Individual Liberty Author(s): Hillel Steiner Reviewed work(s): Source: Proceedings of the Aristotelian Society, New Series, Vol.

75 (1974 - 1975), pp. 33-50 Published by: Blackwell Publishing on behalf of The Aristotelian Society Stable URL: http://www.jstor.org/stable/4544864 . Accessed: 02/02/2012 11:17
Your use of the JSTOR archive indicates your acceptance of the Terms & Conditions of Use, available at . http://www.jstor.org/page/info/about/policies/terms.jsp JSTOR is a not-for-profit service that helps scholars, researchers, and students discover, use, and build upon a wide range of content in a trusted digital archive. We use information technology and tools to increase productivity and facilitate new forms of scholarship. For more information about JSTOR, please contact support@jstor.org.

The Aristotelian Society and Blackwell Publishing are collaborating with JSTOR to digitize, preserve and extend access to Proceedings of the Aristotelian Society.

http://www.jstor.org

III*-INDIVIDUAL

LIBERTY

by Hillel Steiner
An' individual is unfree if, and only if, his doing of any action is rendered impossible by the action of another individual. That is, the unfree individual is so because the particular action in question is prevented another. In the following essay I shall, by first, briefly defend this 'negative' conception of individual liberty, and then proceed to elicit several of its implicationsparticularly those which touch upon our understanding of the relation between liberty and threats. The natureof my argument will be such as to suggest that many of the kinds of circumstance in which an indiviidualis said, by the proponentsof the negative conception, to lack the liberty to do a certain action, cannot be held to be so without self-contradiction.Arguments about the nature of individual liberty-and they are legion-are usually disputes concerning either the relation between a prevented action and its subject, or that which is to count as prevention. Quite clearly, the two issues are connected. Hence what occasions this essay is my belief that many writers who have argued for what I take to be the correct position on the first issue, have nevertheless failed to draw the appropriate conclusions concerning what is to count as prevention. In so doing they have failed to appreciate an important aspect of the concept of individual liberty itself. My defence of the negative conception will thus be 'brief' inasmuch as I shall only cursorily rehearse the arguments establishing the correct position on the relation between prevented actions and their subjects, and shall refer the reader to those writings in which these arguments are set out in greater detail. I Sir Isaiah Berlin, in the introduction to a revised version of his lecture "Two Concepts of Liberty", undertakesto correct what
1 I am particularly indebted to G. A. Cohen for his comments on an earlier draft of this paper.
* Meeting of the Aristotelian Society at 5/7, Tavistock Place, London, W.C.I, on Monday, iith November 1974 at 7.30 p.m.

33

34

HILLEL STEINER

he considers to be an error in the original version.2 In that earlier version Berlin had argued that liberty, properly understood, consists in not being prevented by other persons from to doing whatever one desires do, and thus that one is free to the degree that one is not prevented by another from doing what ones desires to do. Berlin rightly acknowledges that this formulation permits the unacceptably paradoxical (and positive libertarian) inference that one can increase the extent to which one is free simply by suppressing those of one's practical desiresthe satisfactionof which is prevented by others. It permits the inference that ultimately one is one's own gaoler, so to speak. As J. P. Day has pointed out, ridding oneself of the desire to do an action which is prevented by another, does not render one free to do that action.3 Since the question of whether one is prevented from doing a particular action can always be said to arise in regard to actions of a kind which one is able to do, it is absurd to suggest that the extent of one's liberty can be increased by increasing the number of instances in which the question of whether is free does not arise. The class of cases in one which this question does not arise clearly includes those kinds of action which one is unable to do. The conception of liberty as the absence of prevention of only actuallydesiredactionspermitting, as it does, the aforementionedinference about the expansion of liberty-logically requires that we extend this class to include those actions which one has no actual desire to do. On this suggestion, a necessarycondition of our being either free or unfree to do an action is not merely that we are able to do that kind of action, but also that we in fact want to do it. But to assert this is to confuse the condition of 'being free' with that of 'feeling free'. For if there are personswho make it impossible for me to import cannabis into this country, I am unfree to do so irrespective of whether I want to do so, am indifferent to doing so, or want not to do so.4Being placed in a locked prison cell rendersme unfree to go to the theatre regardlessof whether I want to go to the theatre or not.
2 Published in his FourEssayson Liberty,(Oxford, 3 "On Liberty and the Real Will", Philosophy, vol.
I92; p. 191.

1969), pp. xxxviii-xl. XLV (I 970), pp. 177-

The example is Day's; id., p. 179.

INDIVIDUAL LIBERTY

35

Obviously the extent to which such prevention engenders in me a feeling of frustration,the extent to which I experience it as an obstacle to my satisfaction or contrary to my interests, does depend on what I actually desire or want to do. Perhaps the only freedom that matters to me is the freedom to do what I desire to do or believe I ought to do. But it does not follow from this that I can only be free or unfree with regard to those actions which I want or believe I ought to do. For I can equally be free to do actions which I do not want to do. It is not unintelligible-on the contrary,it makesperfectsense-to assert assert that 'I am free to do A, i.e., am not prevented from doing A, though I have no desire to do so.' Again, it is perfectly intelligible to say that 'I am unfree to do A, and have no desire to do so.' The same may be said of actions whose relation to their subject is defined in normative terms. To ask whether an individual is free to do A, is not to ask a moral question. It is, rather, to ask a factual question the answerto which is logically prior to any moral question about his doing A. Indeed, it is difficult to comprehend how one could perform an action which one ought not to perform-a wrong action-unless one is free to do it, not prevented from doing it. Thus it is mistaken to imagine that "our conception of freedom is bounded by our notions of what might be worthwhile doing".5 For such an argument implies inter alia that "incomprehension, not hostility, is the first obstacle to toleration" (ibid.) Whereas, apart from the tautologous character of the suggestion that comprehending the (possible) value of an action is a reason for finding it worth while, there is absolutely no reason to suppose that we are incapable of tolerating actions the worthwhileness of which we do not accept. It follows from these considerations that statements to the effect that 'X is free to do A' do not imply or presuppose statements to the effect either that 'X wants to do A' or that 'X has no obligation to do not-A'.Nor, therefore, do they imply or presuppose statements about what X 'really' wants or about what it is in his 'real' interest to do or have done to him. Judgments about whether an individual is
5 S. I. Benn and W. L. Weinstein, "Being Free to Act and Being a Free Man", Mind, vol. LXXX (I 97I), pp. I94-21 1; p. 195.

36

HILLEL STEINER

free to do a certain action do not presuppose any judgment concerning either his desires or his obligations. II Suppose that I am offered a teaching post at a university other than the one which at present employs me. Suppose, further, that the duties and privileges attached to the offered post are quite similar to those pertaining to my present post, except in this respect: that the offered salary is considerablygreater than my present one. Suppose, finally, that I am not averse to receiving a higher salary and, indeed, would positively welcome it. Is there some significant sense in which this offer has rendered it lmpossible me to remain in my present post and for to reject the offered one? Alternatively, suppose that I have no offer of a teaching post at a university other than the one which at present employs me. Suppose, further, that the relevant university authorities have informed me that unless I substantially increase the amount of teaching I am to do in the next academic session and, moreover, undertake to teach several courses in subjectsunrelated to my own, my contract of employment will not be renewed. And suppose, finally, that I entertain considerable doubt as to the conceptual soundnessof these prospective courses, that I am therefore averse to teaching them, and that in any case I am loath to surrender still more of my time to teaching as I much prefer to spend it reading. Is there some significant sense in which this threat has rendered it impossible me to remain in my present post and for to renew my contract? Offers and threats are interventions,by others, in individuals' practical deliberations. They are intended by their authors to influence how a recipient individual behaves, by altering the extent to which he actually desiresto do a particular action of a kind which he is able to do. If the intervener is correct in his assessmentof the desiresof the recipient, and if he has designed his intervention accordingly, he necessarily succeeds in bringing about the intended alteration in the recipient's desire to do the particular action in behalf of which the intervention is made. However, despite this shared characteristic of interventions which are offers and interventions which are threats,

INDIVDUAL LIBERTY

37

few writers who subscribe to the negative conception of personal liberty contend that the making of an offer constitutes a diminution of the liberty of its recipient; while many of them would insist that a threat does so constitute. (Positive libertarians allow that both offers and threats, as heteronomous influences, may diminish personalliberty and they tend to suggest that the distinction between the two is therefore of little moment.) Thus we are faced with four questions. What, if any, are the grounds for distinguishing those interventions which are offers from those which are threats? If such a distinction can be established, does it imply a difference between the ways in which offers and threats, respectively, affect the practical deliberationsof their recipients? If such a difference exists, does it constitute a reason for asserting that threats, but not offers, diminish personal liberty? If such a difference does not exist, can we nevertheless claim-as do positive libertarians-that both offers and threats diminish personal liberty? In pursuing answers to these questions I shall put aside the further complications which could be introduced into the discussion by a consideration of the obvious truth that what counts as a threatening intervention to some individuals may often count as an offer to others. Attaching the intervening consequence, of accommodation in a gaol cell, to the action of sleeping on a park bench at night, may well constitute an offer to vagrants while at the same time constituting a threat to other members of the public. Similarly, what counts as a strong threat or offer to some individuals may constitute only a weak threat or offer to others. Interpersonal variations of these kinds-whether between different recipients or between a recipient and an intervener---though important for the purposes of some discussions, are not relevant to this one. Such considerations can therefore be excluded by adopting the assumption that everyone knows the nature and extent of the desiresof everyone else, and intervenes accordingly. Cinema-goers will doubtless recall a recent popular film concerning the Mafia in which the padrone, periodically confronted with an unco-operative business associate, declares his intention of making the recalcitrant "an offer he can't refuse". The amusing irony of this turn of phrase might

38

HILLEL STEINER

understandablybe taken as proof that we are all reasonably able to distinguish an offer from a threat, because we all know the difference between a benefit and a penalty. But if a distinction of this kind can be drawn, it cannot be done simply upon such grounds as these. For it is true of both offers and threats that compliance promises to make one better off than non-compliance, i.e., that for both offers and threats, there is a clear sense in which compliance is seen to involve beneficial consequences and non-compliance to involve penal consequences. So the differences which must exist if a distinction is to be drawn between offers and threats are those (I) between the benefits conferred by compliance with an offer and a threat, respectively, and correspondingly (2) between the penalties incurred by non-compliance with an offer and a threat, respectively. It is not necessary to rehearse the accounts provided by the growing body of literature on this subject, to appreciate that an affirmation of the existence of such differences logically presupposesa conception of 'normalcy'into which the threatening or offering action is taken to be an extrinsicintrusion.6That such a presupposition is required is evident from the fact that the casual distinction commonly drawn between offering interventions and threatening interventions-that compliance with the former results in an augmentation of well-being while non-compliance with the latter results in a diminution of well-being-tends to obscure the point that non-compliance with offers results in a relative diminution of well-being while compliance with threats results in a relative augmentation of well-being. To establish the distinction between offers and threats it is thereforenecessaryto establish that the complianceconsequences of the former and the non-compliance-consequences of the latter are not merely relative augmentations and diminutions (respectively) of well-being, but absolute ones. And this presupposes a standard or norm from which such
6 Cf. Robert Nozick, "Coercion", in Philosophy, Science Method:Essays and in Honor ErnestNagel, (eds.) S. Morgenbesser, P. Suppes and M. White, of (New York, I969); Harry G. Frankfurt, "Coercion and Moral Responsibility", in EssaysonFreedom Action,(ed.) T. Honderich, (London, I 973); of and the papers by M. D. Bayles, B. Gert and V. Held in Nomos XIV: Coercion, (eds.) J. R. Pennock and J. W. Chapman, (Chicago, 1972).

INDIVIDUAL LIBERTY

39

consequences are judged to be departures. In the literature, the conception of the norm to be employed for this purpose is the description of the normal and predictable course of events, that is, the course of events which would confront the recipient of the intervention were the intervention not to occur. (Thus a shop-keeperis not threatening his customers when he raises his prices during a generally inflationary period.) Given this conception of the norm, we get the following configuration of alternative consequences: for an offer-'You may use my car whenever you like'-the compliance-consequencerepresents a situation which is preferred to the norm, while the noncompliance-consequencerepresentsa situation on the norm, no more or less preferred than it because identical to it; for a threat-'Your money or your life'- the compliance-consequence representsa situation which is less preferredthan the norm (no money), but the non-compliance-consequence represents a situation which is still less preferred (no life). We can, in addition, distinguisha third kind of intervention which I shall call a 'throffer',e.g., 'Kill this man and you'll receive CIoo -fail to kill him and I'll kill you'. Here the complianceconsequence represents a situation which is (let us suppose) preferred to the norm, while the non-compliance-consequence represents a situation which is less preferred than the norm. This configuration can be displayed diagranmmatically: ascending degrees
of .I

.5
.2

norm .3 .4 .6

desirability

where the vertically-ordered pairs of points represent the alternative consequences posed by offers, threats and throffers, respectively; and where the odd-numbered points represent compliance-consequences, even-numbered points representing non-compliance-consequences.Hence it would appear that the answer to our first question is an affirmation that we can distinguish offers from threats, and that the grounds for doing

40

HILLEL STEINER

so consist in the fact that the alternative consequences posed by the former occupy a different position relative to the norm than do those posed by the latter. We may now considerour answerto the second question in the light of this distinction. Does this distinction, between those interventions which are offers and those which are threats, imply any differencebetween the ways in which each affects the practical deliberations of their recipients? The short answer to this question is 'No'. The way in which both offers and threats affect the practical deliberations of their recipients consists in the reversal of the relative desirability of doing a particular action with that of not doing it. Whereas in the normal course of events-in the absence of an intervention-X's desire to do A is greater than his desire to do not-A, in the presence of an intervention his desire to do A is less than his desire to do not-A. Now what is consequentialfor the deliberationsof the recipient of an intervention is not whether the pair of alternatives confrontinghim is above (and on) or below the norm. Rather it is the fact-true of both offers and threats-that compliance leaves him in a more desiredposition than does non-compliance. The modus operandi an intervention-its method of promoting of a compliant response-consists in effecting a positive remainder when the degree of desirability attached to the non-complianceconsequence is subtracted from that of the complianceconsequence. This is true irrespective of whether that pair of consequences lies above (and on) or below the norm, that is, irrespectiveof whether that intervention is an offer or a threat. And while it is necessarily true that an action complying with an offer is more desired than an action complying with a threat, it is very far from being necessarily true that the difference in desirability between compliance and non-compliance with offersis of a lesser magnitude than the correspondingdifference pertaining to threats. This means, as will be shown, that it is not necessarily true that offers are more resistible or exert less influence than threats. With respect to any intervention, it is the existence of this difference which affects the practical deliberations of the recipient, and not the kind of interventioninvolved. If (and only if) this argument is correct, it should be true that the factor determining the strength of a recipient's desire to comply with an intervention is the magnitude of this difference,

INDIVIDUAL LIBERTY

41

and not the position of either of its consequencesrelative to the norm. That this is indeed the case can be seen by comparing the following threatening interventions: (i) Give me /Jioo or I shall kill you; (ii) Give me Li ooo or I shall kil you; (iii) Give me Liooo or I shall kill you and your brother; (iv) Give me /ioo or I shall kill you and your brother. Making all the usual (though by no means incontrovertible) assumptionsabout individuals' relative preferences concerning money, personal survival and fraternal welfare, we can readily see that the desire of a recipient to comply would be greatest in the case of (iv) and least in the case of (ii). (Whether his desireto comply would be greater or less in the case of (i) than of (iii) is undecidable on these preferenceassumptions.)What this indicates is that the strength of a threat is not a function of the desirabilityof its compliance-consequencerelative to that of the norm: (ii) is weaker than both (i) and (iv). Nor is the strength of a threat a function of the desirability of its non-complianceconsequence relative to that of the norm: (iii) is weaker than (iv), and (ii) is weaker than (i). Differences in degree of desirability between consequences and the norm are utterly irrelevant in assessing the strength of a threat. All that is relevant is the difference in degree of desirability between compliance- and non-compliance-consequences. that respect, In it is not strictly mistaken-as it is in the case of threats-to claim that the strengthof an offer is a function of the desirability of its compliance-consequencerelative to that of the norm. But this is not a reason to suppose that the strength of offers is determined by considerations different from those of threats, i.e., that their respective strengths are incommensurable. It is merely an analytic fact that the non-compliance-consequenceof an offer lies on the norm. Its strength, like that of other interventions, is purely a function of the difference in desirability between the two alternative consequences.That this is indeed a rule covering all interventionsis also to be seen in a comparison of the strength of the following throffers: (i) Do A and I shall give you Ji oo-fail and I shall kill
you;

(ii) Do A and I shall give you ?iooo-fail

and I shall kill

42

HILLEL STEINER

(iii) Do A and I shall give you Lioo-fail and I shall kill you and your brother; (iv) Do A and I shall give you ?Iooo-fail and I shall kill you and your brother. Again, making all the usual assumptions about relative preferences,it is clear that the greatest desire to comply arisesin (iv) and the least in (i), with (ii) and (iii) in the middle position (and not susceptible of mutual ranking on these assumptions). This ranking, in terms of capacity to affect the desire to comply, exactly correspondsto the ranking of these throffersin terms of the difference of desirability between their alternative consequences. It does notcorrespond to their ranking in terms either of the differenceof desirabilitybetween their complianceconsequences and the norm, or of the difference of desirability between their non-compliance-consequencesand the norm. There is one further point which requires to be made. The preceding discussion of offers and threats has been in terms of how they affect their recipients' desiresto do and not do a particular action. It is equally possible, however, to reformulate the discussionin termsof the effect of such interventionson their recipients' obligations do and not do a particular action. to Whereasin the normal course of events X may have a duty to do A, in the presence of a circumstancecreated by an intervention, he may have a duty to do not-A. The only difference between the descriptive account and the prescriptiveone is that, in the latter case, the reversalin the desirabilityof the two alternatives is not a matter of degree: interventions, in the prescriptive account, do not make compliance more desirable and noncompliance less desirable. Rather, compliance becomes obligatory and non-compliance prohibited. The reversal in the prescriptive account is, as it were, one of quality rather than quantity. This, however, does not alter the point that whether interventions are spoken of as affecting desires or obligations, the ways in which these are affected are the same-namely, by the reversalof the desirabilityof a complying action with that of a non-complying action. of Brieflythen, both the modus operandi an intervention and its strength are specifiable without reference to the norm. Since it is in the concept of the norm that the distinction between offering and threatening interventions is grounded, we may

INDIVIDUAL LIBERTY

43

conclude-in answer to the second question-that there is no difference between the ways in which offers and threats respectively affect the practical deliberationsof their recipients. And this provides us with the answer to the third question, as well: since no such difference exists, it cannot constitute a reason for asserting that threats, but not offers, diminish since there appearsto be no other personalliberty.Furthermore, way that threats can be said to affect personal liberty-other than through their effect on the deliberationsof their recipients -there is no reason to believe that, if they do affect it, these effects are different from those of offers. We have now to consider the answer to the fourth question which asks whether, in the absence of such a difference, it is neverthelesspossible to claim-as do positive libertarians-that both threats and offers diminish personal liberty. We have already seen that statementsto the effect that 'X is free to do A' do not imply or presupposestatements to the effect either that 'X wants to do A' or that 'X has no obligation to do not-A'. Interventions of an offering or threatening kind effect changes either in individuals' relative desires to do certain actions or in the evaluative status assigned to their doing certain actions. Whereas in the normal course of events it might be the case that 'X wants to do A' or 'X has no obligation to do not-A', the occurrence of an intervention may cause it to be the case that 'X wants to do not-A' or 'X ought to do not-A'. But neither of these latter two statements, nor the fact that they are true as a consequence of another's intervention, entails that 'X is unfree to do A'. They do not imply that 'X doing A' is rendered impossible. It is, of course, not disputed that the truth of the first of these two statements rules out the possibilityof 'X doing A eagerly' and that the truth of the second rules out the possibilityof 'X doing A justifiably'. But that is another matter. Hence it would appear that neither the making of threats nor that of offers constitutes a diminution of personal liberty. Intervention does not count as prevention. The argument to the contrary-that r's intervening action B, in behalf of 'X doing not-A', does render 'X doing A' impossible-presupposes that rendering a compliant action (not-A) more desirable than its non-compliant alternative (A), entails rendering the latter impossibleand the former,therefore,

44

HILLEL

STEINER

necessary.And this in turn presupposesthat only that one which is the more desirable of two alternative courses of action, can be done. But if this were true, then r's intervening action .B must have been more desirable than not-B. And this would imply that ' Ydoing not-B' was impossible and that 'T doing B' was necessary. But if this were so, then 'r doing B'-as a necessary occurrence-must itself be part of the normal and predictable course of events, since it is analytically true that all necessary events are inevitable events and all inevitable events are predictable events. In which case, however, 'Y doing B' cannot be construedas an intervention. Thus the argument that intervention is prevention is self-contradictory, because its proponents are logically committed both to affirming and to denying that an intervening action is part of the normal and predictable course of events. This contradiction seems to me to be implicitly present in the political writings of many of those who defend the positive conception of individual liberty. It is thereforeall the more surprisingthat it is also to be found in the opposed conception presented by some negative libertarians. III The preceding arguments have been brought in support of a single claim: that since an individual is unfree to do-is prevented from doing-a particular action if and only if the action of another renders it impossible for him to do it, an action on the part of one individual in behalf of intervening another's not doing an action does not render the latter unfree to do that action. The interventiondoes not count as the prevention of his doing that action. We have now to consider what doescount as prevention. Prevention is a relation between the respective actions of two (or more) individuals such that the occurrence of one of those actions rules out, or renders impossible, the occurrence of the other (or others). If there are two individuals' actions which can both occur, neither can be preventative of the other. Hence what we want to know is the kind of condition under which either of two individuals' actions can occur, but not both. Acknowledging the immense diversity of actions and of the circumstancesof their prevention, can we neverthelessspecify a

INDIVIDUAL LIBERTY

45

universally-valid description of the conditions of prevention? The grounds for an affirmative answer to this question should furnish us with the conceptual equipment to formulate more positively what it is to be free to do a particular action. Consider the case of an individual incarcerated in a locked gaol cell which is ten feet high, wide and long, which is devoid of any furniture or fittings, and for the lock of which he lacks a key. There is, we might say, an indefinitely long list of actions which this individual is prevented from doing. It is also true that there is an indefinitely long list-though not as long as the previous one-of actions which this individual is not prevented from doing. He is not prevented from jumping up and down, nor from singing Waltzing Matilda,nor from twiddling his thumbs in a clockwise direction, nor from twiddling his thumbs in a counter-clockwisedirection, and so forth. Now consider the change that would be wrought, in the extent to which he is subject to prevention, were his gaolers to place in his cell a (ventilated) mummy-case and to lock him inside it. We should say that his list of prevented actions, however indefinitely long it had been, would lengthen; and his list of unprevented actions would shorten. It is true, however, that there would now (in the mummy-case) be certain actions possible for him to do do which were not so before. Before, he was prevented from, among other things, rubbing his foot against the inside of a mummy-case. Indeed, one could compile a considerable inventory of actions now open to him by virtue of his access to the mummy-case, which were previously rendered impossible by the denial of such access by his gaolers. Hence, in order to establish a clear-cut comparison between any two hypothetical situations in terms of the relative amount of prevention each would involve, we must eliminate as many differencesbetween them as possible, without rendering them exactly alike. Let us say then, that in the first situation the incarcerated individual finds himself in the aforementioned locked cell, which also contains a mummy-case which is not locked though which he can lock from inside. And in the second situation the individual is locked inside the mummy-case (not lockable/unlockable from inside) which is, in turn, located within the locked cell. It seems clear that however indefinitely long are the lists of prevented and unprevented actions respectively pertaining to

46

HILLEL STEINER

the individual in each of these two situations, the extent of prevention is greater in the second than in the first. Next, compare the extent of prevention obtaining in the case of an individual confined in a cell like the one just mentioned and which is devoid of any furnishings,to that obtaining in the case of an individual similarly confined but who can secure writing materials for limited periods of time when he requests them from his gaolers. We should not hesitate to say that prevention is greater in the former case than in the latter. A similarjudgment would be rendered in comparing the circumstance in which an individual is compelled to pay a fine of /JIooo, with that in which he is fined only JI oo. For even if the money economy in which he lives and works were to cease to exist while he was still in the court-room, there would still be more actions open to him were he to be deprived of only ?ioo than there would be if he were deprived of fiooo. Again, an individual is more free if he is chained to a dungeon wall by a shackle on only one wrist, than if both wrists are shackled. And finally, the number of actions rendered impossible for one individual by another, is less if the preventor has crippled only one of his victim's legs than if he has crippled both of them. In all of these cases we should, of course, be hard pressed to specify preciselythe extent to which one individual's action prevents the otherfromacting. This is because the number of actions which the prevented individual is and is not thereby prevented from doing, is incalculably great. Nevertheless,the fact that this number cannot be specified does not constitute an insurmountable obstacle to any furtheranalysisof the manner in which one action may stand in a preventative relation to others. For the fact that we are able to compare at least some hypothetical situations where prevention occurs, and to form judgments as amounts of prevention respectively obtaining in to the relative these compared situations, indicates that-despite the vast diversity of preventative conditions-there is some limitedly quantifiable common element present in them. The reason why we judge an individual to be subject to less prevention in the cell with the unlocked mummy-case than in the cell with the locked one is, obviously enough, that he is unpreventedfrom doing all those actions which would be open -to him were he to be locked inside the case, as well as others

INDIVIDUAL LIBERTY

47

which would not be open to him were he so confined. Yet upon what grounds is this comparativejudgment made? What is the nature of the difference, between these two situations, which enables us to claim with complete confidence-and in the absence of an actual comparative inventory of prevented actions-that the one allows of greaterfreedomthan the other? The differenceis, simply and solely, that in the formersituation the incarceratedindividual can make use of a greater amount of physical space and material objects than his confinement in the locked case would permit. No other difference exists between these two situations. The same kind of claim can be made about the other hypothetical situations compared above. In other words, the greater the amount of physical space and/or material objects the use of which is blocked to one individual by another, the greater is the extent of the prevention to which that former individual is subject. This is because to act is, among other things, to occupy particularportionsof physical space and to disposeof particular material objects including, in the first instance, parts of one's own body. I shall call the particular portions of physical space occupied in a particular action, and the particular material objects disposed of in that action, the 'physical components' of that action. Thus, pursuing the universally valid description desiderated at the beginning of this section, the kind of condition under which the occurrence of one action renders impossible the occurrence of another is that at least one of the physical components of one action is (simultaneously)identical with one of the physical components of another. If two agents' respective actions (simultaneously) have no common physical components, there is no reason why they cannot both occur. It follows that to prevent an individual from doing a particular action is (simultaneously) to occupy and/or to dispose of at least one of the physical components of that individual's action. To be free to do A therefore entails that all of the physical components of doing A are (simultaneously) unoccupied and/or undisposed of by another. The relation between an agent and a portion of physical space which he occupies, and between an agent and a material object of which he disposes, is commonly called possession. An individual is said to possessan object when he enjoys exclusive

48

HILLEL STEINER

physical control of it, that is, when what happens to that object-allowing for the operation of the laws of physics-is not subject to the determination of any other agent and is therefore subject only to his own determination. Possession is thus a triadicrelation obtaining between an agent, an object and all other agents. Statements about the freedom of an individual to do a particular action are thereforeconstruable as claims about the agential location of possession of the particular physical components of that action. The statement that 'X is free to do A' entails that none of the physical components of doing A is possessed by an agent other than X. The statement that 'X is unfree to do A' entails that at least one of the physical components of doing A is possessedby an agent other than X. My theorem is, then, that freedom thepersonal is possession physical of objects. At least one interesting inference may be drawn from this theorem. It has to do with what is implied by any statement about either the expansion or diminution of his personal liberty that may be experienced by an individual. If X's freedom consistsin the physical objects X possesses,any expansion in his freedom must consist in an increase in the physical objects X possesses. But if a physical object P is in X's possession, it cannot be in the possessionof any agent other than X. In this circumstance, another agent r is prevented from doing any action of which P is a physical component. r is unfree to do any action of which one or more of the physical components are possessed by X. If there were only two agents, X and 1, the extent of X's freedom and of r's unfreedom would both be functions of the extent of X's possessions.Any expansion in the freedom of X would constitute a diminution in the freedom of r: it would extend the list of actions which r is prevented from doing. In a universeof more than two agents, any increasein the number of physical objects controlled by one agent must constitute an increase in the number of physical objects the control of which is denied to other agents. Conversely, any decrease in the number of physical objects controlled by one agent, must constitute a decrease in the number of physical objects the control of which is denied to other agents. This much at least is analytically true and, perhaps, reasonably obvious.

INDIVIDUAL LIBERTY

49

Hence it is often asserted, with some justification, that the paradigminstanceofbeingunfree is thatin which an individualis imprisoned. Certainly it is true that, for most people, imprisonment involves a very considerable decrease in the amount of physical objects they control. (Where it does not, imprisonment may fail to penalize.) And, in the case of any one individual, this decrease implies a correspondingincrease in the amount of physical objects over which other individuals enjoy control. Nevertheless the paradigmatic character of imprisonment is doubtful since, as was noted previously, certain actions are possible even in prison and, to that extent, a prisonerdoes enjoy control over some physical objects. Thereforethe true paradigm of prevention, the condition under which an individual is maximally unfree, is that in which another individual controls his voluntary nervous system and thereby rendersit impossible for him to dispose of the various parts of his body in a manner appropriateto the doing of any action whatever. In such a case it is readily apparent that the diminution in the extent of control enjoyed by the one individual corresponds to the expansion in the extent of control enjoyed by the other. It does not stretch our conceptual capacities too far, even if it is somewhat unidiomatic, to say that the latter possessesthe body of the former. Of course, most instances of prevention are rather less drastic and thus less thoroughgoing. But the paradigm does serve to exemplify the nature of the relation obtaining between the extent of one agent's freedom and that of others. Berlin observes, in a figurative vein, that " 'Freedom for the pike is death for the minnows'" and interprets this epigram literally to mean that "the liberty of some must depend on the restraint of others".7It is thus inconsistent as well as mistaken to suggest, as he does just slightly further on in his argument, that there can be circumstancesin which "an absolute loss of liberty occurs", i.e., that one individual can lose freedom without thereby increasing the individual liberty of others (id., p. 125) . Within the universe of agents, that is, within the class of beings who count as authorsof actions and who are thereforethe
7 Op. cit., p. 124; see also S. I. Benn and R. S. Peters, SocialPrinciples and theDemocratic State, (London, I 966), p. 213.

50

HILLEL STEINER

subjects of statements concerning freedom and prevention, there can be no such thing as an absolute loss of (or gain in) individual liberty.