While on the subject of Robert Nozick, we might note that he’s been written up this week in Slate, in an article by Stephen Metcalf. It’s a pretty feeble piece – gratuitously snotty, philosophically shallow, and lame even as mere journalism insofar as its central “hook” is just wrong. Contrary to what Metcalf supposes, Nozick did not renounce libertarianism. In fact he explicitly denied doing so in an interview with Julian Sanchez given not long before Nozick’s death in 2002 (as Sanchez reminds us in responding to Metcalf). Like too many critics of Nozick, Metcalf also focuses exclusively on his famous “Wilt Chamberlain argument” (and, as Sanchez notes, badly misses the point of it). That argument is indeed important, but Nozick gave other arguments too, some of them no less interesting. Consider, for instance, the argument implicit in his thought experiment “The Tale of the Slave.”
You can find the thought experiment on pp. 290-92 of Anarchy, State, and Utopia. You can also find it online here. It’s brief – give it a read, then come back.
Now, let’s consider Nozick’s closing question: Which transition in the series of cases Nozick describes makes it no longer the tale of a slave? Nozick’s own apparent implicit answer is that none of them does, and that the situation in the last case constitutes a kind of slavery, even if a mild sort. And since that situation is more or less the sort all of us find ourselves in in modern democratic societies, the implication is that such societies constitute slave states, at least to the extent that they make demands of their citizens (in the form of taxation, regulation, and the like) that go beyond the functions of Nozick’s “minimal state” (police, national defense, and courts of law). Naturally, this would also entail that non-democratic governments that make such demands also constitute slave states, presumably even more so.
In my libertarian days I agreed with this judgment, though I no longer do. The usual disclaimer applies: That does not by any means entail that I now think that “anything goes.” Much that modern governments demand of us is unjust, and some of it (such as saddling their citizens with crushing debt) may fairly be described as in some respects comparable to slavery. The point is just that I now believe, on natural law grounds, that it is false to say that requiring citizens to support any functions beyond the minimal state is inherently unjust or comparable to slavery. Some such demands can be justifiable, even if the demands socialists and egalitarian liberals would make are (I agree) not justifiable. (Which demands are justifiable, and under what conditions? There is no glib, one-line answer to that question of the sort a Rand- or Rothbard-style libertarian always seems to want, because the relevant moral considerations are more complicated than they suppose. See my Social Philosophy and Policy article “Classical Natural Law Theory, Property Rights, and Taxation” for the way traditional Thomistic natural law theory would approach the issue.)
But then, where exactly does Nozick’s Tale of the Slave go wrong? The problem with it is that it is underdescribed. In particular, the end result seems like “slavery” only because Nozick leaves out certain crucial moral facts. To see how, let’s first alter the thought experiment a little bit by supposing that the “10,000-headed master” emancipates you, but that, for whatever reason, you go on voluntarily to sign an employment contract with this group of people. Specifically, you agree to be on call at all hours of the night, over the course of a year – perhaps to do errands for various members of the group, or to deal with various overnight emergencies, or whatever. Suppose further that after a few weeks of this you get sick of being awakened in the middle of the night and in other ways start to hate the job. You wish you could quit but given the terms of your contract you cannot. Still, the group has come to rely on you and would be seriously inconvenienced if you didn’t do your job, and after all, you did agree to do it. Now, are you a slave? Nozick would say that in this case, you aren’t a slave, even though you hate your job as much as a slave might.
Why aren’t you a slave? The reason, of course, is that in this case you consented to the job, consented to be on call, for one year, to your 10,000-headed employer; and for the libertarian, consent suffices to make the situation something other than slavery. (Whether agreeing to surrender complete control over yourself for life would generate an enforceable obligation is something libertarians disagree over, but that issue does not affect the present point.)
But notice that this cannot be a complete answer, for two reasons. First of all, while you might have consented initially, you might also come to regret this. You might say “I want more than anything else to be free of this job! It’s so horrible – I feel as if I were a slave! Oh how I wish I had never signed that contract!” All the same, Nozick would say, you are obligated to do the job. And this shows that, whatever it means to say that you are obligated to do what you consent to doing, it does not mean that you have to be in any way willing to do it at the time you have to do it. “Consent” in that sense is not required. Hence Nozick and like-minded libertarians would seem to be committed to the following proposition:
1. A person can be obligated to do something even if, at the time he has to do it, he does not want to do it and feels as if he were being treated like a slave.
But notice that even your having initially consented cannot be the whole story either. For why is it that you are obligated in the first place to do what you have consented to doing? The answer cannot be “Because I consented to do whatever I consent to doing”, on pain of infinite regress. Consent can generate obligations only if we are already obligated, for other reasons, to do what we consent to doing. Consent cannot itself be the ground of all obligation. Certainly the Nozickian libertarian cannot think so: In Nozick’s view, it is because of a Kantian respect for persons as “ends in themselves” that we are obligated not to make slaves of them, even if we have never consented to regarding them as ends in themselves. They just are ends in themselves and that’s that. In that case, though, Nozick and like-minded libertarians are also committed to a further proposition:
2. A person can have enforceable obligations to others that he did not at any time consent to having.
Now the circumstance described in 2 is itself something a person might come to dislike intensely. In the situation I’ve described, where you have consented to be the midnight errand boy for your 10,000-headed employer, you might think to yourself “I hate it, just hate it, that the world is set up in such a way that I have obligations I never consented to putting myself under – for example, obligations like the obligation to honor contracts that I freely consented to signing!” Yet the Nozickian libertarian would say: “Too bad. That’s the way the world is. Deal with it.” More to the point, the Nozickian libertarian would have to say that having obligations like this does not by itself suffice to make you a slave. That is to say, the Nozickian libertarian is committed to a third proposition:
3. It is possible for a person who is not a slave nevertheless to have enforceable obligations to others that he never consented to having and that he finds deeply odious.
Notice further that these obligations may not be merely of a negative sort, but even of a positive sort. In the scenario I’ve described, you are obligated not merely to refrain from harming your 10,000-headed employer, but also to carry out certain acts as your employer demands, given the terms of your contract. It is true that you consented to this, but you did not consent to being obligated to do whatever you’ve consented to doing. And in this sense, your consent is only partial. That means that the Nozickian libertarian is committed also to:
4. It is possible for a person who is not a slave nevertheless to have enforceable positive obligations to others that he never fully consented to having and that he finds deeply odious.
Now, keep in mind that none of what has been said so far rests on any moral premises of my own, natural law or otherwise. All I have been doing is drawing out what is implicit in Nozick’s own position. But look what happens when we return to the original Tale of the Slave and apply to it the points we have made. Nozick, it seems, expects us to regard case 9 – the case that parallels a modern democratic society – as tantamount to a mild form of slavery. But why should we so regard it? The answer cannot be “Because it involves our having obligations to others that we find odious and that we never consented to having.” For given 3, there is nothing necessarily wrong or slavery-like with that.
Note also that Nozick does not tell us in his Tale of the Slave whether the “10,000-headed master” gives the purported slave a right of exit – that is, a right to emigrate from the territory over which the 10,000-headed master rules. (This is one respect in which Nozick’s thought experiment is, as I have said, underdescribed.) But he will have to add such a right to the story if he wants the example to be relevantly parallel to a modern democratic society, since such societies do allow their citizens to emigrate. Now, a right of exit entails that anyone who dislikes the positive obligations a 10,000-headed master (or some government) imposes on him could always escape them by emigrating. Of course, exercising this option might be burdensome, but if a person could still take it and yet refrains from doing so, then his being subject to the positive obligations in question involves at least partial consent, even if not full consent. But in that case, if we ask why we should regard Nozick’s case 9 as tantamount to slavery, the answer cannot be “Because it involves our having positive obligations to others that we find odious and that we never fully consented to having.” For given 4, there can be nothing necessarily wrong or slavery-like with that either.
So, whatever it is about case 9 that is supposed to amount to slavery, it cannot – by Nozick’s own lights cannot – be merely that it involves obligations to others that we find odious and that we never consented to, nor even that it involves positive obligations to others that we never fully consented to. For given 3 and 4, which Nozick himself is committed to, these circumstances simply do not suffice to make a situation tantamount to slavery.
I submit that these points entirely undermine the force of Nozick’s thought experiment. Probably most people who find the Tale of the Slave an impressive piece of libertarian argumentation do so because they are implicitly reasoning in one of two ways. First, they might be reasoning as follows:
Major premise: Slavery is odious and non-consensual.
Minor premise: The demands imposed on us by democratic and other governments (or at least those that go beyond the functions of the minimal state) are also (sometimes) odious and non-consensual.
Conclusion: So demands of this sort amount to slavery.
But this argument is invalid, for (as anyone who has taken a basic logic course can see) it commits the fallacy of the undistributed middle term. Alternatively, we might replace the major premise with its converse, giving us the following valid argument:
Major premise: Any demands made on us that are odious and non-consensual amount to slavery.
Minor premise: The demands imposed on us by democratic and other governments (or at least those that go beyond the functions of the minimal state) are (sometimes) odious and non-consensual.
Conclusion: So demands of this sort amount to slavery.
But as we have seen, the Nozickian libertarian is committed to rejecting this alternative major premise, given his commitment to 3 and 4. Nor would it do the Nozickian libertarian any good to reconsider his commitment to 3 and 4, so as to be avail himself of the proposed alternative major premise. For one thing, 3 and 4 follow, as we have seen, from Nozick’s basic commitments; to abandon them is just to abandon the foundations of Nozick’s libertarianism. For another thing, the claim that “Any demands made on us that are odious and non-consensual amount to slavery” simply begs the question against the non-libertarian, who holds precisely that we have certain non-slave-like obligations we did not consent to, even if we find them odious.
Now, obviously the Nozickian libertarian will insist that, even if not all obligations that are non-consensual and odious amount to slavery, the specific obligations that democratic and other governments impose on us do amount to slavery, at least when they go beyond the functions of the minimal state. The trouble is that we now need a separate argument for this claim; merely appealing to the odious and non-consensual nature of these obligations will not suffice, for the reasons we’ve seen. That is another reason I say that Nozick’s thought experiment is underdescribed. To know whether his case 9 amounts to slavery – and thus to know whether the demands democratic and other governments make of us amount to slavery – we need to know what specifically are the demands that the 10,000-headed master (or such governments) are making of us, and why these specific demands amount to slavery when other odious and non-consensual demands to not. But in that case the Tale of the Slave itself – with its implied emphasis on the non-consensual and odious nature of the obligations the 10,000-headed master imposes on you, abstracting away from the actual contents of the obligations – drops out as irrelevant, since it would not really be doing any work. Rather, this separate argument (whatever it is) would be doing all the work.
So, the implied argument of the Tale of the Slave seems to be either irrelevant, or invalid, or to be committed to a premise which both begs the question against the non-libertarian and which Nozick himself implicitly rejects in any case. Vivid and interesting though the thought experiment is, it thus fails to provide any support for libertarianism. Its appeal is entirely rhetorical, and has no actual logical force.
Again, this does not entail that democratic governments, or any other kind of government, may demand whatever they wish of their citizens without this amounting to slavery. That a certain argument for a claim fails does not mean that the claim itself is false. In any event, like Nozick, I reject socialism and egalitarian liberalism. But the reasons have to do with natural law considerations of the sort outlined in the article of mine linked to above, and not with superficial comparisons to slavery of the sort that the Tale of the Slave rests on.
Nor do I intend any disrespect toward Nozick or his arguments. On the contrary, Nozick was a brilliant philosopher, and the arguments he sets out in Anarchy, State and Utopia are important ones that deserve our consideration even if we ultimately reject them. Certainly they are far more formidable than those of Nozick’s absurdly overrated rival John Rawls, whose main “arguments” are little more than flatulent tautologies. The contrast between the cringe-making hagiography usually afforded Rawls and the condescension toward Nozick one finds in commentators like Metcalf (and Matthew Yglesias) says more about the commentators than it does about the respective merits of Rawls and Nozick themselves. Rawls’s arguments are murky, plodding, and (given their ultimate circularity) anticlimactic, but reinforce liberal prejudices. Nozick’s are clever, clear, and crisp, but challenge those prejudices. Nothing more need be said.
Now, let’s consider Nozick’s closing question: Which transition in the series of cases Nozick describes makes it no longer the tale of a slave? Nozick’s own apparent implicit answer is that none of them does, and that the situation in the last case constitutes a kind of slavery, even if a mild sort. And since that situation is more or less the sort all of us find ourselves in in modern democratic societies, the implication is that such societies constitute slave states, at least to the extent that they make demands of their citizens (in the form of taxation, regulation, and the like) that go beyond the functions of Nozick’s “minimal state” (police, national defense, and courts of law). Naturally, this would also entail that non-democratic governments that make such demands also constitute slave states, presumably even more so.
In my libertarian days I agreed with this judgment, though I no longer do. The usual disclaimer applies: That does not by any means entail that I now think that “anything goes.” Much that modern governments demand of us is unjust, and some of it (such as saddling their citizens with crushing debt) may fairly be described as in some respects comparable to slavery. The point is just that I now believe, on natural law grounds, that it is false to say that requiring citizens to support any functions beyond the minimal state is inherently unjust or comparable to slavery. Some such demands can be justifiable, even if the demands socialists and egalitarian liberals would make are (I agree) not justifiable. (Which demands are justifiable, and under what conditions? There is no glib, one-line answer to that question of the sort a Rand- or Rothbard-style libertarian always seems to want, because the relevant moral considerations are more complicated than they suppose. See my Social Philosophy and Policy article “Classical Natural Law Theory, Property Rights, and Taxation” for the way traditional Thomistic natural law theory would approach the issue.)
But then, where exactly does Nozick’s Tale of the Slave go wrong? The problem with it is that it is underdescribed. In particular, the end result seems like “slavery” only because Nozick leaves out certain crucial moral facts. To see how, let’s first alter the thought experiment a little bit by supposing that the “10,000-headed master” emancipates you, but that, for whatever reason, you go on voluntarily to sign an employment contract with this group of people. Specifically, you agree to be on call at all hours of the night, over the course of a year – perhaps to do errands for various members of the group, or to deal with various overnight emergencies, or whatever. Suppose further that after a few weeks of this you get sick of being awakened in the middle of the night and in other ways start to hate the job. You wish you could quit but given the terms of your contract you cannot. Still, the group has come to rely on you and would be seriously inconvenienced if you didn’t do your job, and after all, you did agree to do it. Now, are you a slave? Nozick would say that in this case, you aren’t a slave, even though you hate your job as much as a slave might.
Why aren’t you a slave? The reason, of course, is that in this case you consented to the job, consented to be on call, for one year, to your 10,000-headed employer; and for the libertarian, consent suffices to make the situation something other than slavery. (Whether agreeing to surrender complete control over yourself for life would generate an enforceable obligation is something libertarians disagree over, but that issue does not affect the present point.)
But notice that this cannot be a complete answer, for two reasons. First of all, while you might have consented initially, you might also come to regret this. You might say “I want more than anything else to be free of this job! It’s so horrible – I feel as if I were a slave! Oh how I wish I had never signed that contract!” All the same, Nozick would say, you are obligated to do the job. And this shows that, whatever it means to say that you are obligated to do what you consent to doing, it does not mean that you have to be in any way willing to do it at the time you have to do it. “Consent” in that sense is not required. Hence Nozick and like-minded libertarians would seem to be committed to the following proposition:
1. A person can be obligated to do something even if, at the time he has to do it, he does not want to do it and feels as if he were being treated like a slave.
But notice that even your having initially consented cannot be the whole story either. For why is it that you are obligated in the first place to do what you have consented to doing? The answer cannot be “Because I consented to do whatever I consent to doing”, on pain of infinite regress. Consent can generate obligations only if we are already obligated, for other reasons, to do what we consent to doing. Consent cannot itself be the ground of all obligation. Certainly the Nozickian libertarian cannot think so: In Nozick’s view, it is because of a Kantian respect for persons as “ends in themselves” that we are obligated not to make slaves of them, even if we have never consented to regarding them as ends in themselves. They just are ends in themselves and that’s that. In that case, though, Nozick and like-minded libertarians are also committed to a further proposition:
2. A person can have enforceable obligations to others that he did not at any time consent to having.
Now the circumstance described in 2 is itself something a person might come to dislike intensely. In the situation I’ve described, where you have consented to be the midnight errand boy for your 10,000-headed employer, you might think to yourself “I hate it, just hate it, that the world is set up in such a way that I have obligations I never consented to putting myself under – for example, obligations like the obligation to honor contracts that I freely consented to signing!” Yet the Nozickian libertarian would say: “Too bad. That’s the way the world is. Deal with it.” More to the point, the Nozickian libertarian would have to say that having obligations like this does not by itself suffice to make you a slave. That is to say, the Nozickian libertarian is committed to a third proposition:
3. It is possible for a person who is not a slave nevertheless to have enforceable obligations to others that he never consented to having and that he finds deeply odious.
Notice further that these obligations may not be merely of a negative sort, but even of a positive sort. In the scenario I’ve described, you are obligated not merely to refrain from harming your 10,000-headed employer, but also to carry out certain acts as your employer demands, given the terms of your contract. It is true that you consented to this, but you did not consent to being obligated to do whatever you’ve consented to doing. And in this sense, your consent is only partial. That means that the Nozickian libertarian is committed also to:
4. It is possible for a person who is not a slave nevertheless to have enforceable positive obligations to others that he never fully consented to having and that he finds deeply odious.
Now, keep in mind that none of what has been said so far rests on any moral premises of my own, natural law or otherwise. All I have been doing is drawing out what is implicit in Nozick’s own position. But look what happens when we return to the original Tale of the Slave and apply to it the points we have made. Nozick, it seems, expects us to regard case 9 – the case that parallels a modern democratic society – as tantamount to a mild form of slavery. But why should we so regard it? The answer cannot be “Because it involves our having obligations to others that we find odious and that we never consented to having.” For given 3, there is nothing necessarily wrong or slavery-like with that.
Note also that Nozick does not tell us in his Tale of the Slave whether the “10,000-headed master” gives the purported slave a right of exit – that is, a right to emigrate from the territory over which the 10,000-headed master rules. (This is one respect in which Nozick’s thought experiment is, as I have said, underdescribed.) But he will have to add such a right to the story if he wants the example to be relevantly parallel to a modern democratic society, since such societies do allow their citizens to emigrate. Now, a right of exit entails that anyone who dislikes the positive obligations a 10,000-headed master (or some government) imposes on him could always escape them by emigrating. Of course, exercising this option might be burdensome, but if a person could still take it and yet refrains from doing so, then his being subject to the positive obligations in question involves at least partial consent, even if not full consent. But in that case, if we ask why we should regard Nozick’s case 9 as tantamount to slavery, the answer cannot be “Because it involves our having positive obligations to others that we find odious and that we never fully consented to having.” For given 4, there can be nothing necessarily wrong or slavery-like with that either.
So, whatever it is about case 9 that is supposed to amount to slavery, it cannot – by Nozick’s own lights cannot – be merely that it involves obligations to others that we find odious and that we never consented to, nor even that it involves positive obligations to others that we never fully consented to. For given 3 and 4, which Nozick himself is committed to, these circumstances simply do not suffice to make a situation tantamount to slavery.
I submit that these points entirely undermine the force of Nozick’s thought experiment. Probably most people who find the Tale of the Slave an impressive piece of libertarian argumentation do so because they are implicitly reasoning in one of two ways. First, they might be reasoning as follows:
Major premise: Slavery is odious and non-consensual.
Minor premise: The demands imposed on us by democratic and other governments (or at least those that go beyond the functions of the minimal state) are also (sometimes) odious and non-consensual.
Conclusion: So demands of this sort amount to slavery.
But this argument is invalid, for (as anyone who has taken a basic logic course can see) it commits the fallacy of the undistributed middle term. Alternatively, we might replace the major premise with its converse, giving us the following valid argument:
Major premise: Any demands made on us that are odious and non-consensual amount to slavery.
Minor premise: The demands imposed on us by democratic and other governments (or at least those that go beyond the functions of the minimal state) are (sometimes) odious and non-consensual.
Conclusion: So demands of this sort amount to slavery.
But as we have seen, the Nozickian libertarian is committed to rejecting this alternative major premise, given his commitment to 3 and 4. Nor would it do the Nozickian libertarian any good to reconsider his commitment to 3 and 4, so as to be avail himself of the proposed alternative major premise. For one thing, 3 and 4 follow, as we have seen, from Nozick’s basic commitments; to abandon them is just to abandon the foundations of Nozick’s libertarianism. For another thing, the claim that “Any demands made on us that are odious and non-consensual amount to slavery” simply begs the question against the non-libertarian, who holds precisely that we have certain non-slave-like obligations we did not consent to, even if we find them odious.
Now, obviously the Nozickian libertarian will insist that, even if not all obligations that are non-consensual and odious amount to slavery, the specific obligations that democratic and other governments impose on us do amount to slavery, at least when they go beyond the functions of the minimal state. The trouble is that we now need a separate argument for this claim; merely appealing to the odious and non-consensual nature of these obligations will not suffice, for the reasons we’ve seen. That is another reason I say that Nozick’s thought experiment is underdescribed. To know whether his case 9 amounts to slavery – and thus to know whether the demands democratic and other governments make of us amount to slavery – we need to know what specifically are the demands that the 10,000-headed master (or such governments) are making of us, and why these specific demands amount to slavery when other odious and non-consensual demands to not. But in that case the Tale of the Slave itself – with its implied emphasis on the non-consensual and odious nature of the obligations the 10,000-headed master imposes on you, abstracting away from the actual contents of the obligations – drops out as irrelevant, since it would not really be doing any work. Rather, this separate argument (whatever it is) would be doing all the work.
So, the implied argument of the Tale of the Slave seems to be either irrelevant, or invalid, or to be committed to a premise which both begs the question against the non-libertarian and which Nozick himself implicitly rejects in any case. Vivid and interesting though the thought experiment is, it thus fails to provide any support for libertarianism. Its appeal is entirely rhetorical, and has no actual logical force.
Again, this does not entail that democratic governments, or any other kind of government, may demand whatever they wish of their citizens without this amounting to slavery. That a certain argument for a claim fails does not mean that the claim itself is false. In any event, like Nozick, I reject socialism and egalitarian liberalism. But the reasons have to do with natural law considerations of the sort outlined in the article of mine linked to above, and not with superficial comparisons to slavery of the sort that the Tale of the Slave rests on.
Nor do I intend any disrespect toward Nozick or his arguments. On the contrary, Nozick was a brilliant philosopher, and the arguments he sets out in Anarchy, State and Utopia are important ones that deserve our consideration even if we ultimately reject them. Certainly they are far more formidable than those of Nozick’s absurdly overrated rival John Rawls, whose main “arguments” are little more than flatulent tautologies. The contrast between the cringe-making hagiography usually afforded Rawls and the condescension toward Nozick one finds in commentators like Metcalf (and Matthew Yglesias) says more about the commentators than it does about the respective merits of Rawls and Nozick themselves. Rawls’s arguments are murky, plodding, and (given their ultimate circularity) anticlimactic, but reinforce liberal prejudices. Nozick’s are clever, clear, and crisp, but challenge those prejudices. Nothing more need be said.
