3 ms·
"Social scientists should never try to predict the future; they have enough trouble predicting the past." - Steven Pinker. From John Gray's: Pinker dismisses
by TheSpiceIsLife 12y ago
"Social scientists should never try to predict the future; they have enough trouble predicting the past." - Steven Pinker.
From John Gray's:
Pinker dismisses the role of nuclear weapons on the grounds that the use of other weapons of mass destruction such as poison gas has not prevented war in the past; but nuclear bombs are incomparably more destructive.
The First and Second World Wars were so shockingly destructive as to make Europe realise that, perhaps, there are better ways of going about things. Like trade.
No serious military historian doubts that fear of their use has been a major factor in preventing conflict between great powers.
Well, between two Great Powers -- the former USSR and the USA. Does anyone really think the only thing preventing China, the USA, and Russia from engaging in all-out war is the threat of nuclear annihilation? Everyone stands more to gain from trade.
Moreover deaths of non-combatants have been steadily rising. Around a million of the 10 million deaths due to the first world war were of non‑combatants, whereas around half of the more than 50 million casualties of the second world war and over 90% of the millions who have perished in the violence that has wracked the Congo for decades belong in that category.
That's disingenuous. Historically whole villages and towns, even cities, were routinely laid to waste.
Gray's article is sufficiently hand-wavey and cherry-picky to not even bother with.
From Taleb's critique:
Loss of the Island Effect: My point now is the loss of island effect, from interconnectedness. The number one danger is a biological agent that can travel on British Air and reach the entire planet. And it can be built in a high school lab, or, even, a garage.
That sounds a bit sensationalist. A biological agent that has the potential to be the number one danger can be built in a high school lab, or, even, a garage. Well, ok, sure, given a sufficiently equipped high school lab or garage -now we might as well call it a Biological Weapons Facility- but built by whom? Surely not high school students or me.
Generating a sufficiently virulent biological agent isn't trivial, but that doesn't mean it's impossible. Maybe at some point in the future I'll be able to '3D print' a biological agent with all the right traits. Let's hope that never happens.
Taleb's statement above trivialises biological weapons, and spreads FUD.
Nuclear Potential: As I explain in Antifragile, risk management is about fragility, not naive interpretation of past data. If Fannie Mae is sitting on a barrel of dynamite I would not use past statistical data for my current analysis. Risks are in the fragility.(Sensitivity to counterfactuals is more important than past history, something Greenspan missed in his famous congressional testimony).
Pinker in 'The Better Angels of Our Nature' does repeatedly point out that we can't assume anything about the future based on the past but that there does appear to be a trajectory. Whether Pinker is wrong about that trajectory is beside the point here because Taleb is taking issue with naive interpretation of past data, yet Pinker wrote in the same book "Social scientists should never try to predict the future; they have enough trouble predicting the past."
Pinker conflates nonscalable Mediocristan (death from encounters with simple weapons) with scalable Extremistan (death from heavy shells and nuclear weapons). The two have markedly distinct statistical properties. Yet he uses statistics of one to make inferences about the other. And the book does not realize the core difference between scalable/nonscalable (although he tried to define powerlaws). He claims that crime has dropped, which does not mean anything concerning casualties from violent conflict.
I'm half way through reading 'The Better Angels of Our Nature' and that certainly isn't something I'd say about the book. Pinker admits that simple weapons and nuclear weapons are different, but that he focus on death rates in war over time. Pinker treats civilian <> civilian deaths separately as 'murders', so surely Taleb isn't conflating murder and simple-weapon-war-deaths.
Another way to see the conflation, Pinker works with a times series process without dealing with the notion of temporal homogeneity. Ancestral man had no nuclear weapons, so it is downright foolish to assume the statistics of conflicts in the 14th century can apply to the 21st. A mean person with a stick is categorically different from a mean person with a nuclear weapon, so the emphasis should be on the weapon and not exclusively on the psychological makeup of the person.
Should it? Why? Saying it should doesn't make it so, and doesn't convince. If we place the emphasis entirely on weapons potential this is most certainly game over. Yet global nuclear (and biological weapons) capacity has decreased since the height of the Cold War, and we're still here. This is one point of 'The Better Angels of Our Nature' -- that despite our ability we eradicate ourselves we've chosen -so far- not to, and Pinker acknowledges this means nothing about predicting the future.
Now, I can't speak much for that part of the world I know little about, geographically & culturally, and also mathematics isn't something I've studied a lot, so I'll take it as given that Tabel's paper 'On the Super-Additivity and Estimation Biases of Quantile Contributions' is correct, and that Pinker's time series data is shaky.
So good, I've learned something from Taleb, and I think it's important to read criticisms.
Tabel's paper doesn't contain the words child, women, race, religion, rights, animals.
Edit: copy paste mistake resulted in only half my comment, corrected within 1 minute of posting.