A Theistic Argument, New York: Routledge, 2008.
Does Neuroscience Undermine Retributive Justice?
77
Further, if consciousness is a causally impotent, accidental byproduct, then its deliverances cannot have been honed by selection, and there is therefore no good reason to suppose that they would track events in the environment or even our own behaviors in reliable ways. Suppose we are attacked by devouring lions. On the assumption that consciousness is epiphenomenal, so long as our bodies evade the lions, it won’t matter if we consciously believe we are enjoying the Queen Mother’s garden party! Yet it is a fact of experience that our conscious states do (normally) track the world in systematic, reliable ways.22 And even physicalists must assume they do, because scientists rely on conscious reports of their observations and behavior when performing experiments and testing theories. Second, consider intentionality, the fact that our thoughts are about something and have a content. Thus, the belief that the Great Wall of China is visible from space is about the Great Wall and its content is that the Wall is visible from space. Intentionality also appears to be irreducible to the physical. All physical relations are between actually existing entities, and these entities cannot be true or false of anything else in the world. But we can have beliefs about non-existent objects (such as a celestial toasted cheese sandwich) and the content of our beliefs can be true or false. If intentionality is irreducible to the physical, then (granted the failure of nonreductive physicalism and emergentism) the physicalist will have to claim that what we believe has no effect on our behavior. It follows that reliable belief-forming mechanisms would be invisible to natural selection: if what we believe is epiphenomenal, then there are close physically possible worlds in which the content of people’s beliefs is redistributed at random. But then a scientist might believe he was testing a prediction of Quantum theory when in fact he was judging Alsatian wines. Under these circumstances, no-one, including the physicalist, can rely on his beliefs.23 Lastly, consider rationality. Rational beings express the reasons for their actions and convictions by citing the content of their beliefs and desires. But these reasons cannot be reduced to the physical, not only because they exhibit intentionality, but also because they are inherently teleological. Thus someone opens the fridge because they desire a cold drink (the goal) and believe they can get one by opening the fridge (the means to the goal). And they believe that A = C because that is the goal (conclusion) you reach if you consider what follows from A = B and B = C. If reasons cannot be reduced to the physical, then (granted the failure of non-reductive physicalism and emergentism) on the physicalist’s account of reasoning, the content of an agent’s reasons is causally irrelevant to the actions he performs and the conclusions he draws. But if so, it becomes an extraordinary, unexplained coincidence that 22 Despite his skepticism about the causal efficacy of conscious will, even Daniel Wegner grants this, since he claims our conscious intentions are previews of our actions. 23 For a more thorough development of this argument, see my Beyond Skinnerian Creatures: A Defense of the Lewis / Plantinga Critique of Evolutionary Naturalism, Philosophia Christi, Vol. 5, No. 1, 2003, 143–165. This paper is a defense of work by Alvin Plantinga, including his Is Naturalism Irrational?, ch. 12 of his Warrant and Proper Function, New York: Oxford University Press, 1993 and his Warranted Christian Belief, New York: Oxford University Press, 2000. Plantinga has responded at length to his critics in Reply to Beilby’s Cohorts in: Naturalism Defeated: Essays on Plantinga’s Evolutionary Argument Against Naturalism, ed. James Beilby, Ithaca, New York: Cornell University Press, 2002.
Angus Menuge
78
people, including scientists, regularly perform actions that make sense in light of their beliefs and desires. Why shouldn’t the scientist do an experiment because he desires to become an opera singer or because he believes that ping pong promotes harmonious international relations? More fundamentally, if reasons are causally irrelevant, then people cannot be credited with reasoning to their conclusions, even if those conclusions happen to be in accordance with reason. The distinction is important. Any pocket calculator reaches conclusions in accordance with the rationality of arithmetic; but the calculator does not reason to its conclusions, because no reasons of the calculator’s (it has none) play any role. By contrast, consider William Hasker’s example of a bright student evaluating a logical argument: If she is skilled in carrying out such assessments, she is said to possess “good logical insight,” an intellectual virtue which is prized, in part, for the specific reason that it enables one to reach good, well-justified conclusions about the arguments one encounters. The entire process makes no sense at all, except on the assumption that a person’s awareness of reasons and her knowledge and application of principles of rationality make a difference to the conclusions that are accepted.24
The idea that we do not reach conclusions because of our own reasoning (we do not reason to those conclusions because of our beliefs about what follows) is absurd. And most decisively for the physicalist, if reasons (qua reasons) are not causes, then the scientist can no longer be credited with reasoning to his conclusions and we no longer have grounds for believing in the rationality of science.25 By excluding the causal power of consciousness, intentionality and rationality, CCP allows no credible explanation of the way these mental characteristics track the world in systematic, reliable ways. CCP should not be used to exclude libertarian free will because it proves too much: it undermines the rationality of action and belief upon which everyone, including the physicalist, depends. Elsewhere, I have argued that CCP faces other difficulties as well, not the least of which is that there is no credible, stable definition of ‘physical’.26 3. INTERVENTION
WITHOUT
VIOLATION
Physicalists, however, remain attached to CCP, because it is closely connected to what they see as a defining principle of science, Methodological Materialism (MM). According to MM, science can only recognize the undirected causes of chance and necessity as described by the fundamental laws of physics.27 Were science to allow 24 William Hasker, The Emergent Self, 73 25 For further development of the argument from reason, see: William Hasker’s The Emergent Self, especially, 64–75; Victor Reppert’s C. S. Lewis’s Dangerous Idea: In Defense of the Argument from Reason, Downers Grove, IL: Intervarsity, 2003; and my Agents Under Fire: Materialism and the Rationality of Science, Lanham, MD: Rowman and Littlefield, 2004. 26 See Angus Menuge, Is Downward Causation Possible? How the Mind Can Make a Physical Difference, Philosophia Christi, Volume 11, No. 1, 2009, 93–110, especially 97–99. 27 For a critique of MM, see Angus Menuge, Against Methodological Materialism, in: The Waning of Materialism, eds. George Bealer and Robert Koons, New York: Oxford University Press, 2009.
Does Neuroscience Undermine Retributive Justice?
79
teleological causes, such as libertarian volitions, then, it is claimed, these laws could be routinely violated; but this cannot be so, for the regularity of the laws is confirmed by masses of sound evidence. Interestingly enough, this argument is really a re-packaging of the standard deistic argument against miracles. The deists argued that miracles were impossible because they would violate the rigid laws of nature; thus even God does not have the libertarian free will to intervene in nature. In response, Christian apologists such as Samuel Clarke and Thomas Sherlock pointed out that our belief in the regularity of nature is founded on experience, but experience of a pattern of cause and effect can never show that the connection is necessary and without exception. 28 Moreover, as C. S. Lewis argued, even if laws are necessary connections, they still make room for miracles. 29 This is because the laws of nature are conditional in nature. They do not say that an effect E (unconditionally) must happen: they say that if there is a cause C and initial conditions I and no interfering factors, then E must happen. But in the case of a miraculous intervention, the cause may be changed to C2, or the initial conditions may be changed to I2, and in that case there may be a different result E2. This response to the deists can also be used to defend mental causation and libertarian free will, as Victor Reppert has pointed out.30 The physical laws that govern brains do not say that a behavior B is unconditionally necessary. They say that if neurological event N occurs under circumstances C and if there are no interfering factors, then B occurs. But the fact that N is sufficient for B under C does not exclude other possible causes of B under other circumstances C*, and it says nothing about what happens if there is an interfering factor. This possibility was recognized by the renowned neuroscientist Wilder Penfield, who produced limb movements in his subjects by stimulating their motor cortex with electrodes.31 Penfield did not conclude that this type of purely physical cause was the only cause of our behaviors, and he personally endorsed a mind-body dualism which allows the mind an independent causal role.32 As Stewart Goetz and Charles Taliaferro point out: One of the things he noticed in his experimental work was that the patients reported being consciously aware of the distinction between being agents and doing things, and being patients and having things done to them. Phenomena such as this led him to endorse dualism.33
Goetz and Taliaferro go on to argue that if we understand the brain as a collection of neural capacities which can be actualized by the causal powers of other entities, then the fact that these capacities can be actualized by the stimulation of electrodes 28 For more details, see chapter 6 of William Lane Craig, Reasonable Faith, Third Edition, Wheaton, IL: Crossway Books, 2008. 29 See C. S. Lewis, Miracles, New York: Macmillan, 1960, 58. Lewis argued that even if the laws of nature have the logical necessity of arithmetic, they do not exclude miracles. Suppose that 5$ is put in a drawer on one day and 10$ on the next. Arithmetic has not been violated if we do not discover 15$ on the third day. Arithmetic says only that if we add $10 to $5 and nothing interferes then we will have $15: it says nothing about what happens if there is intervention by a thief (or a generous benefactor). 30 Victor Reppert, C. S. Lewis’s Dangerous Idea, 110 31 Wilder Penfield, The Mystery of the Mind, Princeton: Princeton University Press, 1975 32 Wilder Penfield, The Mystery of the Mind, 80 33 Stewart Goetz and Charles Taliaferro, Naturalism, 35–36
80
Angus Menuge
(or other purely physical causes) does not show they cannot also be actualized by an agent’s volitions (or other irreducibly mental powers). [O]ne can concede that a neuroscientist…might discover in his experimental work that actualizations of a neural capacity … can be produced by stimulation with an electrode or by exercisings of the causal powers of other neurons. But why think that every actualization of a neural capacity can be produced only in these ways? Why could not an actualization of a neural capacity be caused by an exercising of a mental power …?34
A good example that supports this point is provided by neurological studies of selfregulation in response to emotionally negative pictures. K. N. Oschner’s experiments35, cited by neuroscientist Mario Beauregard, contrasted ‘attend trials’, where ‘volunteers were requested to attend to and be aware of, but not try to modify any feelings induced’ with ‘reappraise trials’ on which ‘volunteers were instructed to reinterpret the negative picture so that it no longer generated a negative emotional response’.36 The fMRI scans revealed that, as expected, the limbic system (specifically the right amygdala) was very active on attend trials. Let us suppose that this happens in accordance with physical laws, so that whenever these stimuli are processed passively, such limbic activation occurs. This is clearly consistent with Oschner’s demonstration that subjects on reappraise trials who are instructed to actively suppress their emotions (by viewing the stimuli more objectively) are able to do so, and thereby produce different physical results in their brain (diminished activation of the limbic system, increased activation of the prefrontal cortex). Even if this is, as it surely seems to be, downward mental causation, it does not violate physical laws specifying what normally happens to the limbic system if no reappraisal of the stimuli occurs. In truth, natural laws prohibit libertarian free will only if one also assumes the causal closure of the physical, which simply begs the question against dualism and is also vulnerable to the objections of the previous section. The foundational issue here is really metaphysical: what kind of causal powers exist in the world? Materialism assumes that we can understand everything which occurs as passive event causation, in which a cause automatically produces its effect. By contrast, according to our common-sense, dualistic self-understanding as rational beings (‘folk psychology’), in addition to event causation, there is also agent causation, in which the exercise of an agent’s causal powers is not automatic but depends on the agent’s unforced active choices?37 But although the underlying issue is metaphysical, the results of neuroscience are relevant to the dispute, because the hypothesis of agent causal power makes predictions about an agent’s behavior which conflict with those of the materialist model. 34 Stewart Goetz and Charles Taliaferro, Naturalism, 41 35 K. N. Oschner et. al., Re-thinking feelings: an fMRI study of the cognitive regulation of emotion, Journal of Cognitive Neuroscience, 14 (2002), 1215–1229 36 Mario Beauregard, Mind does really matter: Evidence from neuroimaging studies of emotional self-regulation, psychotherapy and placebo effect, Progress in Neurobiology Volume 81, Issue 4, March 2007, 218–236, 224 37 For a recent, sophisticated defense of agent causation, see Timothy O’ Connor’s Agent-Causal Power, forthcoming in: Dispositions and Causes ed. Toby Handfield New York: Oxford University Press, 2009, draft available at: http://www.indiana.edu/~scotus/files/ACPower.pdf.
Does Neuroscience Undermine Retributive Justice?
81
If agent causal power exists, then for an identical stimulus, there should be a detectable difference between what the brain does when it processes the stimulus passively, and what occurs when the agent employs active causal power (this, I would claim, is supported by K. N. Oschner’s experiments). By contrast, if there is no such thing as agent causal power, then all of the brain’s processing is ultimately passive, and talk of an agent’s decisions is something to be explained or explained away in the terms of ordinary event causation in the brain. So neuroscience is in principle relevant to determining whether it is materialism or dualism which is the most plausible metaphysic. Indeed, quite a few materialist scientists and philosophers have claimed that specific neuroscientific results undermine the idea of libertarian free will. For example, Daniel Wegner argues that while consciousness may precede and preview our actions, these actions are actually produced by unconscious mechanisms. Thus, the efficacy of conscious will is an illusion of apparent causation produced by the fact that both the conscious preview and the action have causes we are unaware of.38 In the next section, we will consider and respond to these arguments. Then, in the last section, we will show a wide range of medical scientific data which support a role for libertarian free will. 4. THE EMPIRICAL CASE AGAINST LIBERTARIAN FREE WILL In a series of well-known experiments, Benjamin Libet challenged the idea that conscious will is the cause of our behaviors.39 In one experiment, subjects were instructed to perform a simple flexion of their fingers any time they felt the (unplanned) urge. By averaging many trials, he was able to show that a readiness potential (RP) preparing the body to move appears about 550 msecs before the movement, but that the conscious wish to perform the movement is not noticed until 350 msecs later. Libet inferred from this that our conscious will does not initiate actions, and so it might seem that they are the result of ordinary, physical event causation as the materialist claims. However, in further experiments, Libet claimed to show that there was a possible, though diminished, role for free will. In these experiments, subjects were instructed to veto the urge to flex when they consciously detected it. He found that subjects were able to veto or allow the action to proceed. The existence of a veto possibility is not in doubt. The subjects in our experiments at times reported that a conscious wish or urge to act appeared but that they suppressed or vetoed that … They were able to exert the veto within the 100 to 200 msec. before the pre-set time to act …. A large RP preceded the veto, signifying that the subject was indeed preparing to act, even though the action was aborted by the subject.40
Libet’s own conclusion from this was that conscious will serves not to initiate, but to allow or veto a movement which has already been prepared unconsciously. In this 38 Daniel Wegner, The Illusion of Conscious Will 97 39 Benjamin Libet summarizes his key experiments and his opinions about their implications for free will in his Do We Have Free Will? in: The Volitional Brain: Towards a Neuroscience of Free Will, eds. Benjamin Libet, Anthony Freeman and Keith Sutherland, Exeter: Imprint Academic, 1999, 47–57. 40 Benjamin Libet, Do We Have Free Will? 52
82
Angus Menuge
understanding, the agent is rather like the President who may either veto or rubber stamp the legislation that comes across his desk. While Libet’s account might salvage some form of free will41, it seriously undermines the libertarian intuition that agents can at least sometimes be credited with originating their actions, transforming the agent from author of his own actions into an editor of behavioral proposals arising from other sources. Further, many commentators on Libet’s work, holding a variety of opinions on the free will issue, have given good reason to think that he misinterpreted the significance of his results. For example, Henrik Walter, himself a compatibilist, argues that ‘Even if libertarian free will is an illusion, Libet’s reasons for that are wrong.’42 The reason is that Libet mistakenly assumes that ‘a conscious intention … must be viewed as a causally effective event in the brain, which exists immediately prior to the action itself ’.43 And Gilberto Gomes likewise notices the distinction between a general intention to act at some time (e.g. the intention to get a haircut this week) and the intention to act now (e.g. the intention to drive to the barber).44 Al Mele (who has described himself as agnostic about compatibilism) makes a similar distinction between distal and proximal intentions.45 Failure to make such a distinction overlooks the rather obvious fact that subjects of Libet’s experiments have previously been instructed either to flex their fingers, or to veto the urge to do so, and so they have already consciously formed an intention to flex or not to flex even before the RP begins. What would we reply if asked what the reason was for the movement of the test person’s finger? The instructions given prior to the experiment! The test persons were given the instructions to behave, such that in the course of the experiment … they should move a hand whenever they wanted. This information was consciously received and generated a conscious goal representation, namely, to execute a (prescribed) movement under certain conditions …. The test persons were to execute a movement when they felt an urge to do so. I suppose the test person certainly (consciously) felt something, but not their intention to act. What did they feel? There is experimental data supporting the idea that the test persons felt an internal trigger … to which they selectively guided their attention …. If this is true, then a conscious intention immediately prior to the movement is of no importance, at most a conscious feeling of having passed a threshold. The test persons waited for the start signal with an already fixed intention of moving a finger.46
The point is that an action can be initiated by a (distal) conscious intention even though that intention is not the proximal cause of the action. Laura Ekstrom eloquently explains the main point in terms directly relevant to Libet’s experiments: 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 41 This is denied by skeptics, who argue that even the conscious veto is the result of a physical process which by itself terminates the motor program initiated by the RP, so that we become conscious of the veto after it has already been implemented by purely physical processes. Libet attempts to refute this skepticism; see Benjamin Libet, Do We Have Free Will? 52–53. 42 Henrik Walter, Neurophilosophy of Free Will, 245 43 Henrik Walter, Neurophilosophy of Free Will, 251 44 Gilberto Gomes, Volition and the Readiness Potential. In: The Volitional Brain: Towards a Neuroscience of Free Will, eds. Benjamin Libet, Anthony Freeman and Keith Sutherland, Exeter: Imprint Academic, 1999, 59–76 45 See, for example, Al Mele’s Free Will and Luck, New York: Oxford University Press, 2006, 31– 32. 46 Henrik Walter, Neurophilosophy of Free Will, 251–252
Does Neuroscience Undermine Retributive Justice?
83
has become appropriate, generates an immediate executive intention, which, in turn, if all goes well and the agent does not change his mind, leads normally to the act.47
As a result, even if the executive (proximal) intention never became conscious, it would be no reason to think the action was not consciously willed by a distal intention. As Timothy O’Connor points out, in Libet’s experiments, almost all of the essential content of a subject’s intention is already solidified before the experiment begins: A subject agrees at the outset to perform a specific action within a short interval of time. All that is left to determine is the precise time of its occurrence … [W]e should not join Libet … in ignoring a very obvious fact here: in agreeing to cooperate with the experiment as described, the agent has already decided to perform a specific action …. No evidence is adduced that there is a slow build up towards a readiness potential before this decision.48
O’ Connor is aware, however, that Libet might reply that his point makes no difference, because the fact remains that awareness of the proximal intention to act now occurs later than the RP that appears to initiate the movement. Libet assumes that onset of the RP closely correlates with the formation of an unconscious intention to act now which later becomes conscious. However, as Al Mele has pointed out, this is a highly questionable assumption, as Libet fails to carefully distinguish between wishes, desires and urges on the one hand and intentions or decisions on the other. This leaves it open that at -550 ms [the onset of RP], rather than acquiring an intention or making a decision of which he is not conscious, the agent instead acquires an urge or desire of which he is not conscious.49
Urges and desires are not the same as intentions, for one can have an urge or desire without any intention of acting on it (e.g. one desires to buy a sports car but does not intend to do so because one believes it is too expensive).50 As Mele says, ‘To intend to do something is, at least in part, to be settled (but not necessarily irrevocably) upon doing it… Merely desiring to do something is compatible with being unsettled about whether to do it.’51 Further, Mele provides a strong reason for thinking that the RP should be associated with an urge and not an intention to act now by considering the implications of those trials on which subjects are instructed to veto their action. Having accepted these instructions, subjects presumably intend not to flex. But the RP arises on these trials as well, so, if Libet is right and this correlates with the intention to flex, the subjects must intend to flex and intend not to flex at the same time, which though logically possible, is psychologically unreasonable (there is no reason to question the rationality of the subjects):
47 Laura Ekstrom, Free Will: A Philosophical Study, Boulder, CO: Westview Press, 2000, 109–110. 48 Timothy O’Connor, Freedom with a Human Face, Midwest Studies in Philosophy, XXIX (2005), 225 49 Alfred R. Mele, Free Will and Luck, 33 50 And one can intend to do something one has very little desire to do (e.g., attend a faculty meeting). 51 Alfred R. Mele, Strength of Motivation and Being in Control: Learning from Libet, American Philosophical Quarterly, Vol. 34, Number 3, July 1997, 319–332, 321
Angus Menuge
84
Keep in mind that the subjects were instructed in advance not to flex their fingers, but to prepare to flex them at the prearranged time and to “veto” this. The subjects intentionally complied with the request. So what is indicated by the RP? Presumably, not the acquisition or presence of an intention to flex; for then, at some point in time, the subjects would have both an intention to flex at the prearranged time and an intention not to flex at that time.52
By contrast, it is psychologically reasonable (in fact common) to have an urge to do something (e.g. make an offensive gesture) at the same time as an intention not to do it (as any civilized person compelled to drive on the New Jersey turnpike can attest). What is more, such an urge may be the result of a prior intention to act sometime in the future, as O’Connor points out: [B]y asking the subjects to carefully introspect to pinpoint the timing of the impulse to move, the experimenter is inviting the subject to adopt the role of observer in relation to his conscious experience, and specifically to wait for an unplanned urge to occur. This certainly encourages a passive posture …. The result of a pre-formed intention to act upon the right internal ‘cue’ for which one looks, I suggest, is that an unconscious process is triggered that promotes the occurrence … of [an eventually] conscious appetitive state [i.e. an urge] that is not actively formed. In context, the state’s default is to trigger the pre-planned activity, absent a last-second ‘veto’ by the agent.53
Putting these points together, what Libet’s experiments actually seem to show is that a distal conscious intention to act sometime in the future may generate an unconscious urge of which one later becomes conscious, at which time the urge can be consciously vetoed. Whether the veto occurs or not, there is no reason to deny that the action is the result of a conscious choice. In addition to citing Libet, Daniel Wegner provides several additional arguments for the conclusion that conscious will is not causally operative in producing actions. On his view, while folk psychology assumes that conscious will, which he defines as the “feeling of doing,” normally precedes actions which we are inclined to call voluntary and also that this feeling does not normally precede inaction, it ignores two classes of abnormal case. In the first class, called automatisms, subjects perform complex behaviors but with no feeling of doing. One example is alien hand syndrome, in which one of a subject’s hands seems to have a will of its own: for example, it may close a cupboard while the subject consciously attempts to open it with his other hand; other examples include automatic writing, behavior under hypnosis and the table-turning of the spiritualists.54 In the second class, called illusion of control, subjects are convinced they are doing something which they are not. Examples include “playing” an automatic replay on a video-game (in which the subject falsely supposes it is his use of the controls that makes items move) and a series of examples using mirrors and other props in which one thinks one is moving one’s own hand, but it is actually someone else’s hand or a rubber hand that is moving. However, as Timothy O’ Connor has argued, neither kind of example supports the idea that conscious will makes no causal contribution. Wegner seems to set up a straw-man by suggesting that ordinary people instinctively believe that all of their voluntary action is preceded by a conscious intention to act now. This is surely false, because we are all aware that large parts of highly practiced, voluntary behaviors, like 52 Alfred R. Mele, Free Will and Luck, 34 53 Timothy O’Connor, Freedom with a Human Face, 226 54 Daniel Wegner, The Illusion of Conscious Will, 4–8
Does Neuroscience Undermine Retributive Justice?
85
digging the garden or opening doors do not require conscious supervision because they have been automated. Even at the common-sense level we talk of ‘going through the motions’ or ‘working like a machine’ or ‘cranking out another one’ without implying that the actions are involuntary. [I]t is widely appreciated that much of what we do voluntarily is not consciously willed at all, but is simply a more or less unconscious outworking of previously settled intentions to act. Only some of the time do we consciously deliberate about what to do … And we easily recognize on reflection that when we deliberate, our self-knowledge is less than complete.55
But from the fact that conscious will is not involved in constantly supervising every aspect of a voluntary behavior, it does not follow that it has no causal role. Like Libet, Wegner fails to distinguish distal or standing intentions to act from proximal intentions to act now. If someone has already consciously decided to drive to the store, we would normally consider that action voluntary even though many parts of it consist of automated behaviors (the agent is an experienced driver, has driven to the store many times). Consciousness enters the scene here only to negotiate those aspects of the drive which are novel e.g. exactly when the traffic lights change, which cars, pedestrians and potholes to avoid, etc. Frequently as well, consciousness is focused on the outside world, and the details of the responses are largely unconscious, unless they are unusually difficult or novel and so require attention. No-one would say, however, that the action of stamping on the brake pedal when a child runs out in the street was involuntary because the driver’s consciousness is attending to the child but not to their own intention to stop the car. It is no part of the ordinary concept of voluntary action that it must be accompanied by an obsessive Pharisaical scrutiny of one’s own intentions.56 In fact, by considering examples like this, we can provide strong evidence that consciousness does have a vital causal role. John Searle cites important work by Wilder Penfield on epileptic behavior.57 Penfield found that in some cases, patients whose epileptic seizure left them completely unconscious were still able to perform complex tasks: they continued to walk, play the piano and even to drive a car! However, what the patients did was to complete learned, automated sequences initiated before the seizure, with no room for creativity or flexibility in light of unforeseen circumstances or new ideas. Thus, the walker would continue to follow a stereotyped route, the unconscious pianist would play a learned piece with no innovation, and the driver, unable to respond to new variables, drove through red lights. Pretty clearly, some feedback information was entering the patients’ cognitive systems (enough to stop them from falling over or slumping) and was being processed at an unconscious level, but the patients’ lack of consciousness still prevented them from adjusting their behaviors contextually. The evidence then is that, inter alia, consciousness plays a causal role in adjusting and shaping otherwise automatic behavior in light of unpredictable factors. But if consciousness has a causal role here, why 55 Timothy O’Connor, Freedom with a Human Face, 224 56 This is a good thing, since highly practiced actions are often performed less well when consciously attended to, e.g. serving in tennis. 57 See John Searle, The Rediscovery of the Mind Cambridge, MA: MIT Press, 1992, 107–108 and Wilder Penfield, The Mystery of the Mind: A Critical Study of Consciousness and the Human Brain, 1975, 39.
86
Angus Menuge
suppose it has nothing to do with the initiation of voluntary actions in the first place? 58 What is unusual about alien hand syndrome is that the behavior continues even when it is opposed by conscious will. It is a case where a neurological defect in the frontal lobe on the opposite side of the hand renders the conscious will unable to control the hand’s behavior. But this does not show that conscious will is unable to control movements in normal circumstances (including the guidance of the other hand). As for the examples of illusion of control cited by Wegner, they provide no evidence that conscious will is causally redundant. When ‘playing’ the video-game replay, for example, consciousness of the location of items on the screen certainly is causally relevant to the agent’s movement of the controls, and in fact, as O’ Connor notes, if the agent experiences his own willing that the items move, there is no reason to say that this experience is illusory, just because his movements don’t cause the items to move. The illusion of control in the video (and other machine-interaction) examples are merely cases of belief, based in external clues, that one’s actions are determining certain outcomes beyond one’s action. Clearly, these cases lend no support to [the] thesis that the agent’s experience of willing his own action is illusory.59
More generally, while Wegner’s scientific results are fascinating, he makes some elementary philosophical mistakes about their implications. He uses the fact that people are sometimes deceived about what their conscious will is causing as evidence that conscious will never plays a causal role. But this does not follow for the simple reason that causes have their characteristic effects only in certain circumstances. The fact that one can sometimes be deceived into thinking that those circumstances obtain when they do not does not show that they never do obtain. For example, a magician can deceive me into thinking that his sawing action is cutting a woman in half because he has made it appear that the normal circumstances for sawing obtain when, happily, they do not. It does not follow that in the normal circumstances the sawing action would not divide the woman. Again, Wegner’s examples do show the logical possibility that we are always deceived in supposing our conscious will is the cause of our behavior. But Descartes’ evil genius hypothesis also shows it is logically possible that we are deceived in thinking that there is an outer world. In neither case does the mere logical possibility of the skeptical scenario provide evidence that it obtains. To be sure, Wegner’s work is important in showing that introspection is a fallible guide to our agency, because, in abnormal cases, our conscious experience of doing may lead us to believe we are doing something that we are not. But this provides no evidence that conscious volitions play no causal role in normal cases where the intended action does occur.
58 From all one can tell, the epileptic patients initiated their behaviors by conscious choice, since the actions started before the seizure rendered the patients unconscious. This means that in a sense their behavior is voluntary, although their failure to respond to contextual clues after the seizure is not, and certainly would provide grounds for some kind of diminished responsibility. 59 Timothy O’Connor, Freedom with a Human Face, 224 (italics in original)
Does Neuroscience Undermine Retributive Justice?
87
Wegner also fails to provide a plausible explanation of the existence of consciousness. If, as he claims, the experience of conscious will as a cause of action is always illusory, then conscious will cannot have been selected for because of its beneficial effects on behavior. The best that Wegner can suggest is that although consciousness merely previews action, it is useful for social cooperation: The ability to know what one will do, and particularly to communicate this to others verbally, would seem to be an important asset, something that promotes far more social interaction than might be the case if we all had no idea of what to expect of ourselves or of anyone around us.60
But if consciousness can causally influence our behavior by providing information we then communicate by vocalizations, then it does have a causal influence. And if consciousness can play a causal role in informing communication, it can surely inform an agent about how to act. This is precisely the conclusion Searle draws from Penfield’s experiments on epileptic patients. Why, then, should we suppose that conscious will has no role in initiating voluntary actions? Wegner’s best answer is Libet’s experiments, but as we have already seen, they do not provide the necessary support. So, if Libet’s experiments, automatisms and illusions of control constitute the best evidence against the efficacy of conscious will, the libertarian has nothing to fear. But does neuroscience provide any positive support for the idea that conscious will has a causal role? I believe that it does.61 5. THE EMPIRICAL CASE
FOR
LIBERTARIAN FREE WILL
I will now offer three main lines of evidence in favor of the idea that libertarian will has active causal power: (1) cognitive therapies for neurological disorders; (2) the placebo effect; and (3) psychoneuroimmunology. What is particularly compelling about the evidence here is that there is a direct correlation between conscious acts of volition and measurable, physical changes in patients’ health. 5.1 Cognitive Therapies for Neurological Disorders As Jeffrey Schwartz and Sharon Begley have shown, enormous advances in neuroscience have made it possible to map the specific neural pathways employed in many cognitive tasks.62 Among the most surprising results from this field has been the discovery of an unsuspected level of neuroplasticity. The orthodox view in neurophysiology had been that only young children can compensate for deficits in the brain by ‘remapping’, the employment of alternate regions and pathways in the 60 Daniel Wegner, The Illusion of Conscious Will, 97 61 Much of the material in section 5 is adapted from Angus Menuge, Is Downward Causation Possible? How the Mind can Make a Physical Difference, Philosophia Christi Volume 11, No. 1, 2009, 93–110, especially 103–110. 62 Jeffrey M. Schwartz and Sharon Begley, The Mind and the Brain: Neuroplasticity and the Power of Mental Force, New York: HarperCollins, 2002
Angus Menuge
88
brain to accomplish the same cognitive task. After maturation, it was assumed that new pathways could no longer be developed and alternate regions of the brain could not be reassigned to compensate for failure elsewhere. The only neuroplasticity left for adults was Hebbian plasticity, the ability to learn by increasing synaptic strength, “known by the maxim ‘Cells that fire together, wire together’”.63 In the background, what drove the orthodox view was the materialist assumption that the mind was the passive product of the brain and environmental conditioning. The idea that the mind, and particularly consciousness, could act back on the brain, was rejected as inherently unscientific. A materialistic view of the mind led psychiatrists to assume that their patients were passive entities who could only be helped by subjecting them to bottom-up causal processes, such as drugs designed to change their brain chemistry and conditioning designed to reinforce alternative behaviors. Both of these approaches do have a limited amount of success, but there is also a significant number of patients who fail to make progress or who refuse to accept these treatments. Jeff Schwartz was among the minority of scientists who wondered if patients could learn alternative behaviors by harnessing the power of their own mind. Schwartz’s own specialization is therapy for Obsessive Compulsive Disorders (OCDs). An OCD is ‘a condition marked by a constant barrage of intrusive thoughts and powerful urges’.64 These urges are ‘ego-dystonic: they seem apart from and at odds with, one’s intrinsic sense of self ’.65 Unlike depressives who may genuinely believe they should harm themselves or retreat from others, patients with OCD experience an urge to do X even when they know there is no good reason to do X. Examples of OCD include: Excessive and ritualized hand-washing…alphabetizing the contents of a pantry, repeatedly checking to see whether a door is locked or an appliance is turned off, checking over and over to see whether you have harmed someone…following rituals to ward off evil … touching or tapping certain objects continuously, being unable to resist counting … or even excessively making lists …. obsessions about order or symmetry … an obsession about hoarding …66
The more the patient yields to these irrational urges, the worse the OCD becomes. Any therapy must therefore find a way to help the patient resist the urges. Following a materialist paradigm, psychiatrists have used a method of behavioral conditioning called ‘exposure and response prevention’. The idea is to expose the patient to what they are reacting to, and then to remove the ability to carry out the obsessive response so that the patient learns to withstand the urges. However, this can be demeaning or even dangerous: patients with hand-washing obsessions are required to touch soiled toilet seats and rub themselves (exposure) and are then deprived of hand-washing facilities (prevention)67; patients who obsessively check the mirror for (non-existent) people they fear they have run over are required to drive (exposure) without a rearview mirror (prevention)!68 Rather than subjecting patients to these 63 64 65 66 67 68
Jeffrey M. Schwartz and Sharon Begley, The Mind and the Brain, 107 Jeff Schwartz, The Mind and the Brain, 1 Jeff Schwartz, The Mind and the Brain, 55 Jeff Schwartz, The Mind and the Brain, 56 Jeff Schwartz, The Mind and the Brain, 3 Jeff Schwartz, The Mind and the Brain, 5
Does Neuroscience Undermine Retributive Justice?
89
ethically questionable treatments, Schwartz wondered if patients could actively choose strategies to weaken the grip of their obsession. He hypothesized that if patients could gain an objective perspective on their own disorder and then consciously refocus on alternative behaviors, they might be able to overcome OCD. From a materialist perspective, this hypothesis looks quite incredible. Indeed, advances in brain imaging techniques (using Positron Emission Tomography or PET scans) showed that OCD patients have a detectable physical abnormality, ‘brain lock’, causing them to repeatedly feel that something is wrong (even though they know there is not) which requires a response, and to be unable to move on to another thought and related behavior: Our PET scans had shown that the orbital frontal cortex, the caudate nucleus, and the thalamus operate in lockstep in the brain of an OCD sufferer. This brain lock in the OCD circuit is undoubtedly the source of a persistent error-detection signal that makes the patient feel that something is dreadfully wrong.69
However, Schwartz’s results provide evidence that purely mental events, such as conscious attention, actually change the physical structure of the OCD circuit. Given the powerful arguments from the philosophy of mind that consciousness does not itself reduce to physical processes in the brain70, this result is not plausibly interpreted as one part of the brain gaining control of another, which could be explained as the result of materialistic, bottom-up processes. Rather, Schwartz’s work is best explained by theorizing that consciousness has a downward causal influence on the brain, that attention can actually reconfigure the brain’s structure. Pursuing this idea, Schwartz developed a 4-step program for OCD patients that requires conscious mental attention. First, following the Eastern idea of ‘mindfulness’, patients must try to distance themselves from their OCD, so that they can see it as an abnormal condition afflicting them, rather than part of who they are. They are then able to ‘Relabel their obsessions and compulsions as false signals, symptoms of a disease’.71 Second, patients try to Reattribute these thoughts as deriving from bad brain circuitry, not an objective need. The third and hardest task, which requires the most conscious attention, is to Refocus the mind on more constructive behavior. Finally, patients try to Revalue their OCD thoughts, seeing them as without significance or power. Each of the four steps, and especially the third, is actively initiated by the conscious mental effort of attention—an act of conscious will or will power, deliberately aimed at the goal of alleviating OCD. Yet, remarkably, using PET scans, Schwartz was able to show that after 10 weeks of therapy, these mental acts resulted in observable changes to the OCD circuit in many of his patients’ brains: PET scans after treatment showed significantly diminished metabolic activity in both the right and left caudate … There was also a significant decrease in the abnormally high, and pathological, correlations among activities in the caudate, the orbital frontal cortex, and the thalamus in
69 Jeff Schwartz, The Mind and the Brain, 85 70 See, for example, David Chalmers, The Conscious Mind, New York: Oxford University Press, 1996. 71 Jeff Schwartz, The Mind and the Brain, 14
Angus Menuge
90
the right hemisphere … [T]herapy had altered the metabolism of the OCD circuit. Our patient’s brain lock had been broken.72
The obvious question is how such dramatic physical change to the brain is possible. Part of the answer is that adults have a much greater degree of neuroplasticity than previously believed. While the orthodox materialist view had held that the adult brain has a fixed map of sensory and motor homunculi in dedicated parts of the brain, neuroscientists Edward Taub, Michael Merzenich and Jon Kaas showed, by experiments on monkeys, that when nerves connecting the cortex to a hand are severed, the cortex rapidly responds to signals from other parts of the hand.73 They also showed that an organism’s use of its limbs dynamically changes its cortical map. If fingers are joined (‘artificial syndactyly’) the map is changed so that only a single digit is represented; if the fingers are separated, the map is redrawn to represent both digits separately. Applying these lessons to humans, it was shown conclusively that many stroke victims can regain use of their affected arm by exploiting cortical reorganization. One of the most effective therapies is constraint induced movement (CI) therapy, in which the good arm is restrained most of the time, and the patient is rewarded for any progress in moving the affected arm. Changes in the motor cortex were observed in as little as two weeks. While it is normal for an arm to be controlled by the motor cortex on the other side of the brain, cortical reorganization through CI therapy allowed these patients to move their arms with the motor cortex on the same side: ‘when the patients moved their affected arm, the motor cortex on the same side crackled with activity’.74 An equally important part of the answer is that successful therapies for various neural deficits and disorders involve patient’s conscious attention, and a conscious effort of will. Patients are not passively conditioned or merely medicated but must consciously try to perform a new behavior. In the case of OCD, the most important step is the consciously willed effort to refocus on a more constructive alternative behavior that enables them to resist the compulsion to respond to a baseless sense of unease. For stroke victims, therapy requires the patient to consciously and deliberately try (will) to perform basic, but now very difficult tasks, such as placing a peg in a hole.75 The same approach is used to cure focal hand dystonia where a pianist loses the ability to move two fingers independently, through the reception of near simultaneous signals from both fingers when playing fast, complex pieces: therapy required “highly attended, repetitive, nonsimultaneous movements.”76 Another successful application is therapy for dyslexia, which is now known to derive from the inability to distinguish spoken phonemes. The therapy involves carefully and consciously attending to artificially elongated phonemes so that they are properly separated.77 Other examples include treatment for Tourette’s syndrome (manifested by stereotypical outbursts or behavioral tics) and depression, which is treated by help-
72 73 74 75 76 77
Jeff Schwartz, The Mind and the Brain, 89–90 See the extended discussion in ch. 5 of Jeff Schwartz, The Mind and the Brain. Jeff Schwartz, The Mind and the Brain, 193 Jeff Schwartz, The Mind and the Brain, 190 Jeff Schwartz, The Mind and the Brain, 219 Jeff Schwartz, The Mind and the Brain, 226–236
Does Neuroscience Undermine Retributive Justice?
91
ing patients to attend to thoughts that trigger a cascade of negative associations and to refocus on more positive alternatives. It is important to see that none of these therapies can be explained by the model of passive behavioral reconditioning. Consciously willed attention is vital because it actively changes how the brain processes information. As Schwartz and Begley explain, we are normally bombarded by a large number of parallel stimuli, competing for attention, which has the effect that the stimuli mutually suppress each other. While some stimuli are naturally stronger than others, functional magnetic resonance imaging (fMRI) shows that the observer’s interests make a large difference to subsequent processing. When we look for a specific target, such as a particular profile in a crowd, or one voice in the babble of a cocktail party, ‘neurons that respond to [the] target (the image attracting your attention) fire more strongly than neurons that respond to a distraction’.78 Attention biases the brain so that the suppression of target stimuli by competing distracters is reduced. While the information impinging our cortex is not under our control, the brain’s response to that information is selective and depends on conscious attention. ‘An activity usually determined to be a property of the mind—paying attention—determines the activity of the brain.’79 If it is a fact that consciously willed selective attention changes how the brain processes information, and deliberate, conscious focusing of the mind on therapeutic behaviors can cause cortical reorganization, it is reasonable to conclude that conscious will has a downward causal influence on the brain (and hence on behavior). Schwartz’s conclusion from his work with OCD is that the mind is not an impotent shadow thrown up by the brain. While the mind is obviously strongly influenced by brain events in a bottom-up fashion, as is evident from the mental effects of brain deficits, medication and behavioral conditioning, the scientific evidence shows that our conscious will also has the power to reconfigure the brain. [W]illful, mindful effort can alter brain function, and … such self-directed brain changes—neuroplasticity—are a genuine reality … In other words, the arrow of causation relating brain and mind must be bidirectional.80
Mario Beauregard has offered additional support to Schwartz’s claim. In an important article surveying the use of cognitive (mind-based) therapies to alter brain function, Beauregard summarizes his conclusions as follows: The results of these [neuroimaging] studies strongly supports the view that the subjective nature and intentional content…of mental processes (e.g. thoughts, feelings, beliefs, volition) significantly influence the functioning and plasticity of the brain…mentalistic variables have to be seriously taken into account to reach a correct understanding of the neurophysiological bases of behavior in humans.81
In addition to the examples of cognitive therapies cited by Schwartz, Beauregard lists the examples of downward suppression of sexual arousal and negative emotions (such as fear and sadness), and treatments for panic disorder, unipolar major depressive disorder, social phobia, spider phobia and the placebo effect.
78 79 80 81
Jeff Schwartz, The Mind and the Brain, 328 Jeff Schwartz, The Mind and the Brain, 329 Jeff Schwartz, The Mind and the Brain, 94–95 Mario Beauregard (note 36) 219
Angus Menuge
92 5.2 The Placebo Effect
The placebo effect provides particularly powerful evidence of downward mental causation. By definition, a placebo is ‘any treatment – including drugs, surgery, psychotherapy and quack therapy – used for its ameliorative effect on a symptom or disease but that is actually [physically] ineffective or not specifically effective for the condition being treated’.82 The placebo therefore has no physical power to heal the patient, and yet patients who trust in the placebo have a statistically greater chance of improvement. Doctors have long been aware of the power of the placebo effect: The placebo effect depends on a patient’s trust in the physician. I’ve become convinced that this relationship is more important, in the long run, than any medicine or procedure. Psychiatrist Jerome Frank of Johns Hopkins University found evidence for this belief in a study of ninetyeight patients who had surgery for detached retinas. Frank assessed the subjects’ independence, optimism, and faith in their doctors before the operations, and found that those with a high level of trust healed faster than the others.83
The placebo’s power evidently derives not from its physical properties, but from the patient’s mental attitudes of trust and hope. Initially, this could only be supported indirectly by surveying patient outcomes. For example, Norman Cousins cites the work of Drs. Sheldon Greenfield and Sherrie Kaplan of the UCLA School of Public Health on ulcer diseases, hypertension, diabetes and breast cancer: Drs. Greenfield and Kaplan found that increased patient control, more expression of affect by doctor and patient, and greater information provided by the doctor in response to patient questions, were related to better patient health status as measured by audiotapes of office visits, questionnaires, and physiological measurements.84
More recently, however, improved brain scanning techniques have allowed a more direct scientific study of the effects of psychological attitudes on health. Most remarkably of all, there is evidence to show that the placebo effect was at least as effective as the drug apomorphine in treating the chronic underproduction of dopamine in patients with Parkinson’s Disease: The magnitude of the placebo response was comparable to that of the apomorphine … These results constitute … evidence for considerable release of endogenous dopamine in the striatum of PD patients in response to placebo… Garris et al. … have provided evidence that it is the expectation of reward that elicits dopamine release.85
5.3 Psychoneuroimmunology More generally, there is an emerging field of psychoneuroimmunology, which studies how mental attitudes affect the immune system via the brain. For example, it has 82 A. K. and E. Shapiro, The Powerful Placebo: From Ancient Priest to Modern Physician, Baltimore: MD: Johns Hopkins University, 1997, cited in Mario Beauregard (note 36) 227. 83 Bernie S. Siegel, M.D., Love, Medicine and Miracles: Lessons Learned About Self-Healing From a Surgeon’s Experience with Exceptional Patients, New York: Harper & Row Publishers, 1986, 37 84 Norman Cousins, Head First: The Biology of Hope, New York: E. P. Dutton, 1989, 234 85 Mario Beauregard (note 36) 227–228
Does Neuroscience Undermine Retributive Justice?
93
long been known that psychological stress is correlated with negative effects on physical health. One successful therapy for stress related to chronic illness uses an approach called ‘Mindfulness Based Stress Reduction’ (MBSR) derived from Eastern meditation. Like Schwartz’s approach to OCD, this approach also relies on consciously willed attention. The University of Massachusetts Medical School describes the approach as follows: Mindfulness is a way of learning to relate directly to whatever is happening in your life, a way of taking charge of your life, a way of doing something for yourself that no one else can do for you — consciously and systematically working with your own stress, pain, illness, and the challenges and demands of everyday life. In contrast, you’ve probably encountered moments of “mindlessness” — a loss of awareness resulting in forgetfulness, separation from self, and a sense of living mechanically. Restoring within yourself a balanced sense of health and well being requires increased awareness of all aspects of self, including body and mind, heart and soul.86
More specifically, as recounted by Mary Jane Ott et. al., a 2004 study explored the effect of MBSR on cancer patients who are hospitalized for a long time with stem cell / autologous bone marrow transplants, and found ‘a statistically significant decrease in pain … and increases in the levels of relaxation … happiness … comfort … reduced heart rate … and respiratory rate’, while other studies have shown benefits from MBSR in ‘decreasing anxiety, depression, anger, demoralization, and symptoms of somatic fatigue in male and female cancer patients’.87 Cognitive therapies for neurological disorders, the placebo effect and psychoneuroimmunology provide strong evidence for the causal role of conscious will (and more generally, for downward mental causation) in promoting the health of patients. Conclusion Scientific materialists deny the possibility of libertarian free will because they think it is incompatible with the causal closure of the physical, it would violate physical laws, and that it is (or can be) empirically refuted by neuroscience. We have seen that none of these arguments survives scrutiny. We have excellent philosophical and scientific reasons for thinking that libertarian free will is a reality and therefore that (at least sometimes88), criminals deserve punishment. If patients with debilitating neurological and other physically based conditions can choose to work against their afflictions, healthy murderers certainly deserve to be punished.
86 University of Massachusetts Medical School, Stress Reduction Program, available at: http:// www.umassmed.edu/Content.aspx?id=41254. 87 Mary Jane Ott, Rebecca L. Norris and Susan M. Bauer-Wu, Mindfulness Meditation for Oncology Patients: A Discussion and Critical Review, Integrative Cancer Therapies 5, no. 2 (2006), 106 88 There is a huge literature on cases of diminished responsibility, and nothing I say here implies the simplistic idea that all criminals are equally responsible for their actions. Excellent discussions of diminished responsibility in cases of defects of will caused by addiction are provided by R. Jay Wallace, Addiction as Defect of the Will: Some Philosophical Reflections, Law and Philosophy 18 (1999), 621–654 and Stephen J. Morse, Hooked on Hype: Addiction and Responsibility, Law and Philosophy 19 (2000), 3–49.
JOHN WARWICK MONTGOMERY, STRASBOURG1 SOME REMARKS ON PUNISHMENT AND FREEWILL AND CLASSICAL CHRISTIAN THEOLOGY
IN
LEGAL THEORY
Can – and should – societal punishment operate in the absence of freewill on the criminal’s part? Should punishment exist only if rehabilitation can be achieved? In this paper, we contend (1) that Christian theology answers these questions in the negative, and (2) that a proper jurisprudence does likewise. THE CRIMINOLOGICAL SCENE Utilitarian theories of punishment, so popular in the latter half of the 20th century, do not rely for their justification on the freewill of the criminal. Just as social philosophies such as Marxism and Environmentalism believe that altering the physical or natural climate will change human behaviour, rehabilitative theories of punishment maintain that an enlightened punitive system can per se lead to positive change in the criminal. Retributive theories, however, are based squarely on the reality of freewill. Classically, Immanuel Kant argued: ‘Juridical punishment (poena forensis) … can never be administered merely as a means for promoting another Good either with regard to the Criminal himself or to Civil Society, but must in all cases be imposed only because the individual has committed a Crime. … The Penal Law is a Categorical Imperative; and woe to him who creeps through the serpent-windings of Utilitarianism to discover some advantage that may discharge him from the Justice of Punishment, or even from the due measure of it, according to the Pharisaic maxim: ‘It is better than one man should die than that the whole people should perish.’2
Fleischacker comments: ‘Retributive punishment serves a moral function for Kant by making the criminal live under the law he implicitly sets up in his criminal act. The criminal acts on a maxim that he would not will as a universal law; we apply the law of that maxim to him, as thought he had willed it universally. … We are merely following out the rational interpretation of his irrational act, and he should have no reason to complain.’3
Even though there are serious problems with the logic of Kant’s categorical imperative,4 a steady movement away from utilitarian to retributive approaches to punishment can be observed in contemporary criminology. Easton and Piper describe this shift in the following terms:
1
2 3 4
Ph.D. (Chicago), Th.D. (Strasbourg, France), LL.D. (Cardiff, Wales). Distingnished Research Professor of Philosophy and Christian Thought, Patrick Henry College; barrister-at-law, England and Wales; avocat à la cour, barreau de Paris. Immanuel Kant, The Philosophy of Law: An Exposition of the Fundamental Principles of Jurisprudence As the Science of Right [The Metaphysics of Morals, Pt. II, 6:331], trans. William Hastie, 1887 Samuel Fleischacker, Kant’s Theory of Punishment, 79 / 4 Kant-Studien (1988), 442 John Warwick Montgomery, Tractatus Logico-Theologicus, 2002, sec. 5.5–5.6
96
John Montgomery ‘In the UK and the USA criticism [of the utilitarian approach] focused on the ‘inequities’ and ineffectiveness of rehabilitation, and on wide judicial discretion. … The leading voice for modern retributivist theory … was von Hirsch who argued that fairness and justice should be the key elements of a coherent penal theory. In Doing Justice (1976), he maintained that the aim of the penal system should, then, be to ‘do justice’ rather than to maximise utility. In other words he construed justice—in line with classical retributivism—as giving offenders punishments in proportion to their crimes and, in doing so, recognising them as moral agents possessing autonomy.’5
Necessarily, a retributivist view of punishment requires as its justification the belief that the criminal is an autonomous entity, possessing a freewill that he or she has employed in a socially deleterious manner. Retributivism thus presupposes that the criminal has knowingly or recklessly committed a serious fault and thus deserves an appropriate and proportional punishment. In point of fact, when one analyzes the modern penal law of any civilised nation, the retributive basis of the legislation is clearly seen.6 Take, as a single but typical example, the French criminal law of intentional harm to others (‘wilful attacks on the integrity of the person’), as set out in Articles 222–1 through 222–5 of the new (1994) Code pénal: Art. 222–1 Subjecting a person to torture or barbarous acts is punishable by fifteen years of imprisonment. Art. 222–2 The offense in Article 222–1 is punishable by imprisonment for life when it precedes, accompanies, or follows a felony other than murder or rape. [Murder and rape carry their own severe penalties, set out elsewhere in the Code pénal.] Art. 222–3 The offense in Article 222–1 is punishable by twenty years imprisonment when it is committed: 1. On a minor less than fifteen years old; 2. On a person whose special vulnerability, due to age, sickness, infirmity, physical or mental deficiency, or pregnancy, is apparent or known to the perpetrator; 3. On an ascendant, either legitimate or natural, or on a father or mother by adoption; 4. When the status of the victim is apparent or known to the perpetrator, on a magistrate, juror, lawyer, public or ministerial officer, officer of the gendarmerie, agent of the national police force, customs official, prison administration official, or any other person exercising governmental authority or entrusted with a mission of public service, in the performance or on the occasion of performing his or her duties or mission; 5. On a witness, victim, or civil party [in a legal action]; 5b. On a victim who is thought to belong, or not to belong, to a given race, nationality, ethnic group, or religion—whether he or she does in fact so belong; 5 6
Susan Easton and Christine Piper, Sentencing and Punishment: The Quest for Justice, 2d ed. 2008, 63; italics ours. It should not be necessary to point out that a retributivist philosophy does not necessarily entail acceptance of capital punishment. ‘Just deserts’ must be determined in a proportionate manner, taking into account all factors relative to the given case and the perpetrator. Cf. John Warwick Montgomery, Capital Punishment, in his Christ As Centre and Circumference [forthcoming]).
Some Remarks on Punishment and Freewill
97
6. By the spouse or concubine of the victim; 7. By a person exercising governmental authority or entrusted with a mission in the public service, in the performance or on the occasion of performing his or her duties or mission; 8. By several persons acting as perpetrators or accessories; 9. With premeditation; 10. By using or threatening to use a weapon. Art. 222–4 The offense in Article 222–1 is punishable by thirty years imprisonment when it is committed habitually on a minor less than fifteen years old or on a person of special vulnerability. Art. 222–5 The offense in Article 222–1 is punishable by thirty years imprisonment when it results in a mutilation or permanent infirmity. It will be observed that an effort is made here to relate penalties directly to the actual harm caused, i.e., to give the perpetrator a sentence proportionately reflecting the seriousness of his or her volitional act. Granted, the judge may as a general rule reduce a given sentence in light of mitigating circumstances, but even that discretion is being continually circumscribed—as in the 10 August 2007 ‘Peines-plancher’ law, incorporated into the Code pénal as Articles 132–18–1 and 132–19–1, which largely eliminates sentence reductions and mandates jail time for repeat offenders. Apart from the assumption of freewill exercisable and exercised by the criminal, such penalties would be meaningless at best and immoral at worst. Indeed, the Code pénal expressly declares as a principle underlying all of its provisions: “A person is criminally responsible only for his or her own conduct” (Article 121–1). A THEOLOGICAL PERSPECTIVE What is the biblical view of punishment and freewill and the connections between them? Holy Scripture—the formal basis of all Christian theology—presents God’s human creation as morally responsible and subject to punishment for violations of the Creator’s revealed will. From the fall of mankind’s first parents in the Garden of Eden to the casting of Satan into the Lake of Fire at the end of time, God holds his creatures responsible for their acts. Jesus weeps over Jerusalem: ‘O Jerusalem, Jerusalem, thou that killest the prophets, and stonest them which are sent unto thee, how often would I have gathered thy children together, even as a hen gathereth her chickens under her wings, and ye would not! Behold, your house is left unto you desolate’ (Matthew 23:37–38; italics ours). The Apostle Paul begins his Epistle to the Romans with a sad description of the fallen human race: Gentile peoples volitionally chose to violate the moral law written on their hearts, committing idolatry and engaging in unnatural practices such as homosexuality, and the Jews volitionally broke the revealed law given to them by God in the Old Testament (Romans 1–2); in sum, ‘All have sinned, and come short of the glory of God’ (Romans 3:23). The consequence follows inexorably: ‘Be not deceived; God is not mocked: for whatsoever a man soweth, that shall he also reap’ (Galatians 6:7).
98
John Montgomery
Throughout the Bible, human freewill and moral choice are asserted, and the violation of God’s will leads inevitably to proportionate punishment—if not in this life then at the Last Assize when all evils will be judged and all wrongs righted.7 To be sure, this theology has met with strong objection ever since the rise of modern secularism during the 18th century Enlightenment. ‘An eye for an eye and a tooth for a tooth’ has been condemned as barbaric, and 20th century liberal theological ethicists such as Joseph Fletcher have attempted to replace the alleged ‘prescriptive legalism’ of biblical revelation with forms of ‘situation ethics’ which rely not on principle but on existential decision-making and ill-defined notions of ‘love’.8 But such efforts have devolved into relativism and subjectivism, leaving Christian believers with no ethical moorings in the face of more and more agonising moral dilemmas (stem cell research on embryos, gun control, capital punishment, etc.). And serious philosophical and jurisprudential defences of ‘eye for an eye’ retributive punishment have come on the scene. Thus University of Michigan law professor William Ian Miller argues: ‘The deuteronomic talion adds the notion of ‘teaching a lesson’ to the notion of ‘getting even’ that characterizes the formulations in Exodus and Leviticus, just as in our own speech we will often find both idioms—getting even and teaching a lesson—to be equally appropriate to explain the ministering of justice.’9
But are there not serious theological objections to position here described? Let us briefly consider three such problem areas. Firstly, does not the transmission of original sin from Adam to his descendants rule out the principle of personal responsibility and therefore the legitimacy of individualised punishment? True, according to clear biblical teaching, the sin of Adam passed to all his descendents (Romans 5), but this simply means, as Augustine put it in his phrase, non posse non peccari, that human beings in this fallen world never reach perfection (1 John 1:8); it does not mean that one is forced by one’s humanity to commit any particular sin. If one does choose to commit a sin or do an illegal act, one’s personal responsibility for it remains. Secondly, did not Jesus replace the Old Testament lex talionis by a new, loving, constructive, forgiving approach to punishment? Did he not say to the judgmental crowd ready to stone an adulteress, ‘He who is without sin, cast the first stone’ and to the adulteress, ‘Go and sin no more’(John 8)? But the fact that Jesus condemns mob justice and offers a new way of life to a fallen woman does not in any way suggest that he is discarding the Old Testament law or its standards of justice. He plainly stated in the Sermon on the Mount: ‘Think not that I am come to destroy the law, or the prophets: I am not come to destroy, but to fulfil. For verily I say unto you, Till heaven and 7
8 9
Indeed, the source of the legal concept of mens rea is a sermon by St Augustine. ‘Coke, Third Inst. 6, gives ‘Et actus non facit reum nisi mens sit rea.’ Coke knew the Red Book of the Exchequer which contains the Leges Henrici where the maxim stands ‘Reum non facit nisi mens rea.’ The original source is S. Augustinus, Sermones, No. 180, c. 2, Patrol. vol. 38, col. 974): ‘Ream linguam non facit nisi mens rea.’ This passes into the Decretum, c. 3, C. 22, qu. 2. The author of the Leges took it from some intermediate book”: Sir Frederick Pollock and Frederic William Maitland, The History of English Law Before the Time of Edward I, 2 vols., 2d ed. 1959, II, 476. Cf. Joseph Fletcher and John Warwick Montgomery, Situation Ethics: Is It Sometimes Right to Do Wrong? A Debat, 2d ed., 1999. William Ian Miller, Eye for an Eye, 2006, 68
Some Remarks on Punishment and Freewill
99
earth pass, one jot or one tittle shall in no wise pass from the law, till all be fulfilled. Whosoever therefore shall break one of these least commandments, and shall teach men so, he shall be called the least in the kingdom of heaven: but whosoever shall do and teach them, the same shall be called great in the kingdom of heaven. For I say unto you, That except your righteousness shall exceed the righteousness of the scribes and Pharisees, ye shall in no case enter into the kingdom of heaven’ (Matthew 5:17–20). The point is well made by the late American theologian Carl F. H. Henry: ‘The specific references make it apparent that Jesus is not changing the Law, but rather unveiling its inner requirements. The prohibition against murder ([5:]21ff.) and adultery ([5:]27ff.) apply to the life of thought as well as of deed; the moral obligation they impose is spiritual, and not merely external. Jesus does not set forth a higher law of his own to discredit the Old Testament law, but declares that the requirement of Old Testament law was more exacting than the current tradition taught.’10
Thirdly, does not the atoning death of Christ for the sins of the world unjustly shift the subject of punishment from the deserving sinner to a sinless victim—thus countermanding the principle of proper retribution? In biblical perspective, one here encounters Grace as the fulfilment of the Law: the Creator God, in his infinite mercy, comes down from heaven and takes the sins of the fallen world on himself, expiating them and saving all those who do not reject his gift. Anselm, in Cur Deus Homo?, persuasively argued that this was in fact not an abrogation but a cosmic illustration of proper retribution, Christ being both God and man: as a human being, he could represent the entire race (as Adam had done), whilst, as God, he had the capacity to cancel out the penalty of sin for all mankind through his sacrificial death – thereby fulfilling the legal condition that ‘without shedding of blood is no remission’ (Hebrews 9:22).11 Personal responsibility and freewill remain, for the effectiveness of redemption for the individual depends on his or her not rejecting God’s grace: ‘Without faith it is impossible to please him [God]: for he that cometh to God must believe that he is, and that he is a rewarder of them that diligently seek him’ (Hebrews 11:6). These difficulties bring us quite naturally to a more profound comment relating to the theology of punishment—specifically, a word about the relationship between Law and Gospel. The lex talionis functions as an aspect of the Schöpfungsordnungen, or ‘Orders of Creation’ imbedded in our world by God to permit human survival in our fallen state of radical self-centredness.12 The politico-legal order requires proportionate, retributive justice; otherwise, pragmatism and naked power prevail. But on the model of God’s redemptive order, there is place for mercy and forgiveness. This is the foundation for Equity in the Anglo-American common law tradition, and the basis in the criminal law in all civilised nations for the employment of mitigation in sentencing, judicial discretion, amnesty, and the fitting of the punishment to the condition of the offender (individualisation of penalty).13 10 Carl F. H. Henry, Christian Personal Ethics, 1957, 307 11 Cf. John Warwick Montgomery, Chytraeus on Sacrifice, 2d ed. 2000, especially 139–46; also, Milton S. Terry, The Mediation of Jesus Christ, 1903, passim. 12 See especially, Werner Elert, The Christian Ethos, trans. Carl J. Schindler, 1957, 101 ff. 13 In marked contrast, we have the recent example of the application of Shari’a law in Saudi Arabia: A 75-year-old widow has been condemned to forty lashes and four months in prison followed by expulsion from the country for having allowed two young men not of her immediate
John Montgomery
100
Central to all this, however, is the relationship of the utilitarian / rehabilitative factor to the retributive. It is the biblical view – and, in our judgment, the necessary jurisprudential approach – to make retribution the foundation and rehabilitation the second-, not the first-, storey of the punitive structure:
Rehabilitation
Retribution
(applied where possible)
(fundamental)
The chief reasons why one must not make utilitarian / rehabilitative considerations primary in a justice system are: (1) Without clear evidence that justice is done through ‘making the punishment fit the crime’, the society loses its moral foundation and there will inevitably be more and more creative attempts to circumvent the law through utilitarian techniques.14 (2) Rehabilitative theories invariably reduce the level of personal responsibility for wrongdoing by deemphasising the importance of freewill in the performance of criminal acts. (3) Rehabilitation simply does not work in the majority of cases, the root cause being that the self-centredness of the criminal can only be changed by a radical, spiritual conversion. Only the gospel of the grace of God in Jesus Christ—not any human system of punishment—has proven capable of achieving this.
family to visit her in her home (they were doing shopping for her). A local lawyer offered as justification of the penalty that the Shari’a is the Shari’a and even though a women of 75 years of age is ‘not normally regarded as seductive, nevertheless age is not a sufficient condition for acquittal’ (Figaro, 18 March 2009). 14 It will be noted that our position here agrees in its essentials with that of C. S. Lewis, as presented in his essay, ‘The Humanitarian Theory of Punishment’ and ‘On Punishment: A Reply to Criticism’, included in Lewis’s God in the Dock, ed. Walter Hooper, 1970, 287–300. Concludes Lewis: ‘All I plead for is the prior condition of ill desert; loss of liberty justified on retributive grounds before we begin considering the other factors.’
FRIEDRICH TOEPEL, BONN ‘DEEP RESPONSIBILITY’ IN CRIMINAL LAW
AND
I. PUNISHMENT
OF
AS A
FORM
ALTERNATIVE POSSIBILITIES
ADDRESSING PERSONS
In this paper I shall address the issue of whether alternative possibilities are necessary for imputing responsibility to an agent in a ‘deep’1 sense and whether a compatibilist account of free will allows for such responsibility. I shall come to the conclusion that alternative possibilities are indeed necessary for imputing responsibility to an agent in the ‘deep’ sense needed for moral or legal imputation. I shall also defend the position that a compatibilist account of free will cannot accommodate such responsibility within its system. Why would alternative possibilities become relevant at all in criminal law or moral theory? I accuse someone that he has brought about a state of affairs which is wrong according to legal or moral standards, and he answers me he couldn’t help bringing it about. Is it unjust if I impose a sanction on the perpetrator for bringing about the state of affairs he could not avoid? The theories of punishment answer this question differently: The theory of negative general prevention or deterrence2 could point at least to a positive effect of a sanction in case of lacking alternative possibilities because the sanction could function as a threat for citizens concerning possible future violations of the law. The message could be understood by the citizens as: ‘Do your utmost to avoid bringing about the state of affairs we do not wish to see!’ This could even be the case if the perpetrator would not understand why the sanction was imposed in his particular situation because he had no possibility to avoid bringing about that state of affairs. Nor would we receive a satisfactory answer from a theory of special prevention3 why we should not impose a sanction in case of lacking alternative possibilities. We might ask whether the perpetrator does not have to be able to understand what he has done in order to be educated. However, a theory of special prevention may take the view that the criminal is in any way unable to understand what he has done. His education is only a means to enable him to avoid future violations of the law. It is still possible to instruct someone how to avoid future violations of the law, even if the cause for the treatment was a behaviour which could not have been avoided.
1
2 3
Just to avoid any doubts about what I mean when using the term deep responsibility, let me state that I in accordance with the common usage in philosophy want to understand deep responsibility as the responsibility traditionally needed for moral or legal imputation, cf. for example the interview between Galen Strawson and Tamler Sommers, http://www.naturalism. org/strawson_interview.htm. Cf. Paul J. A. Feuerbach, Lehrbuch des gemeinen in Deutschland geltenden peinlichen Rechts, 14th ed. 1847, 38. Cf. Franz v. Liszt, Der Zweckgedanke im Strafrecht, ZStW 3 (1885), 1–47.
Friedrich Toepel
102
Even positive general prevention, the preventive theory developed most recently,4 is not capable to provide a satisfactory answer why someone lacking alternative possibilities should not be punished. Positive general prevention means that the public is informed about the law by imposing the sanction. Thus, trust in the law is strengthened, and the sanction prevents that the violation is learned as acceptable behaviour. Positive general prevention only thinly veils the threat dominating the negative preventive theory: The best learning effect is obtained by a sanction that is as psychologically disagreeable for the learner as possible. The decisive clue why we need alternative possibilities to impute responsibility is provided by the retributive theory of punishment.5 Retribution here means that a legally wrong behaviour is sanctioned to ensure that the perpetrator receives exactly the sanction which is provided in a particular society for the wrong committed and which shows how serious the wrong is estimated to be in that society.6 The only reason why the sanction must be exactly the one provided in this society, is that the society treats its citizens as persons. A person is someone treated as a holder of rights and must be enabled to exercise these rights. Therefore, the law not only treats citizens as persons, but also addresses them as persons. In such a society a punishment cannot be just if the wrongdoer cannot understand it as a just punishment. And he cannot understand it as just if he was not able to avoid the sanction. If the wrongdoer was not able to avoid the sanction, possible future delinquents may still be deterred or educated, but only by using the punished wrongdoer as a means, not as an end in the Kantian sense.7 II. TWO KINDS
OF
ABILITY
AS
REQUIREMENTS
FOR
RESPONSIBILITY
The possibility for the criminal to avoid the sanction under the retributive theory therefore is a necessary presupposition for imposing a just sanction. The sanction can be avoided if the criminal is capable to act in accordance with the law. He is capable to act in accordance with the law if he has 1. the technical ability to act in accordance with the law (Handlungsfähigkeit)8 and 2. the ability to motivate himself to act in accordance with the law (Schuldfähigkeit).9 Both kinds of ability presuppose alternative possibilities: If you wilfully decide to cause an accident by turning your car to the right, in my opinion the accident cannot be imputed to you if – maybe unbeknownst to you – the steering apparatus 4 5 6
7 8 9
Günther Jakobs, Strafrecht Allgemeiner Teil, Die Grundlagen und die Zurechnungslehre, 2nd ed. 1991, 1/4 Georg W. F. Hegel, Grundlinien der Philosophie des Rechts, 1821, § 99; also Günther Jakobs, Norm, Person, Gesellschaft, Vorüberlegungen zu einer Rechtsphilosophie, 3rd ed. 2008, 112 Retribution should not be confused with the idea of retaliation that a wrong must be requited with a sanction which must be of the same kind as the wrong like “eye for eye, tooth for tooth”. I define retribution in the present context as the idea that the sanction imposed for a crime must mirror exactly the negative value which the legal system attributes to that crime. Immanuel Kant, Metaphysik der Sitten (Akademieausgabe) vol. 4, 1968, 453 (§ 49, Allgemeine Anmerkung E) Urs Kindhäuser, Gefährdung als Straftat, 1989, 46 Urs Kindhäuser (note 8) 46; cf. also Harry Frankfurt, Three Concepts of Free Action: II. In Moral Responsibility, ed. John M. Fischer, 1986, 113–116.
Free Will and Alternative Possibilities
103
is broken and your car would have gone to the right in exactly the same way even if you had tried to guide the car in any other direction. In this respect you lacked the technical ability to act in accordance with the law. The technical ability is the capability to choose the means to achieve a state of affairs compatible with the ideal world envisaged by the law. Further, you need the motivational ability: If you are under the influence of a psychosis and the steering apparatus of your car is intact, you may still know how to avoid an accident but you may not be able to motivate yourself to act in accordance with the law, for example because you are driven by fear under the hallucination that the other car which you hit with yours is driven by a deadly enemy whom you must fight. In this case you lack the motivational ability to act in accordance with the law. The motivational ability is a higher-order ability because it consists in a motivational capability to make use of the (first-order) technical ability to act in accordance with the law. The sort of power or control we need for the motivational ability insofar involves a radically free will. I am of the opinion that genuine or deep responsibility necessary for imposing a just sanction demands both kinds of ability. This is a controversial issue. In particular, the necessity of a motivational ability for the imputation of a moral or legal wrong has been denied by philosophers and criminal law scholars. The motivational ability is not well-liked by some philosophers because it can only exist in a strict sense if our choices are not fully predetermined. The majority of modern philosophers seeks to avoid the requirement of a motivational ability for the reason that these philosophers are compatibilists.10 Compatibilists think that a plausible theory of moral or legal imputation must be compatible with the hypothesis that the world may not wholly be predetermined (at least partial indeterminism) as well as with the hypothesis that the world may be fully necessitated by the past together with the natural laws (causal determinism). A motivational ability in the sense described above is not compatible with causal determinism and therefore is rejected by compatibilists. In the following sections I shall try to show that compatibilists are not able to integrate the motivational ability needed for moral or legal responsibility consistently into their concepts. I think it is important to show this because the majority of German criminal law scholars endorses a kind of compatibilism.11 They are of the opinion that they may bypass the question whether free will exists because the traditional concepts of responsibility and guilt would be valid irrespective of the fact whether causal determinism would be true or not. III. THE CONSEQUENCE ARGUMENT The decisive objection against compatibilism is that allowing determinism to be a realistic option compels compatibilists to make presuppositions under the presump10 See only Richard Double, The Non-reality of Free Will, 1991, 48; John M. Fischer, The Metaphysics of Free Will, 1994, ch. 8. 11 See for example Günther Jakobs (note 4) 1 / 17, Claus Roxin, Strafrecht Allgemeiner Teil, vol I, 4th ed. 2006, 19 / 36–39.
Friedrich Toepel
104
tion of determinism which can only be true in an indeterministic world. To show this is the objective of the so-called Consequence Argument:12 Compatibilists are of the opinion that causal determinism does not exclude a sort of free choice sufficient to impute a decision to someone. However, supposing that causal determinism is indeed true, a choice (to commit a crime, for example) is entailed by true propositions about the past and the laws of nature. If I were free to choose otherwise, I therefore must have been able to change the past or the natural laws, too, for in a causally deterministic world my choice flows from that past and those natural laws. Yet, in a causally deterministic world, I can change neither the past nor natural laws. The deterministic flow of events is an unbroken chain. Even if I would be able to change the past (making use of a sort of backward causation via an odd time machine), the decision to change the past would have to be free and not determined. Otherwise, my ability to change the past would not show anything. The past would prompt my decision to change the past which again would cause the past to be as it is. I would not really have changed the past, because I could not have chosen to choose otherwise than I did. Thus, I cannot be able to choose otherwise if causal determinism is true. In my opinion, the Consequence Argument is flawless. IV. FRANKFURT-TYPE EXAMPLES
AND THE
KANE / WIDERKER OBJECTION
Compatibilists who accept the soundness of the Consequence Argument13 try to plead for the possibility of a free choice without alternative possibilities. The bestknown way to point out such possibility of a free choice without alternatives is the description of so-called Frankfurt-type cases, named after Harry Frankfurt who first invented them.14 I wish to deal with the following prototype of these examples:15 A contemplates to shoot a politician at a meeting but postpones a definitive decision until the last moment at the meeting itself. Thanks to recent advances in science, a neurosurgeon B has implanted a chip in A’s brain which allows B to monitor A’s thoughts and to influence his decision-making process. B observes A 12 Cf. John Martin Fischer, Compatibilism. In Four Views on Free Will, J. M. Fischer, R. Kane, D. Pereboom, M. Vargas, 2007, 53–56; Peter van Inwagen, The Incompatibility of Free Will and Determinism. In Free Will, ed. Gary Watson, 1982, 96–110. The Consequence Argument does not refute philosophers who take the view that only the attitude of people evaluating decisions is decisive. It is sufficient for these philosophers that people live under the illusion of free will. Sometimes such philosophers are also called compatibilists, like Peter F. Strawson, Freedom and Resentment, Proceedings of the British Academy, xlviii (1962), 1–25, and after him Saul Smilansky, Freedom and Illusion, 2000, passim. However, I shall not examine those theories here. I think that they face a different kind of difficulty, namely that the practice does only function as long as the fact has not become known that free will is only an illusion. Lawyers could not ignore conclusive evidence from the natural sciences, for example, that there is no free will. Otherwise they risk not to be accepted as rational anymore by people who know better. 13 For example John M. Fischer (note 12) 56 14 Harry Frankfurt, Alternate Possibilities and Moral Responsibility, Journal of Philosophy 1969, 829–839 15 The example is parallel to John M. Fischer’s (note 12) 58.
105
Free Will and Alternative Possibilities
deliberating whether he should shoot the politician or not. B has decided to make use of his ability to stimulate A’s brain in favour of shooting the politician if he should foresee that A will decide not to kill the politician. If it should become clear that A will decide by himself to shoot the politician, B will not intervene but only observe. As it happens, A decides by himself to shoot the politician and B remains inactive. Was A’s decision a free choice? Did he have alternative possibilities? The compatibilists would answer in the affirmative to the first question, but in the negative to the second question: Yes, A’s decision was a free choice, and no, he did not have alternative possibilities. Therefore, a free choice for the compatibilist does not presuppose alternative possibilities.16 The famous counter-argument is the ‘Kane / Widerker objection’:17 The possibility to monitor A’s brain and to act exactly in the moment in which A would have made the decision presuppose causal determinism. If indeterminism is true, however, then B will not be able to foresee before the actual decision which decision A is going to make. Consequently, B will not be able to replace A’s decision at the exact moment at which A is going to make it. A will always only be able to act a split-second after the decision. Therefore, under the presumption of indeterminism, real alternative possibilities are open to A at the moment at which he makes the decision which are not excluded by B’s device. V. A PROBABILISTIC TWIST
TO THE
FRANKFURT-TYPE EXAMPLES
Does the Kane / Widerker objection present a convincing rebuttal? I had argued in my Krakow paper18 that it is possible to conceive of an event as the result of a free choice and at the same time as being fostered by certain characteristics which raise the probability of that result. These characteristics can make it possible in some situations to predict behaviour even if the agent has freely chosen to behave in the way predicted. A wellknown example is the stock market where good stock brokers predict how investors will behave although the investors are not perceived as determined but as making free decisions. Would it therefore be a valid rejoinder to the Kane / Widerker objection that we might think of a case in which B is able to foresee on account of a preponderant probability which choice A is going to make while a relevant (though maybe very small) complementary probability remains that A resists making the choice foretold by the probabilistic prognosis? Then the relevant probability that A does not act in the way pointed out by the preponderant probability ensures that the choice is not predetermined while the preponderant probability makes it possible to foretell A’s choice with sufficient certainty. Consequently, we get the impression that it has been tried to refute a highly artificial Frankfurt-type example with a highly formal Kane / Widerker objection. However, our impression is misleading: What did the Frankfurt-type example pur16 John M. Fischer (note 12) 60 17 Robert Kane, Free Will and Values, 1985, 51; Robert Kane, The Significance of Free Will, 1996, 142–144; David Widerker, Libertarianism and Frankfurt’s Attack on the Principle of Alternative Possibilities, The Philosophical Review 104 (1995), 247–261 18 Friedrich Toepel, Free Will and Infinite Regress, above, 36–37
106
Friedrich Toepel
port to show? That a choice may be called free even if no alternative possibilities are open to the chooser. What did the Kane / Widerker objection purport to show? That alternative possibilities were open to the chooser even in the highly artificial Frankfurt-type examples. Now, the Kane / Widerker objection may not have taken probabilistic reasoning serious enough and therefore may have wrongly assumed that the madcap neurosurgeon could not foresee A’s choice. Yet, that was not a mistake which made the Kane / Widerker objection useless. The point of the objection was that A did have alternative possibilities at the moment of his choice, and that is true even in my probabilistic variation of the Frankfurt-type example: We had described the complementary probability as ‘relevant’ that A will resist acting in the way pointed out by the preponderant probability. The complementary probability has to be relevant just because we wanted to ensure that A does make a real choice, that the outcome of his reasoning is not predetermined. I had emphasized in my Krakow paper that indeterminism requires a practically relevant possibility that the actor chooses an alternative to his actual behaviour.19 However, there is no practically relevant possibility for the actor to choose an alternative if he does not have real alternative possibilities. Thus, the neurosurgeon may mostly guess rightly what choice the actor is going to make, but the actor still must have alternative possibilities at the time of his choice. We have included that by using the word ‘relevant’ in our variation of the Frankfurt-type example. If the probability to resist the tendency of the preponderant probability would sink to a merely theoretical level, the whole world of the example would have changed. As a merely theoretical possibility to act differently is not enough to call a choice free, the indeterministic gap would have closed, and a probabilistic determinism would reign. Therefore, we come to the conclusion that the Kane / Widerker objection was correct insofar as either a choice is free and the chooser does have genuine alternative possibilities, or a choice is not free and the chooser does not have genuine alternative possibilities. Accordingly, the Frankfurt-type examples have not served the purpose for which they were construed. In our variation of the Frankfurt-type example above, A by himself made a free choice to kill the politician and he did have genuine alternative possibilities, even if B was determined to bring about A’s resolution to kill the politician with the help of his stimulating device should there be any indications that A was going to decide not to kill the politician. B’s determination does not reduce the probability that A will decide not to kill the politician to an irrelevant level, for B lacks sufficient knowledge to ensure that A will choose as B has determined. This is so because predictability of an event e automatically means unpredictability of the complementary event not e. Whatever degree a probability must have to make an event predictable, it has to be at least greater than 0.5. But if the probability p (e) > 0.5, then the complementary probability p (¬ e) = 1 – p (e) < 0.5. Assuming therefore that we construe the Frankfurt-type example above in such a way that A’s choice to kill the politician is predictable, this means automatically that the complementary probability that A will make the choice not to kill the politician is so small that such event is not predictable. If the preponderant probability motivates B to predict that A will decide 19 Friedrich Toepel (note 18) footnote 13
107
Free Will and Alternative Possibilities
by himself to kill the politician, B will not be prepared for the rare event that the complementary probability materializes. Consequently, a practically relevant alternative possibility is open to A in this example when he chooses to kill the politician. B could only ensure A’s choice to kill the politician by activating his stimulating device regardless of what the preponderant probability predicts. But in this case – if one wants to concede at all that A makes a choice – no practically relevant probability to resist the tendency of the preponderant probability would be left over for A. A’s choice therefore would have to be considered as fully determined by B.20 The variation of the Kane-Widerker objection shows us that the world of indeterminism may just be separated from the world of determinism by a thin veil marked by a distinction between a probability yielding a merely theoretical possibility to decide otherwise and a practical possibility to decide otherwise. The degree of the probability necessary for a practical possibility to decide otherwise is even a matter of evaluation and may be controversial among expert psychiatrists. VI. A RELEVANT DIFFERENCE
BETWEEN
DETERMINISM
AND
INDETERMINISM
Compatibilists have used the small difference between determinism and indeterminism with respect to the degree of probabilities as an argument in favour of their view: Fischer thinks it bizarre to give up what he calls ‘apriori metaphysical truths’ that support one’s views about free will – simply because the theoretical physicists have established that the relevant probabilities are 100 percent rather than 99 percent.21 Such a statement will be debunked as misleading by philosophers well-versed in probability theory. It is correct that a small change of statistical probabilities is never relevant per se to mark a borderline so important as the one between determinism and indeterminism. Yet, the insignificance of the difference cannot be attributed to the fact that it is not even possible that such difference becomes relevant. If the relevant probability that the agent cannot do otherwise is 100 percent, causal determinism reigns with a certainty that cannot be topped. However, if that probability is only 99 percent, this may not be sufficient to exclude determinism because a one percent probability that the agent will act otherwise than predicted on account of the preponderant probability may be only theoretical and not allow for a practically relevant chance of the actor to choose otherwise. But it is a question of evaluation whether a one percent lower probability is relevant or not. Psychiatrists, for example, may thus interpret the outcome of a double-blind study, that the one percent signifies a practically relevant possibility that an agent may resist the tendency of being drawn to make a certain choice. In terms of probability theory, these numbers interpreted as subjective probabilities or de-
20 Consequently, the popular philosophical argument that God’s foreknowledge does not impede man’s free will must be qualified: Man is only free as long as God has decided only to watch, but not to interfere with man’s decisions. If God would decide to replace any decision of a man which God does not accept by one better fitting God’s plans, then man could not have free will. 21 John Martin Fischer (note 10) 47
108
Friedrich Toepel
grees of (an experienced psychiatrist’s) confirmation may gain relevance for the difference between determinism and the indeterminism relevant for free will. I do not know whether it ever will be possible to assign numerical values to the difference between the relevant degrees of confirmation. However, this cannot be decisive. It is rather essential that in principle such difference exists, even if it may only be practicable to assign vague interval probabilities to it. To say that the difference is small and therefore insignificant thus means to beg the question and to jump to a conclusion not supported by the mere fact that the numerical difference between the probabilities mentioned is small. The mere numerical difference never can justify a decision as to relevance because numerical differences must be evaluated and say nothing without such evaluation. VII. ANOTHER TWIST: PEREBOOM’S TAX EVASION EXAMPLE Pereboom has proposed another variation to salvage the point of the Frankfurt-type examples.22 He thinks the Kane / Widerker objection can be refuted by dividing the Frankfurt-type examples into two stages. According to Pereboom’s variation, A first is able to make a choice to do a or to be prepared to do not a, and only if he has chosen to be prepared to do not a, he will later be able to choose not a. Psychiatrist B only influences the second stage with his technical device. Pereboom illustrates this constellation by the following example: A is considering claiming a tax deduction which he knows to be illegal. He has a strong but not overriding desire to decide to act in an illegal way. He could only act differently if he would let himself be motivated by moral reasons. Before he could let himself be motivated by moral reasons, he necessarily would have to raise his level of moral attentiveness. If, however, he would have raised his moral attentiveness, this would not be sufficient for him to abstain from acting in an illegal way. Rather, he still could decide to act in an illegal or a legal way. Raising the level of moral attentiveness only enables him to choose to act in a legal way. This time, unbeknownst to A, psychiatrist B implants a stimulating device in A’s brain that ensures A’s choice of the illegal way once he has raised his moral attentiveness. So, A will always end up choosing the illegal way: Either he chooses not to raise his moral attentiveness. Then he will slide into acting in an illegal way. Or he chooses to raise the level of moral attentiveness. Then he won’t be able to make a further decision because B’s stimulating device ensures that A’s neural centers will make the decision to act in an illegal way. Pereboom assumes that A does not choose to raise his moral attentiveness. He simply chooses to evade taxes. Therefore, B’s device remains inactive. Was A’s choice to evade taxes free? Pereboom thinks that yes, it was free, though no, A did not have what he calls robust alternative possibilities, for by raising his moral attentiveness A would not have chosen to act legally. He only would have thought he had postponed his decision, while in reality he would have run into a trap which would prevent any further decision concerning this matter. 22 Derk Pereboom, Living Without Free Will, 2001, 18–28
Free Will and Alternative Possibilities
109
Kane has objected that here like in the other Frankfurt-type examples, B either not immediately intervenes when A has decided to raise his level of moral attentiveness.23 Then B under the presupposition of indeterminism will not be able to prevent A from making a second choice. Or B intervenes before the choice is made, and then A is not able to make a second choice. However, this objection does not get to the core of Pereboom’s example. Pereboom’s point was that the decision to act in an illegal way must be considered as a free choice though A had no robust alternative possibilities. That A had no robust alternative possibilities is true just because A would not have had a chance to make a second choice. VIII. SOLUTION OF PEREBOOM’S EXAMPLE: ALTERNATIVE POSSIBILITIES IN THE MIND Has Pereboom therefore convincingly shown that there may be free choices without robust alternative possibilities? I think that Pereboom’s description of the case conceals an impossibility: It is misleading when he writes that A “chooses” to evade taxes. I shall contend that A in Pereboom’s example did not make any choice at all because he did not even think he had alternative possibilities. Even if in principle it would be possible to make a free choice without having alternative possibilities – which I deny –, the decider must at least think that he has alternative possibilities when he prepares his free choice. Otherwise, we couldn’t say that his choice was reasons-responsive. Even a compatibilist must concede this: There are two kinds of reasons why I may not think I have alternative possibilities in a situation. There may be a voluntative or a cognitive shortcoming which prevents me to think that I have alternative possibilities. On the one hand, I lack the voluntative element of a choice when I possibly see that hypothetically two alternatives exist but I think that I lack the willpower to carry out the alternative. I may see that I am hypothetically strong enough, for example, to kill my cat by wringing its neck, but I think that I cannot bring myself to overcome the love I have for my cat. Then I do not decide to let the cat live because I do not see a practical possibility to kill it. On the other hand, I lack the cognitive element of a choice when I do not think of alternative possibilities. If I walk through an ancient corridor of a castle and there is a secret door in the wall, for example, I only choose between two possible ways when I know of the secret door. When I do not know of the secret door, I do not choose between two ways if I walk straight on, for in my mind there is no alternative. I do not see the secret doorway. I just might think that hypothetically I could choose between two alternatives if there would be a secret doorway. However, the possibility to pass through the secret doorway must remain a theoretical possibility as long as I do not have the requisite knowledge. But why may not A in Pereboom’s example think he has alternative possibilities? He would have to think that at the time at which he decides to evade taxes. I 23 Robert Kane, Response to Fischer, Pereboom and Vargas. In Four Views on Free Will, J. M. Fischer, R. Kane, D. Pereboom, M. Vargas, 2007, 171
110
Friedrich Toepel
think that the situation is different depending on whether A lacked the necessary voluntative element or the necessary cognitive element. If A lacked the necessary voluntative element, he might miscalculate his own willpower and think that he is strong enough to make the alternative decision. However, if he lacked the necessary cognitive element, he does not know of the alternative, and as long as this knowledge is missing, there is no way to choose. I am of the opinion that A in Pereboom’s example can only be conceived as lacking the necessary cognitive element. A is not able to choose to act in a legal way because he first must raise his moral attentiveness. That means that A must confront himself with the facts more clearly before he can make the choice to act in a legal way. But when he has confronted himself with the facts, he is able to choose. He therefore does not lack the willpower to make the choice. All that is needed is a heightened awareness of the facts. A lacks a cognitive element, and such lack of knowledge prevents him to see practically relevant alternatives. Therefore he does not make a choice at all in Pereboom’s example. Would it not at least be possible that A chooses to raise his moral attentiveness in order to choose to act in a legal way later? No, because that presupposes that he at least can choose to use the choice to raise his moral attentiveness as a means to make himself able to choose to act in a legal way later. However, he who chooses a means to achieve an end also chooses the end itself. And that is precluded by Pereboom’s description that A is not able to choose to act in a legal way before he has raised his moral attentiveness. Cannot A at least choose to raise his moral attentiveness in order to keep open the possibility of choosing to act in a legal way later? No, for choosing possibly to choose to act in a legal way later means letting oneself be motivated by the possibility to act in a legal way if the possibility is practically relevant and not merely theoretical.24 He who cannot be motivated to act in a legal way, also cannot be motivated possibly to act in a legal way. Consequently, if Pereboom writes that A cannot be motivated to act in a legal way before his moral attentiveness is raised, then A also cannot be motivated possibly to act in a legal way before his moral attentiveness is raised. Therefore, A would lack alternative possibilities in this variationt as well. Now, we are able to see more clearly: A’s acting in an illegal way before he has raised his moral attentiveness cannot be the result of a choice but only the result of eventually being caught by his ‘strong but not overriding’ desire. Again, it is misleading when Pereboom uses the term ‘not overriding’. The desire may not prompt immediate action, but if a person cannot choose the negation of evading taxes, it is not understandable how he should motivate himself to raise his moral attentiveness. He can only say: ‘I do not want to act illegally now.’ Yet, it will remain a coincidence whether he not only postpones the gratification of his desire but also raises his moral attentiveness. Postponing the gratification of the desire and raising the moral attentiveness are two totally unconnected things. It is imaginable that A simultaneously with his decision to postpone acting in an illegal way experiences something (a disaster, for example) which motivates him to raise his moral attentiveness. It is 24 If the possibility is merely theoretical, it cannot be decisive for the motivation to raise his moral attentiveness. Someone who only sees a theoretical possibility to act differently which in his opinion is not practically relevant, does not see an alternative possibility.
Free Will and Alternative Possibilities
111
possible that after this has happened, A sees the whole situation in a different light and decides consciously to act in a legal way. Whether his life takes such a turn, however, is not in A’s hands, not if he is the person as whom Pereboom describes him. Therefore, Pereboom correctly writes that A has no robust alternatives in his example, but the description of the situation also precludes that a choice has been made at all. IX. A LAST ATTEMPT
TO
AMEND PEREBOOM’S ENDEAVOUR
We have discovered in the last section that Pereboom’s endeavour to find a convincing Frankfurt-type variation, failed because A lacked a cognitive element which prevented him to see alternative possibilities. But perhaps it would be possible to amend Pereboom’s example if we substitute the cognitive shortcoming by a voluntative shortcoming? A then would not have the necessary willpower to choose to act in a legal way if he, for example, did not first take an energy drink. However, he may believe he has the necessary willpower to choose to act in a legal way without taking the energy drink. Has he made a free choice if he chooses to evade taxes instead of first choosing to take an energy drink? Here, it is interesting that the energy drink is no necessary element of a plan to act in a legal way because A falsely believes he can choose to act in a legal way without taking the energy drink. Insofar a false belief concerning a voluntative shortcoming is always connected with a cognitive shortcoming. However, the cognitive shortcoming always excludes a free choice with respect to the facts which are concealed to the decider. Therefore, the cognitive shortcoming in this energy drink variation cannot help to amend Pereboom’s example. It should be disregarded, and we should concentrate on A’s false belief. A believes that he now is able to choose to act in a legal way, while in reality he lacks sufficient willpower to make the decision. However, with respect to the lacking willpower our variation is similar to a classic Frankfurt-type example like the one discussed under section IV in the constellation that B actually activates his stimulating device. The difference between our variation and the example in section IV is that in section IV the device implanted in A’s brain is stimulated at the exact moment of the decision, while in our variation A’s brain is influenced in one direction because of lacking will power already before the decision. I think that the solution of such a variation should not be controversial: If A’s brain is thus influenced in advance or if A lacks sufficient will power, the decision is not A’s own. He is helplessly exposed to the circumstances and therefore not making a free choice. Therefore, my attempt to amend Pereboom’s example has failed.
Friedrich Toepel
112 X. GUIDANCE CONTROL
VS.
REGULATIVE CONTROL
John Martin Fischer has proposed a criterion for free choices without alternatives which he terms ‘guidance control’.25 A person acts with guidance control when he makes reasons-responsive choices though no alternative possibilities may be open to him. However, as we have already noted above, it is incomprehensible how someone should choose in a reasons-responsive way if no alternative possibilities are open to him. What significance can a reason have if it cannot guide a person to decide differently than he would have decided without the reason? Fischer offers the example that you turn to the right with your car while – unbeknownst to you – the steering apparatus of your car is broken. The car would have gone to the right anyway if you had tried to guide the car in any other direction. Fischer thinks that the defect of the steering apparatus does not prevent you to have guidance control over the car’s movements as long as you turn to the right and the trajectory actually travels that way.26 In my opinion, we have to remember the distinction between the technical ability to act in accordance with the law and the ability to motivate oneself to act in accordance with the law from section II above. The control over the car when turning to the right in Fischer’s example is similar to the technical ability to act in accordance with the law, while your decision to turn right is similar to the motivational ability to act in accordance with the law. The decision to turn right is not impaired by the broken steering apparatus. You have real alternatives insofar and therefore can make a free choice. However, the ability to turn your car into a certain direction is impaired by the broken steering apparatus. You may not successfully turn the car to the left. You are only lucky that you want to turn the car in exactly the direction in which it is able to go. The trajectory does not travel the way it does because of your decision. It is not reasons-responsive, only your decision to travel that way is. Therefore, to exercise guidance control over your car means that you exercise no control at all. Fischer himself insists that in order to exhibit guidance control, the actual-sequence mechanism must be the agent’s own.27 If there are no alternative possibilities to the agent’s behaviour, the actual-sequence mechanism cannot be the agent’s own. That Fischer argues to the contrary, shows that he relies on an inadequate notion of an agent. As a compatibilist he must choose a notion of ‘being active’ which is completely consistent with causal determinism. Such notion in Fischer’s mind is what is required for moral responsibility. But a notion of being active which is consistent with causal determinism would also include ‘active’ animal behaviour where the question of moral responsibility is not appropriate. Animals are not held responsible exactly because they do not freely choose between alternative possibilities and are drawn to a certain behaviour by their instincts. Choices therefore are not free if the decider does not have alternative possibilities open to him or ‘regulative control’ as Fischer terms it. Guidance control is not enough.
25 John M. Fischer (note 12) 57–59, 78–80 26 John M. Fischer (note 12) 57 27 John M. Fischer (note 12) 59
Free Will and Alternative Possibilities
113
XI. SUMMARY ‘Deep responsibility’ needed to impute an act legally or morally to a person requires a radically free will: If criminal law addresses the criminal as a person, it must make it possible for him to understand the punishment as a just reaction to the crime committed. The criminal can only understand the punishment as a just reaction if he sees the possibility that he could have acted in accordance with the law instead of committing the crime. Therefore, a genuine alternative possibility to act in accordance with the law must have been open to the criminal. Genuine alternative possibilities were only open to the criminal if he had a free choice to act otherwise than he did. Such free choice is not compatible with causal determinism. There cannot be a ‘deep’ responsibility required for moral or legal imputation without this freedom of choice. Frankfurt-type examples are no argument against the necessity of this freedom of choice. Compatibilists are not able accommodate this freedom of choice within their systems. The substitute compatibilists offer for the freedom of choice is not sufficient to justify moral or legal imputation.
MATEUSZ KLINOWSKI, KRAKÓW1 WHAT CAN LOGIC OF ACTION TELL US ABOUT RESPONSIBILITY AND FREE WILL? ABSTRACT Logic of action may be understood as a distinct approach to philosophical problems concerning agency. As it has turned out recently, some powerful calculi are able to express not only propositions about doing and refraining, but also about conditions of responsibility. In my paper I will try to demonstrate that the calculus based on branching-time structures and known as sees to it theory can be used as a convincing argument for choosing among rival proposals concerning conditions of responsibility. This calculus supports also so-called agent causation theory and the indeterministic view of our reality. INTRODUCTION Free will is certainly one of the most debated philosophical topics. Surprisingly, quite recently the topic has lost some of its metaphysical connotations and started to be an object of scientific analysis thanks to the progress in neurosciences and various neuroimaging techniques. Yet in this paper my aim is not to deal with the nature of free will. Rather, I would like to examine the conceptual side of the socalled ‘free will debate’ and point at the fact that contemporary modal logic can help us to solve some problems that emerged during the debate. FREE
WILL, DETERMINISM AND CONDITIONS OF RESPONSIBILITY
Let me start with some preliminary definitions. I assume that free will is a trait that every agent has to have if he / she is to be held responsible for his / her actions. Thus, we can define free will as follows: [1] Free will: the necessary condition for being responsible for action. The ‘free will debate’ can be viewed as a disagreement on how free will should be understood – what is an exact content of this necessary condition. A significant number of authors doubt that it is ever possible to find such a concept, since in our reality, at the deepest level, there is no place for responsibility at all. It is because according to their opinions our world is deterministic. Let me define what the latter term means. In order to do that we need to define the term ‘deterministic system’ first. 1
Chair of Legal Theory, The Faculty of Law and Administration, Jagiellonian University
Mateusz Klinowski
116
[2] Deterministic system: it is a system having this property that every complete, momentary state of the system, taken together with the set of law governing the system, suffices to predict / determine / describe the state of the system in every other moment of time (no matter where located – in the future or in the past). Now, to believe in determinism is to believe that: [3] (Physical / causal) determinism: our reality is a deterministic system. In my opinion there are many good reasons not to believe in [3]. One of the most important is the fact that we don’t have at hand any deterministic physical theory that gives us a comprehensive description of the reality. However, some authors claim that the key argument here is rather our social or psychological (introspective) experience – the fact that we consider ourselves as rational free agents. And because freedom and free will seem to be incompatible with determinism, [3] cannot be true. In other words, we know that people are sometimes held responsible (and we don’t know for sure that [3] is true). Of course the force of this argument is disputable, because what we see or we know is not always true. The largest group of authors taking part in the ‘free will debate’ escape all those perplexes and simply argue that [1] and [3] are not contradicting each other, being both true. Those authors are proponents of the standpoint in the debate called compatibilism. [4] Compatibilism: [1] and [3] are consistent with each other. It seems to me that compatibilism focused attention and attracted so many supporters mainly because of the prima facie implausibility of its rival. This implausibility is expressed as the so-called Mind Argument. The central point of this argument consists in the fact that if we choose to deny [4], it seems that we have no other option than to deny agency altogether. The reason is that we seem to be forced to assume, by the denial of [4], that free will depends on undetermined events. And such events hardly can be viewed as controlled by anyone, including a supposed agent. So the denial of [4] seems to lead to the completely implausible concept of agency. AGENT
CAUSATION THEORY
According to the dominant view on incompatibilism the supporters of this standpoint assume that responsibility requires truly undetermined events. And this particular assumption is puzzling, because agency obviously requires an agent’s control over events. This line of criticism misses one point. Incompatibilism is a contradictory stance as long as we suppose that the terms undetermined events and spontaneous events are synonymous. As soon as we assume that indeterminacy could be understood in more limited way: as a lack of physical causes, it is possible to shield incompatibi-
What can Logic of Action Tell Us About Responsibility and Free Will?
117
lism from the Mind Argument. One of the proposals designed to do the job is called the agent causation theory (ACT). ACT is one of the rival theories explaining the nature of action.2 According to the theory action is not an event, but rather an agent causing an event. So action is a special type of causation. ACT can also be used to explain how to reconcile indeterminism and responsibility. Since it is an agent that causes events described as his actions, he exercises control over them, even if these events lack any physical causes. Thus, according to ACT there are events physically undetermined, but at the same time not spontaneous in an absolute sense. Assuming ACT, it seems that we still have a reason to believe that requirements for ascribing responsibility are met, even if we disagree with [4]. We need only to postulate that it is an agent that determines events described as his actions. REQUIREMENTS
FOR ASCRIBING RESPONSIBILITY
Let’s now focus on the conditions of responsibility that have been proposed by different authors so far during the debate. Robert Kane claims that the most fundamental requirement for ascribing responsibility and, a fortiori, free will is the condition of ultimate responsibility (UR).3 This requirement means that it is an agent and only an agent who determines whether such and such event takes place. UR: a person is morally responsible for an event A only if he can be considered as an exclusive source of A. UR seems to be congruent with ACT and, obviously, it excludes entirely the possibility of holding [4] true. But for some authors this condition might seem to be too general and not instructive at all when taken as an explanation of the phenomenon of responsibility. In their opinion the latter should rather be described in terms of possibility and necessity. The first proposal exploring this stance is based on an assumption that agency, responsibility and free will means that an agent have power (ability) to do otherwise than he / she actually did. This condition is called the principle of alternative possibilities. PAP: a person is morally responsible for what he / she has done only if he / she could have done otherwise. There are a lot of scholars who claim that this condition not necessarily implies UR – many supporters of [4] (compatibilists) are keen on assuming that they are able to identify the meaning of could have done otherwise that is in accordance with [3]. Naturally, for Kane all arguments they use could be rather interpreted as arguments for
2
3
The theory was proposed for the first time by Richard Taylor and Roderick M. Chisholm, see Richard Taylor, Action and Purpose, 1966; Roderick M. Chisholm, The Agent as Cause, in: Action Theory, ed. M. Brand, D. Walton, 1976, 199–211. Robert Kane, The Dual Regress of Free Will and the Role of Alternative Possibilities, Philosophical Perspectives 14 (2000): 59–79
118
Mateusz Klinowski
the proposition that PAP is not a sufficient or a necessary condition for responsibility, than for truth of [4]. The first way to question PAP is due to J.L. Austin and it is founded on the observation that could have done otherwise might stand for a concept of ability which is based on so-called normal (or standard) circumstances, not on actual possibility (or a lack of it) for action. So there are cases when a particular event was outside of the scope of agent’s voluntary control, even though an agent could have done otherwise in a sense of having ability in general.4 Hence, when one takes a suitable meaning of PAP, one can demonstrate that PAP is not sufficient nor is necessary for responsibility. The second way to support the skepticism about PAP is thorough so-called Frankfurt-style examples. I will not discuss them here, although this enormously analyzed argument is in my opinion flawed. Kane was not the only author criticizing on PAP. Also Linda Zagzebski opposed PAP arguing that responsibility requires rather causal contingency of an event.5 CC: An agent is morally responsible for what he has done only if his act is causally contingent. Causal contingency of an event A is understood as the existence of the alternative future (to the actual) for A, given the same causal history of A. Zagzebski thinks that an event could be contingent causally and thus free (ascribed to an agent) even if an agent in fact is unable to do anything else than he actually did or will do. Such a situation will take place when there are determining factors that are lying outside of causal history of the ascribed event. Zagzebski proposed God’s foreknowledge as one of the examples of such determining factors. Finally, Peter van Inwagen proposed yet another requirement for responsibility. According to his proposal an agent can be held responsible only if he is able to prevent an event to happen.6 This condition is known as the principle of possible prevention (PPP). PPP: An agent is morally responsible for what he has done only if it is possible for him to prevent the event from happening. To sum up, there are at least four, probably partially different, major proposals of conditions of responsibility. And it is an interesting task to ask a question about their adequacy. In the following paragraphs I will use one of logical calculi suitable to express action sentences and concepts of agency (logic of agency) to examine plausibility of described requirements for responsibility and also plausibility of ACT itself.
4 5 6
John. L. Austin, Ifs and Cans. In Philosophical Papers, J. O. Urmson / G. J. Warnock eds., 3rd ed. 1979, 205–232 Linda Zagzebski, Does Libertarian Freedom Require Alternate Possibilities? Philosophical Perspectives 14 (2000): 231–248. P. Van Inwagen, An Essay on Free Will, 1989
What can Logic of Action Tell Us About Responsibility and Free Will?
LOGIC
119
OF ACTION: SEES TO IT THEORY
During my research I figured out that many of the problems listed above can be expressed in logical language of one particular calculus. It is N. Belnap’s and J.F. Horty’s calculus known as sees to it theory (stit theory).7 The name is due to the fact that both authors decided to represent every sentence about agency as (where “A” represents any descriptive sentence): α (an agent) sees to it that ‘A’. For this expression, abbreviated as ‘α dstit: A’, they proposed semantic theory based on so called branching time structures. Skipping all technical details, these structures are meant to represent indeterministic universe, with many different versions of future history and the one settled past. Branching time frames were designed to give a background for temporal logic, but Belnap and Horty adopted them to serve as a backbone of their logic of action. The idea was simple – add to the existing semantics for temporal logic a function which operates on histories and represents choices of an agent, then stipulate that ‘α dstit: A’ is true in a moment m of a particular history h only if ‘A’ is true in every history lying in the same partition of the set of histories at m that history h belongs to, and there is a history (at least one) at m that ‘A’ is not true in it. Fig.1 illustrates how this definition works. At moment m we have five different histories represents different chains of event that are the potential future for m. Three of them belong to the option O1, which represents one way the agent may behave, two others belong to the option O2 represents a different action (a lack of it). In h1, h2, h3 sentence ‘A’ is true and in h5 it is not. Thus, we can assume that in this case, in a moment m of histories h1, h2, h3 it is an agent α that guarantees the truth of ‘A’. Since he / she can restrict the future course of events, by choosing O1, in a way that only those histories that support ‘A’ will be actualized.
7
J.F. Horty, N. Belnap, The Deliberative Stit: A Study of Action, Omission, Ability, and Obligation, Journal of Philosophical Logic 24 (1995): 583–644.
Mateusz Klinowski
120
Using modal operator ‘dstit’ to express doing is convenient, since it is easy to combine the operator with other modal operators or with nesting. For example, we can combine modal operator for historical possibility ‘◊’ with ‘dstit’ and as a result we will get the formula: ‘◊(α dstit: A)’ which represents ability to do. Since the truth of ‘◊A’ in a particular moment of a history means that sentence ‘A’ is true in at least one history branching at this moment, similarly ‘◊(α dstit: A)’ means that in a particular moment / history an agent is able to do ‘A’, because ‘(α dstit: A)’ is true in at least one of the histories branching at m. Combining ‘dstit’ with negation and nesting we can also express inaction: ‘¬(α dstit: A)’ and refraining: ‘α dstit: ¬(α dstit: A)’. And using a deontic modal operator representing ‘it is ought to’ we can construct deontic logic that, as it turns out, is free of well known paradoxes. All of those facts prove that stit is a successful and very plausible story about how our action sentences are understood and what the logical relations between them are. Let’s now examine what stit can tell us about the plausibility of ACT and all of the requirements for responsibility listed above. STIT
AND AGENT CAUSATION
Proponents of [4] are denying that in an indeterministic set-up UR could be met. But it is exactly the statement that is to be dismissed on the ground of stit.
What can Logic of Action Tell Us About Responsibility and Free Will?
121
It was already said that stit is based on an indeterministic model of branching histories against which plausible truth-conditions for action sentences are defined. In this model a sentence ascribing agency is represented by the formula: ‘α dstit: A’. We can understand that expression as equivalent to the statement that an agent α is an author (a source) of an event described by a sentence ‘A’. And – according to ACT – that represents an action. For compatibilists such an interpretation is implausible since in their opinion genuine agency is impossible in an indeterministic model (there is no room for an agent’s voluntary control over events). So in their opinion, even in stit, it should be impossible to infer that an agent is the source of his action from the statement that he is a source of a particular event A. This proposition can be expressed in terms of stit as: ¬ (α dstit: A ⊃ α dstit: (α dstit: A)). But this is never a case according to stit, since the following formula is a theorem of this calculus: (α dstit: A) ≡ (α dstit: (α dstit: A)). Hence, it seems that we have a good reason to believe that ACT is at least a plausible explanation of how responsibility can be ascribed in an indeterministic reality – we have a calculus at hand that proves that directly. STIT
AND CONDITIONS OF RESPONSIBILITY
In this paragraph I will examine one after another the conditions for ascribing responsibility that were presented above. At the beginning we have to decide what formula will represent the sentence: ‘α is responsible for A’. I think it is quite obvious that in most of the cases if ‘α is responsible for A’ then ‘α did or is doing A’. Having this in mind we can assume that conditions that are required for responsibility are also required for agency. And conversely, if meeting a particular condition is not enough for agency, it is also not enough for being responsible for an action. In terms of stit PPP and CC can be expressed as: PPP: (α dstit: A) ⊃ ◊(α dstit: ¬A) CC: (α dstit: A) ⊃ ¬ (α dstit: A) The latter implication is a universal truth for stit, but not the former. PPP seems flawed downright – it is possible to have such a division of histories that there is no single choice cell contains only histories with ¬A, even if at the same time there is a cell with all histories contain A. So the left side of PPP could be true even if the right side is false. In case of CC the truth of the left side of the implication guarantees the truth of the right side, since according to the definition of validity for ‘α dstit: A’, there must be a history that renders ‘α dstit: A’ false.
Mateusz Klinowski
122
Next, if we take into an account the fact that in stit we have a theorem: ◊(α dstit: A) ≡ ◊(α dstit: ¬(α dstit: A)) we can express could have done otherwise as an ability to refrain from doing. Hence, PAP would be formalized as: PAP: (α dstit: A) ⊃ ◊(α dstit: ¬(α dstit: A)), which is also an universal truth. We already know UR is to be formalized as: UR: (α dstit: A) ⊃ (α dstit: (α dstit: A)). As we see CC, PAP and UR are necessary conditions for ascribing agency (responsibility). But are they also sufficient? This question seems to be much more important. UR can play a role of a sufficient condition for agency because if ‘α dstit: (α dstit: A)’ is true for a moment / history ‘α dstit: A’ is also true for the same pair. If you are the author of an action, there must be an action that takes place. In case of PAP situation is less obvious, since the truth of ◊(α dstit: ¬(α dstit: A)) doesn’t determine that there was an action that took place (maybe it was not, because an agent refrained from doing it). However, still in stit the possibility of refraining means that doing is also possible (there are histories that contain ‘α dstit: A’). But in the case of CC even a sheer possibility of an action taking place cannot be supposed. It is because it can be demonstrated that even when ‘¬ (α dstit: B)’ holds at the particular moment m, still we can imagine that there is no single history through m containing ‘α dstit: B’. Fig.1 is again an example of this situation. Because B is true in every history, ‘α dstit: B’ is, by definition, false in all of them (and, simultaneously, ‘¬ (α dstit: B)’ holds at m). SUMMARY Relying on logic of action it is possible to formalize some disputable statements from the field of philosophy of agency. By using a particular calculus that has already proved its relevance and value as a tool for analyzing philosophical problems regarding agency, we can see that ACT is a coherent story about action and responsibility and it suits the UR requirement. It is rather optimistic, since it can be also proved that UR is a necessary as well as sufficient condition for holding someone responsible. At the same time we can also prove that PPP is not a condition for responsibility at all and that CC and PAP are only the necessary one. PAP seems to be an insufficient condition for holding someone responsible, since there is still up to an agent to decide if he / she wants to exercise his / her ability.