Malcolm Gladwell's response to Steven Pinker's review of his new book | Hacker News Reader

Malcolm Gladwell's response to Steven Pinker's review of his new book

gladwell.typepad.com

As long as we are returning to the "Gladwell is full of shit" meme, I recommend the Vanity Fair satire of Gladwell (http://www.vanityfair.com/culture/features/2009/12/gladwell-...) for anyone in search of a quick laugh.

Wow. That is hilarious. Takes the piss out of a large swath of middlebrow journalism along with Gladwell, too.

“The conclusion is both remarkable and inescapable but also—most importantly—counter-intuitive,” Dr. Bunquum told me over a glass of organic lemonade in his stunning waterstulp, or waterside studio, near Rotterdam.

I'm trying hard not to giggle because the person next to me is actually getting work done. Oh yeah, work...

I'm surprised that all responses seem to be on Pinker's side, but here is my take on it

>>What do Sailor's beliefs about an unrelated issue have to do with his analysis of NFL

Well, the credibility of a source (and his past "research" conclusion about racial superiority) is a factor in reviewing conclusions.

>>Gladwell seems to rely on ad-homenim.

This is subjective, but imo calling Pinker an "IQ fundamentalist" is a pretty mild response to the much more harsh attacks from Pinker.

>>I was absolutely shocked when Pinker's writing revealed that Gladwell in his book referred to an "igon value"

imo questionable spelling is better than questionable data.

If you're writing entire chapters on 'igon values' then obviously you've never read a single page on the topic. I'm sorry but this is not about forgetting to click 'spell check'.

>>If you're writing entire chapters on 'igon values'

Pinker's review says thatthe book "quotes an expert speaking about an 'igon value' (that’s eigenvalue,.."

I haven't read the book.

Are you sure that Pinker was downplaying the "igon value" issue and that Gladwell really wrote "entire chapters on 'igon values'" ?

Which entire chapters are about "igon values" ?

There aren't entire chapters. It appears in a quote by a mathematician. Gladwell isn't trying to explain mathematics poorly, he's transcribing poorly.

Here's the original New Yorker article:

http://www.gladwell.com/2002/2002_04_29_a_blowingup.htm

I agree, the "igon value" thing was a more revealing error than a minor typo. Gladwell is a talented quick study, by necessity. In the New Yorker article, he was just using "gaussian" and "igon value" for color, to give a sense of the mathematical sophistication of the quants.

In this case, the ploy was obvious to people with a modicum of a math background, because "eigenvalues" and "gaussian distributions" are not particularly obscure topics - they're something almost any undergrad in a quant-ish field will encounter (early, in lower division classes). Furthermore, by mis-spelling it, Gladwell does suggest that he implies a greater understanding of a field than he actually has (something quick studies are good at, including many progrmmers, which is probably why folks on this board are so attuned to it).

That said, it was a minor part of the article. This shouldn't have made it past the editors.

Pinker's review: http://www.nytimes.com/2009/11/15/books/review/Pinker-t.html

Steve Sailer has a reponse up on his blog: http://isteve.blogspot.com/2009/11/gladwell-strikes-back.htm...

He has another explanation of his point here: http://isteve.blogspot.com/2009/11/pinker-v-gladwell-on-nfl-...

Overall, I think Sailer's critique is quite valid: What Berri is doing, in effect, by using his "per-play" measure is comparing quarterbacks taken at the top of the draft (most of whom get a lot of plays in the NFL) to those taken lower in the draft who turned out to be surprisingly better than expected, and thus got a lot of plays. He's essentially leaving out of his analysis all those lower drafted quarterbacks who turned out to be as mediocre as expected and thus didn't get many plays. In other words, his methodology is pre-rigged to produce the conclusion that Malcolm likes.

Hypothesis: The reference to Sailer is mainly a hit, intended to link Pinker and Sailer, and drag Pinker down. Therefore, Sailer's response is largely irrelevant.

(Agreed that Sailer's response is valid. Further support for my hypothesis: Gladwell spends the rest of the article denigrating blogs for their names and associations with Fantasy Football, as opposed to Science.)

This does a great job refuting one of Pinker's fact based counter examples. It in no way refutes the core thesis of Pinker's review, that Gladwell is full shit.

That was my thought exactly. He zoomed in on the one wrong thing he could find and tried to associate the whole review with that thing.

An honest reply would be to say: Pinker got everything right except this football thing.

In this case, the "everything else" is of much higher importance than the football thing.

He zoomed in on the one wrong thing he could find

Not necessarily: the key to getting a letter to the editor published is to make a single coherent point, and to express yourself well. A letter to the editor is not the right venue for a point-by-point refutation of someone else's work.

I agree, but he did have options.

Why didn't he do it on his blog, then? Or say at the end of a letter: For a more complete response to pinker, see [URL].

We can conclude that he either had lots of other good arguments but chose to focus on that minor one, or that he didn't have good rebuttals for the other things.

The fact that he pointed out a error in analysis in no way affects your opinion of the conclusion?

The whole point of Gladwell's assertion is that if Pinker is this haphazard in gathering supporting data, perhaps his entire review is nothing more than opinion.

While there's nothing wrong with that - being a review and all - it certainly isn't a factual debunking of data.

(btw I'm not debunking the assertion, just your reasoning)

When he spends a whole essay refuting less than a third of a sentence in the original criticism the impact is diluted to very little indeed. The fact that he doesn't respond to any of the other factual assertions is more telling to me.

I agree he's full of shit, but to be fair, that was one of Pinker's relatively few factual examples. I think it was good form on Gladwell's part to address only facts and not Pinker's thesis.

I was absolutely shocked when Pinker's writing revealed that Gladwell in his book referred to an "igon value" (instead of eigenvalue, which I learned about in the second week of my Linear Algebra class.)

How the hell does someone lecturing us on social sciences have little-to-no math background?

How the hell does someone lecturing us on social sciences have little-to-no math background?

Are you being intentionally humorous? I thought it was a sad but accepted fact that, on the aggregate, the Social Sciences were largely rather unscientific.

Gladwell is called an "statistical detective"; he should know how to spell eyegin value.

"eigen value", in case you weren't being clever.

Bertrand Russell had his own take on the "Igon Value Problem":

A stupid man's report of what a clever man says is never accurate because he unconsciously translates what he hears into something he can understand.

A binary distinction between "stupid" (Gladwell?) and "clever" (domain expert) isn't very helpful here. Gladwell's ability to take obscure scientific results and present them in a relevant, insightful way is very powerful. If anything, I think the lesson is about being careful when drawing conclusions outside your domain of expertise, and that it is always helpful to get outside reviewers for your work (astonishing that a mistake of this magnitude slipped into a published book!).

When you fill gaps in your understanding with imagination, the result is that these 'relevant, insightful' ways are in fact made up. They are powerful in the way of teaching the public even more crap. One cannot possibly understand what eigenvalues are about if one doesn't know how to write the word. One may however think he understands what they are about, because one has filled in the gaps with imagination, and get it published, because reviewers think the author understands his stuff. In exactly the same way, all kinds of people, from complete crackpots to usually reasonable journalists, publish all kinds of crap because they think they understand quantum mechanics.

QM, interestingly, relies heavily on those eigenvalues.

Hadn't heard this quote before but I do so love it now. I can just see a flow diagram of some complex concept inside a talk bubble next to one person, and a homer-simpson-esque caveman picture in a thought bubble next to the other.

If the term "igon value" had been replaced with "eigenvalue", would the usage or context have been incorrect?

In other words, a mere misspelling -- while embarrassing -- doesn't demonstrate that someone has "little-to-no math background".

People who know what eigenvalues are don't misspell "eigenvalue" as "igon value," ever. Malcolm Gladwell doesn't know what eigenvalues are. That's the only point, really, but it's a pretty big one -- Gladwell's stock in trade is the popularization of counterintuitive statistical studies and their conclusions, and if he's never learnt any linear algebra he's not in a good position to actually understand those studies.

Just addressing your first sentence -- since I don't know Gladwell well enough personally to respond to the others -- it's not beyond the realm of imagination that he uses a ghost writer, or that someone transcribes for him, or even that he uses voice recognition software (as one of my clients does, which prompted my initial response).

My point was whether the usage was incorrect, because it's not impossible to misspell it and still know the meaning of the term.

Come on, Malcolm Gladwell isn't being ghostwritten. Anyone that good doesn't need a ghost writer, and if his ghost writer is that good, he's not going to be ghost writing instead of writing his own stuff and building his own name. Even if he were using a ghost writer, Gladwell didn't go back and read through?

It's may be more correct to say that Malcolm Gladwell does ghostwriting for the scientists he interviews.

Misspelling that word so badly does. I'm sorry. Eigenvalues are very fundamental to the types of math social scientists have to do. They are a core concept of linear algebra, which itself is all over modern statistics.

Imagine if you get a book about geometry, and the author spells pi as "pie." What would you think of his competency?

Confusing 'pi' with a homophone can only come from an incredibly superficial and lore-based understanding of the mathematics. Even if he offered an explanation was not technically wrong, it wouldn't do much to dissolve the suspicion that his understanding is based on lore. Most likely, he heard someone superficially explain "pie," recorded it or wrote it down in his notes, and then just used that explanation in his book.

Actually our "pie" author would be in a better place than Gladwell, because he could blame it on a spell checker. But it's not possible for bad copy-editing to turn 'eigenvalue' into "Igon Value."

There are others, and more information here: http://scienceblogs.com/pontiff/2009/11/igon_value_problems....

A weak defense, insofar as he picked only one point to go all-in on, and it's probably the least significant one. Also, notice his attempt to trivialize "igon value" as "my spelling"; the appropriate phrase would have been "my ignorance". I'd say this rebuttal does more to expose Gladwell as a sleight-of-hand artist than Pinker's actual critique did. It doesn't mean that Gladwell isn't a superbly entertaining writer and storyteller, though; and at his best, a popularizer of damned interesting stuff.

The thing is he's strictly a popularizer of flattering stuff. It's all designed to make his bourgeois readership feel like they have special insights, or feel good about themselves. One would think a lot of uncomfortable truths or null results (common sense or the experts are actually right) would come up, but they never do.

Pinker's core point is that Gladwell is a shallow dilletante that writes well, but makes profound conclusions based on this basic dabbling.

Gladwell does nothing but make one of Pinker's specific challenges appear much shakier (and yes, I'm going to be stereotypical here and guess that Pinker probably is much less comfortable with Football than Gladwell). If I hadn't ever heard of Pinker before, this would do something to my trust in his argument.

However, knowing Pinker's track record and background, Gladwell's "rebuttal" seems a frail attempt that tries to hit at the one weak point he found. Really? Nothing more substantial to refute?

Malcolm Gladwell has a rather long history of being corrected by experts. See here: http://www.fastcompany.com/magazine/122/is-the-tipping-point... and here: http://www.dr5.org/judge-posner-on-blink-the-power-of-thinki...

Full disclosure: I am a Gladwell fan, though I recognize that he is more storyteller than scientist.

What do Sailor's beliefs about an unrelated issue have to do with his analysis of NFL draft positions and performance? Gladwell's retort would have been better without that.

How about the fact that Steve Sailer has absolutely no domain knowledge of NFL draft picks besides being (maybe) an American football fan?

I'm familiar with Sailer myself, and I'd say anyone who uses him to back up their hypothesis is suspect. Every angle Sailer plays is somehow tied to race and the genetics of race; it's neither objective nor logical.

Sailer's introduction to many, including myself, was his posting of mile-long screeds, primarily regarding race and IQ, in the comments of Gladwell's blog. He was banned from same and his comments removed in late 2006:

http://gladwell.typepad.com/gladwellcom/2006/12/the_lunatic_...

He seems to be trying to make a name of himself in part by carrying on a one-sided debate with Gladwell, which in itself makes him a suspect source even if he wasn't already familiar to the blog's long-time readers.

Sailer's posts were not screeds. They were well-argued and relevant, just like his contributions in the current debate. He was not banned. Gladwell made his readers vote whether or not he should stay, and they overwhelmingly voted in support of Sailer. Nor were his comments removed.

This is an instance of poisoning the well: http://en.wikipedia.org/wiki/Poisoning_the_well

This is a very interesting comparison of how to do research well with how to do research poorly, with the journalist coming out better than the academic.

After edit: I'll have to check a Pinker book reference after I return from PZ Myers's debate this evening to give a more focused example of what Pinker can miss while doing research.

Unfortunately the original research paper by Berri and Simmons is locked up behind a paywall. But Sailer does seem to make a good point - if you measure QB performance on a per-play basis, the low-end of the draft order is going to be dominated by the few Tom Brady's of the world (draft number 199, 3 Super Bowl rings, lots of plays) because most late draft picks don't even get any NFL snaps.

It would be like saying IQ is not correlated with Nobel Prize awards, if you only look at people who have earned PhD's.

Actually the Pro Bowl thing might not be bad.

Am I missing something or is Gladwell not taking in account the fact that higher draft pick QBs play on worse teams then QBs taken later in the round? Also, QBs taken later in the round will probably be allowed a season to learn and mature as opposed to first round picks who will play right away because of the large salary.

Gladwell seems to rely on ad-homenim. Calling Pinker an IQ "fundamentalist" is a bit over the top, too, especially in light of Pinker's moderate writings that point to the relevance of both environment and heredity.

No, he doesn't. The closest thing to a true ad-hominem fallacious point in his reply would be his mentioning of Sailer's opinion that blacks are intellectually inferior to whites. However, he immediately follows that up by addressing Sailer's actual data, cited by Pinker.

I need a fancy-sounding Latin term for "trotting out 'ad hominem' when there is no such thing"...

In real life, virtually no one uses true ad hominem. No on says, "he is a jerk, therefore he's wrong." In my book, using insults and connotative (negative) language rather than strictly denotative is on the ad hominem scale. It is absolutely designed to undermine the credibility of the person you're debating while not engaging the argument at all.

This new meme that started on reddit claiming ad hominem isn't ad hominem is weird. We shouldn't defend argument techniques that make truth-seeking harder. There's a time and a place for ad hominem and that time and place is entertainment writing.

Mmm, a debate about debates. How meta.

I have a couple of problems with this. First, by broadly sweeping "everything I don't like about arguments" under the label "ad hominem", you're powerfully eroding the usefulness of the term. It has a specific use and purpose, and it's useful to understand the various fallacies in debate (http://www.nizkor.org/features/fallacies/).

Also, you admit that this style of debate has a place in entertainment writing -- in a thread about his letter to the editor. This was not a formal debate, nor was it the appropriate forum for such; he was writing a response that would reach a large number of readers, and they were his audience.

Besides, aren't you committing the selfsame fallacy that you're accusing him of, according to your own rules? i.e., "Gladwell used methods I don't like, so he's wrong."

I wasn't aware that this was a Reddit meme. I stopped visiting Reddit some months ago, and my usage fell off even before then. I'm glad it is, though. If someone's going to try and bolster their criticism of someone else by using fancy Latin phrases, they really ought to get called out on it by someone that knows better.

If something is "a lot like ad hominem, but not ad hominem", then referring to it as "ad hominem" is deliberately distorting the truth.

We shouldn't defend argument techniques that make truth-seeking harder.

I don't see anyone claiming we should. What is being claimed is that this isn't ad hominem, which is a point you've all but conceded in your opening two sentences.

If you are going to request that people debate "properly", it would be rather nice if you didn't reserve a special place for yourself and your favourite misinterpretations.

Sailer's opinion of blacks has nothing to do with the discussion. However, that information, along with the claim that Pinker is an "IQ fundamentalist" is enough to lead readers unfamiliar with Pinker to suspect that he is racist and dismiss his arguments without further inspection.

That is exactly what ad-hominem is about.

I was left feeling completely dissatisfied with this response. It came across as childish.

http://scienceblogs.com/pontiff/2009/11/igon_value_problems....

malcolm owned that biatch.

Lost me at football.

So two soft scientists are fighting? Who cares?

Pinker is not a soft scientist.

Oh? What is he then? (I read evolutionary psychologist off his wikipedia bio)

Read his papers from the 80s -- they are highly technical linguistics manuscripts. Pinker knows Unix and uses emacs. Yes, really.

I took a class from him at MIT (9.00). He is a very, very smart person.

I took that too, which is also how I knew :)

Gladwell is a journalist, not a scientist.

I didn't like the implied ad hominems much 8/ Respect for Gladwell--;

Am I to take the downvotes to mean we like to undermine the credibility of people by saying things like "...who is best known for his belief that black people are intellectually inferior to white people" before we address the actual data?

It's cute, yeah, but it's also fallacious argumentation. Even if the smear is true, which Galdwell doesn't substantiate, it's not relevant to the data. He points out why the data is weak--and that part is fine--my issue was with his extraneous ad hominem.

But apparently here on HN, ad hominems are a-ok, as long as they are clever and make for a good read, right?

Reply on news.ycombinator.com