Something that's caught my attention over the past few months is that out of the strongest thinkers who I'm familiar with, a very low proportion record their informal ideas online, and discuss their thoughts in the public domain. 

The domain that I'm most familiar with is pure math, so I'll focus on that, but I imagine that my remarks apply more broadly.

A very large fraction of pure mathematicians (including a large fraction of elite mathematicians) post their papers to ArXiv, a database which stores preprints of mathematical papers. The invention and widespread use of ArXiv has been very valuable, in that researchers can easily notice and access the new papers in their fields as soon as they become available. 

That not withstanding, the fraction of mathematical thinking that's in the public domain is vastly smaller. Math papers generally don't include explanations of why the researchers are interested in the questions that they're writing about or how they came up with the proofs of the theorems. Even when an author does attempt to explain his or her thinking, a reader will often find parts of it opaque, and want clarification. The need to write to the author for clarification poses a trivial inconvenience which, in practice, discourages a large fraction of questions that people ask.

Informal mathematical thoughts are extremely important for doing mathematical research, and it's generally difficult to learn it without in-person contact with authors of papers.

A natural solution to these issues is for researchers to spend more time blogging about their informal thoughts, and for a commenting system to be enabled for readers to offer suggestions or request clarification.

The mathematicians who have the most to offer in the way of ideas and insight are the elite mathematicians. The fraction of them who blog is tiny. Terence Tao and Timothy Gowers do a considerable amount of blogging, but they're nearly alone in this.

The opportunity cost to mathematical research here seems to be enormous. Some explanations for the phenomenon are:
  1. Many of the prize winners who don't blog are elderly, and blogging is relatively uncommon among elderly people.
  2. There's adverse selection coming from the best people having the least to gain from public online discourse (in light of the fact that the fraction of participants who are of similar quality being tiny).
  3. Inertia.
Point #1 is relevant, but the fraction of young mathematicians who blog is also very small (though larger than that of the older mathematicians).

To elaborate on #2 — when I first started using MathOverflow, I found it to be a very useful resource, and learned a lot from it. In the beginning, there was a substantial number of very high quality mathematicians who answered a lot of important questions. Since then, the number has dwindled (though I would emphasize that it remains true that some of the contributors are outstanding). In more recent times, I've had the experience of asking very natural questions that I'm sure that many mathematicians have thought about, without anyone answering (here too, I would emphasize that in recent times I've gotten some helpful answers: I'm thankful to the current MathOverflow community for this). I think that the drop off in participation by high quality mathematicians can be partially attributed to [the posters who were contributing the most getting relatively little helpful feedback on their own questions and answers]. In such a situation, corresponding with one's colleagues privately starts to look more attractive than posting publicly.

It seems as though it might be possible to get around the issues #2 and #3 by simultaneously involving a substantial number of very high quality mathematicians, and having them write for and respond to each other, so that putting their thoughts into the public domain would become attractive. But doing so is beyond my pay grade :-)

I suspect that the phenomenon that I describe extends well beyond pure math, and that across fields, there's a large opportunity cost associated with high quality thinkers not putting their thoughts into the public domain. Changing the status quo here could have enormous value. I think that there might be low hanging fruit in this area, insofar as what efforts there have been seem to be few and far between, relative to the landscape of people who think about ideas. 

 

New Comment
42 comments, sorted by Click to highlight new comments since: Today at 9:23 PM

Blogging technical material at a level where it won't embarrass you can be a pretty demand and time-consuming task (especially if you want to provide any sort of graphs or visual accompaniment, and equations aren't hugely better either since you're not writing PDFs).

What makes you think that blogging wouldn't crowd out high-level contributions, and that the elites are not elite in part because they don't blog?

Blogging technical material at a level where it won't embarrass you can be a pretty demand and time-consuming task (especially if you want to provide any sort of graphs or visual accompaniment, and equations aren't hugely better either since you're not writing PDFs).

I still think that many people could do more at the margin without great cost to themselves.

What makes you think that blogging wouldn't crowd out high-level contributions, and that the elites are not elite in part because they don't blog?

I agree about there being opportunity cost, but I think that at the margin there should be more of an effort to put things in the public domain.

I still think that many people could do more at the margin without great cost to themselves.

Why do you privilege your view over theirs? They know how much technical writing costs them.

  1. I'd recur to my remarks about inertia and adverse selection.
  2. Cost needs to be thought of as relative to benefits, and the public externality benefits may not be salient to them.

Cost needs to be thought of as relative to benefits, and the public externality benefits may not be salient to them.

No, it doesn't need to be thought of that way when you are asking why private agents do not engage in privately-costly public-goods. Who is paying or otherwise rewarding these cognitive elites into taking the substantial time and effort to write up their thoughts and engage with the many low-quality commenters etc? Has anyone ever gotten tenure for a blog post, or any number of blog posts?

Incentives matter.

This could be fixed by someone providing infrastucture to blogging experts. A service that would do the following:

  • receive the article in whichever form (word, latex, e-mail, paper)
  • convert it to HTML
  • check spelling
  • publish the article online
  • moderate the forum
  • send relevant feedback (individually, or daily digests) by e-mail to author
  • publish the author's response

This way the burden to authors could be kept at minumum. And it could be done by a few volunteers.

So, basically - a journal.

[-][anonymous]11y30

That would be nice, but the software to do steps two and three doesn't exist yet. Modern spellcheckers more or less fail at parsing math without serious mucking about.

aspell handles tex just fine.

[-][anonymous]11y00

No, it doesn't. Specifically, it fails on words that have accents.

No, it doesn't.

Rather, I meant that it works fine with math mode.

Specifically, it fails on words that have accents.

It doesn't yet understand tex accents, but if you set the encoding using the tex package, you can directly enter è, é, ê, ...

Writing good popular essays on technical material still takes many hours of thought and work. These kinds of steps could be somewhat helpful for people who were very technically illiterate or who wanted to use a lot of pictures or equations, but even then they're still a long way from fixing the problem.

The costs of blogging this way would not be reduced to zero, but still could be reduced significantly, and that could make a difference between blogging and not blogging to some poeple.

Don't underestimate the inconvenience of setting up and maintaining a website. Especially if someone is not in a web programming business. For many people just setting up an open-source blog software is too much of an obstacle.

There could be people for whom hours of thought and writing of the article are fun, but all the remaining web-related work is an obstacle.

but still could be reduced significantly

I disagree in the general case, though I grant that it could possibly be true for extremely computer-illiterate people. Creating a Blogspot account and using it is already much less work than setting up a custom blog.

Creating a Blogspot account and using it is already much less work than setting up a custom blog.

It is not just how much time you spend creating the account.

First, there are decisions involved (what service to use, what design to use, whether customize the design, how to call the blog,...), which would all be removed by someone telling you "send me the document, I will publish it; see, other people are using this service too".

Second, by creating a blog you kind of precommit to moderate spam and trolls in your comments. Unless you want to have one of those blogs where the first 10% of the page is the article and the remaining 90% is spam. But even if the discussion is reasonable, you kind of precommit to read it and to respond.

As an extra explanation - maybe a small percentage of people are bloggers, period - be they elite mathematicians or groups that are more stereotypically online, like programmers, Finns, libertarians, assholes, etc.

I'm not sure you'll find many more blogs by elite mathematicians than by other equally small groups, like piano repairmen, frisbee champions, etc.

Surely this partially accounts for the phenomenon. But I didn't mean to claim that elite mathematicians are unusually unlikely to blog: what I meant to convey is that they have the most to offer (mathematically), so that the opportunity cost (to the world) associated with them not blogging is highest.

You missed two blogging Fields medalists.

Ngo Bao Chau (with appropriate accent marks) has a blog. It is in Vietnamese, because he is Vietnamese. http://thichhoctoan.net/ I don't know how much he writes about math because I don't read Vietnamese. I know he writes a lot about politics, or at least used to.

Cedric Villani writes about math for Le Monde sometimes and has a blog on his website. http://cedricvillani.org/ I am not good at French, so I don't read it.

I think 10% for living Fields medalists who blog is pretty high, and many of those living are older. (And two have basically withdrawn from mathematical society.)

Thanks for pointing these out, which I had not seen before.

Neither blog seems to have almost any mathematical content, and the posts seem few and far between, so I don't think that they budge the bottom line. Even in the case of Gowers, the average number of mathematical posts is about one a month.

Also, the category "Fields Medalists" is in some sense cherry-picked, in that it happens to include Gowers and Tao in particular: if one broadens consideration to all of the prizes that I listed, the fraction of bloggers is much smaller (though the winners of the other prizes also tend to be older).

John Baez seems worth mentioning when it comes to math bloggers.

John Baez has made outstanding contributions to the mathematical community via blogging, and there are others. I restricted myself to major prize winners so as to avoid considering a larger and more amorphous reference class.

[-][anonymous]11y90

There are about one to two hundred living people in the world who care about my sub-field of mathematics.

About four of them could be classified as "elite", though none of them have won any of the big prizes. None of them are Gold IMO medalists, so it's a good thing we're not working on existential risks!

What's the value of them blogging to such a small audience, when most of us see each other two or three times a year at conferences?

Out of curiosity, what is your subfield?

What's the value of them blogging to such a small audience, when most of us see each other two or three times a year at conferences?

Probably about the same benefit as having a mailing list, except that outsiders can access it. (I have no particular opinion on how useful a mailing list would be, but it seems like an appropriate reference class.)

None of them are Gold IMO medalists, so it's a good thing we're not working on existential risks!

http://lesswrong.com/r/discussion/lw/hj5/research_is_polygamous_the_importance_of_what_you/

[-][anonymous]11y40

None of them are Gold IMO medalists, so it's a good thing we're not working on existential risks!

http://lesswrong.com/r/discussion/lw/hj5/research_is_polygamous_the_importance_of_what_you/

I take none of what Diego writes as canon.

The original quote was a not-very-veiled reference to MIRI's inexplicable love of the IMO.

I wonder how much of it is not wanting to have to deal with the peanut gallery, and spammers.

Leaving the comments off would hurt a little in popularity but would absolutely do wonders for the timesinks.

I share your views. Notice that the case for having public conversations applies not just to conversations between elite researchers, though of course the associated opportunity cost there is highest.

My background is in philosophy, not in math, but in this profession there are a few group blogs, such as PEA Soup, that feature many top academics in the relevant subfields as regular participants (in the case of PEA Soup, the subfield is moral philosophy) . These blogs seem to succeed largely because they manage to exclude comments below a certain quality threshold. This of course connects to point #2 above.

these blogs succeed ... because they ... exclude comments whose quality falls below a certain threshold.

I see an opportunity for philanthropy. Identity the elite people that one hopes will blog, and then pay for somebody else to do the comment moderation for them.

The problem I foresee is that this turns out to be big-money philanthropy. Who do you hire as your moderator? They probably need a PhD in mathematics, and the right personality: agreeable yet firm. People like that have lots of well paid options in which they are not playing second fiddle. The philanthropist backing this may have to come up with $150,000 a year to pay the wage bill.

Turning this round, it answers the question of why so few elites blog. Hoping that they take on the task of doing their own comment moderation and community building is hoping that they engage in some serious philanthropy and tolerate getting little credit (because they are paying in kind rather than in big wodges of cash).

Who do you hire as your moderator? They probably need a PhD in mathematics, and the right personality: agreeable yet firm.

Or, you get grad students to do it. The experience is mostly there, the personality perhaps not.

The problem I foresee is that this turns out to be big-money philanthropy. Who do you hire as your moderator?

Money isn't the only thing that motivates people. Most people who moderate forums on the internet aren't payed to do so.

Although those other motivations might not be the kind you want. You probably don't want a moderator driven by a desire to lord their power over other people.

Two other possible selection biases:

  1. Someone exceptionally good at something is likely to spend almost all of their free time on that something. If so they won't have time to routinely blog about it.

  2. Almost everyone who blogs a lot eventually posts something very stupid. If you strike someone off your list of elite thinkers when they say something really dumb, you'll end up with a list dominated by non-bloggers. (See also this old taw comment.)

Someone exceptionally good at something is likely to spend almost all of their free time on that something. If so they won't have time to routinely blog about it.

Yes, this is indeed so. Still, it may be possible for them to blog at least some more without very much cost to themselves.

Almost everyone who blogs a lot eventually posts something very stupid. If you strike someone off your list of elite thinkers when they say something really dumb, you'll end up with a list dominated by non-bloggers. (See also this old taw comment.)

  • This may be a reason not to blog a lot, but not a reason not to blog at all.
  • It's awfully silly to strike someone off of your elite thinkers for things that they say in a few blog posts, especially if they redact their positions later on. But the issue that you describe could be a deterrent in practice.
  • I think the world would be a better place if people worried less about their reputations.

To clarify, my comment was meant to be descriptive, not prescriptive. I mostly agree with what you've just said!

For an up-and-comer in an intellectual field, insightful thoughts are a limited resource, and should be carefully allocated. Once can say interesting things in public, or one can devote the effort to more professionally productive writing. Or, one can save the effort to be insightful for professional writing, and write less insightful comments for the public. Which carries the risk of saying something that is dreadfully wrong, instead of merely a little bit boring.

It always astonishes me to see the volume of high-quality blog writing by academics in my field (i.e. law professor and lawyer blogs). I can barely write something interesting twice a month, and some folks write better stuff twice or more per week. If I really wrote that frequently, my writing would be an order of magnitude worse, with attendant risk of writing something foolish. I don't see my incentives justifying that kind of risk.

But if you wrote much more frequently, wouldn't your writing get better that much faster?

The writing might but the ideas probably wouldn't, and the latter might be the limiting factor.

(That's why I don't do top-level posts here. I'm content with my writing's quality, at least when I proofread my writing carefully, but I don't have any thoughts about rationality worth a fully fledged post.)

I agree that blogging more isn't the right thing for everyone to do. But for people who are highly productive and whose reputation is sufficiently secure, I think that it can be a very good choice.

I think people who blog normally expose inconclusive thoughts or drafts, but not complete solutions. Or wants to teach more people, and build a community. In academic format, this is not so easy.

An implementation of a solution to your problem #2 and #3 for MO: instead of corresponding with people through email, write that email as a MO question and email them that question. Two problems with this though are that you have a more restrictive format - you have to make it up to MO's standards - and that you have to convince other people to do it.

Gil Kalai also has a nice blog.

Actually, I hadn't checked this site in a while. There is some awesome stuff there, including some questions probably of interest to many LWers.