Recently I was invited to write an article on the philosophy of Bayesian statistics. For a long time I’ve been unhappy with the discussions of philosophy offered by Bayesian statisticians and also with the perspectives on Bayesian statistics coming from philosophers. I’d been planning for about fifteen years to write an article on the topic but had never gotten around to it, so I welcomed this opportunity.
I thought it made sense to do some reading, and I thought I’d start with Lakatos, whom I think of as a sort of rationalized Popper (Lakatos actually attributes some of his own ideas to a hypothetical Popper_2). In retrospect, I think this was a good choice. I like a lot of what Lakatos had to say–even though he didn’t write much about statistics, or Bayesian statistics, most of the ideas transfer over fine, I think.
But that’s not the reason for this note.
I’m writing here to tell you what happened when I ordered the two volumes of Lakatos’s collected writings, published by Cambridge under the titles, “The Methodology of Scientific Research Programmes” and “Mathematics, Science, and Epistemology,” paperbacks selling on Amazon for about $50 bucks each. I eagerly awaited their arrival in my mailbox, but when they finally came, and I opened them . . . they were really hard to read! The type was blurry.
I guess they took the original book and did some sort of crappy photoimaging . . . Hey! This is Cambridge University Press we’re talking about, reprinting a classic academic book and not even taking the trouble to do it right! What’s with that??? I can see that it might be a pain to retype the original book, but can’t they scan in the text and reset it? Or, maybe even simpler, take their photoimaged text and run it through some software to unblur it? The current version is a joke, and I was embarrassed to even have it in my office.
I returned the volumes to Amazon and ordered the books from the Columbia library. (That was a pain too, but that’s another story. I doubt the readers of my blog need to hear about my problems with the Columbia library.) These original hardcovers are fine. Not the greatest print job in the world, actually–I find the font pretty hard to read–but much better than the blur-o-matic that Cambridge was charging $100 for. (Oddly enough, the printing in my paperback copy of Proofs and Refutations is fine.)
P.S. Yes, yes, I know this is unimportant compared to all the hunger and strife in the world, etc etc. But still . . . what ever happened to professionalism?
The fact that you ordered it first through Amazon rather than from the library says enough about the utility of Amazon, the pain of a traditional library, and the value of your time. I haven't been to a brick-and-mortar library in years.
Too bad the used hardcovers are over US$100 on Amazon.
Once you read a book on typesetting and another on graphic design, you realize most of the production team for most books wasn't paying attention in class.
If you're keen on good fonts, check out the film Helvetica.
Yeah, this sort of thing is irritatingly common. Heinemann's latest reprints of Landau and Lifshitz's "course of theoretical physics" also appear to have been scanned in incompetently. There are smudges all over the place and half the subscripts are illegible: it's easier, in some bits, to rederive equations than to decipher them.
But as everyone knows, academic publishing is a racket. Heinemann is owned by Reed Elsevier, and Elsevier's got better things to do with their good printers, like publish fake medical journals.
abebooks.com should be your friend :)
<a href="http://www.abebooks.com/servlet/SearchResults?an=Lakatos%2C+Imre&sts=t&x=49&y=12" rel="nofollow"> <a href="http://;http://www.abebooks.com/servlet/SearchResults?an=Lakatos%2C+Imre&sts=t&x=49&y=12” target=”_blank”>;http://www.abebooks.com/servlet/SearchResults?an=Lakatos%2C+Imre&sts=t&x=49&y=12
The universe of small used book stores is a wonderful thing!
I heard of one high-prestige publisher that, in the 90s, used to put out anthologies of papers by sending someone to the Stanford library and *xeroxing* articles. And that was their camera-ready copy.
If you decide to write a book about bayesian statistics, make sure to include "poker philosophy" into it.
These professional gamblers are the best type of applied statisticians. Their MO is to maximize wealth, with the aid of Bayesian statistics, and maximizing wealth is the most important thing.
I'd also recommend Bill Chen's book the Mathematics of Poker to be a good start. He only briefly touches upon bayesian statistics, but you see the parallels.
I also wrote a super quick blog post about bayesian statistics and poker at:
http://www.investorgamblereconomist.com/2009/05/b…
Shameless plug, but I think it's relevant.
Hi Andrew,
I'm interested in what your philosophy has to say about Don Fraser's views, which seem to be very similar to Larry Wasserman's views: an interval estimate ought to have the claimed coverage to as good an approximation as possible. The work on matching priors shows that Bayesian credible intervals only have claimed coverage to a good approximation for special cases, and cannot provide accurate coverage for more than one scalar parameter at a time (I think).
My personal response to this view is to fall back on the complete class theorems, but I get the impression from my reading of your books and papers that you wouldn't look to statistical decision theory to justify Bayesian inference.
Corey: Please take a look at this article and search on "everybody wants to go to heaven but nobody wants to die." The first three paragraphs on that page pretty much cover my view on this.
Here is what I want to know: Why is this field still called Bayesian Analysis? Yes of course Bayes Theorem is important, but he might not even agree that what you all are doing is valid. He isn't really the true "father" of the field like Darwin, and it is odd in Math to name a field after a person rather than a construct or theorem. It always strikes me as odd. It ought to be called Modeling Structuralism or some goofy thing like that. Maybe that is why it still has an old name, I can't offhand think up a better. But then, I don't get paid for it either.
Markk, the term "Bayesian" seems to have originated in the 1950s. Prior to that, the method was called "inverse probability". Here are some links: When did Bayesian inference become "Bayesian"?, Earliest known use of some of the words of mathematics.
Markk: I've never liked the term "Bayesian." If I had to use a name, I'd prefer "Laplacian." Or something more descriptive involving the term "probability modeling." We thought long and hard about this when titling Bayesian Data Analysis but finally decided that Bayesian is the term that everybody is familiar with.
There does seem to be a hope for theorems that will set out what must be done in practice rather than just rhetoric about what should be done.
This might explain some of the reluctance – in the past ;-) – of Bayesians to doubt and check their priors. With a given joint distribution there is an error free way (a "must") to get the posterior – as long as you don't question that joint distribution.
Peirce's take was to some how maintain inquring communities that practice science in manner that they continually gets less and less wrong (in expectation).
Keith