Why Wars Break Out

 

I’ve undertaken the unhappy task of arguing that the probability of the outbreak of a major war in the next president’s term or terms is greater than it has been since the end of the Cold War, and greater than at many points during the Cold War.

An interesting argument ensued at the tail end of the comments on my Trump–>Armagedon post. Beginning at comment 88, Joseph Kulisics and Chuck Walla debated whether it was possible to disambiguate my predictions about the probability of this kind of event from my unfounded opinion.

This, roughly, was my answer (I’ve rewritten it to make it clearer):

I in no way believe I can predict the future well enough to offer a formal proof that Trump–>Armageddon. I’m responding to Genferei’s request that I sketch out scenarios under which this might happen without leaving out steps, handwaving, amateur psychology, or appeals to authority.

Predicting the behavior of dynamical systems is very tricky, and prone to modelling and initial-condition errors. I make no claim of an ability to predict the future of international relations to a degree of accuracy that would satisfy the standards of a climate scientist, no less a physicist. Nor do I wish to dress up “my opinion” in purely ornamental mathematics.

Leave all claims about Trump aside for the moment.

The first part of the task is to satisfy readers that no matter who is elected, the likelihood of major war is greater than it has been since the end of the Cold War, and greater than at many points during the Cold War. And to do so in a way that’s an improvement on “unfounded opinion.”

First question: Do you think anyone has an unusually good record of making generally reliable predictions about when war is apt to break out and why? Many have tried, that’s for sure. Volume upon volume has been written by international relations theorists who have tried to find patterns in the history of warfare that might suggest how better to predict wars. We’ve discussed, for example, Graham Allison’s work on China and the Thucydides trap at length here.

For another example, my father recently wrote what amounts to an extended review of Stephen Pinker’s book, The Better Angels of Our Nature, Pinker, as you probably know, Whiggishly argued that the world is getting more peaceful by the day. My father didn’t think much of Pinker’s skill as a historian.

If you read it, you’ll understand the question I’d put to Joseph Kulisics: Does he believe wars have causes, or do they just break out, as Lewis Fry Richardson proposed, following a Poisson distribution?

I suspect they have causes. I think they’re more likely to break out under — roughly — these circumstances:

Screen Shot 2016-03-19 at 15.06.41

Would you agree that such a model could be empirically tested? If not, then perhaps you’re defining the question as unanswerable.

But assume it can be, and that it’s a generally reliable heuristic. If I could demonstrate that to your satisfaction, would you agree that my first paragraph isn’t an unfounded opinion, but a reasoned one?

There are 49 comments.

  1. 1
  2. 2
  1. ToryWarWriter Thatcher

    I am about to read your piece Claire, and till opened your link to your fathers post, I never triggered on who he was.

    Anyways, I will read his piece then get back to you, but my first thought is to say that peace is an unnatural state for mankind and that war is the constant, the thing we are always doing.

    • #1
    • March 19, 2016, at 8:13 AM PDT
    • Like
  2. Profile Photo Member

    ToryWarWriter: peace is an unnatural state for mankind

    Portfolio advice: buy stock in companies that make flags for Gold Star mothers. (They’re probably Chinese companies.) Americans who think peace is natural state for humankind (Americans have moved on from “mankind”)have the memories and historical understanding of stones. Fukuyama didn’t have a clue.

    • #2
    • March 19, 2016, at 8:44 AM PDT
    • Like
  3. Michael Collins Member

    Claire Berlinski, Ed.: Would you agree that such a model could be empirically tested? If not, then perhaps you’re defining the question as unanswerable.

    By definition “empirical testing” means waiting to see whether war actually breaks out under the defined conditions. If preventing war is your goal empirical testing is useless, because it means breaking the egg first to find out whether it is going to be bad.

    Modeling and gaming through different scenarios can be helpful in forming good judgement, but they are useless if you don’t have a strong foundation of experience and good judgement to begin with.

    I agree with the poster about two things. 1) We live in dangerous times. 2) Trump does not have the “foundation of good judgement” required in a president.

    • #3
    • March 19, 2016, at 8:55 AM PDT
    • Like
  4. I Walton Member

    Russ Roberts questions whether economics can be empirically tested and in economics every transaction leaves numbers behind and lots of tested hypotheses about those transactions. War and peace? That’s a tough one I seriously doubt its causes can be statistically tested. Kagan’s “The causes of War” updates Thucydides, his specialty, through relatively contemporary wars and near wars. It’s a great book on the subject, but Kagan draws on wisdom, judgment and history, probably as good as we can do. There are simply too many variables, too many unknowns, too many partially knowns and too few wars in any sample to provide statistical testing. Another point Roberts frequently makes is that statistical results don’t change minds even when carefully crafted. So judgement and wisdom is ok and what you’re good at. Will now read your father.

    • #4
    • March 19, 2016, at 8:59 AM PDT
    • Like
  5. Claire Berlinski, Ed. Editor
    Claire Berlinski, Ed. Post author

    Michael Collins: By definition “empirical testing” means waiting to see whether war actually breaks out under the defined conditions

    Not necessarily. You can test it against the actual historic record.

    • #5
    • March 19, 2016, at 9:06 AM PDT
    • Like
  6. Hang On Member

    I would submit that a major war has already broken out and has been ongoing since the Iranian Revolution in 1979. It is a war between Sunni and Shia and the non-Islamic world. It has the intricacies, complexity and confusion of the Thirty Years War and is a religious-political one with many different stages, some of which we have already seen. At some point historiography will put all of these together.

    • #6
    • March 19, 2016, at 9:11 AM PDT
    • Like
  7. Jules PA Member

    I am enjoying your father’s article. Please tell him, “Thank you.”

    What is the date of that publication?

    This passage strikes me,

    ROBERT SOUTHEY was a poet, a dreamer, and a dunce. Born in 1774, he died in 1843. Like Wordsworth, he had been a supporter of the French Revolution; and, like Wordsworth, he had been disappointed by the revolution that he had supported.

    We people often discover people who seem to give voice to our dreams, then give them our full-fledged support in an effort to make dreams come true. But we are often disappointed.

    We need to learn from history, and read the signals it gives us.

    • #7
    • March 19, 2016, at 9:17 AM PDT
    • Like
  8. Jules PA Member

    And if your father’s description of war horrors wasn’t Armageddon, the ultimate in human destruction, I fear for the future, as it is unimaginable.

    Two hundred and thirty one million men, women, and children died violently in the twentieth century, shot over open pits, murdered in secret police cellars, asphyxiated in Nazi gas ovens, worked to death in Arctic mines or timber camps, the victims of deliberately contrived famines or lunatic industrial experiments, whole populations ravaged by alien armies, bombed to smithereens, or sent to wander in their exiled millions across all the violated borders of Europe and Asia.4

    • #8
    • March 19, 2016, at 9:19 AM PDT
    • Like
  9. ToryWarWriter Thatcher

    A couple years ago I was at a US war college event. We were discussing how to make wargamming into a bit more than just an intellectual curiosity and formalizing the ideas.

    Anyways one of our people mentioned that he has been running simulations with mid ranking American army officers. And he was running into a big problem. It was next to impossible to get American officers to initiate hostilities. That the only way he could get them to attack was to assassinate (in game terms), the senior US leadership and humiliate the troops.

    I can only imagine that in the senior levels of the Military it has only gotten worse under Obama.

    My own gaming experience in a ISIL game where I was playing Iran against an American army officer playing the US forces in Iraq. He kept cooperating with me, despite the fact that I kept trying to make him look bad. It was just not in his worldview that I might have different victory conditions than him.

    • #9
    • March 19, 2016, at 9:38 AM PDT
    • Like
  10. Claire Berlinski, Ed. Editor
    Claire Berlinski, Ed. Post author

    Jules PA: I am enjoying your father’s article. Please tell him, “Thank you.”

    I will. I’m sure he’ll be pleased to know you found it worthwhile.

    • #10
    • March 19, 2016, at 9:39 AM PDT
    • Like
  11. genferei Member

    Claire Berlinski, Ed.:

    Michael Collins: By definition “empirical testing” means waiting to see whether war actually breaks out under the defined conditions

    Not necessarily. You can test it against the actual historic record.

    Then I can find an infinite number of models that will perfectly fit the historical record. That you find them more or less convincing will have nothing to do with their mathematical properties, but their narrative ones. So dispense with the maths and do history/social science the way it has always been done: with (competing) stories.

    • #11
    • March 19, 2016, at 9:47 AM PDT
    • Like
  12. Jules PA Member

    from the David Berlinski article,

    So long as men have memories, nothing done in the present can ever be independent of the past. Men are haunted by memory and desire; influence seeps down through history, a viscous, lava-like flow.

    • #12
    • March 19, 2016, at 9:48 AM PDT
    • Like
  13. ToryWarWriter Thatcher

    Having read a bit now of the previous piece and your fathers piece I think we need to define terms here.

    Please define major war?

    Your previous piece includes limited nuclear exchange. The chances are high for such an occurrence no matter who is in charge is higher, though still not as high I would think during the Carter Presidential years as I have previously argued.

    • #13
    • March 19, 2016, at 9:54 AM PDT
    • Like
  14. Instugator Thatcher

    ToryWarWriter:

    My own gaming experience in a ISIL game where I was playing Iran against an American army officer playing the US forces in Iraq. He kept cooperating with me, despite the fact that I kept trying to make him look bad. It was just not in his worldview that I might have different victory conditions than him.

    True this. It is an American trait to seek our own victory conditions and not worry so much about how we are perceived.

    Wargames are not designed to be won, but to explore the decision making process used to deal with situations the adjudicators wish to explore.

    • #14
    • March 19, 2016, at 9:55 AM PDT
    • Like
  15. TG Thatcher
    TG

    Did you send a “Katie rule” notice to Joseph?

    Claire Berlinski, Ed.:

    Michael Collins: By definition “empirical testing” means waiting to see whether war actually breaks out under the defined conditions

    Not necessarily. You can test it against the actual historic record.

    Even when you “test” it against the actual historic record, a great deal of uncertainty remains – problems with fairly small data sets, building models (simplified by necessity!) based on those data sets, and then testing the skill of your model against an even smaller data set that you excluded from your “calibration” set … Others can express this better than I can. The end point I’m getting to is even if the “test against the historic record” looks to provide perfect results, there are reasons why there will be a non-infinitesimal probability that the results of your model “matched” reality by accident. Difficulties abound.

    • #15
    • March 19, 2016, at 10:00 AM PDT
    • Like
  16. Claire Berlinski, Ed. Editor
    Claire Berlinski, Ed. Post author

    genferei:

    mathematical properties, but their narrative ones. So dispense with the maths and do history/social science the way it has always been done: with (competing) stories.

    The mathematical properties are relatively simple, but we do require some mathematical terms. Such as “one, two, three.” Okay with you?

    • #16
    • March 19, 2016, at 10:02 AM PDT
    • Like
  17. I Walton Member

    Peter Robinson’s interview with David Berlinsky on uncommon knowledge a year or more ago, dealt with some of the material in the sited article. There are two phenomenal Robinson interviews with Belinsky.

    • #17
    • March 19, 2016, at 10:05 AM PDT
    • Like
  18. Tim Wright Inactive

    Reading the biography of Stalin by Stephen Kotkin can be a chilling exercise. Not least, his explanation of the runup to World War I, and how each participant carefully calculated how it was to their benefit and how they would succeed. Nobody “sleepwalked” into that war.

    Is the same thing happening now? I don’t think you can model these things. I think it comes down to a political analysis of a country’s leadership and their aims. For example, I believe both Russia and China are carefully calculating how far they can push it, and how weak the response will be.

    Tim

    • #18
    • March 19, 2016, at 10:05 AM PDT
    • Like
  19. Claire Berlinski, Ed. Editor
    Claire Berlinski, Ed. Post author

    TG: Difficulties abound.

    No argument from me. I went out to get milk this morning and asked myself why I’d committed myself to writing a dissertation-length thesis about this when I’m apt to change far more minds by leaving out steps, handwaving, amateur psychology, and appeals to authority. Arguably, since I’m sure I’m right about this, my moral obligation is to change as many minds as I can, via any method that works.

    • #19
    • March 19, 2016, at 10:07 AM PDT
    • Like
  20. Jules PA Member

    from the David Berlinski article

    What makes the Great Terror terrible, Schlögel correctly observed, “is not merely the number of its victims,” but the ambitions of the state that undertook their destruction. The Great Terror was intended to reduce one sixth of the world’s population to a state of gibbering hysteria. In this, it was successful.

    We like to think in America that the “state” is operated by the people, the citizens. But in the end, doesn’t any “state” consist of individuals given positions of leadership to make decisions as THEY see them.

    The person who controls the deployment of America’s nuclear triad has tremendous power to sway the path of the civilized world.

    • #20
    • March 19, 2016, at 10:07 AM PDT
    • Like
  21. Barfly Member

    Claire Berlinski, Ed.:If you read it, you’ll understand the question I’d put to Joseph Kulisics: Does he believe wars have causes, or do they just break out, as Lewis Fry Richardson proposed, following a Poisson distribution?

    Every event at a larger scale than the subatomic does have a direct and unambiguous cause. In fact the web of directly causal links is what defines perceptible realms of events, particularly at different scales. Those different perceptible scales aren’t limited to physical dimensions.

    Fuzzy thinking conflates our perceptions of different realms. Flawed analogies apply the lessons drawn from one realm to another. It’s idiocy for sophomore philosophers to argue that societal differences are governed by the principle of relativity, but they do it anyway.

    I think your question assumes a false equivalence of two different views into two distinct realms. Every war has proximate direct causes, without which the conflict would not have started. As we investigate those we find they branch and proliferate beyond our limited bookkeeping; we’ve moved from a local to a broader point of view. In that more global realm details of each war are supplanted by awareness of all the others.

    There we have statistical knowledge with its own strengths and limitations. It’s only our limited human working set that differs between the two realms; wars have proximate direct causes and are statistically predictable.

    • #21
    • March 19, 2016, at 10:08 AM PDT
    • Like
  22. Claire Berlinski, Ed. Editor
    Claire Berlinski, Ed. Post author

    TG:

    Did you send a “Katie rule” notice to Joseph?

    No, you’re right, I should.

    • #22
    • March 19, 2016, at 10:10 AM PDT
    • Like
  23. Michael Collins Member

    Claire Berlinski, Ed.:

    TG: Difficulties abound.

    No argument from me. I went out to get milk this morning and asked myself why I’d committed myself to writing a dissertation-length thesis about this when I’m apt to change far more minds by leaving out steps, handwaving, amateur psychology, and appeals to authority. Arguably, since I’m sure I’m right about this, my moral obligation is to change as many minds as I can, via any method that works.

    Educating a few minds is of more value than changing many.

    • #23
    • March 19, 2016, at 10:14 AM PDT
    • Like
  24. civil westman Inactive

    I do not believe historical modeling can result in a testable method of prediction. To paraphrase, “What did the President know and when did he know it?” – applying this question to every consequential decision maker in the causal chain leading to initiation of hostilities, I believe, presents an insurmountable problem in understanding why any past war started.

    Whatever we know today about the antecedents of any past war can only be a gross approximation. To paraphrase again, there are far too many “unknown unknowns” – as who knew (was it true?) and thought (reasonably or not?) what and when and why – to permit any reliable approximation of “the truth.”.

    • #24
    • March 19, 2016, at 10:40 AM PDT
    • Like
  25. Jules PA Member

    Claire Berlinski, Ed.:

    Jules PA: I am enjoying your father’s article. Please tell him, “Thank you.”

    I will. I’m sure he’ll be pleased to know you found it worthwhile.

    and a technical comment, I found the font of that article easy to read. I can’t explain it, but there was a roundness that permitted my eyes to keep moving forward.

    Your dad has a fantastic writing tone, it is serious, but he can put in the jabs at stupidity in a very clear way, without being mean. That’s my take.

    • #25
    • March 19, 2016, at 10:41 AM PDT
    • Like
  26. Sabrdance Member

    Well, at the risk of contradicting both Drs. Berlinski, I do think wars are, at least partly, random. The question of WWI is not why did WWI happen, but why did it happen in 1914 instead of 1905, or during any of a thousand other international crises. It was the year in which the stars aligned. No one thought the people running the empires were mediocrities at the time. Our judgment of them as such is only possible because we know how their actions ended.

    Much the same can be said of the 2008 housing crisis. Why did it happen then and not some other time? That was the year the stars aligned. No one thought the brokers making these bets were idiots. We can look at their own documents and see that they recognized the risks they were taking, but they thought the risks were small (maybe they even were, and this was just the year for small risks).

    Life is a series of risks. Sometimes you lose the bet. The mistake, then, is too make additional foolish risks and to hold nothing back. That is when the great catastrophes happen. Maybe Mars rolled the dice on 1914, and Mercury the dice on 2008. But it took human action to make bad luck into horror.

    I do, however, agree on the larger point that an analysis of the world must be systemic. It is no use to ignore systemic violence and then declare the world has fewer murders.

    • #26
    • March 19, 2016, at 12:38 PM PDT
    • Like
  27. Midget Faded Rattlesnake Contributor

    Sabrdance:Well, at the risk of contradicting both Drs. Berlinski, I do think wars are, at least partly, random.

    To some extent, random is as random does. The humble admission that we don’t know the specific causes of an event – and perhaps may never know them – is different from saying specific causes didn’t exist.

    Many things are well-modeled as if they are random, even if, by spending more time and money, we could locate a cause for the apparent randomness. For example, “random failures” in a manufacturing process may be caused by a specific defect that it’s not worth us looking for – from a “God’s eye” view, the failures are not random, but we’re actually better off treating them as if they were.

    Very often, we can get by with conflating what we can know about the world with “what’s really out there”. That is, we can get away with the “mind projection fallacy”:

    Physicist and Bayesian philosopher E.T. Jaynes coined the term mind projection fallacy to refer to this kind of failure to distinguish between epistemological claims (statements about belief, about your map, about what we can say about reality) and ontological claims (statements about reality, about the territory, about how things are).

    And perhaps if you’re more interested in epistemological claims than ontological claims to begin with, harm is rarely done by being fairly relaxed about the fact that we do project our minds. For example, I think my default intuition is to treat claims as epistemological rather than ontological. This annoys people who want to talk ontology, but often works out fairly well otherwise.

    • #27
    • March 19, 2016, at 1:40 PM PDT
    • Like
  28. Quietpi Member

    In a word, Claire, I agree with you. We, the United States, are in the greatest danger we’ve faced since at least the Cuban missile crisis. Or, if not that, then the American Civil War. Yes, that great.

    I won’t argue that randomness has nothing to do with it, but the world is currently realigning, not along random lines, but with the realization that we are actively emasculating ourselves. Our friends are losing confidence in us, and our enemies see that the time is now to strike, in whatever manner they see as best for them. None of that is random at all. No crystal ball is necessary. Only a look at the newspaper (as an example – I hardly ever look at a newspaper anymore).

    • #28
    • March 19, 2016, at 1:57 PM PDT
    • Like
  29. Claire Berlinski, Ed. Editor
    Claire Berlinski, Ed. Post author

    Quietpi:In a word, Claire, I agree with you. We, the United States, are in the greatest danger we’ve faced since at least the Cuban missile crisis. Or, if not that, then the American Civil War. Yes, that great.

    I won’t argue that randomness has nothing to do with it, but the world is currently realigning, not along random lines, but with the realization that we are actively emasculating ourselves. Our friends are losing confidence in us, and our enemies see that the time is now to strike, in whatever manner they see as best for them. None of that is random at all. No crystal ball is necessary. Only a look at the newspaper (as an example – I hardly ever look at a newspaper anymore).

    I’m not going to “like” this except in so far as one always likes to hear that someone approves of one’s judgment. I’d exchange that vain pleasure quickly for the pleasure of realizing I’ve been alarmed for no reason at all. But basically, I’m astonished that I need to make this argument explicitly. It seems obvious to me: As you say, look at a newspaper.

    I believe I can explain why this is happening, or some aspects of why it’s happening (I’ve got no answer to the deeper problems in theodicy), but I don’t quite know what to say to someone who says, “It’s not happening.”

    • #29
    • March 19, 2016, at 8:30 PM PDT
    • Like
  30. Wes046 Member

    The Fourth Turning | Wm Strauss & N. Howe provides some history and “empiricism” into crisis. Nutshell: every 80 to 100 years there has been upheaval in western civilization. Authors trace this back to the War of Roses. Few saw it coming [difficult to predict]. Came anyway. The authors died in the ’90s. Made some decent predictions for the 2000’s. This is not an easy read. … Yet it had a very calming effect on me because it took most of the “personalities” out of the current situation.

    [like T Sowell’s Conflict of Visions ] .

    Predicting war is a “low percentage move”. Predicting Crisis is not.

    wbajr tbc

    • #30
    • March 19, 2016, at 8:45 PM PDT
    • Like
  1. 1
  2. 2