The footnote and return anchor links don't work. They're formatted as absolute links to somewhere on slatestarcodex.com. They should be formatted as relative works so that they'll work regardless of where the page is located. (As in, the link href should only contain "#anchortolinkto", not the rest of the URL.)
We have all been spoiled by warlessness for so long, we have forgotten how anyone could live differently then this. The symptom of wealth can only exist in an era of peace, at least on a local level. Every good quality arises naturally while at peace.
Prediction and fulfillment, the favourite pastime of the brain, handles all the accounting and grumpiness for us on it's own. Provided there's not fights breaking out over every petty grievance, that is.
Every single example cited aims at having peaceful interactions for many generations, which is really hard if you're talking about weeks at see with lots of pointy sticks at your disposal, for example.
I think he explains the symptom, not the source.
(also, good god this not only is long, it feels long... very interesting but very few lighthearted moments that usually are the stamp at every base camp)
Yeah, I'm not buying the hypothesis. It seems like a noble savage stereotype targeted at rural communities translated into a hypothesis.
I've experienced more than a few of these close knit, traditional societies, both in the US and outside of it. Perhaps more importantly, I was in one of them as an insider and another as a quasi-insider. These communities were certainly orderly but because the enforcers of that order were private they in turn could use that authority for personal projects. A completely social society is a society without objective standards, which in turn means everything is a matter of social dominance. This might sound horrible but the advantage is that there is no dispassionate enforcement. If the law says X, and X is monstrous, you can at least be sure there's no bureaucrat paperclip maximizing X. The private enforcer will realize this is stupid and stop. And no one gossips, "Look at Scott, he stopped maximizing paperclips!"
But make no mistake. These societies can be very cruel. They can also be very kind. But they don't necessarily have to be. For example, a bunch of Southern plantation owners instead would also use this kind of enforcement and not to good ends. Lynching was violence aimed at social control very famously. This is without getting into problems inherent the system, like how they have forces inherently pushing them towards insularity, conformity, and (often hereditary) hierarchy.
These kind of rural, neighborly societies solve many problems caused by deracinated urban life. You'll never feel alone, or out of place, or like you don't have a purpose or place in life. But they have other issues.
I like the case studies better than the theorizing. Learning about what people actually do in these circumstances is fascinating. The theorizing seems inconclusive at best?
One of the bullet points towards the end of the review links to a post on the author's blog (or at least claims to do so). This would seem to violate the anonymity premise of the contest?
You could dismiss this as liberal-moralistic pearl clutching by the anthropologists, but the criticism is widespread, and often based on first hand experience.
I've taken to saying things like "The law only matters if someone's willing to go to court". I've knowingly been the victim of lawbreaking - even in contexts where correcting it would be fairly painless and modestly profitable to me - and gone along with it because I didn't think they were doing anything immoral. (tl;dr, the one union job I've ever had, my union somehow managed to bargain away my legally guaranteed holiday pay. But the base pay was high enough that the deal still seemed more or less fair.)
Laws exist for the big offenses, the sort where you'd solve it by rounding up your family members to bring guns and ropes to bear if the law didn't exist. The sort where corporations have millions (or billions) at stake, or where you want to ensure that nobody screws up and kills a bunch of people. For all the smaller issues in the world, the time-honored rule of "Don't be a dick" is the true law. And it's mostly for the best.
I think your distinction between global and local maxima is a good articulation of why both totalitarianism and anarchism fail as (large scale) social structures: you need coordinated, organized efforts to overcome collective action problems and overcome high cost social transition - but you also need organic, bottom-up structures (like norms) to help optimize for welfare within a given local range of solutions.
I see that it has now been over a decade since I read & blogged about this book, but my blogging didn't add enough value to be worth linking to.
Turnbull is somewhat notorious now that he's not around to defend himself. It has been said that he liked groups that let him indulge his proclivity toward pederasty and hated those (like the Ik) that didn't, thus motivating his depiction of the latter.
There's a lawyer with a youtube channel called "Fudd Busters" devoted to debunking incorrect beliefs about gun laws. The "Fudd" comes from Elmer Fudd, and it's not just backwoods hunters who are ignorant. Federally licensed firearm dealers frequently don't know the actual law they deal with as part of their jobs.
I found the summary of the book and its claims to be interesting, useful, and readable. The tone is a bit "inside baseball" by SSC standards, expecting that readers will be familiar with concepts from LessWrong and certain fields of econ. It didn't help that I had no idea what the subject and thesis of the book were until the middle of the review. Only at the very end did the practical relevance of Ellickson's thesis really click for me. The reviewer engages with Ellickson in a technical, nuts-and-bolts kind of way that I found unhelpful and a bit tedious, mostly due to my unfamiliarity. I did come away feeling like I understood the book, though, which is a win.
in East Texas a saw a loose horse beside the highway. I pulled over and nabbed the horse. when a pickup came by I waved it down. they just shook their heads. "Naw, that's so-and-so's horse, and he don't help us with our loose stock and we don't help with his."
sounds to me like I've got a basis for 450 page book.
Re the lying part. I think societies have an incentive to channel/control topics and methods of dishonesty. In the immortal words of not only Shakespeare but also Neal Peart, “all the world’s indeed a stage/and we are merely players/performers and portrayers.” Dishonesty is (imho) an unavoidable side effect of this multiple-masks aspect of human nature. So societies say “don’t lie” but what they mean is “lie about only certain things in certain ways.” Deciding what the acceptable lies are in any given social context is sort of illuminating. Hence the bafflement of some serial philanderers; that territory used to be a safe lie.
A comment on the system of an informal "balance of favors" across the entirety of the interactions between two people:
A major advantage of accounting like this is that there's no need to agree on the value of any particular favor. At all times, you have a balance of trade with the other party. Either you've done more for them than they have for you, and your balance is negative, or they've done more for you than you have for them, and your balance is positive. But that balance can never be zero; it's a real number.
So instead of expecting an unachievable balance, the expectation is that the balance shouldn't be too far from zero in either direction. If you think it's too far toward you, you fix that by asking for a favor. If that doesn't work, you complain to your friends.
And at this point, it doesn't matter whether both parties agree on what the balance is! Even if I think you owe me a minor favor and you think I owe you one, that is fine - eventually someone will ask for a favor, and the balance will shift in their direction. No problems arise unless several recent favors all have sharply asymmetric valuations, such that each party's estimation of the overall balance is wildly different from the other's.
"Or, what about dueling traditions? We might again say the “workaday” assumption (that brings rules against murder) is violated, but that seems like a cheat. My vague understanding, at least of pistol dueling as seen in Hamilton, is it was less lethal than we might expect; and fell out of favor when better guns made it more lethal. But neither of these feels enough to satisfy, and we should demand satisfaction. Did the group gain something that was worth the potential loss of life?"
Leaving aside the "non-lethal" part and aiming straight at duels to the death - yes, there is a social benefit to having the accepted means of resolving (serious) disputes be a duel to the death, primarily based around social unity.
1) A duel to the death doesn't leave an aggrieved party who might want to relitigate matters. There's a winner (who isn't aggrieved) and a loser (who is dead).
2) In particular, in reasonably-small societies where individual charismatic politicians are a big deal, this tends to defang rabble-rousers and cut down on civil strife. A would-be revolutionary is going to wind up getting challenged to a duel at some point (or will themselves challenge existing leaders), which can end in three ways: a) they win, and are now in charge, b) they lose, and are dead, c) they repeatedly refuse to fight, lose credibility, and can't mount any kind of insurrection or resistance. All of those horns will generally avoid ongoing strife.
It should also be noted that duelling has mostly involved men, and not women. On societal timescales, men are quite expendable - losing 10% of your women loses 10% of the next generation, but this is not true of men.
There's still a link to the author's blog after the phrase “my previous thoughts on ” -- depending on how much you care about preserving anonymity, you might want to remove it or something.
Three comments: 1) it's very long and rather boring; 2) 'Good fences make good neighbours' is right up there with 'To thine own self be true' in the rhetorical irony top 10 - literary quotations often used by politicians, management and lifestyle guru / charlatans, unaware or unheeding their authors put them in the mouths of hypocrites and fools (Polonius, Frost's uptight WASP neighbour, etc.); 3) beef farming is unsustainable and the meat industry is killing the planet - in my anarchist paradise, it would be outlawed and all ranchers in California, Brazil and anywhere would be rounded up to work in the organic orchards and bug farms, and everyone would be much happier 😀
A fishery is basically the term for a specific group of fish in a specific place. The Grand Banks Cod Fishery, or the Antarctic Wale Fishery or whatever. You can have multiple different fisheries in the same place, but they're going to be different kinds of fish being pursued with different methods. Traditionally, each fishery was operated by a specific group, which had its own norms. These days that's less the case.
Also, it's probably worth pointing out that virtually all American whalers were out of New England, even the ones in the Pacific.
"Ellickson mentions a tribe where they “cut a finger from the hand of each of a man’s close female relatives after he dies”; what medical knowledge are they lacking that makes this seem welfare-maximizing?"
Did he give no further detail on what tribe this is from where, or are you leaving it out? An initial notion that comes to mind is the fear of witchcraft, so by social sanctions such as this you are trying to prevent bad outcomes - a potential witch wanting to curse someone to death has to bear in mind the loss of a finger, and other female family members who are at risk will try to prevent the witch from doing anything. Substitute "poison or otherwise encompass the murder of an abusive male" and the same applies. But not knowing the social and cultural background leaves a lot of information out - what is the position of unmarried daughters/sisters, of widows and mothers, when it comes to "close female family members"? Is this about sacrificing for the benefit of the deceased man in the afterlife - the custom of sati was religiously sanctioned in Hinduism as examples of chaste wives who would benefit their families and the souls of their husbands by virtue of their sacrifice, as well as widows being considered inauspicious and capable of bringing misfortune to others. There may be a lot of reasons for such a practice and I'd like to know more about the tribe in question.
Which brings us on to the naive belief in Science and Progress doing away with the Darkness of Superstition, and quite frankly I'm surprised that even a book from the 80s would be holding on to that kind of hopefulness:
“A tribe that used to turn to rain dancing during droughts thus is predicted to phase out that ritual after tribe members learn more meteorology. Tribes are predicted to abandon dangerous puberty rites after members obtain better medical information. As tribe members become more familiar with science in general, the status of their magicians and witch doctors should fall. As a more contemporary example, faith in astrology should correlate negatively with knowledge of astronomy. These propositions are potentially falsifiable.”
A tribe that knows to tune in to the nightly weather forecast on the TV news may still engage in rain dances as a matter of cultural heritage, and because it's fun to dress up and have dances as a social activity. Magicians and witch doctors will still have places as sources of traditional wisdom and advice on matters that "take two aspirin and call me in the morning" can't address - how do I propitiate my mother-in-law's ghost because my marriage is going badly and she always threatened to come back and haunt us because she never liked me? 'Dangerous puberty rites' may be tidied up by having trained medical doctors carry out FGM https://reproductive-health-journal.biomedcentral.com/articles/10.1186/s12978-019-0817-3 and for other rites, adolescents will always find ways to risky behaviour involving sex, drugs, fights, jumping off heights, driving very fast, and so on.
As for astronomy and astrology - tell that to all the psychic hotlines, ghost hunter TV shows, Tarot readings and astrology websites that are still up and running even though we have The Hubble Telescope etc.
I think one thing about norm-enforcement is that there's a Keynesian beauty contest thing going on. My actions (norm-violating, norm-enforcing, whatever) are going to be seen by others as more or less reasonable, and that will affect my reputation and my interactions with them. At the extreme unreasonableness end, it will even lead others to punish me for my norm violations. I think that's the driver for a tendency to maximize welfare (visible well-being).
When the healthy guy doesn't shovel his walk because he can't be arsed, everyone knows he's just being a lazy bum. When the 70 year old widow who uses a cane to walk doesn't shovel her walk, nobody thinks she's violating any norms. (And indeed, it's quite likely that some neighborhood kid or other will just go ahead and shovel her walk while he's shoveling his own, in my neighborhood.) This works well because it's visible. If the apparently healthy guy actually has some kind of back problem and shoveling the walk will leave him incapacitated for the next week, he'll probably get a lot of unjust pushback from the neighbors, because it sure *looks* like he could do it.
If I yell at the little old lady for not shoveling her walk, everyone is going to think I'm a jerk. If I yell at the apparently-healthy guy for not shoveling his walk, everyone is going to think I've got a point. And I am instinctively weighing that kind of consideration when I decide whether or not to yell at my neighbors for not shoveling their walks. (Well, okay, I'm probably never going to yell at them for that, but I guess I *could*.)
Norm-violations and norm-enforcement that takes into account visible stuff like this works passably well, right? You can tell I'm a jerk if I yell at the little old lady for not shoveling her walk, and others who see you treating me like a jerk can see why you did it and it makes sense, and....
Evolution does a lot of optimizing, but also often falls into weird and counterproductive situations where, say, 90% of your energy budget is spent on your spectacularly long and useless tail because that's what gets mates, or your species' evolution paints you into a corner that drives you to extinction as soon as some aspect of your environment changes. I think we should be unsurprised to see norms evolve in much the same way--sometimes giving us stunningly elegant solutions to problems, sometimes giving us solutions that more-or-less work but look like they're held together with duct tape and zip ties (think of DNA copying in eucaryotes and why we need telomeres), sometimes stuff that works according to the hill-climbing, mutation/selection/drift/flow logic of evolution but is honestly a fairly shitty solution.
It seems to me this kind of analysis is the basis for a lot of science fiction. Take Heinlein: many of his books (e.g. The Moon is a Harsh Mistress) run with the central point that future humans living in space will develop norms totally different from ours because of the radically different environment, and a lot of the narrative is built on his predictions of what those norms would be. Another example is 'The Expanse' with its Belter culture.
Keep in mind that ranches are big and require a long-term investment. Cattle ranching requires lots of land, and while a herd can be moved, but the land cannot. Not only that, but thanks to surveys, land law and easements, it is usually pretty clear who holds title to a given parcel of land at any given time, and who is permitted to enter on which parcel of land and for what purpose.
Further, because ranches are big long-term enterprises, there can only be so many ranches and ranchers in a given area, and it is likely that all the ranchers know one another. Their families probably all go way back. Moreover, all the ranchers will need favors or forbearances from one another from time to time (not necessarily business related), so it is best not to make long-term enemies.
Therefore, law or no law, the ranchers either have to get rid of each other (whether by violence or buyout), or they have to learn to live with one another. Wild West Game of Thrones aside, violence ends badly in the long-term for most participants, and even at best is unpredictable, especially in the age of firearms. Buyout is expensive. Learning to tolerate each other is probably the least-cost solution, not to mention the one least likely to attract heat from bankers, banditti, lawyers, and law enforcement.
In a similar vein to ranchers, dorm room anarchists assume that "primitive" tribes have no laws. In fact, such tribes typically have laws, in the sense that they have codes, often quite convoluted, governing who is allowed to marry whom, who is permitted to eat what and with whom, who is in the right in a particular type of dispute, who answers for the conduct of whom, etc.. Just that these codes are not written down or lawyered over.
Margaret Mead aside, a lot of these codes pertain to sex and family, probably because arguments over who is permitted to sleep with whom, over "is that really my kid?", over which kid is entitled to what, are the fastest way to tear a tribe apart or simply to lead to unnecessary intra-tribal conflict. And outside the tribe, you may be fair game for anyone. This is why tribes are so tribal. in the sense that it's "us against them" and at the same time, why tribal norms are so relentlessly enforced. They have to be.
For instance, in spite of his wartime prowess, Crazy Horse lost his status as Shirt Wearer in the Oglala Sioux, because he had an affair with a married woman. If he had abducted a woman from an enemy tribe, the outcome probably would have been different.
As to whether such a not-quite-anarchy solution works in other environments, one in which the participants are not incentivized to at least learn to tolerate one another - well, that is probably why city dwelling businessmen are much more litigious than ranchers or tribesmen, and city drug gangs are rumored at times to resort to even less cooperative solutions.
If the Code Duello was clear about nothing else, it was that only a nobleman could institute a duel, and a nobleman only need respond to a challenge from another nobleman.
Peons need not apply and cannot apply. For it is one thing if the Duke of Buggeringham gets into a fight with Lord FrouFrou over some perceived slight, but the social order cannot tolerate the son and heir of the duchy getting into a one-on-one fight with a peon on anything remotely resembling fair odds.
For if history and the fairy tales teach us nothing else, it is this: beware an ambitious poor boy with nothing to lose.
re: academia my sense was always that this was a kind of unofficial PR/advertising interest in publishers refraining from making a big stink. Indeed, since professors often decide what books whole classes use (often with much better profit margins than the academic books profs use) and are the ones often providing the content it would be idiotic for any publisher with an interest in having their books choosen to make a fuss about this. They also know that (more so pre-digital but scans are still inferior) established profs with the money to spare will probably buy copies of the books but while the grad students who lack the cash will favor the copies but, if they find the book appealing, will likely buy a copy at some point in their career.
It's more that the overhead of trying to do this via a formal legal process wouldn't be workable. For one, if an official copy lacks features (search) or has bad quality the publisher gets blamed and any rule they tried to articulate would run the risk of being abused at scale.
This isn't to say that it's not the kind of system held together by informal norms but that the publishers are part of this informal norm system.
My loose thinking about dueling is that it does actually serve the community pretty well. First, duels stop infinite vendetta cycles, which otherwise can seriously harm whole tribes. Everybody accepts that the point of the duel was to settle the matter once and for all. And second, when you're dueling over honor, it gives the community a certain number of people who have proven, skin-in-the-game, that they're willing to die to preserve their reputation.
I'd just like to say I appreciate the careful summaries and attention to detail of this review. I admit I'm glad the other reviews aren't this long, but I think this one justifies its length by giving me what I feel is a good sense of the book and its strengths and weaknesses.
The summary of Ellickson’s discussion of campus photocopying is a little ambiguous about fair use. For example, “fair-use doctrine is quite restrictive unless they get explicit permission,” can be read as suggesting that the fair use determination depends in some way on asking for permission. But the whole point of fair use is that it doesn’t require permission.
What’s really going on here, which the author gets at by referring to professors “avoiding learning the details of fair use doctrine,” is that professors almost universally believe that their photocopying is fair use. Making multiple copies for classroom use is one of the illustrative fair purposes in the statute. Professors rely on that fact and tend either not to know that it’s not conclusive or to assume that even if you do the rest of the analysis, it will obviously cash out in favor of fair use.
Ah yes, slime molds. As per usual with claims of solving NP-hard problems, the issue is that they are solving a slightly different problem than claimed.
Let's look at what they are actually solving for a moment. (Note: I'm using equality in the below in an arguably incorrect manner. Heads up.)
First, note that a slime mold (and indeed, any physics-based solution for pretty much any problem) has a finite maximum resolution. (You might argue what exactly said resolution is, but ultimately you can agree hopefully that there exists some floor here.)
Second, note that a slime mold has Θ(area) computational power (at least in theory. Biological constraints may reduce this).
Third, note that a slime mold has a finite maximum speed of information propagation (trivial bound: speed of light; practical bound - I don't know the actual figure but probably substantially less than the speed of sound). And hence requires O(max distance) time to compute a result. (Be it exact or α-approximation - if you require less than O(max distance) time, consider two points at (-L, 0) and (L, 0), and a third point at choice[(0, L/sqrt(3)), (0, -L/sqrt(3))], with L chosen to be large enough that the third point cannot communicate with the others before time is up. Without communication, you are forced to pick some common point for the Steiner node - but no matter what point you pick there's an adversarial choice where you are incorrect. Your best bet is to pick (0,0), but this results in a distance of (2 + 1/sqrt(3))L. Given that the optimal tree for both has a distance of 2L, this means you cannot approximate better than a factor of 1/(2 sqrt(3))... which is required for both exact solutions and α-approximation.)
And finally, note that the slime mold is not solving the problem exactly - assuming it is solving this particular problem exactly, it is solving the original problem within an error of... let's see.
Let's treat things as a grid, and work in distance units of our resolution. We have n nodes. Scale the problem such that the maximum dimension of the resulting problem is `s`. With this solution, we're moving every Steiner node by up to a half of a unit distance from the optimal tree; an optimal Steiner tree has at most n-1 nodes (and hence at most 2n-1 edges), so the cost changes by at most 0.5*(2n-1)=n-0.5 units. The optimal tree trivially has a distance of at least o(s). So overall our factor of error is (n-0.5)/o(s), or O(n/s).
Meanwhile, to get said factor of error we had O(s^2) area of slime mold computing for at a minimum of O(s) time [in general; obviously if a problem is restricted e.g. entirely 1d this does not happen]. Or O(s^3) total computational steps.
All told this works out to O(n^3 ε^-3) complexity. (...ish. I don't know how much of the slime mold is actually active during this entire time.)
Now, how does this compare to standard algorithms?
Steiner trees are NP-hard, yes, but Euclidean Steiner trees have α-approximation schemes for constant dimension (that is, for any given desired `ε>0`, you can construct a polynomial-time algorithm that is guaranteed to produce a result that is no more than `1+ε` times the optimal result).
The best I could find at a quick glance was O(n (log n + ε^(-C/ε))). You'll note that this is substantially better than the slime mold for large n.
(Interesting, the α-approximation schemes for Euclidean Steiner trees are also largely based around discretizing to a grid first - but they can form a finer grid because they don't spend anywhere near O(s^3) steps on the resulting grid.)
Or to put it another way: I'm not sure if we _care_ what the slime molds are doing, because we can already do better by simply exploring the entire area, sending back information about any nodes to a central spot, and doing the calculation there. (Well, we may care. But not for this particular reason.)
Seems like a problem that's better solved bottom up.
As far as I see it tit-for-tat morality and gossip as a solution to the free rider problem are well established results. They're sufficient to explain the observed phenomena without overreaching.
Elinor Ostrom got a Nobel Memorial prize for analyzing exactly this kind of stuff. I think she did a pretty good job. The book was written 30 years ago so I can understand if he doesn't mention her, but a review should show some basic familiarity with the way Ostrom described these norms.
You finally did it. You finally published something that is so long and meandering that my curiosity withered and died, even after trying to skip topics that just can't be that interesting (and were not). This was a rough start for community book reviews.
I wonder, as a thought experiment, if we applied a computer to defining counties, and states in the US, with the optimum defined as 'counties will have the smallest surface area (or some other simple geometric rule to minimize opportunity for gerrymandering in support of some group over some other group) to hold 150 people (Dunbar's number). States will be proportioned to hold as close as possible the an equal number of counties.'
Other than the absurdly high number of counties in dense cities (some buildings in New York or other dense urban core might have multiple counties, by floor perhaps?) and weird relationships for utilities (a thousand counties need to band together to pay for the sewer system for example), I wonder what would happen to governance if we limited the #'s to human understandable human walkable?
I suspect a lot of people would be really mad, because they would have to participate in governance and politics... I think many people (more urban than rural) are governmental free riders who enjoy nor having to worry about governance and such unless something triggers them for their cause du jour.
I suspect politics would be radically different (better?) when limited to human-understandable groups you could have actual human relationships with - the monthly 'County 101' dinner where all 150 voters get together for a meal and a conversation would be very different than the politics of today.
Many things would work better, and many things would be harder/worse. Interesting...
Great read! Although I'm still pretty confused on what the definition of "welfare" here is, which is obviously pretty fundamental. Could someone hopefully help and provide a more concrete definition? If it has to do with what the preferences an ordinary person holds, then how are we defining "ordinary person"- slightly different definitions (say mean vs. median vs. modal) seem to have the potential to provide radically different answers. I must be missing something.
I suppose I have three slightly different questions:
1. What's Ellickson's definition?
2. What definition does the author of this review think is best?
3. Is there a definition that anyone can think of that best steelmans the hypothesis.
I'm a bit of a topic expert on the history of dueling, so I noticed your aside about that.
There is a norm-centered explanation for dueling with explains it in terms compatible with Ellickson's theory. The duel (or rather the threat of injury or death in one) is often a form of norm enforcement that flourishes when recourse to formalized legal recourse is unavailable.
Thus, for example, a very prevalent and widely accepted cause of duels clear up to the end of the 19th century was incidents in which the challenged party was thought to have cheated at cards or other gambling games. More generally, betrayals outside the perceived scope of the law tended to trigger challenges that were socially supported.
Two factors that decreased the incidence of dueling were (a) increasing availability of justice through a court system and (b) increasing lethality of personal weapons. One significant transition point was reached in the 1840s when the sealed percussion cap greatly reduced the odds of a misfire or squib in damp European weather. The sword duel actually outlasted the pistol duel; the last recognizable formal sword duel was actually fought in 1967 in Paris!
I was astonished reading this that no consideration was given to the cost or reliability of turning to the legal system. There seems to be an underlying assumption by some legal economists that laws provide smooth and frictionless incentive structures. Litigation is expensive, mentally and emotionally taxing, can drag on for years, and can be uncertain in outcome. Lawyers are expensive and the legal system is extremely hostile to non-specialists. Invoking the legal system in many disputes could be viewed as a form of altruistic punishment - choosing the lose-lose option in order to deter future transgressors. In this case, the legal system would be best reserved for extreme edge cases – financially ruinous breaches of contract, life ruining torts, or reliably and investigating and prosecuting extreme examples of criminal self help (murder, assault).
Given the length, I assume you secretly entered your own contest.
The footnote and return anchor links don't work. They're formatted as absolute links to somewhere on slatestarcodex.com. They should be formatted as relative works so that they'll work regardless of where the page is located. (As in, the link href should only contain "#anchortolinkto", not the rest of the URL.)
" I’ll ask you to vote for your favorite,"
Could we use an alternative voting system?
I was surprised not to see any reference to Elinor Ostrom's work. This seems very related.
“ They might actually injure or kill it .”
I’m not a vegetarian so I can’t have a problem with killing it. But to injure it is to torture it, right? That’s not OK.
Also TL;DR. The review can’t be longer than the book you’re reviewing.
It's not welfare. It's peace.
We have all been spoiled by warlessness for so long, we have forgotten how anyone could live differently then this. The symptom of wealth can only exist in an era of peace, at least on a local level. Every good quality arises naturally while at peace.
Prediction and fulfillment, the favourite pastime of the brain, handles all the accounting and grumpiness for us on it's own. Provided there's not fights breaking out over every petty grievance, that is.
Every single example cited aims at having peaceful interactions for many generations, which is really hard if you're talking about weeks at see with lots of pointy sticks at your disposal, for example.
I think he explains the symptom, not the source.
(also, good god this not only is long, it feels long... very interesting but very few lighthearted moments that usually are the stamp at every base camp)
Yeah, I'm not buying the hypothesis. It seems like a noble savage stereotype targeted at rural communities translated into a hypothesis.
I've experienced more than a few of these close knit, traditional societies, both in the US and outside of it. Perhaps more importantly, I was in one of them as an insider and another as a quasi-insider. These communities were certainly orderly but because the enforcers of that order were private they in turn could use that authority for personal projects. A completely social society is a society without objective standards, which in turn means everything is a matter of social dominance. This might sound horrible but the advantage is that there is no dispassionate enforcement. If the law says X, and X is monstrous, you can at least be sure there's no bureaucrat paperclip maximizing X. The private enforcer will realize this is stupid and stop. And no one gossips, "Look at Scott, he stopped maximizing paperclips!"
But make no mistake. These societies can be very cruel. They can also be very kind. But they don't necessarily have to be. For example, a bunch of Southern plantation owners instead would also use this kind of enforcement and not to good ends. Lynching was violence aimed at social control very famously. This is without getting into problems inherent the system, like how they have forces inherently pushing them towards insularity, conformity, and (often hereditary) hierarchy.
These kind of rural, neighborly societies solve many problems caused by deracinated urban life. You'll never feel alone, or out of place, or like you don't have a purpose or place in life. But they have other issues.
It looks like there are links to the authors home page towards the bottom of this essay, which compromises anonymity.
> Patrons of a singles bar at O’Hare Airport
Are there really singles bars inside airports? How would that work?
I like the case studies better than the theorizing. Learning about what people actually do in these circumstances is fascinating. The theorizing seems inconclusive at best?
"with Marxism, which sees norms as serving only a small subset of a group"
Citation needed.
One of the bullet points towards the end of the review links to a post on the author's blog (or at least claims to do so). This would seem to violate the anonymity premise of the contest?
Turnbull's work on the Ik has been criticized by later anthropologists, who found it tendentious and even racist:
https://en.wikipedia.org/wiki/Ik_people#Criticism_of_Turnbull's_work
You could dismiss this as liberal-moralistic pearl clutching by the anthropologists, but the criticism is widespread, and often based on first hand experience.
Let the reader decide, I guess.
I've taken to saying things like "The law only matters if someone's willing to go to court". I've knowingly been the victim of lawbreaking - even in contexts where correcting it would be fairly painless and modestly profitable to me - and gone along with it because I didn't think they were doing anything immoral. (tl;dr, the one union job I've ever had, my union somehow managed to bargain away my legally guaranteed holiday pay. But the base pay was high enough that the deal still seemed more or less fair.)
Laws exist for the big offenses, the sort where you'd solve it by rounding up your family members to bring guns and ropes to bear if the law didn't exist. The sort where corporations have millions (or billions) at stake, or where you want to ensure that nobody screws up and kills a bunch of people. For all the smaller issues in the world, the time-honored rule of "Don't be a dick" is the true law. And it's mostly for the best.
I think your distinction between global and local maxima is a good articulation of why both totalitarianism and anarchism fail as (large scale) social structures: you need coordinated, organized efforts to overcome collective action problems and overcome high cost social transition - but you also need organic, bottom-up structures (like norms) to help optimize for welfare within a given local range of solutions.
I see that it has now been over a decade since I read & blogged about this book, but my blogging didn't add enough value to be worth linking to.
Turnbull is somewhat notorious now that he's not around to defend himself. It has been said that he liked groups that let him indulge his proclivity toward pederasty and hated those (like the Ik) that didn't, thus motivating his depiction of the latter.
There's a lawyer with a youtube channel called "Fudd Busters" devoted to debunking incorrect beliefs about gun laws. The "Fudd" comes from Elmer Fudd, and it's not just backwoods hunters who are ignorant. Federally licensed firearm dealers frequently don't know the actual law they deal with as part of their jobs.
https://www.youtube.com/c/FuddBusters
Brief review-of-the-review:
I found the summary of the book and its claims to be interesting, useful, and readable. The tone is a bit "inside baseball" by SSC standards, expecting that readers will be familiar with concepts from LessWrong and certain fields of econ. It didn't help that I had no idea what the subject and thesis of the book were until the middle of the review. Only at the very end did the practical relevance of Ellickson's thesis really click for me. The reviewer engages with Ellickson in a technical, nuts-and-bolts kind of way that I found unhelpful and a bit tedious, mostly due to my unfamiliarity. I did come away feeling like I understood the book, though, which is a win.
in East Texas a saw a loose horse beside the highway. I pulled over and nabbed the horse. when a pickup came by I waved it down. they just shook their heads. "Naw, that's so-and-so's horse, and he don't help us with our loose stock and we don't help with his."
sounds to me like I've got a basis for 450 page book.
Re the lying part. I think societies have an incentive to channel/control topics and methods of dishonesty. In the immortal words of not only Shakespeare but also Neal Peart, “all the world’s indeed a stage/and we are merely players/performers and portrayers.” Dishonesty is (imho) an unavoidable side effect of this multiple-masks aspect of human nature. So societies say “don’t lie” but what they mean is “lie about only certain things in certain ways.” Deciding what the acceptable lies are in any given social context is sort of illuminating. Hence the bafflement of some serial philanderers; that territory used to be a safe lie.
A comment on the system of an informal "balance of favors" across the entirety of the interactions between two people:
A major advantage of accounting like this is that there's no need to agree on the value of any particular favor. At all times, you have a balance of trade with the other party. Either you've done more for them than they have for you, and your balance is negative, or they've done more for you than you have for them, and your balance is positive. But that balance can never be zero; it's a real number.
So instead of expecting an unachievable balance, the expectation is that the balance shouldn't be too far from zero in either direction. If you think it's too far toward you, you fix that by asking for a favor. If that doesn't work, you complain to your friends.
And at this point, it doesn't matter whether both parties agree on what the balance is! Even if I think you owe me a minor favor and you think I owe you one, that is fine - eventually someone will ask for a favor, and the balance will shift in their direction. No problems arise unless several recent favors all have sharply asymmetric valuations, such that each party's estimation of the overall balance is wildly different from the other's.
"Or, what about dueling traditions? We might again say the “workaday” assumption (that brings rules against murder) is violated, but that seems like a cheat. My vague understanding, at least of pistol dueling as seen in Hamilton, is it was less lethal than we might expect; and fell out of favor when better guns made it more lethal. But neither of these feels enough to satisfy, and we should demand satisfaction. Did the group gain something that was worth the potential loss of life?"
Leaving aside the "non-lethal" part and aiming straight at duels to the death - yes, there is a social benefit to having the accepted means of resolving (serious) disputes be a duel to the death, primarily based around social unity.
1) A duel to the death doesn't leave an aggrieved party who might want to relitigate matters. There's a winner (who isn't aggrieved) and a loser (who is dead).
2) In particular, in reasonably-small societies where individual charismatic politicians are a big deal, this tends to defang rabble-rousers and cut down on civil strife. A would-be revolutionary is going to wind up getting challenged to a duel at some point (or will themselves challenge existing leaders), which can end in three ways: a) they win, and are now in charge, b) they lose, and are dead, c) they repeatedly refuse to fight, lose credibility, and can't mount any kind of insurrection or resistance. All of those horns will generally avoid ongoing strife.
It should also be noted that duelling has mostly involved men, and not women. On societal timescales, men are quite expendable - losing 10% of your women loses 10% of the next generation, but this is not true of men.
There's still a link to the author's blog after the phrase “my previous thoughts on ” -- depending on how much you care about preserving anonymity, you might want to remove it or something.
I haven’t counted, but it seems a bit longer than 500 word limit in the comp t&cs ... or are those ‘rules’ like the fences of shasta county? 😉
The link about open-source reveals the author. Scott should probably remove that one.
Three comments: 1) it's very long and rather boring; 2) 'Good fences make good neighbours' is right up there with 'To thine own self be true' in the rhetorical irony top 10 - literary quotations often used by politicians, management and lifestyle guru / charlatans, unaware or unheeding their authors put them in the mouths of hypocrites and fools (Polonius, Frost's uptight WASP neighbour, etc.); 3) beef farming is unsustainable and the meat industry is killing the planet - in my anarchist paradise, it would be outlawed and all ranchers in California, Brazil and anywhere would be rounded up to work in the organic orchards and bug farms, and everyone would be much happier 😀
The writing it's tedious, I'd rather had read the book. In fact, given the length, I feel like I have...
Granted, it's long, but it's a fantastic review that made me go right out and buy and book.
A fishery is basically the term for a specific group of fish in a specific place. The Grand Banks Cod Fishery, or the Antarctic Wale Fishery or whatever. You can have multiple different fisheries in the same place, but they're going to be different kinds of fish being pursued with different methods. Traditionally, each fishery was operated by a specific group, which had its own norms. These days that's less the case.
Also, it's probably worth pointing out that virtually all American whalers were out of New England, even the ones in the Pacific.
"Ellickson mentions a tribe where they “cut a finger from the hand of each of a man’s close female relatives after he dies”; what medical knowledge are they lacking that makes this seem welfare-maximizing?"
Did he give no further detail on what tribe this is from where, or are you leaving it out? An initial notion that comes to mind is the fear of witchcraft, so by social sanctions such as this you are trying to prevent bad outcomes - a potential witch wanting to curse someone to death has to bear in mind the loss of a finger, and other female family members who are at risk will try to prevent the witch from doing anything. Substitute "poison or otherwise encompass the murder of an abusive male" and the same applies. But not knowing the social and cultural background leaves a lot of information out - what is the position of unmarried daughters/sisters, of widows and mothers, when it comes to "close female family members"? Is this about sacrificing for the benefit of the deceased man in the afterlife - the custom of sati was religiously sanctioned in Hinduism as examples of chaste wives who would benefit their families and the souls of their husbands by virtue of their sacrifice, as well as widows being considered inauspicious and capable of bringing misfortune to others. There may be a lot of reasons for such a practice and I'd like to know more about the tribe in question.
Which brings us on to the naive belief in Science and Progress doing away with the Darkness of Superstition, and quite frankly I'm surprised that even a book from the 80s would be holding on to that kind of hopefulness:
“A tribe that used to turn to rain dancing during droughts thus is predicted to phase out that ritual after tribe members learn more meteorology. Tribes are predicted to abandon dangerous puberty rites after members obtain better medical information. As tribe members become more familiar with science in general, the status of their magicians and witch doctors should fall. As a more contemporary example, faith in astrology should correlate negatively with knowledge of astronomy. These propositions are potentially falsifiable.”
A tribe that knows to tune in to the nightly weather forecast on the TV news may still engage in rain dances as a matter of cultural heritage, and because it's fun to dress up and have dances as a social activity. Magicians and witch doctors will still have places as sources of traditional wisdom and advice on matters that "take two aspirin and call me in the morning" can't address - how do I propitiate my mother-in-law's ghost because my marriage is going badly and she always threatened to come back and haunt us because she never liked me? 'Dangerous puberty rites' may be tidied up by having trained medical doctors carry out FGM https://reproductive-health-journal.biomedcentral.com/articles/10.1186/s12978-019-0817-3 and for other rites, adolescents will always find ways to risky behaviour involving sex, drugs, fights, jumping off heights, driving very fast, and so on.
As for astronomy and astrology - tell that to all the psychic hotlines, ghost hunter TV shows, Tarot readings and astrology websites that are still up and running even though we have The Hubble Telescope etc.
I think one thing about norm-enforcement is that there's a Keynesian beauty contest thing going on. My actions (norm-violating, norm-enforcing, whatever) are going to be seen by others as more or less reasonable, and that will affect my reputation and my interactions with them. At the extreme unreasonableness end, it will even lead others to punish me for my norm violations. I think that's the driver for a tendency to maximize welfare (visible well-being).
When the healthy guy doesn't shovel his walk because he can't be arsed, everyone knows he's just being a lazy bum. When the 70 year old widow who uses a cane to walk doesn't shovel her walk, nobody thinks she's violating any norms. (And indeed, it's quite likely that some neighborhood kid or other will just go ahead and shovel her walk while he's shoveling his own, in my neighborhood.) This works well because it's visible. If the apparently healthy guy actually has some kind of back problem and shoveling the walk will leave him incapacitated for the next week, he'll probably get a lot of unjust pushback from the neighbors, because it sure *looks* like he could do it.
If I yell at the little old lady for not shoveling her walk, everyone is going to think I'm a jerk. If I yell at the apparently-healthy guy for not shoveling his walk, everyone is going to think I've got a point. And I am instinctively weighing that kind of consideration when I decide whether or not to yell at my neighbors for not shoveling their walks. (Well, okay, I'm probably never going to yell at them for that, but I guess I *could*.)
Norm-violations and norm-enforcement that takes into account visible stuff like this works passably well, right? You can tell I'm a jerk if I yell at the little old lady for not shoveling her walk, and others who see you treating me like a jerk can see why you did it and it makes sense, and....
Evolution does a lot of optimizing, but also often falls into weird and counterproductive situations where, say, 90% of your energy budget is spent on your spectacularly long and useless tail because that's what gets mates, or your species' evolution paints you into a corner that drives you to extinction as soon as some aspect of your environment changes. I think we should be unsurprised to see norms evolve in much the same way--sometimes giving us stunningly elegant solutions to problems, sometimes giving us solutions that more-or-less work but look like they're held together with duct tape and zip ties (think of DNA copying in eucaryotes and why we need telomeres), sometimes stuff that works according to the hill-climbing, mutation/selection/drift/flow logic of evolution but is honestly a fairly shitty solution.
The bit about tomatoes made me suspicious. You _can't_ just replant tomatoes, because you plant in the spring. This author seems dodgy.
It seems to me this kind of analysis is the basis for a lot of science fiction. Take Heinlein: many of his books (e.g. The Moon is a Harsh Mistress) run with the central point that future humans living in space will develop norms totally different from ours because of the radically different environment, and a lot of the narrative is built on his predictions of what those norms would be. Another example is 'The Expanse' with its Belter culture.
Keep in mind that ranches are big and require a long-term investment. Cattle ranching requires lots of land, and while a herd can be moved, but the land cannot. Not only that, but thanks to surveys, land law and easements, it is usually pretty clear who holds title to a given parcel of land at any given time, and who is permitted to enter on which parcel of land and for what purpose.
Further, because ranches are big long-term enterprises, there can only be so many ranches and ranchers in a given area, and it is likely that all the ranchers know one another. Their families probably all go way back. Moreover, all the ranchers will need favors or forbearances from one another from time to time (not necessarily business related), so it is best not to make long-term enemies.
Therefore, law or no law, the ranchers either have to get rid of each other (whether by violence or buyout), or they have to learn to live with one another. Wild West Game of Thrones aside, violence ends badly in the long-term for most participants, and even at best is unpredictable, especially in the age of firearms. Buyout is expensive. Learning to tolerate each other is probably the least-cost solution, not to mention the one least likely to attract heat from bankers, banditti, lawyers, and law enforcement.
In a similar vein to ranchers, dorm room anarchists assume that "primitive" tribes have no laws. In fact, such tribes typically have laws, in the sense that they have codes, often quite convoluted, governing who is allowed to marry whom, who is permitted to eat what and with whom, who is in the right in a particular type of dispute, who answers for the conduct of whom, etc.. Just that these codes are not written down or lawyered over.
Margaret Mead aside, a lot of these codes pertain to sex and family, probably because arguments over who is permitted to sleep with whom, over "is that really my kid?", over which kid is entitled to what, are the fastest way to tear a tribe apart or simply to lead to unnecessary intra-tribal conflict. And outside the tribe, you may be fair game for anyone. This is why tribes are so tribal. in the sense that it's "us against them" and at the same time, why tribal norms are so relentlessly enforced. They have to be.
For instance, in spite of his wartime prowess, Crazy Horse lost his status as Shirt Wearer in the Oglala Sioux, because he had an affair with a married woman. If he had abducted a woman from an enemy tribe, the outcome probably would have been different.
As to whether such a not-quite-anarchy solution works in other environments, one in which the participants are not incentivized to at least learn to tolerate one another - well, that is probably why city dwelling businessmen are much more litigious than ranchers or tribesmen, and city drug gangs are rumored at times to resort to even less cooperative solutions.
Re: duelling.
If the Code Duello was clear about nothing else, it was that only a nobleman could institute a duel, and a nobleman only need respond to a challenge from another nobleman.
Peons need not apply and cannot apply. For it is one thing if the Duke of Buggeringham gets into a fight with Lord FrouFrou over some perceived slight, but the social order cannot tolerate the son and heir of the duchy getting into a one-on-one fight with a peon on anything remotely resembling fair odds.
For if history and the fairy tales teach us nothing else, it is this: beware an ambitious poor boy with nothing to lose.
re: academia my sense was always that this was a kind of unofficial PR/advertising interest in publishers refraining from making a big stink. Indeed, since professors often decide what books whole classes use (often with much better profit margins than the academic books profs use) and are the ones often providing the content it would be idiotic for any publisher with an interest in having their books choosen to make a fuss about this. They also know that (more so pre-digital but scans are still inferior) established profs with the money to spare will probably buy copies of the books but while the grad students who lack the cash will favor the copies but, if they find the book appealing, will likely buy a copy at some point in their career.
It's more that the overhead of trying to do this via a formal legal process wouldn't be workable. For one, if an official copy lacks features (search) or has bad quality the publisher gets blamed and any rule they tried to articulate would run the risk of being abused at scale.
This isn't to say that it's not the kind of system held together by informal norms but that the publishers are part of this informal norm system.
My loose thinking about dueling is that it does actually serve the community pretty well. First, duels stop infinite vendetta cycles, which otherwise can seriously harm whole tribes. Everybody accepts that the point of the duel was to settle the matter once and for all. And second, when you're dueling over honor, it gives the community a certain number of people who have proven, skin-in-the-game, that they're willing to die to preserve their reputation.
I'd just like to say I appreciate the careful summaries and attention to detail of this review. I admit I'm glad the other reviews aren't this long, but I think this one justifies its length by giving me what I feel is a good sense of the book and its strengths and weaknesses.
The summary of Ellickson’s discussion of campus photocopying is a little ambiguous about fair use. For example, “fair-use doctrine is quite restrictive unless they get explicit permission,” can be read as suggesting that the fair use determination depends in some way on asking for permission. But the whole point of fair use is that it doesn’t require permission.
What’s really going on here, which the author gets at by referring to professors “avoiding learning the details of fair use doctrine,” is that professors almost universally believe that their photocopying is fair use. Making multiple copies for classroom use is one of the illustrative fair purposes in the statute. Professors rely on that fact and tend either not to know that it’s not conclusive or to assume that even if you do the rest of the analysis, it will obviously cash out in favor of fair use.
Ah yes, slime molds. As per usual with claims of solving NP-hard problems, the issue is that they are solving a slightly different problem than claimed.
Let's look at what they are actually solving for a moment. (Note: I'm using equality in the below in an arguably incorrect manner. Heads up.)
First, note that a slime mold (and indeed, any physics-based solution for pretty much any problem) has a finite maximum resolution. (You might argue what exactly said resolution is, but ultimately you can agree hopefully that there exists some floor here.)
Second, note that a slime mold has Θ(area) computational power (at least in theory. Biological constraints may reduce this).
Third, note that a slime mold has a finite maximum speed of information propagation (trivial bound: speed of light; practical bound - I don't know the actual figure but probably substantially less than the speed of sound). And hence requires O(max distance) time to compute a result. (Be it exact or α-approximation - if you require less than O(max distance) time, consider two points at (-L, 0) and (L, 0), and a third point at choice[(0, L/sqrt(3)), (0, -L/sqrt(3))], with L chosen to be large enough that the third point cannot communicate with the others before time is up. Without communication, you are forced to pick some common point for the Steiner node - but no matter what point you pick there's an adversarial choice where you are incorrect. Your best bet is to pick (0,0), but this results in a distance of (2 + 1/sqrt(3))L. Given that the optimal tree for both has a distance of 2L, this means you cannot approximate better than a factor of 1/(2 sqrt(3))... which is required for both exact solutions and α-approximation.)
And finally, note that the slime mold is not solving the problem exactly - assuming it is solving this particular problem exactly, it is solving the original problem within an error of... let's see.
Let's treat things as a grid, and work in distance units of our resolution. We have n nodes. Scale the problem such that the maximum dimension of the resulting problem is `s`. With this solution, we're moving every Steiner node by up to a half of a unit distance from the optimal tree; an optimal Steiner tree has at most n-1 nodes (and hence at most 2n-1 edges), so the cost changes by at most 0.5*(2n-1)=n-0.5 units. The optimal tree trivially has a distance of at least o(s). So overall our factor of error is (n-0.5)/o(s), or O(n/s).
Meanwhile, to get said factor of error we had O(s^2) area of slime mold computing for at a minimum of O(s) time [in general; obviously if a problem is restricted e.g. entirely 1d this does not happen]. Or O(s^3) total computational steps.
All told this works out to O(n^3 ε^-3) complexity. (...ish. I don't know how much of the slime mold is actually active during this entire time.)
Now, how does this compare to standard algorithms?
Steiner trees are NP-hard, yes, but Euclidean Steiner trees have α-approximation schemes for constant dimension (that is, for any given desired `ε>0`, you can construct a polynomial-time algorithm that is guaranteed to produce a result that is no more than `1+ε` times the optimal result).
The best I could find at a quick glance was O(n (log n + ε^(-C/ε))). You'll note that this is substantially better than the slime mold for large n.
(Interesting, the α-approximation schemes for Euclidean Steiner trees are also largely based around discretizing to a grid first - but they can form a finer grid because they don't spend anywhere near O(s^3) steps on the resulting grid.)
Or to put it another way: I'm not sure if we _care_ what the slime molds are doing, because we can already do better by simply exploring the entire area, sending back information about any nodes to a central spot, and doing the calculation there. (Well, we may care. But not for this particular reason.)
Seems like a problem that's better solved bottom up.
As far as I see it tit-for-tat morality and gossip as a solution to the free rider problem are well established results. They're sufficient to explain the observed phenomena without overreaching.
Elinor Ostrom got a Nobel Memorial prize for analyzing exactly this kind of stuff. I think she did a pretty good job. The book was written 30 years ago so I can understand if he doesn't mention her, but a review should show some basic familiarity with the way Ostrom described these norms.
You finally did it. You finally published something that is so long and meandering that my curiosity withered and died, even after trying to skip topics that just can't be that interesting (and were not). This was a rough start for community book reviews.
I wonder, as a thought experiment, if we applied a computer to defining counties, and states in the US, with the optimum defined as 'counties will have the smallest surface area (or some other simple geometric rule to minimize opportunity for gerrymandering in support of some group over some other group) to hold 150 people (Dunbar's number). States will be proportioned to hold as close as possible the an equal number of counties.'
Other than the absurdly high number of counties in dense cities (some buildings in New York or other dense urban core might have multiple counties, by floor perhaps?) and weird relationships for utilities (a thousand counties need to band together to pay for the sewer system for example), I wonder what would happen to governance if we limited the #'s to human understandable human walkable?
I suspect a lot of people would be really mad, because they would have to participate in governance and politics... I think many people (more urban than rural) are governmental free riders who enjoy nor having to worry about governance and such unless something triggers them for their cause du jour.
I suspect politics would be radically different (better?) when limited to human-understandable groups you could have actual human relationships with - the monthly 'County 101' dinner where all 150 voters get together for a meal and a conversation would be very different than the politics of today.
Many things would work better, and many things would be harder/worse. Interesting...
Great read! Although I'm still pretty confused on what the definition of "welfare" here is, which is obviously pretty fundamental. Could someone hopefully help and provide a more concrete definition? If it has to do with what the preferences an ordinary person holds, then how are we defining "ordinary person"- slightly different definitions (say mean vs. median vs. modal) seem to have the potential to provide radically different answers. I must be missing something.
I suppose I have three slightly different questions:
1. What's Ellickson's definition?
2. What definition does the author of this review think is best?
3. Is there a definition that anyone can think of that best steelmans the hypothesis.
I'm a bit of a topic expert on the history of dueling, so I noticed your aside about that.
There is a norm-centered explanation for dueling with explains it in terms compatible with Ellickson's theory. The duel (or rather the threat of injury or death in one) is often a form of norm enforcement that flourishes when recourse to formalized legal recourse is unavailable.
Thus, for example, a very prevalent and widely accepted cause of duels clear up to the end of the 19th century was incidents in which the challenged party was thought to have cheated at cards or other gambling games. More generally, betrayals outside the perceived scope of the law tended to trigger challenges that were socially supported.
Two factors that decreased the incidence of dueling were (a) increasing availability of justice through a court system and (b) increasing lethality of personal weapons. One significant transition point was reached in the 1840s when the sealed percussion cap greatly reduced the odds of a misfire or squib in damp European weather. The sword duel actually outlasted the pistol duel; the last recognizable formal sword duel was actually fought in 1967 in Paris!
I was astonished reading this that no consideration was given to the cost or reliability of turning to the legal system. There seems to be an underlying assumption by some legal economists that laws provide smooth and frictionless incentive structures. Litigation is expensive, mentally and emotionally taxing, can drag on for years, and can be uncertain in outcome. Lawyers are expensive and the legal system is extremely hostile to non-specialists. Invoking the legal system in many disputes could be viewed as a form of altruistic punishment - choosing the lose-lose option in order to deter future transgressors. In this case, the legal system would be best reserved for extreme edge cases – financially ruinous breaches of contract, life ruining torts, or reliably and investigating and prosecuting extreme examples of criminal self help (murder, assault).