A year ago I sent Mike Huemer the following comments on his book draft, then titled Freedom and Authority.  He’s incorporated many of my suggestions, but if you’re curious to peer inside the revision process, enjoy.  (Detailed comments using the draft’s pagination are below the fold).

Comments on Freedom and Authority

Finished the chapters you sent me.  I’m in awe. This is the best book on the philosophy of libertarianism ever written – and I’ve read almost all of them.  I feel honored to live in a world where your book exists.  You genuinely practice what you preach: Finding broadly acceptable premises, then validly reasoning from them to novel, important conclusions.  Perhaps most amazingly, you’re turning philosophy into a teachable method for avoiding absurdity and evil.

That said, the best book on the philosophy of libertarianism ever written can still be improved.  The big two:

1. The parts where you reply to specific modern philosophers (especially Gauthier and Rawls) are well-done, but sharply detract from the flow and tone of the book.  Only professional philosophers will want to read these parts, and they will alienate the rest of your audience.  (This is especially bad because your discussions of specific modern philosophers are highly front-loaded).  You should focus exclusively on statist positions with broad plausibility, not the tortured question-begging of high-status academic philosophers.

If you must address specific modern philosophers, you should either (a) distill their strongest arguments, with name-dropping and exegesis confined to the footnotes; or (b) move the parts on specific modern philosophers to a technical appendix.

The technical sections are the single biggest barrier between you and lots of readers.  And you deserve lots of readers!

The deliberative democracy part works better than the Gauthier and Rawls stuff; at least many people might find deliberative democracy independently plausible.

2. Part II is excellent, but I think you can try even harder to meet skeptics on their own ground.  When I teach anarchism, I begin with the “minimal state” with a monopoly
on police, courts, and law, then point out that existing democracies don’t have monopolies on any of the three!  I then try to whittle down the plausible “minimal” scope of the state by helping students imagine the expansion of existing private alternatives to the state.  Once we’re down to almost zero, then it’s much more plausible to imagine finishing the job.

http://econfaculty.gmu.edu/bcaplan/e854/pc13.htm, sections III-VII.

This might entail major re-writing, but I think it’s worth it.  For most of the book, you determinedly keep yourself in the shoes of the reasonable skeptic.  Part II moves too fast.

Detailed comments:

1. You introduce libertarian jargon (“libertarian,” “anarchism,” “anarcho-capitalism,” etc.) en passant at almost random locations in the book.  Put yourself in the shoes of the reasonable skeptic who’s never heard of these terms – or has unfavorable preconceptions.  Either explain the terms early on, or wait and explain them the first time you use them.

2. p.35, 1st full graf.  “would not be justified in demanding that board members pay one dollar to express their objection.”  But even verbally objecting or raising your hand requires some effort.  Many would disvalue that effort more than spending a dollar.

3. p.41, 2nd to last graf.  “usually with imprisonment.”  No.  Monetary fines are vastly more common than imprisonment.  Almost everyone’s gotten a ticket at some point, but most people never go to jail.

4. You should talk more about the condition that you shouldn’t have to give up your own stuff to refuse consent.

5. p.54, 1st full graf.  The last sentence is too long and confusing.

6. p.56, first sentence.  This would flow better if you said “one is to appeal to the very simple argument that the police may catch you if you violate the law.”

7. p.63.  Is it really true that philosophers lack consensus?  As you probably know, some researchers have been surveying philosophers about their views.  Worth checking:  http://philpapers.org/surveys

8. p.65.  Using “efficacious” in the title is confusing.  Maybe make it “morally efficacious”?  Or is there a better word than “efficacious”?

9. p.66, bottom.  You suddenly start talking about “natural rights.”  Why not just stick to “rights”?  The “natural” prefix is just obscurantism.  If you want to distinguish between legal and moral rights, say “moral rights.”

10. p.67.  “the particularly lenient moral attitudes we take toward government.”  This point should be more prominent.

11. p.71.  Robin Hanson’s done a good job of arguing that disagreement wouldn’t persist if people accepted other’s disagreement as a form of evidence.  Not a perfect argument, but better than I expected by far:


12. p.72.  Why say “even the staunchest realist should recognize”?  Why not “even the staunchest realist holds”?

13. p.87.  I think you should emphasize the unselfishness of voters a lot more.  Voters really are amazingly unselfish, and this fact seems to stack the deck against you.  Then you can reply with something like the following:


14. p.100, fn 106.  Do you mean Caplan (2006) or (2007)?

15. p.115, last full graf.  “only feasible source” is too strong.  How about “best source” or “only reliable source”?

16. p.121.  You can make rule consequentialism more plausible by emphasizing the probability of making a difference.  And if the outcome is important, even a low probability could suffice.

17. p.125, bottom.  Double check your claims about the illegality of requiring payment in gold or foreign currency.  Are these really still completely illegal?

18. p.127.  You overuse Latin.  Even I had to look up “pro tanto.”

19. p.129.  It’s Cthulhu, not Cthulu.

20. p.131.  “one need not obey the law, if there are more socially valuable uses.”  Shouldn’t this be “one need not obey the law, if you make a more socially valuable choice instead”?  The mere existence of a more socially valuable use that you have no intention of pursuing hardly seems relevant.

21. p.135.  It seems like the consequentialist argument plausibly legitimizes (7) as well as (1) and (2).

22. Chapter 6 neglects what I consider the best argument for potentially wide-ranging obedience to government:

Premise 1: People can rightfully coerce others if their coercion would have very good consequences.

Premise 2: There are very good consequences when government coerces people to do X.

Premise 3: If someone can rightfully coerce you to do X, then you should do X voluntarily.

Premise 1 and 2 imply that government can rightfully coerce given sufficiently good consequences.  Premise 3 then implies that you should voluntarily comply with government edicts with very good consequences.  As long as the latter factual claim checks out, you’re morally obliged to obey.

I’m tempted to say that this argument also explains why you’re obliged to obey laws with very good consequences even if your behavior has little effect on the outcome.  But that’s not totally clear.

23. This chapter on the “Psychology of Authority” is magnificent.  You should try publishing a shorter version of it in a major magazine like The New Yorker as a tie-in to the book.  It could also conceivably be a Saturday Essay in the Wall St. Journal.  I’d be happy to introduce you to possible outlets – I know a lot of people first-hand, and practically everyone important second-hand.

24. p.141. If acting on your arguments had bad consequences, wouldn’t the consequentialist defense of authority be true – in which case your book is wrong?

25. p.153, fn 156.  Percent or percentage points higher?

26. p.182.  Is tobacco seven times more deadly per user?  Or total?

27. p.184, 1st full sentence.  Cut “In my view,” plus about half the occurrences of similar expressions throughout the book. You’re overusing them.

28. p.185.  On immigration, you should address an argument that most people sometimes seem to believe: That the government (or its citizens) collectively own the country.  Ridiculing conservatives for suddenly turning into radical Leninists is appropriate.

Related: Does government rightfully own the roads?  If so, can’t they legitimately force us to do virtually anything by denying us permission to use the roads?

29. p.186, second graf.  Why use the expression “generally acceptable”?  It’s imprecise, and tempts people to say, “Since almost everyone wants to restrict immigration, it is ‘generally acceptable.'”

30. p.193, 1st graf.  “Murray cites numerous statistics in support of this argument.”  Cut this sentence, it adds little.

31. p.193, fn 209.  Since domestic welfare programs are a common argument for restricting immigration, I’d say the net effect on even the current poor is extremely negative.  Also check out my paper on the effect of welfare on the poor:


31. p.194.  You should look at some stats on administrative costs.  For big programs, they’re very small in percentage terms.  But often (especially for medical spending) this is because they have very weak cost control.

You should discuss the fact that the people Americans consider poor are obese, own cars, have cable t.v., etc.

32. p.202, fn 223.  Isn’t the cabin break-in a good time to invoke hypothetical consent?

33. p.203, 2nd full graf.  The argument would also justify robbing the government!

34. p.210.  You mention terrorism and revolution.  What about tyrannicide?  Von Stauffenberg deserves a page in your book.

35. p.211.  You cite Pape later in the book, but isn’t he relevant here, too?  My memory is that Pape argues fairly plausibly that terrorism often does work.

36. p.221-4.  These are excellent pages.  You have indeed separated yourself from the dogmatic, implausible, tortured pack of other political philosophers – libertarian and otherwise.  But your sentence about “relatively narrow range of special circumstances” is overly strong.  You seem to argue from the weaker premise that aggression would be justified if it led to sharply better consequences – and take seriously the possibility that this often happens.

37. p.229.  See my critique of Tetlock?  He said it was the best.


38. p.244.  You should consider the possibility that predators want to develop reputations for cruelty.  Genghis Khan supposedly murdered everyone alive in any city that resisted him.  Costly in the short-run, but in the long-run cities surrendered without resistance.  The key variable seems to be exit costs.  If exit is expensive, reputations for cruelty become extremely valuable.  I discuss this issue here in section VI:


39. p.246.  Did you see me on Hobbes here?


40. p.248.  It’s not clear that weapons on net increase the effective equality of fighting ability – especially when you remember that many are expensive or inconvenient to
carry.  Consider machine guns.

41. p.252.  Try to find better cites than Rummel.

42. p.256, last sentence.  Don’t you mean “the expenditure is not prudent”?

43. p.256, fn 41.  You might want to cite the broader lit on probability of decisiveness, to emphasize that you’re using a high estimate of p.  If people get more detailed information before each election, the p will be much lower because there’s a bell curve with an expected mean not equal to 50/50.

44. p.258. “no fear of incurring disfavor” is too strong.

45. p.258, fn 46.  Should there be a page #?  Same for various other fns.

46. p.260.  You
overlook one of voters’ best enforcement mechanisms: increasing the severity of punishment to compensate for low probability of detection.  (Economists often call this “Beckerian punishment”).  In the real world, of course, voters mostly punish for silly scandals.  But if voters wanted to monitor politicians at low cost, they have an effective mechanism.  If you object that this doesn’t work well with division of powers, that’s all the more reason to switch to a Parliamentary-type system.  For more, see section VI:


47. p.260.  It’s unclear that women’s suffrage was a “major improvement.”  It’s pretty clear that Indian independent wasn’t.  The death toll of the partition was massive, and was followed by decades of further wars and very slow economic growth.
From Wikipedia:


Massive population exchanges occurred between the two newly formed states in the months immediately following Partition. Once the lines were established, about 14.5 million people crossed the borders to what they hoped was the relative safety of religious majority. Based on 1951  Census of displaced persons, 7,226,000 Muslims went to Pakistan from India while 7,250,000 Hindus, Sikhs and Jews moved to India from Pakistan immediately after partition. About 11.2 million or 78% of the population transfer took place in the west, with Punjab  accounting for most of it; 5.3 million Muslims moved from India to West Punjab in Pakistan, 3.4 million Hindus and Sikhs moved from Pakistan to East Punjab in India; elsewhere in the west 1.2 million moved in each direction to and from Sind.[citation needed]

The newly formed governments were completely unequipped to deal with migrations of such staggering magnitude, and massive violence and slaughter occurred on both sides of the border. Estimates of the number of deaths range around roughly 500,000, with low estimates at 200,000 and high estimates at 1,000,000.[11]

48. p.262-3.  Your first reason is weak.  Journalists work long hours no matter what they cover, so laziness is a red herring.  You should put reason #3 – consumer demand –

49. p.262.  Wikipedia suggests that your Wikileaks story omits key facts.  Are you confident you’ve been fair here?


50. p.267.  The first few sentences are a little confusing.  I can see why a sudden increase in poverty would lead to more government action.  But high levels of poverty are usually found in poor countries, which basically accept poverty as unavoidable.

51. p.269.  You should say “federal government” or “U.S. federal government” rather than “U.S. government.”

52. p.280.  I’d emphasize the fact that under anarcho-capitalism, you can change your defense agency without changing your location.  Even with an HOA, the HOA can change its agency without changing the HOA’s location.

Why do you say “is thought to render it more just”?  Why not just say “renders it more just” or “makes it more just”?

53. p.281, graf 2.  You neglect the possibility of frequency-dependent selection.  Quick version: If everyone else is peaceful, it pays to be aggressive.  If everyone else is aggressive, it pays to be peaceful.  In equilibrium, you get both types of behavior.  Details:

Section VI:


Sections V and VI:


54. p.282, graf 1.  Don’t the Milgram experiments suggest that even employees of private firms might follow awful orders?  It’s even worse with government orders, but still.

55. p.294.  Have you looked at victimization data?  They supposedly solve the problem of unreported crimes.

56. p.296, fn 90.  Why couldn’t the family sue the DEALER?

57. p.298, 3rd full graf.  “in whatever ways they had available” is way too optimistic.

4th full graf: It can also be beneficial for someone to established a bad precedent in their favor.

58. p.300.  You should more strongly emphasize the democratic credentials of the government you’re comparing to anarchy.  The last graf is good, but readers deserve more like it.

59. p.303-4.  No greater efficiency in operating ten automobile factories than one?  How many automakers only have one factory?!  Also, I’m pretty sure these factories cost more than “millions” and produce more than “thousands” of cars.  Check.

60. p.305, 1st graf.  You are hardly the only one who holds this view.  Good textbooks mention it, too.  The economist John Haltiwanger produces massive evidence that selection is the main source of productivity growth – sometimes the sole measurable source.

61. p.307, 1st full graf.  It seems like the existence of a cartel  would reduce the cost of violent conflict, because they have a better chance of eliciting surrender.

62. p.310.  You should include cites on HOAs.  Alex Tabarrok’s done neat work here.

63. p.316.  I think that most people would agree to “unfair” arbitration contracts.  Why?  Firms have strong reputational incentives to treat consumers and workers well, so they rarely have valid complaints.  (See your Target example).  As a result, a second layer of formal protection just isn’t very valuable to consumers and workers.  For firms, though, formal protections are extremely costly – lots of time and trouble to address largely bogus complaints of a handful of litigious folks.  From an economic point of view, of course, this just means that lawsuits are a waste of resources – and it’s better if potential litigants waive their rights to sue.

64. p.319, last graf.  “generally accepted values of his society” should probably be “generally accepted values of his customers”

65. p.320.  Why do you say “This extra compensation would provide a deterrent to crime”?  ALL required compensation is a deterrent, no?

66. p.321.  You seem to forget the probability of punishment.  Punishments conditional on getting caught need to exceed the value gained by crime, or else it’s a “heads I win, tails I break even” regime.

67. p.323.  Another check on the severity of punishment: Parents don’t want their kids to be harshly punished for youthful indiscretions.

68. p.325, graf 1.  “there remain more” should be “there remains more.”  As long as civil suits don’t have spiraling punishments, you can always piggyback remaining criminal cases onto the flourishing civil system.

69. p.326, graf 1.  Wrongly Convicted says “Because she was serving five years for another crime, Victoria Banks did not seek her immediate resentencing.”  Still bad, but it sounds like the government is more lazy than evil here. http://forejustice.org/db/fc/Banks–Victoria-Bell.html

70. p.327, last full graf.  Isn’t “separately” more essential than “sequentially”?

71. p.329, 2nd full graf.  The SBA estimate sounds high compared to others I’ve heard.

72. p.330, 1st graf.  $284/hour seems way too high.  The average lawyer doesn’t make $500+k a year, not even close.  Even subtracting overhead, it’s not credible.

The sentence “simply cannot afford… unless he is forced” makes no sense.  How does forcing you to buy something make it more affordable?

73. p.331.  Data on prison rape:



74. p.331, last graf.  Are there genuine experiments on the effect of rehab programs on recidivism?

75. I think my argument for pacifism is worth inserting somewhere in chapter 12:


Guerrilla warfare may be effective, but it’s so awful it’s usually better just to endure occupation.

76.  p.339.  You should distinguish between “nonviolent” and “noncoercive.”  MLK’s tactics were often highly coercive, as would be clear if his followers occupied a private home.

77. p.344.  On Hitler’s motives, see:


Lebensraum is capitalized, since it’s a German noun. 78.  On “realism,” see:



79. p.360, 2nd full graf.  Terrorist leaders say what they think their audience wants to hear.  That might be the truth, but quite possibly not.  Still, if their demands are minor, it’s worth a try.  See section 3.3:


80. p.361, fn201.  Do you really want to call these “martyr videos”?

81. p.365, 3rd full graf.  The guerilla warfare response, to repeat, is weak.  It often works, but at a terrible cost.