Copyright © 1992, 1999, Peter Suber.

The Paradox of Liberation
Peter Suber, Philosophy Department, Earlham College

Understanding freedom is hard enough. What is much harder is to understand liberation, or how we become free. I am concerned here with a paradox which many philosophers seem to depict in the process of liberation: that we may freely choose to become free. In some the paradox is stark enough to be a vicious circle: that we must be free in order to become free. On occasion we find the paradox so starkly that the unfree are criticized, blamed, or held responsible for their unfreedom, as if they were free to become free even when unfree.

It seems paradoxical to expect the unfree to use the freedom they lack to liberate themselves. Are philosophers who make this demand simply confused? Isn't it unfair to demand that the unfree free themselves? Why isn't it like asking the unjustly imprisoned to release themselves? Or is liberation essentially a bootstrapping process which would be falsified if not described in this circular way?

The question is important for two reasons. Exegetically, it is important because the philosophers who imply that we must be free in order to become free do not always seem to realize that they imply this paradoxical position. Exploring the issues directly, then, helps us figure out what these thinkers might have meant. Metaphysically, the question is important because if we discover that liberation is a vicious circle, then we will have to conclude that freedom is unattainable (at least through liberation). If we discover that the circle is not vicious, then we will have learned more about the complexity of freedom —and of the means of acquiring it.

To show that the paradox has been skirted, if not affirmed, and to introduce its issues, I discuss three philosophers. Kant directly asserts something very like the paradoxical position that the unfree are responsible for their status, and he blames them if they do not exert themselves to change. Next, to show that this position is not limited to 18th century pietists given to blame the most minute deviations from a severe moral code, I consider a more contemporary liberal thinker, Daniel Dennett. Dennett comes to a position very similar to Kant's from a quite different set of premises. Finally, I discuss Mill, to show that the problem arises for political liberty, not just for metaphysical freedom of the will.

Case Study:  Kant

Negative freedom is freedom from heteronomy, or freedom from determination by external causes; positive freedom is autonomy, or legislation by reason (KdrV B.581-82; Gr.446-47, 452, 458; KdpV.29, 33, 42-43, 47-48).[Note 1] Since people can lack both kinds of freedom, liberation would be to acquire both kinds.

But if one lacks both kinds of freedom, then one is heteronomously determined and without capacity for self-legislation. If to reach positive freedom one must pass through negative freedom (KdpV.20-21), then the first step of liberation must be to attain negative freedom or overcome heteronomy. But ex hypothesi one is heteronomously determined. Hence if liberation is possible, then it seems that one cannot liberate oneself; at least one could not do so freely. For the same reason, unfree or heteronomous people should not be accountable for their actions.

These implications may be implausible, or perhaps merely regrettable. But they have the saving grace of sparing Kant from the paradox of liberation. If the heteronomous are not able to liberate themselves and are not accountable for their actions, then they are not responsible for liberating themselves and cannot be blamed for failure to do so.

But other parts of Kant's doctrine pull in the other direction. Ultimately, he wants to hold the heteronomous responsible for their unfreedom and blame them both for immoral acts they commit heteronomously and for failure to liberate themselves.

The life of heteronomy for Kant is a life of temptations indulged, resistance overcome, appetites pursued, and inclinations followed. It is a life of being pushed and pulled, never acting. It a life without self-control, under the control of various, and ever varying, impulses. For short, let me call it the life of dissipation. If we emphasize the way in which temptations are indulged, and inclinations followed, then we may suggest that heteronomy is suffered, tolerated, permitted, or granted by —what might as well be— a free decision. And this is roughly how Kant sees it (see Gr.457-58).

Here is Kant discussing the case of a culpable dissipate (from the first Critique, B.582-83; the italics are mine).

[L]et us take a voluntary action, for example, a malicious lie by which a certain confusion has been caused in society. First of all, we endeavor to discover the motives to which it has been due, and then, secondly, in the light of these, we proceed to determine how far the action and its consequences can be imputed to the offender. As regards the first question, we trace the empirical character of the action to its sources, finding these in defective education, bad company, in part also in the viciousness of a natural disposition insensitive to shame, in levity and thoughtlessness, not neglecting to take into account also the occasional causes that may have intervened. We proceed in this enquiry just as we should in ascertaining for a given natural effect the series of its determining causes. But although we believe that the action is thus determined, we none the less blame the agent, not indeed on account of his unhappy disposition, nor on account of the circumstances that have influenced him, nor even on account of his previous way of life; for we presuppose that we can leave out of consideration what this way of life may have been, that we can regard the past series of conditions as not having occurred and the act as being completely unconditioned by any preceding state, just as if the agent in and by himself began in this action an entirely new series of consequences. Our blame is based on a law of reason whereby we regard reason as a cause that irrespective of all the above-mentioned empirical conditions could have determined, and ought to have determined, the agent to act otherwise. This causality of reason we do not regard as only a co-operating agency, but as complete in itself, even when the sensuous impulses do not favour but are directly opposed to it; the action is ascribed to the agent's intelligible character; in the moment when he utters the lie, the guilt is entirely his. Reason, irrespective of all empirical conditions of the act, is completely free, and the lie is entirely due to its default.

In this passage we see the paradox of liberation starkly. The man's lie was determined by empirical conditions, yet despite his unfreedom he is blameworthy. His reason was free throughout to override these empirical determinants, and his failure to turn to his reason for this purpose is culpable. "When we say that in spite of his whole previous course of life the agent could have refrained from lying, this only means that the act is under the immediate power of reason, and that reason in its causality is not subject to any conditions of appearance or of time" (KdrV B.584).

Kant takes several steps toward making this position less paradoxical. One is to claim that reason can overcome feeling (KdrV B.562, B.830; cf KdpV.73). One way he often puts this is that respect for the moral law is a feeling, in fact a pleasure, the one feeling determined by reason (Gr.401.n, 410-11, 427, 460; KdpV.30, 38-39, 75, 78, 80, 117). Hence, while most feelings pull in some heteronomous direction, another feeling (respect for the law) may pull in the opposite direction. In this way reason can fight heteronomous influences on their own level. Hence, even the dissipate, wallowing in heteronomy, can summon reason to help fight his dissipation. The paradox of liberation is greatly mitigated if reason is permanently available to aid the unfree in becoming free.

Kant says the malicious liar had an empirical character which was determined and an intelligible character which was free (also see KdrV B.567). For Kant, the empirical character of a person is entirely determined (KdrV B.578, B.581) and, in principle, perfectly predictable (KdrV B.578; KdpV.99). The intelligible character is not causally determined at all, either because it is guided by reason, or because it is not in time, or both (KdrV B.567, B.569, B.579-80, B.581). So from one standpoint, we are determined, and from another we are free. But despite these profound differences, the two characters are not independent of one another. The empirical character is "determined in the intelligible character" (KdrV B.579; cf Gr.453), and the empirical is in turn the schema (KdrV B.581) and sensible sign (KdrV B.574) of the intelligible. The link is tight enough that a "different intelligible character would have given a different empirical character" (KdrV B.584). It follows that the intelligible character can cultivate a different empirical character, further supporting the view that reason can be a cause even at the level of feeling and desire and, to that extent, further mitigate the paradox.

The distinction between the empirical and intelligible characters grounds what has been called Kant's dual-aspect theory of freedom and determinism. The very same act can be viewed as empirically determined and transcendentally free (KdrV B.564, B.565, B.569, B.831). The true moral quality of the act (whether it is good or bad) lies in the intelligible character, which is unknown to us; and for this reason, all imputation is limited to the empirical character (KdrV B.579.n). The dual-aspect theory per se also mitigates the paradox of liberation. Unfreedom is not quite so unfree and hopeless if it always coexists with the free intelligible character. Whether this coexistence can be made consistent and intelligible is a separate question.

In the passage on the malicious liar, Kant also invokes his principle that ought implies can (see also KdrV B.576, B.835-36). If reason should have intervened to reverse the empirical tide of heteronomous influences, then it could have. If ought implies can in a strong sense, then the paradox can never arise; we can always draw upon the freedom required for the acts we ought to perform. The paradox asks, in effect, whether this assurance isn't mysterious, or whether an expectation based on the assurance isn't unfair.

There are two problems with Kant's analysis of the malicious liar. First, Kant is obscure, if not mysterious, on the mechanism by which reason or the intelligible character becomes an empirical cause among others fighting for dominance and control of the inclinations, helping the dissipate advance toward self-discipline and self-control and hence toward autonomy. Kant can say that the 'mechanism' cannot be intelligible if it works at the seam of the phenomenal and noumenal (Gr.458-89, 459-60, 461-62; KdpV.72, cf. 133), and that would be consistent with his doctrine. But if this is his reply, then it emphasizes the convenient and conclusory way in which reason functions like a 'ghost in the machine' to make moral imperatives stick to the heteronomous.

The second problem is that if reason is always free and available to help pull us out of our heteronomy, then the distinction between autonomy and heteronomy begins to break down. This is not a problem for theory so much as it is a problem for Kant's text. His theory of heteronomy seems inconsistent with the permanent liberating accessibility of reason. One part of his theory which he cannot square with the permanent presence of free reason is his view that negative freedom is logically, and perhaps temporally, prior to positive freedom.[Note 2] Another is simply his view that heteronomy exists and consists of determination by external forces. If we take Kant seriously when he says "it is in the power of freedom to pass beyond any and every specified limit" (KdrV B.374), then we can no longer take heteronomy seriously as an existing form of unfreedom. Instead of asserting the interesting position that a heteronomous dissipate is free to avoid lying, Kant would be saying that we are always free and that liberation consequently is never necessary.

Kant must reconcile two strands in his theory: that heteronomy is the unfreedom which must be overcome before higher kinds of freedom can be attained, and that heteronomy is only unfreedom from one standpoint and coexists at all times with a significant kind of freedom.

In less technical language, Kant is saying to the heteronomous dissipate: You are not free now, but you could have cultivated the self-control which would have prevented your dissipation and preserved your freedom; you are blameworthy for not having done so. Moreover, it is never too late to start, for the spark of reason is inextinguishable. The temptations you have indulged will, perhaps more often than not, overpower the voice of reason which you have allowed to become weak and small. But if you start now and work hard, which is certainly within your power, then the voice of reason will gradually, inevitably drown out the other voices bidding for control of your will. You were responsible for letting this happen to you, and you are responsible for taking control of your life again.

This is very likely a helpful sermon to preach to the dissipated. It is roughly what Synanon, the drug rehabilitation movement, extracted from Emerson's "Self-Reliance" —finding Kant less useful for the purposes of exhortation and therapy. But the body of the sermon contradicts its first sentence: you are not free now....

Appendix to the Case Study on Kant

The case study of the malicious liar appeared in both the first and second editions of the first Critique (1781, 1878). In between these two editions, Kant published his Foundations of the Metaphysics of Morals (1785), in which he included a similar case study, with intriguing but obscure detail on the 'mechanism' question (italics mine).

When we present examples of honesty of purpose, of steadfastness in following good maxims, and of sympathy and general benevolence even with great sacrifice of advantages and comfort, there is no man, not even the most malicious villain (provided he is otherwise accustomed to using his reason) who does not wish that he also might have these qualities. But because of his inclinations and impulses he cannot bring this about, yet at the same time he wishes to be free from such inclinations which are burdensome even to himself. He thus proves that, with a will free from all impulses of sensibility, he in thought transfers himself into an order of things altogether different from that of his desires in the field of sensibility. He cannot expect to obtain by that wish any gratification of desires or any condition which would satisfy his real or even imagined inclinations, for the idea itself, which elicits this wish from him, would lose its pre-eminence if he had any such expectation. He can expect only a greater inner worth of his person. He imagines himself to be this better person when he transfers himself to the standpoint of a member of the intelligible world to which he is involuntarily impelled by the idea of freedom, i.e. independence from the determining causes of the world of sense; and from this standpoint he is conscious of a good will, which on his own confession constitutes the law for his bad will as a member of the world of sense. He acknowledges the authority of this law even while transgressing it. (Gr.454-55)

Here the malicious villain concededly cannot attain the virtues he admires, but this does not stop him from judging himself by the standard of those virtues. He is not materially aided in his escape from dissipation, but he is nevertheless susceptible to justified blame and reproach, even from himself, as if he were responsible for his negative freedom and failure to liberate himself.

In some passages Kant asserts that the dissipate is unfree, and yet is blameworthy for sliding into that state and for remaining in it. In this sense the dissipate is both free and unfree, and Kant discusses several qualified kinds of freedom which begin to make this possible. But in this passage and in others like it he suggests another possible solution: the dissipate is unfree —period— but acting as if he were free can effect his liberation. How this can happen opens a new vista on the 'mechanism' problem. How can free reason, or the mere idea of freedom, have effects which materially aid the dissipate achieve negative freedom? Or, how can acting as if we were free make us free?

Kant might want us to here to take the step which Charles Renouvier inspired William James to take:[Note 3]

I think that yesterday was a crisis in my life. I finished the first part of Renouvier's second Essais and see no reason why his definition of free will —"the sustaining of a thought because I choose to when I might have other thoughts"— need be the definition of an illusion. At any rate, I will assume for the present —until next year— that it is no illusion. My first act of free will shall be to believe in free will.

Or he might mean that attaining freedom when we don't have it is like attaining moral virtues when we don't have them: believing the fiction is part of creating the reality. (Or as Alcoholics Anonymous sometimes puts it, "fake until you make it.") Kant argues that human hypocrisy and vanity, "like everything else that comes from nature, must finally contribute to good ends" (KdrV B.775-76). What good office do they perform? Putting on a false show of goodness, and concealing our shortcomings, "has, undoubtedly, not only civilized us, but gradually, in a certain measure, moralized us" (KdrV B.776; italics are Kant's own; cf. B.776-78).

Kant says little more on the benefits of this dissimulation. But his theory has been picked up and fleshed out by others. Hans Vaihinger sees Kant as a forerunner of the philosophy of 'as if', even in ethics.[Note 4] James Branch Cabell, though emphasizing the humorous side of the phenomenon, is as serious as Kant in his claim that hypocrisy and vanity can improve us, and even free us. Cabell refers to the imagination, or the power to act as if something were the case when it is not, as the "demiurgic spirit of romance". Since it leads us to believe falsehoods which depict us as stronger and better than we really are, it is partly a function of vanity and dullness. In his chief work of non-fiction, Beyond Life (1919), Cabell makes the following observations.

[W]e are one and all of us created with large talents in the way of vanity and dullness: and by utilizing these invaluable qualities the demiurgic spirit of romance will yet contrive a world 'as it ought to be'.[Note 5]

The things of which romance assures him are very far from being true: yet it is solely by believing himself a creature but little lower than the cherubim that man has by interminable small degrees become, upon the whole, distinctly superior to the chimpanzee.[Note 6]

So it is that man's vanity and hypocrisy and lack of clear thinking are in a fair way to prove in the outcome his salvation.[Note 7]

For Cabell, romance effects both our improvement and our liberation. In another work he calls it "the untruth which makes [us] free."[Note 8]

Note how Kant is unwilling to apply his analysis to the character so depraved that he no longer possesses or uses his reason (Gr.454). This gives a clue to the 'mechanism' problem. If dissipation has gone so far as to silence the voice of reason, then Kant suggests that blame and reproach would no longer be justified. Heteronomy this total is no longer culpable. But most villainy is not that dissipated. Indeed, most villainy is comparatively prudent, which requires an exercise of reason. For such villains we may hold out hope, but also blame, because reason permits them to escape their heteronomy.

The possibility that the depraved may lose the use of their reason strengthens Kant's position by putting a limit on the blame we cast on the unfree. Nevertheless, it contradicts his analysis of the malicious liar, in which he insisted that reason is perpetually present and accessible even to the heteronomously determined.

Earlier in the Foundations, Kant seemed to reach some clarity on the mechanism of escaping heteronomy by appeal to reason (my italics):

The pure conception of duty and of the moral law generally, with no admixture of empirical inducements, has an influence on the human heart so much more powerful than all other incentives which may be derived from the empirical field that reason, in the consciousness of its dignity, despises them and gradually becomes master over them. (Gr.410-11)

But this is deceptive; for it merely repeats his oft-repeated position that reason affects feeling (citations above), without shedding light on the mechanism by which this could happen.

We don't get anything like detail on the mechanism question until the Critique of Practical Reason of 1788:

Certainly it cannot be denied that in order to bring either an as yet uneducated or a degraded mind into the path of the morally good, some preparatory guidance is needed to attract it by a view to its own advantage or to frighten it by fear of harm. As soon as this machinery, these leading strings, have had some effect, the pure moral motive must be brought to mind. This is not only because it is the sole ground of character...but also because, in teaching a man to feel his own worth, it gives his mind a power, unexpected even by himself, to pull himself loose from all sensuous attachments (so far as they would fain dominate him) and, in the independence of his intelligible nature and in the greatness of soul to which he sees himself called, to find himself richly compensated for the sacrifice he makes. We should prove, by observations which anyone can make, that this property of our minds, this receptivity to a pure moral interest and the moving force in the pure thought of virtue when properly commended to the human heart, is the strongest incentive to the good and indeed the only one when it is a question of continual and meticulous obedience to moral maxims. (KdpV.152-53)

A little later he elaborates this idea, showing how the idea of freedom can exert an upward pull on the dissipated character.

If it is well-established, so that a man fears nothing more than to find himself on examination to be worthless and contemptible in his own eyes, every good moral disposition can be grafted on to this self-respect, for the consciousness of freedom is the best, indeed the only, guard that can keep ignoble and corrupting influences from bursting in upon the mind. (KdpV.161)

In the same work, Kant again gives a strikingly similar case study.

A man may dissemble as much as he will in order to paint his recollected unlawful behavior as an unintentional error, as mere oversight, which can never be entirely avoided, and consequently as something to which he was carried along by the stream of natural necessity, and in this way try to make himself out as innocent. But he finds that the advocate who speaks in his behalf cannot silence the accuser in him when he is conscious that at the time when he committed the wrong he was in his senses, i.e., he was in possession of his freedom. Nevertheless, he explains his misdeed by some bad habits which he has grown into by gradual neglect of attention to such a degree that he can regard the act as a natural consequence of them, but this cannot protect him from the blame and the reproach he casts upon himself. (KdpV.98)

Here Kant is again emphasizing the justification of blame directed at the heteronomous. But he reverts to the lesson of the liar in the first Critique in finding the dissipate free in his unfreedom, and blameworthy for precisely that reason. But by this point Kant may well think that he has shown some promising and effective ways for turning dissipates around, both through external causes benevolently applied (incentives and disincentives), and through the inward aspiration to virtue which never leaves even the most vicious character —except the one who has ceased to reason.

Case Study:  Daniel Dennett

In Elbow Room, Daniel Dennett describes a kind of free will sufficient for moral responsibility, but entirely compatible with a naturalistic world-view (37, 49, 76, 100, 137, 158, 165, 167, 169, 171),[Note 9] in which all things —except at the quantum level— are causally determined. Though we are determined, determinism does not negate control or self-control (72). Moreover, the control we can exert means that the past does not control us (72), that we can use our control to enhance our control (62, 63, 73, 85, 98-99, 100, 169), that our actions cannot be predicted, at least by ourselves (112, 113, 115, 152), and that the future is (epistemologically if not metaphysically) open (112, 113, 118, 139, 152).

In rejecting Kant's dual-aspect theory (26f, 28, 156) in favor of a nuanced determinism in which we are among the determinants of our own conduct, Dennett concludes that our characters are partly the product of determination by environmental and genetic causes, and partly the product of our own skills in making plans, pursuing them, evaluating and changing ourselves, and seeking and avoiding certain situations. These skills in turn have the same dual explanation. While our deliberation about our options and plans plays an ineliminable role in determining which life we live (102, 106, 108, 112, 129, 139), an element of luck cannot be denied either (92ff). Some are born with more opportunities, and some catch more lucky breaks during life. The element of luck, however, does not rule out moral responsibility (153, 164, 169).

We make ourselves and should be held responsible for the selves that we are as adults (74ff). But clearly luck helps some and hinders others in the task of making a moral self with a desirable measure of self-control and fellow feeling. How do we evaluate the case of a man who through bad luck and perhaps also bad will made himself more vicious than virtuous? This is the Kantian dissipate in Dennettian dress. Here is how Dennett discusses the case (italics are his):

"There but for the grace of God go I." (John Bradford (1510? - 1555) A very curious sentiment. It is often repeated in our own times with an even more curious meaning: indeed I could be so unlucky; I may yet suffer some such misfortune. But it would take an extraordinarily unlikely conspiracy of accidents (see chapter six) to turn me into (say) a murderer —while quite a likely, normal, everyday turn of events could make all the difference to a tough young hoodlum who has so far stayed out of serious trouble only by the skin of his teeth. The day someone happens to make the mistake of insulting him is the day his life of violent crime begins. If the sort of temptation that would turn him into a murderer were to flash before my eyes, however, I would almost certainly resist it. If I failed to resist it, I would have no one but myself to blame. I am supposed to be good at resisting such temptations —and in fact most of them are child's play. I'm so good at them, I don't even notice them as opportunities. Consider fine chess players, who never even notice the stupid move opportunities....It is not just luck that keeps them from making "patzer" plays. (99)

The hoodlum is less free than Dennett because the hoodlum can resist fewer temptations and provocations. This lower threshold of resistance is itself a product of their separate life stories: a mix of deliberate choices and luck. From his less free station in life, can the hoodlum lift himself? Can we blame him for not doing so?

The question is analogous to Kant's, but Dennett's answer differs, at least at first:

If the hoodlum is a patzer at life (and he is), this is too bad for him, and it may be just his bad luck that he is so bad at decision making. We do make allowances, however. We set the threshold of expectations lower for patzers. For children, for instance. We keep them out of situations in which their juvenile powers of deliberation might lead them into horrible, regrettable errors. What counts as luck for them is something that would not count as luck for us adults.

That is all very well —comes the objection— but what of those unfortunates who look like adults, and wander onto the adult playing field and into difficult circumstances where they make extremely regrettable decisions? If we can identify them, we ought to. And then we ought to treat them like children and excuse them. (99)

Dennett does not blame "patzers at life" for their bad decisions. But it is not crystal clear that he wouldn't blame the hoodlum. If the hoodlum's poor decision-making can be traced to his own bad decisions —rather than to bad luck— then Dennett might well blame him. Certainly some hoodlums fall into this category. Dennett is worried by the phenomenon of "creeping exculpation" (156ff) in which close examination of the "micro-details" of the life and times of every criminal defendant excuses them all (162) or leaves no one to blame (99, 157, 161).

The solution to the problem of creeping exculpation is not "Cosmic Responsibility" in the manner of Sartre or even Kant; for sometimes we are responsible and sometimes we are not. When we are responsible, "it will have to be a modest, naturalized, slightly diminished responsibility, for we are no angels" (158).

One way that Kant domesticated the paradox of liberation and prevented it from becoming a vicious circle was to posit an effect on the empirical character either from reason in general or from the idea of freedom in particular. Dennett tracks the same logic of the problem; for without following Kant in other details, he comes to a very similar conclusion. Dennett asks whether deliberation about our options can ever be effective, for it seems that determinism makes all deliberation futile (101, 102). He distinguishes determinism from fatalism (104, 123), and finds deliberation futile only for fatalism (106). Since there is no reason to think fatalism is true (104), he counsels us to act as if deliberation were effective (112, 113, 115). The result for one who takes Dennett's advice is the exercise of control and self-control, which is the analogue in his system of achieving negative and positive freedom or, in a word, liberation.

Case Study:  Mill

In On Liberty (1859), Mill shows that the paradox of liberation applies to political liberty as much as to individual freedom of will.[Note 10]

For Mill a free society is one which has overcome both the tyranny of bad laws and the tyranny of public opinion (4, 13, 28, 30f, 53, 64, 65, 67, 68, 69, 73, 78, 107). Tyranny in each case is to infringe a person's liberty except in order to prevent or punish harm to unconsenting others (9, 11). For the purpose of applying this "harm principle," offended sensibilities do not constitute harm (81-82, 83). To avoid this tyranny we must tolerate many things which are often regulated or prohibited: for example, actions of consenting adults which harm their own health, safety, happiness, virtue, or salvation (9). We must tolerate all forms of speech and expression (15ff) except those which are equivalent to actions in the harm they cause (53). We must tolerate the free circulation of falsehood, advocacy of crime, and all sorts of corrupt and immoral ideas, partly because they might be true and justified (16-33) and partly because the suppression of ideas known to be false or dangerous is worse than their circulation (33-43). We must also tolerate liberty in personal conduct (Mill would like the word "lifestyle"), even experimentation (53-4).

The benefit of all this toleration is greater likelihood that we will discover truth and correct error (18, 19, 20, 32, 36), reduced likelihood that hard-won truths will atrophy into "dead dogmas" (34, 38, 39, 40, 62), diversity (5, 12, 16, 19, 36, 43-46, 49, 54, 69, 71, 104, 108) and the free development of individuality (54-60, 60-67)

Mill was harsh on the England of his own day, 1859 (1, 8, 15, 21, 29, 37, 40, 43, 58, 63, 64, 71, 82, 84), even though he was aware of many more unfree societies from history (1-3). We may take any unfree society as the test case for liberation, but might as well take his own. His society had solved one of the ancient problems of liberty by replacing tyrants with elected and accountable public officials (2). But it had still not learned to prevent the tyranny of the majority (3-6) or the tyranny of public opinion (4, 30, 31, 32, 58).

So how does a society like Mill's —or our own— achieve (further) liberation? How do we advance the next increment of toleration and liberty?

Mill's answer is that the small-mindedness which criticizes individuality and difference, rather than tolerating them, slowly melts in the heat of vigorous public discussion. If many different standpoints are expressed, and many lifestyles are lived, then this form of public education will slowly overcome provincialism and its accompanying inclinations to censorship, intolerance, and paternalism (11, 61, 67-70, 104f, 113). In short, liberty improves us (10, 12, 15, 16, 18, 19, 32, 41, 46, 54, 60, 67, 94).

From another standpoint, Mill believed that intolerance was more natural than tolerance (8, 13, 30, 82). This made him somewhat pessimistic about human nature. But he was an optimist about what could be accomplished through liberty, especially in the way of cultivating toleration (8, 56, 63), even altering and improving human nature (56, 57, 63, 65, 67).

Liberty, then, is the path to greater liberty. Our best chance for expanding the scope of toleration and liberty in the future is to practice toleration and liberty now. Since every individual is a source of potential tyranny (through opinion), every individual must be educated to wider toleration; but this is best done by exposure to a diversity of standpoints and examples. We must be free to become freer.

The process is clearly circular. The only reason it is not vicious is that Mill does not call on the utterly unfree to exercise their freedom. We needn't be free before we are free. Instead, we must use the modicum of freedom we have to nurture the enlarged freedom we desire.

The alternative —coerced or involuntary liberation— Mill rules out for civilized societies. In calling for toleration of Mormons, Mill condemns a "civilizade" (90) aimed at their forced conversion to more liberal ideas. He disapproves "hero-worship" which calls on "the strong man of genius" to seize control of government for the good of all (64). He also condemns a moralistic temperance movement ready to use coercion to improve public morals (66). He praises those who resist coercive attempts to improve them (60-61, 81). Finally, if a civilized society is threatened by the spread of immoral ideas, and if its popular morality and educational institutions are not strong enough to protect it, tempting its leaders to turn to censorship and other forms of coercion, then Mill's judgment is merciless: good riddance. "[T]he sooner such a civilization receives notice to quit, the better" (91).

This may seem to do away with the "Leninist" model of involuntary liberation. But just as his theory of liberty applies only to adults, or to the mature (9, 10, 12, 55, 64, 74, 78, 80, 95, 100, 103), it applies only to societies which meet certain standards of advancement and civilization. For more primitive societies "there is nothing for them but implicit obedience to an Akbar or a Charlemagne if they are so fortunate as to find one" (10).

So Mill's position is subtly two-sided. For the civilized, or partly free, liberation requires freedom, manifesting the paradox; for the rest, unfree people are lucky if they can find their Akbar to liberate them.

For the civilized, liberation lies through paradox, but the paradox is not strong. We use our freedom to enhance our freedom —and if we don't, then good riddance to us. Over time, the increase in liberty transforms human nature from intolerance to tolerance. People in his day are "but starved specimens of what nature can and will produce" (56).

This suggests other ways in which the paradox of liberation might arise for political liberty. If the transformation of human nature induced by signing the social contract, and forming a society, were in fact needed to explain why people signed the contract in the first place,[Note 11] or if the transformation of human nature induced by a communist revolution were in fact needed to explain the motivation to revolt,[Note 12] then the liberation represented by the contract or revolution would presuppose just the freedom it makes possible. It is strictly analogous to the dissipate liberating himself with the freedom he doesn't yet have.

The Alternatives

If it is paradoxical to expect those in need of liberation to use their freedom to pursue it, then consider the alternatives. Is liberation really something we think can be accomplished for someone involuntarily?

There are many models of involuntary liberation. It could be done to or for the benighted by the enlightened. We can call this the Leninist model. In this sense, good parents and teachers are also Leninists. Or the liberator can be the benighted one, like the angry father who shouts to his formerly dependent, now pregnant, daughter, "You're on your own!"

Involuntary liberation does seem possible, even if faintly perverse, like commanding love, practicing spontaneity, or faking sincerity. Mill captured this ambivalence by wishing an Akbar on some societies but not on others.

But sometimes involuntary liberation is as unparadoxical as a warden releasing a prisoner who (regardless of his wishes) could not have done so himself. The Fifteenth Amendment to our constitution (Black Suffrage) had to be ratified by whites, and the Nineteenth Amendment (Women's Suffrage) had to be ratified by men. In both cases, however, long struggle in which blacks and women were prominent made the amendments politically urgent; they were not gratuitous.[Note 13]

But when freedom is a gift, it is 'involuntary' in the relevant sense; the recipient did not exercise her freedom to seize or create it. Descartes is only one philosopher who thinks his free will was a gift of God, and one he did not earn.[Note 14] Freedom can be a gift in another, more paradoxical way as well.

As the former colonies of England gained their independence, by war or negotiation, England looked for a way to solemnize the transfer of sovereignty. This is how a law-loving country admits defeat, formalizes its consent, and leaves the age of colonialism. In 1931 Parliament adopted the Statute of Westminster. Subsequently, when a colony or dominion was to be given its liberty, England proclaimed (under the terms of the statute) that it would never legislate for the colony or dominion again without its explicit request and consent. This is much tidier than simply losing on the battlefield or bargaining table, and that was the point.

It soon occurred to lawyers in England and in the former colonies that the Statute of Westminster was merely a statute, and could be repealed by Parliament. What would happen to a colony liberated under the statute if England later decided to repeal the statute? To many legal minds it looked as if, technically, sovereignty would revert to England. Are the liberated colonies, then, free only under the revocable command of England? England was in a dilemma. It wanted in good faith to relinquish sovereignty in the most complete and irrevocable way. But if the gift of liberation were done in law, then it could be undone; and if were not done in law, it would not be lawful.[Note 15]

This is the first layer of the paradox of liberation as a gift. The second layer occurred to those who tried to solve the problem at the first layer. It was tempting to interpret the Statute of Westminster as irrevocable, that is, beyond the power of Parliament to repeal. That would guarantee that the emancipated countries would stay emancipated. The problem is that irrevocable acts or immutable laws would contradict the independence of the English people, their sovereignty in their own country, and deny them the power to change their own laws. For these reasons, both England and the United States follow the fundamental constitutional rule that one Parliament or Congress cannot bind its successors irrevocably. Following this line of reasoning Parliament decided in 1935[Note 16] that the Statute of Westminster could in principle be repealed. So the paradox of liberation remains.[Note 17]

Deciding that the Statute of Westminster could be repealed was a solution satisfactory to England but not to the former colonies, of course. But in one case, it did some good. The South African constitution was written by the English Parliament in 1909 while South Africa was a colony of England. One clause (Section 35) prohibited racial discrimination. Section 35 was specially protected so that it could only be repealed by two-thirds of each house of the South African Parliament sitting in joint session. In 1952 the South African Parliament tried to repeal Section 35 by a simple majority, in order to institute apartheid. It argued that its intervening independence from England nullified the English restrictions on its sovereignty. The South African Supreme Court disagreed, citing the paradox of liberation: since the Statute of Westminster could be repealed, South Africa was not entirely independent of England.[Note 18]

If the Statute of Westminster could be repealed, then the burden fell to the former colonies to find ways to make their liberty more complete and final. The most elegant solution devised first by Rhodesia (now Zimbabwe), and subsequently imitated, was to insert a small irregularity into the ceremony of liberation so that the newly liberated nation could later claim that its independence derived, not from the Statute of Westminster or the grace of England, but from peaceful revolution.[Note 19]


On examination, Kant, Dennett, and Mill do not assert the paradox in an objectionable form. But they show that there are good reasons for holding that liberation is a circular process in which freedom can aid in its own realization, and should. Their reasons for coming close to paradox, and their strategies for evading it in the end make their positions more sophisticated and attractive than alternatives which evade the paradox by a country mile, e.g. by denying that liberation is ever necessary or by assuming that the unfree must always be liberated involuntarily by benevolent liberators.

There are many abstract approaches to the paradox of liberation. Looking them over, one begins to appreciate, perhaps more than before, the subset actually favored by Kant, Dennett, and Mill.

1. If no liberation is possible, then the paradox vanishes. If we are always free, or if we are never free, then liberation is impossible. If we act accordingly, then we won't blame people for not liberating themselves, and we won't exhort them to try. Our three authors all believe that both freedom and unfreedom are possible, in fact actual, and so do not avail themselves of this solution to the paradox.

Nevertheless, in some passages Kant seems to assert strongly that we are always free, minimally, such that liberation from zero is never necessary. These passages are contradicted by others, however, in which Kant seems to acknowledge that freedom and reason may decline to zero in the utterly depraved.

2. If responsibility does not require freedom, then much of the paradox vanishes. We may then hold the unfree responsible for anything —their dissipate crimes, their unfreedom itself— without violating basic norms of fairness. Our three authors do not go this far, but Kant and Dennett make moves in this direction. Insofar as Kant holds even the admittedly heteronomous dissipate responsible for his malicious lie, he is letting responsibility coexist with serious unfreedom. Insofar as Dennett wants to resist creeping exculpation in the teeth of the deterministic "micro-details" of the hoodlum's —and everyone else's— life story, the same conclusion follows.

Avoiding the paradox, then, is one reason to separate freedom and responsibility. However, this tactic carries the price that, beyond a certain point, it unfairly holds the unfree responsible for their actions. Rather than see this as a dangerous erosion of self-evident concepts of freedom and responsibility, however, we could see it as a case of multiple stories or discourses. For some purposes, we can see the hoodlum's crime as determined; as Dennett says, the micro-details supporting the heteronomy narrative are always available. For other purposes, we may want to hold people responsible for their actions regardless of these micro-details. We may use different explanations or descriptions of the same action to suit our different purposes at different times. Kant's dual-aspect theory is the classic of this genre.

Although Dennett wishes to distance himself from Kant's dual-aspect theory (Dennett 26f, 28, 156), he is very close to it. If we say (as a convenient simplification) that Kant's theory posits both a natural and a supernatural explanation for every action, then Dennett recognizes the natural explanation as well as the need for two explanations, although his second explanation is also a natural one. Kant and Dennett both assert that the "micro-details" of a defendant's life will always tell an exculpatory story by revealing that a given act was caused or determined. Further, like Kant, Dennett insists that there must be a second story to tell about the action, or else no act could be described as free; creeping exculpation would occupy the field. Dennett's theory is compelling chiefly because he takes pains to show that this second story need not be a supernatural one.[Note 20]

3. Involuntary liberation provides a way around the paradox, unless it is itself a greater paradox. Since the paradox only arises if one freely becomes free, then it is avoided if one is "determined" or even "doomed" to become free. Freedom might exist for competent adults, but not for children, because it depends on the skills, judgment, and self-control cultivated by socialization, which in its earliest stages is an unfree or involuntary process. Then there is the Leninist-Akbarite model of involuntary liberation. Dennett believes something like the former happens to all of us; Mill believes something like the latter happens, or ought to happen, to some of us. But even Dennett and Mill believe that free liberation is also possible, so they do not rely on unfree liberation as a way to evade the paradox.

Some forms of unfree liberation are paradoxical in their own right (liberation through the Statute of Westminster), but some are not (men ratifying the Nineteenth Amendment). We have no trouble acknowledging that the latter, and perhaps even the former, actually lead to an important kind of freedom. But they only solve the paradox of free liberation if we rely on the unfree methods exclusively, which Kant, Dennett, and Mill persuasively argue that we cannot and should not do.

4. Unfreedom can sometimes be the result of a free choice (if anything can be the result of a free choice). For example, one may choose to enter a contract, to marry, to enlist in the Army, to join a monastery, to become a heroin addict, or to commit suicide. This explains why we allow a "diminished capacity" defense to the insane but not to the person who voluntarily became drunk. Both had authentically diminished capacity, and to that extent, diminished responsibility. But the voluntary drunk chose his incapacity and can be held responsible for that choice.[Note 21] To the extent that we are beginning to allow diminished-capacity defenses for crimes committed while drunk, it is because we have begun to think of alcoholism as a disease rather than a vice —as involuntary rather than voluntary.

When unfreedom is chosen, as in voluntary intoxication, then it makes sense to blame the unfree for their unfreedom and to demand that they liberate themselves. Kant implies that this is always the case with adult dissipates, since they certainly had the use of reason earlier in life and can be blamed for the failure to cultivate it. On the other hand, if they lose the use of their reason through radical neglect, then he stops holding them responsible, even if their decline was initially voluntary.

Mill famously believes that "[i]t is not freedom to be allowed to alienate [one's] freedom" (Mill 101) and would not enforce contracts in which a person voluntarily became a slave. He is unclear, however, whether he believes that free alienation of freedom is descriptively impossible or normatively intolerable (or both). He says that the analysis which condemns slavery contracts has "far wider application" (Mill 102) but he doesn't spell out the implications of his position very far. It is clear from many other passages that he would both permit and encourage contracts, marriage, and enlistment, if not other forms of voluntary unfreedom.

Pliny the Elder even holds that the capacity for freely chosen unfreedom is distinctively human. In fact, it shows that we are freer than the gods, because we mortals can commit suicide.[Note 22]

5. In principle one could argue that freedom is stratified by a theory of types. Freedom1 is the kind we should choose to have for moral action; freedom2 is the kind needed for choosing freedom1; freedom3 is the kind needed for choosing freedom2, and so on. To evade the paradox we must also add that a person can possess freedom2 without freedom1 —or in general, freedomn+1 without freedomn. But this theory is even sillier applied to freedom than it was when Russell applied it to logical expressions. On this model, either the paradox will be replicated at level of freedom2 (or freedomn), or an infinite regress of freedoms will be needed. So either we don't solve the paradox or we commit ourselves to an absurdity greater than the paradox to be solved.

6. If the unfree can become free by acting as if they were free, then self-liberation would be possible and we could avoid the vicious circle of requiring freedom in order to have freedom. This position is difficult to flesh out in detail, although I've shown how Cabell tries to do so. Kant implies that something like this occurs when he argues that the idea of freedom (as opposed to freedom itself) can inspire actions or feelings which tend to counteract heteronomous impulses and work toward the achievement of negative freedom, an important kind of liberation. Dennett implies that something like this occurs when he argues that we should act as if deliberation were effective (even in a deterministic world), because only in this way can we gain the benefits of deliberation, which amount to control and self-control, his forms of freedom.

7. Unfreedom has degrees, or perhaps kinds. Hence, some unfree people have enough freedom to finish liberating themselves, and ought to do so.

But some do not. Aristotle judiciously discusses the two-sidedness of this in a case study in the Nicomachean Ethics:[Note 23]

Let us assume the case of a man who becomes ill voluntarily through living a dissolute life and disobeying doctors' orders. In the beginning, before he let his health slip away, he could have avoided becoming ill: but once you have thrown a stone and let it go, you can no longer recall it, even though the power to throw it was yours, for the initiative was within you. Similarly, since an unjust or self-indulgent man initially had the possibility not to become unjust or self-indulgent, he has acquired these traits voluntarily; but once he has acquired them it is no longer possible for him not to be what he is.

The paradox of liberation is only strongly paradoxical if we blame the radically heteronomous for not liberating themselves. But it is not nearly so strange to conclude that they bear some responsibility for their condition, or that at milder stages they are able to lift themselves up.

Kant, Dennett, and Mill reveal this aspect of their positions when they answer the basic fairness question. Why isn't it unfair to demand that the unfree free themselves? Kant's answer is that it is unfair to make this demand of people so unfree or dissipated that they have lost the use of their reason. But for others, however heteronomous they may be in other respects, the use of reason is enough to make self-liberation a realistic possibility; and this is enough to justify blame and exhortation.

Dennett's answer, similarly, is that the demand of self-liberation is unfair for the hoodlum who is like a child in relevant respects. But to act as if it were equally unfair for others, and to cease to demand self-liberation, would simply contribute to creeping exculpation.

For Mill, similarly, the demand is unfair for uncivilized societies who really need an Akbar or Charlemagne more than moralizing sermons. For the rest, self-liberation is a perfectly fair demand and part of treating civilized people as civilized.

To return to one of my original questions, liberation is a kind of bootstrapping process for these three philosophers, and it would be falsified if described as non-circular. The chief risk of the bootstrapping position —unfair demands on the unfree, or victim-blaming— can be mitigated if we recognize degrees, or kinds, of freedom, and if we only expect self-liberation from the middle degrees upwards, rather than from zero. For each of these three philosophers, there is a gift of freedom which gets us off zero, and puts us in the arena in which self-liberation is a legitimate demand. For Kant it is the use of reason. For Dennett it is luck. For Mill it is Akbar (for societies) and developmental maturity (for individuals).

The bootstrapping position has the virtue of denying some patently oversimplified views of freedom and liberation: that our freedom is always perfect, that our unfreedom is always perfect, and that our decisions and actions have no bearing on our degree of freedom. In short, the bootstrapping position is superior on the merits to its chief alternatives, that liberation is unnecessary, impossible, or involuntary. Whether a less paradoxical position has the same virtues, I leave to others to explore.


1. I will cite the following editions of Kant:

Foundations of the Metaphysics of Morals, trans. Lewis White Beck, Bobbs-Merrill, 1959 (original 1785). I refer to this work as "Gr." (for Grundlegung) and cite the Prussian Academy page numbers, which are included in the margins of this edition.

Critique of Practical Reason, trans. Lewis White Beck, Bobbs-Merrill, 1956 (original 1788). I refer to this work as "KdpV" (for Kritik der praktischen Vernunft) and cite the Prussian Academy page numbers, which are included in the margins of this edition.

Critique of Pure Reason, trans. Norman Kemp Smith, Macmillan, 1968 (original 1781, 1787). I refer to this work as "KdrV" (for Kritik der reinen Vernunft) and cite the second or "B" edition page numbers, which are included in the margins of this edition. [Resume]

2. Kant asserts this view in the second Critique at KdpV 20-21. However, in the first Critique he seems to deny it. After defining "transcendental" and "practical" freedom, essentially, as positive and negative freedom (B.561-62), Kant asserts that practical freedom is based on transcendental freedom, or negative on positive, and not the other way around (ibid.). Hence, another way for Kant to evade the paradox is to return to this doctrine from the first Critique —and at the same time to give it additional explanation and support. [Resume]

3. From an entry in James' diary for April 30, 1870, quoted by Ralph Barton Perry, The Thought and Character of William James, "Briefer Version," Harvard University Press, 1948, p. 121. [Resume]

4. Hans Vaihinger, The Philosophy of 'As If': A System of the Theoretical, Practical and Religious Fictions of Mankind, trans. C.K. Ogden, Routledge & Kegan Paul, second ed., 1935 (original 1911), pp. 47-48, 271-318. [Resume]

5. James Branch Cabell, Beyond Life: Dizain des Démiurges, Robert McBride and Co., 1929 (original 1919), p. 239. [Resume]

6. Cabell, ibid., p. 269; cf. p. 35. [Resume]

7. Cabell, ibid., pp. 34-35. [Resume]

8. James Branch Cabell, Sonnets from Antan, The Fountain Press, 1929, frontispiece. [Resume]

9. In this case study, all page numbers in the text refer to Daniel C. Dennett, Elbow Room: The Varieties of Free Will Worth Wanting, MIT Press, 1984. [Resume]

10. In this case study, all page numbers in the text refer to John Stuart Mill, On Liberty, Hackett Pub. Co., 1978 (original 1859). [Resume]

11. In The Social Contract (trans. Donald A. Cross, Hackett Pub. Co., 1987, original 1762), Rousseau suggests that humanity in the state of nature would perish if it did not undergo a change of its fundamental nature (23). One specific way to describe this transformation is to say that the "general will" came about as the result of making a social contract (23, 30.n). On the other hand, it is plausible to argue that the contract would never have been created if humanity had not already undergone this transformation. [Resume]

12. In The German Ideology (unabridged edition, trans. anon., Moscow: Progress Publishers, 1976, original 1845-46) Marx and Engels posit a transformation of human nature after the revolution (53, 59, 89, 97).

[A]s long as man remains in naturally evolved society [as opposed to a society created by human labor and will], that is, as long as cleavage exists between the particular and the common interest, as long, therefore, as activity is not voluntarily, but naturally, divided, man's own deed becomes an alien power opposed to him, which enslaves him instead of being controlled by him. (53)

For it is the association of individuals (assuming the advanced stage of modern productive forces, of course) which puts the conditions of the free development and movement of individuals under their control —conditions which were previously left to chance and had acquired an independent existence over against the separate individuals.... (89)

Communism differs from all previous movements in that it overturns the basis of all earlier relations of production and intercourse, and for the first time consciously treats all naturally evolved premises as the creations of hitherto existing men, strips them of their natural character and subjugates them to the power of united individuals. (89-90)

Only at this stage [of the abolition of private property] does self-activity coincide with material life, which corresponds to the development of individuals into complete individuals and the casting-off of all natural limitations. (97)

On the other hand, it is plausible to argue that proletarians would not rouse themselves to the energetic act of liberating revolution unless this transformation of the natural into the voluntary had already taken place. [Resume]

13. Actually, the issues of paradox here are more complex. For it has been argued that the Fifteenth Amendment (with the Thirteenth and Fourteenth) was imposed on the defeated southern states by the victorious north, after the Civil War, without the valid consent normally manifest by a vote of ratification. So to the black beneficiaries, the liberation by amendment was involuntary in one sense, and to the white southerners who resisted ratification, it was involuntary in another. See my Paradox of Self-Amendment, Peter Lang Pub. 1990, at p. 23, 99f, 430-31. [Resume]

14. Descartes, Meditations, trans. Laurence Lafleur in his collection Descartes' Philosophical Essays, Bobbs-Merrill, 1964 (original 1641), pp. 112, 115-16 (Meditation IV). [Resume]

15. In fact, Alf Ross has argued that sovereigns cannot release subjects, or parents children, by gift or command in this way. Alf Ross, "On Self-Reference and a Puzzle in Constitutional Law," Mind, 78 (1969) 1-24, at pp. 22-23. [Resume]

16. British Coal Corporation v. The King (A.C. at 520). [Resume]

17. See e.g. W.N. Harrison, "The Statute of Westminster and Dominion Sovereignty (II), Australian Law Journal, 17 (1944) 314; Geoffrey Marshall, Parliamentary Sovereignty and the Commonwealth, Oxford University Press, 1957, Chapter VI; Ilmar Tammelo, "The Antinomy of Parliamentary Sovereignty," Archiv für Rechts- und Sozialphilosophie, 44 (1958) 495-516, esp. 510-12; Geoffrey Marshall, Constitutional Theory, Oxford University Press, 1971, pp. 61ff.

Note that the statute written by the United States to grant independence to the Philippines suffers from the same problem as the Statute of Westminster. See The Paradox of Self-Amendment, op. cit., p. 258. [Resume]

18. Harris v. Dönges (1952) 1 T.L.R. 1245. For more detail, see The Paradox of Self-Amendment, op. cit., at pp. 82f. [Resume]

19. J.M. Finnis, "Revolutions and the Continuity of Law," in A.W.B. Simpson (ed.), Oxford Essays in Jurisprudence, Second Series, Oxford University Press, 1973, pp. 44-76, at p. 52. See The Paradox of Self-Amendment, op. cit., at pp. 131, 258. [Resume]

20. All Kant scholars I know would agree that, if both of Dennett's two stories are naturalistic, then Dennett disagrees to that extent with Kant. But in one passage, usually overlooked, Kant shows himself open to Dennett's thesis.

Whether reason is not...itself again determined by other influences, and whether that entitled freedom, may not, in relation to higher and more remote operating causes, be nature again, is a question which in the practical field does not concern us, since we are demanding of reason nothing but the rule of conduct. (KdrV B.831; italics his.)

Here I take Kant to be willing to leave the door ajar to the possibility that reason is "nature again" or that the freedom which is the causality of reason is just another form of natural causation. (On the possible "naturalism" of reason in Kant, see also KdrV B.xv, B.826, B.829, B.844-45.) [Resume]

21. Aristotle reports that in Athens "the penalty is twice as high if the offender acted in a state of drunkenness, because the initiative is his own: he had the power not to get drunk." Nicomachean Ethics, 1113.b.32f (Martin Ostwald trans., Macmillan, 1962).

One wonders: if voluntary drunks are held responsible for their unfree, drunken acts on account of their free meta-decision to become drunk, then are the involuntarily liberated not to be held responsible for their free acts on account of their lack of freedom at the meta-level? [Resume]

22. Pliny is quoted to this effect by Montaigne, "Apology for Raymond Sebond," The Complete Essays of Montaigne (original 1580, 1588), trans. Donald Frame, Stanford University Press, 1958, p. 393. [Resume]

23. 1114.a.15f, Martin Ostwald trans., Macmillan, 1962. [Resume]

Ribbon] Peter Suber, Department of Philosophy, Earlham College, Richmond, Indiana, 47374, U.S.A. Copyright © 1992, 1999, Peter Suber.