Nassim Nicholas Taleb: “The math comes first, words later.”

https://www.facebook.com/permalink.php?story_fbid=10152921768278375&id=13012333374

by nileshtrivedi • 11 years ago

71 39 11 years ago

Selfcommit• 11 years ago

Dawkins has never struck me as being particularly intellectual. The stance he takes on his definition of science often comes across as extreme as many religious fundamentalist perspectives.

BinaryIdiot2• 11 years ago

Any particular reason why? He holds a PhD and will happily present scientific facts if you wish to debate him in something.

msandford5• 11 years ago

From an article linked elsewhere in the comments, a comment by Dawkins:

"I greatly admire EO Wilson & his huge contributions to entomology, ecology, biogeography, conservation, etc. He’s just wrong on kin selection."

I get that you can slam someone like that about very, very well established stuff like physics; if someone says F = M * a * v then rip into them all you want (there is no v in the eq.). That stuff is well established, I think in physics you have to get 5 sigma to claim "significance" ( p = 0.0000003 ) which is pretty hard to do on accident.

But when you're dealing with biology thus far significance is still a p value of around 0.05 which means there's a 5% chance it could have happened by accident. That's a pretty substantial risk, at least relative to the world of physics.

So for someone to proclaim certainty on a subject which is inherently much, much less certain you have to wonder what's going on. How did a person get so confident of something without the data telling them to be that confident? Is it possible that they're choosing to believe something further than what the data show? If they're willing to exaggerate their confidence on one subject, who's to say they aren't doing it elsewhere?

Unfortunately in a lot of realms of science people end up with beliefs about how things work that aren't rigorously backed up. That doesn't make said beliefs inherently wrong, but pretending that they're provably correct when they're not smacks of charlatanism.

nemo1618• 11 years ago

Can someone summarize the linked paper? What result does it propose w.r.t altruism, and why are Dawkins/Pinker/others challenging it?

EDIT: after some digging, I found this article: http://www.theguardian.com/science/2012/jun/24/battle-of-the..., which summarizes the disagreement like so:

  For lay spectators, the row is a symptom of the long and controversial
  evolution of the very idea of evolution.  At root it is a dispute about
  whether natural selection, the theory of "the survival of the fittest" first
  put forward by Charles Darwin in 1859, occurs only to preserve the single
  gene. Wilson is an advocate of "multi-level selection theory", a development
  of the idea of "kin selection", which holds that other biological, social and
  even environmental priorities may be behind the process.
  
  But Dawkins is far from convinced: "Wilson now rejects 'kin selection' and
  replaces it with a revival of 'group selection' – the poorly defined and
  incoherent view that evolution is driven by the differential survival of
  whole groups of organisms."

jessaustin3• 11 years ago

Dawkins [EDIT: remove adjective] believes in kin selection, based on results of inclusive fitness theory, which makes predictions about fitness when interacting organisms are related. (e.g. ant workers share 3/4 of their genes, and ants are remarkably social.) Wilson et al. contend that inclusive fitness theory is a bit of a distraction, since it only applies in rare cases, and in those cases it's just a specialization of direct fitness, which can be calculated without respect to relatedness. Later in the supplement they build a model which has the features discussed by kin selection without relying on relatedness.

MichaelGG• 11 years ago

Are there accessible lay-person "popular science" type books that contradict Dawkins and are highly regarded? My very limited understanding of evolutionary biology mostly comes from reading Dawkins books, which were highly enjoyable. And the concepts seem pretty solid and playing around with my own toy throwaway simulations, don't seem implausible. I also understood that other math had backed up the general points of Dawkins.

If this isn't the case, what's the recommended pop-sci reading to help make me less wrong on all this?

panarky• 11 years ago

"How can genetically prescribed selfless behavior arise by natural selection, which is seemingly its antithesis?"

"This problem has vexed biologists since Darwin, who in The Origin of Species declared the paradox — in particular displayed by ants — to be the most important challenge to his theory."

"Eusociality is not a marginal phenomenon in the living world. The biomass of ants alone composes more than half that of all insects and exceeds that of all terrestrial nonhuman vertebrates combined. Humans, which can be loosely characterized as eusocial, are dominant among the land vertebrates."

http://www.ncbi.nlm.nih.gov/pmc/articles/PMC3279739/

hammerandtongs• 11 years ago

The paper itself is pretty readable and interesting even for a layperson.

jessaustin3• 11 years ago

Having spoken with Prof. Wilson on numerous occasions (long ago), and thus knowing what a charming gentleman he is, I can't help but suspect that he would have preferred this work escape the attention of combative public intellectuals like Dawkins, Pinker, and Taleb.

The actual paper with math in it, rather than this gloss of an abstract: http://www.ncbi.nlm.nih.gov/pmc/articles/PMC3279739/bin/NIHM...

dragonwriter2• 11 years ago

Having spoken with Prof. Wilson on numerous occasions (long ago), and thus knowing what a charming gentleman he is, I can't help but suspect that he would have preferred this work escape the attention of combative public intellectuals like Dawkins, Pinker, and Taleb.

Dawkins and Wilson have been directly exchanging attacks on this issue for some time [0], and however charming Wilson might be personally, he's hardly been any less combative than Dawkins.

[0] e.g., http://www.theguardian.com/science/2014/nov/07/richard-dawki...

jtkme• 11 years ago

Because countering verbal BS with ad hominem attacks is clearly the answer ...

Luc3• 11 years ago

I think this article from last year sums up the context of the situation pretty well:

http://www.theguardian.com/science/2014/nov/07/richard-dawki...

To me, the simplest explanation for why Taleb is getting involved is because he can't stand Dawkins. Not exactly pure motives.

SiVal2• 11 years ago

Pure motives? Do you imagine that all Dawkins cares about is to find the truth, regardless of what it turns out to be? The reason Dawkins is involved is because he can't stand EO Wilson casting doubt on the ideas for which Dawkins became famous.

Pure motives in science are nice to have but aren't necessary. It shouldn't matter why a scientific argument is made as long as opponents are free to counter it. Taleb is saying that instead of just spinning verbal yarns like some social "scientists" do to disguise their political advocacy as actual science, Dawkins (who is a real scientist) needs to show where Wilson's math is either mistaken or incorrectly models the data (by, for example, abstracting away something important). If he can't or won't do so, it doesn't mean he's wrong, but it means his rebuttal is just rhetoric.

Luc3• 11 years ago

It shouldn't matter why a scientific argument is made as long as opponents are free to counter it.

Yeah, sure thing, if we all had unlimited time and brain power to devote to these things. But I can't weigh the issues myself on their scientific merit since I don't have the background (ignoring the fact I might easily fool myself too) and I'm too old to wait 50 years for everything to get sorted out.

So on the balance of things, the debate appears to be ongoing and reasonable people reasonably disagree, but I don't think I'm wrong saying that Taleb's 'contribution' to the 'scientific argument' is mainly heckling Dawkins.

mathattack3• 11 years ago

Taleb is pretty heavy duty with these attacks. I'm a big fan of his, but isn't it extreme to demote Dawkins [0] to just a journalist? He's got a Phd in animal decision making (close to the field in question here) and has been on the faculty of Oxford for 45 years.

[0] http://en.wikipedia.org/wiki/Richard_Dawkins

dragonwriter2• 11 years ago

I'm a big fan of his, but isn't it extreme to demote Dawkins to just a journalist?

He's just adopting Wilson's attack on Dawkins [0], when Wilson said: "There is no dispute between me and Richard Dawkins and there never has been, because he's a journalist, and journalists are people that report what the scientists have found and the arguments I've had have actually been with scientists doing research."

[0] http://www.theguardian.com/science/2014/nov/07/richard-dawki...

mathattack3• 11 years ago

Interesting. It's clear that this goes way beyond merely the paper in question.

misiti3780• 11 years ago

Taleb comes of as an asshole here, and a lot of the time really, but after reading all of his books (except dynamic hedging) i find myself agreeing with almost all of his core hypotheses. The Black Swan and Anti-fragile are some amazing reads.

thekemkid2• 11 years ago

Could someone explain this to me like I'm a 5 year old? I feel like it went over my head. How does maths relate to altruism?

one-more-minute2• 11 years ago

It's about having theories that explain/predict altruism, in the same way you'd have theories which explain/predict chemical reactions, electrons or the motion of blocks on slopes. You start with some basic assumptions and derive (mathematically/logically) the outcomes.

You should check out kin selection for more detail on Dawkin's side of the debate, but basically the idea is that a "gene" for helping out your brother (for example) is likely to spread through the population because your brother probably also has the same "gene". You don't have to survive personally for your genes to live on, so your genes don't focus wholly on your personal wellbeing.

If you want to be less hand-wavy you can work out the probabilities based on what we know about genes, and therefore work out how much altruism you expect to show towards members of your family vs. random strangers. For example, your relatedness to your siblings is a half, to your cousins 1/8th; hence the quip:

    I'd lay down my life for two brothers or eight cousins.

      — J.B.S. Haldane

My knowledge of Wilson's argument is limited but I think the crux is that he (or at least someone) believes in selection on the level of groups of individuals (as opposed to just genes) and uses that to explain altruism instead / as well. Dawkins and others disagree. Taleb has basically argued "the maths checks out", but that's pretty much irrelevant when it's the starting assumptions that are under scrutiny, not the logic that follows.

thekemkid2• 11 years ago

Thank you very much.

sjg007• 11 years ago

Groups of individuals are shared sets of genes anyway.

stolio4• 11 years ago

If 100 biologists sign a petition saying your model is wrong that should be a sign that it's time to double-check some things. And here we have an economist lecturing scientists about how to properly view and use math? True, biology is on the "softer" side of the scientific spectrum but I'm not convinced economics is even on it.

"Math comes first" is a bad strategy. I'd (softly) argue that just like your probability of randomly hitting a rational number on the number line is zero, the probability of a given framework of math describing reality is also zero. You must explain why you think the math is relevant and descriptive of reality. Or you can demonstrate that your framework allows you to make better predictions than anything else. Either way you'll have to use your words.

one-more-minute2• 11 years ago

This is exactly the thing. Your mathematics may be absolutely watertight, but if your starting assumptions aren't based in reality then your result isn't going to be either. Garbage in, garbage out – and that's going to be true regardless of how "hard" or "soft" your subject is.

You could derive perfectly sound results about the population dynamics of pigs assuming that they can fly, but biologists are going to tell you you're wrong regardless of the maths – and rightly so.

msandford5• 11 years ago

Welp, I guess we're going to have to pack up the modern world and go home then. Because many of the major advances in science thus far have been widely rejected by credentialed people at first only to eventually be proven correct.

First Semmelweis, then Pasteur:

http://en.wikipedia.org/wiki/Ignaz_Semmelweis#Conflict_with_...

http://en.wikipedia.org/wiki/Louis_Pasteur

Copernicus and Galileo vs a lot of other people:

http://en.wikipedia.org/wiki/Galileo_Galilei#Controversy_ove...

http://en.wikipedia.org/wiki/Nicolaus_Copernicus

Einstein and relativity:

http://en.wikipedia.org/wiki/Criticism_of_the_theory_of_rela...

I could do this for days if you'd like. Credentials are worth something of course, but credentials don't have metaphysical powers that prevent a person from being wrong once they have their credentials.

EDIT:

That's not to say I necessarily believe that these folks are correct and Dawkins is wrong. Just that the credentials don't mean as much as people think they should. It's not like getting a PhD prevents you from ever having a bias ever again. It'd be great if it was true, but it's not.

stolio4• 11 years ago

Those scientists all had ideas that made demonstrably superior predictions than other ideas at the time. Evidence outweighs credentials, arguments can take a walk.

So what evidence is Taleb bringing to the table? I ask honestly, because when he says "I spent some time scrutinizing the math: it is impeccable, though unsophisticated by mathematical finance standards." I just think Please...spare me If all he has is a mathematical critique that was roundly rejected by 100 experts in the field it's time to put up or shut up. Einstein predicted Mercury's precession, Galileo predicted new laws of motion. What's Taleb predicting?

mathattack3• 11 years ago

Here's a 2013 Facebook comment from Taleb that highlights the grudge...

Richard Dawkins, in his statement about the number of Nobels granted to Moslems, showed a total ignorance of probability. A primitive violation. You never get an idea about the mean from measuring the tail (number of Nobels per capita).

https://www.facebook.com/permalink.php?story_fbid=1015169418...

macmac• 11 years ago

As Taleb points out himself it is in fact possible for the Wilson paper to be wrong even if the math is correct. This is exactly what this: http://journals.plos.org/plosbiology/article?id=10.1371/jour... paper demonstrates to be the case.

Quote:

"The misleading conclusions all resulted not from incorrect math but from overgeneralizing from narrow assumptions or parameter values. For example, all of their models implicitly assumed high relatedness, but modifying the model to allow lower relatedness shows that relatedness is essential and causal in the evolution of eusociality. Their modeling strategy, properly applied, actually confirms major insights of inclusive fitness studies of kin selection."

shas3• 11 years ago

I think the principal critique of Nowak et al.'s work is based on how it gets the biology wrong, not the math. Chicago evolutionary biologist (and prolific public intellectual) Jerry Coyne, reviews a recent paper that challenges Nowak et al.'s paper [1]. The answer to Taleb's weird post is within Coyne's review of Liao et al.:

As I (Coyne) noted at the time (of Nowak et al.'s publication), their (Nowak et al.'s) dismissal of relatedness and kin selection from their model seemed bizarre, since they didn’t vary relatedness in their model. If you don’t do that, how can you say it’s unimportant in evolving eusociality?

"Math checks out" is a questionable approach to understanding science.

On a possibly related note, Taleb has recently been on a strange anti-Dawkins spree on Twitter and elsewhere [2]. Taleb has also done some weird anti-GMO bashing in the past [3].

[1] https://whyevolutionistrue.wordpress.com/2015/03/27/new-pape...

[2] twitter.com/nntaleb

[3] http://articles.economictimes.indiatimes.com/2014-11-18/news...

tensorproduct• 11 years ago

//"Math checks out" is a questionable approach to understanding science.

Very much agree. By this standard, one could put "1+1=2" as an appendix to every article you ever write and then defend them as having valid math, regardless of the content of the rest of the paper.

I find this most surprising coming from Taleb, because I thought that one of his principal arguments in the world of finance was that rigourous mathematical models are of limited use in understanding complicated real world systems. Maybe I've misunderstood him, but he seems to have changed his tune.

Nassim Nicholas Taleb: “The math comes first, words later.”

Related Stories