Sunday, November 25, 2012

Googling for Usage

I am currently working on a chapter for the third edition of my first book. In it I make repeated references to a concept I usually refer to as a Schelling point, after Thomas Schelling who came up with it. An alternative term for the same concept is "focal point." It occurred to me that perhaps I should use it instead. How to decide?

One obvious way is usage—and nowadays, there is a quick and easy way to check that. I googled for "Schelling Point" and got an estimate of about 5300 results. I tried "focal point" and got an estimate of over thirty million. That seemed to settle the question—until I looked at the first page of the second search and realized that many of the results were for entirely different meanings of the term.

So I tried googling for ["Focal Point" AND Schelling], hoping that Google's search language was adequate to understand that. Apparently it is—at least, the result was down to 32,400. As a check, I tried comparing the result for ["Schelling Point" AND "game theory"] to the result for ["focal point" AND "Game Theory"]. The first got me 2710 results, the second 134,000. 

None of those searches provides perfect information, both because Google's estimate of the number of results is, in my experience, a very uncertain one, and because my search strings do not perfectly identify contexts where the two terms are being used to mean the same thing. Further, I don't really know how sophisticated a search language Google understands—it may interpret my AND as asking for the word "and" rather than limiting the results to pages that include both terms. But the results were sufficiently strong to make me reasonably confident that my preferred terminology is, by a substantial margin, the less common one.

It may occur to readers familiar with Schelling's idea that the story I have just told is relevant to the idea as well as its label. A Schelling focal point is a result that two or more people coordinate on because of its perceived uniqueness. Schelling's initial example involved two students offered a reward if they managed, without any communication, to both be at the same place at the same time in New York city the next day; they ended up under the clock in Grand Central Station at noon. One of my favorite examples involves two bank robbers arguing over how to divide the loot. Each believes he did more than half the work and is entitled to more than half the money, but they agree on a fifty-fifty split because that is the one division that both see as unique; if they argue too long over who is entitled to more and how much, the cops may show up. In the first example the students coordinate on what both see as the unique place and time because they are not permitted to communicate. In the second, the two robbers coordinate on the unique division because although they can talk they are unable to communicate—each has an incentive to claim that he will only be satisfied with the larger share, whether or not it is true.

Language, word usage, also involves a problem of coordination without communication, since it is not practical for me to discuss with all other speakers of the English language what words we will use for what ideas. One possible Schelling point is usage—everyone agrees to whatever terminology the larger number of people currently use. Enough people acting that way could considerably simplify the language—and, arguably, has done so. Google, as I have just demonstrated, makes it much easier to find out what terminology is in more common usage. If enough people use it that way, the effect on the language could be significant.

Not, of course, an entirely positive effect. Speaking as an economist, I regard consistent terminology as on the whole a good thing. Speaking as a poet, on the other hand, there is much to be said for having three different words that mean the same thing—the first two you try might not fit the meter or rhyme scheme.

Readers curious as to what focal points have to do with the subject of my book (The Machinery of Freedom) can find the answer in an old article with an earlier version of the argument. It includes footnotes crediting some of the people whose ideas influenced it. One thing I forgot to include was thanks to the late Earl Thompson, the person who persuaded me of the importance of commitment strategies in understanding human behavior. I plan to remedy that error in the chapter I am now working on—and have just done so here.

At a considerable tangent ... . I first started thinking about the problem of coordination without communication as a college student coordinating plans with my parents at a time when long distance calls were expensive. It occurred to me then that that problem is the central feature of one of the world's most popular games: bridge. From a game theory point of view it is a two player game, since each pair of partners has interests entirely in common. But it is a two player game in which each player consists of two people, permitted to talk to each other in only a very restricted fashion.

Which led me to suspect that perhaps each of the world's great games can be viewed as designed to teach a particular skill. The case of chess is obvious as is that of Diplomacy, Go perhaps a little less so; I never carried the idea much beyond that. But readers are invited to offer additions to the list.

13 comments:

Eric Goldman said...

Not completely responsive, but you might also try Google nGram http://books.google.com/ngrams Eric.

dWj said...

I tend to use the term Schelling point precisely because of the first issue you encountered; when I say "focal point" I sometimes think it seems ambiguous, but even where it could be determined by context, if I'm talking to someone to whom the context is new, they're going to have better luck finding more information googling "Schelling point" than "focal point".

(On a barely related note, my name is much more "google friendly" than yours.)

DoJ said...

I must confess some shock at how few Google search results there are for "Schelling point"; I had thought it was a standard, widely used term.

Anonymous said...

I tend to agree with dWj: the phrase "Schelling point" is much more distinctive, and less likely to be confused with completely unrelated phenomena, than is "focal point".

Anonymous said...

It seems to me that the same Schelling-point phenomenon can be seen in dogs choosing places to leave their "p-mail": they very seldom choose a featureless stretch of grass, but more likely an unusual upright feature, particularly corners of buildings. Of course, there's also a positive-feedback loop: dogs urinate in the same places as others have already done. But the first dog to urinate in a particular area will almost certainly choose a "distinctive" place.

Joe said...

I have two things I'd like to see in a 3rd edition of MOF

1. Examination of government spending in non-US countries. My personal preference is the UK and Canada. But Sweden might be a good idea, since it is often used as an example of socialism working.

2. More references for some of the facts - It helps with copying the arguments and using them as my own. An example from memory is the relationship between calories consumed and income in the US. Or perhaps I'm just searching for these things incorrectly...

Anonymous said...

Google automatically AND's things, so you don't need the AND in there.

One trick to double check the estimation is to skip to the last page of results. Often you hit the 1000th result (Google doesn't give you any more), but sometimes you'll find out that there are only 700 results even though Google estimated 12000 or something.

Tibor said...

David Friedman & Joe:

I think a very interesting place to examine might be present Iceland. I am quite confused about what the situation in that country. There is little information in media that I would consider at least reasonably unbiased.

What I found was pretty much from rather dubious webpages and all somewhat along the story that Iceland had a very free banking market which collapsed because it was not regulated and now "the Icelandic people" got together and made a new constitution more in sync with the scandinavian socialist model.

What seems unlikely to me is first that there actually was a free market. I would guess there were some goverment subsidies or something distorting the market like that, however not regulation - I found that a lot of people on the left understand corporate welfare system as a "free market" so that is why I suspect something along those lines. And I have some doubts about "people" creating the new constitution too.

And if those websites are actually true, it makes Iceland an even more interesting case to be discussed.

Perhaps light can be shed on that in the book...or here? By anybody who knows more tham me, really :)

I apologize if this is just too much off topic.

Anonymous said...

How long will it take to rework the book?

David Friedman said...

"How long will it take to rework the book?"

I don't intend to rework the book. My plan for the third edition, as for the second, is to leave what already exists mostly unchanged, and add a new section with new material.

I expect to have that written in less than a year.

jimbino said...

Well David,

I hope you hire an strict editor to review your grammar, as there is nothing worse than trying to read a book written by someone who gets his grammar by googling.

For example, if you get your grammar by googling, you'll end up writing stupidities like "try and eat..." when common sense would dictate "try to eat...."

RJ Miller said...

This is something I will definitely anticipate for the next several months!

For a while now I have had the intention of making some kind of outline or study guide for the book but was unsure of when the third edition would come out. Glad to know the wait will soon be over!

Anonymous said...

I agree with dWj and hudebnik. "Schelling point" is a much more distinctive term, and less likely to cause confusion.

Economics already causes enough confusion through using ordinary terms in its own special way. It would be better to avoid adding "focal point" to that list. And even if "Schelling point" isn't now a widely used standard term, it deserves to become one.