TCS Daily

The Scribe's Problem

By Tim Worstall - December 22, 2005 12:00 AM

I found Robert McHenry's recent piece on the superiority of Britannica over Wikipedia to be fascinating, for I think he's allowed himself an error of logic that we more usually encounter in economics. It was also a little unkind of the publishing gremlins to schedule his piece the day before Nature came out with that research into the relative accuracies of the two approaches. The, umm, research that showed roughly comparable levels of errors in the amateur thing thrown together on the web and the one expensively and carefully produced by multiple levels of editors.

The error of logic I mean is the one we normally see when people note that there is a problem with a market, a failure perhaps. The instinctive reaction is then that this must be a job that we get Government to do as we can actually see the faults in the market approach. What usually takes some years of good tutelage in a decent economics department (otherwise known as common sense, you knew this before you went to High School but had it educated out of you) is to look for what is hidden. That the Government can (and often does) screw it up quite as successfully as any failing market.

That experts, each responsible only for that part of the encyclopedia in which they are experts, should, under close supervision and proper editing, lead to a better and more reliable product than any old yahoo surfing the web seems like a logically impeccable position. Yet as we see from Nature this isn't (despite the Mnestheus point...a name I must admit I usually associate with the email address of a fellow writer here at TCS) necessarily so. On science subjects at least, and I assume that such were chosen because it is in fact possible to be entirely objective about what is right or wrong.

The thing I think that is hidden here, the Bastiat Point if you wish, is that articles in Britannica are not necessarily written by the experts and Wikipedia ones sometimes are.

As our example for the day I shall present myself. I am, in my day job as head of the international scandium oligopoly (yes, really, someone's got to do that job) the expert on the mining of, volumes, trade in and uses of scandium. That's "the" not "a" or "an" expert. I have not been asked to write an encyclopedia entry as yet but I have modified the Wikipedia entry. Quod erat demonstrandum perhaps?

Well, not quite. Leaving aside the Britannica's entry for scandium, which as far as I can see looks absolutely fine if less informative than Wikipedia's, and look instead at the Columbia's. There are a couple of things still not quite right with the Wikipedia article:

Thortveitite is the primary source of scandium

This is untrue as it says later:

The main source of scandium is from military stockpiles from the former Soviet Union, which were themselves extracted from uranium tailings.

Which is true and is part of what I added as I did this:

About 80 kg of scandium is used in lightbulbs globally per year.

(Well, to be completely honest I'm such a technological fool that I emailed someone else the information and he added it).

Both those facts (and believe me I know exactly how much scandium I sell each year to the lighting industry) are only to be found at Wikipedia. No other encyclopedia, industry magazine or database, not even the Minerals Department of the US Geological Survey, includes these as yet.

On the other side, from the Columbia:

Scandium oxide (scandia) finds use as a catalyst and in making crucibles and other ceramic parts. Scandium sulfate in very dilute aqueous solution is used in agriculture as a seed treatment to improve the germination of corn, peas, wheat, and other plants.

There is a very limited use in some ceramic parts and there's been talk of use as a catalyst but crucibles? In agriculture? No and no.

When I first started out on this chase for the plutocratic wealth that my iron fisted control of one of the world's precious natural resources would give me I did, as you might imagine, some research. This included reading all of the encyclopedia entries I could find and many of them included the above information. One reason it would appear multiple times would be that it is true. But it isn't, so scratch that. In fact, in its original form, it stated that scandium nitride was used in making crucibles for the casting of gallium, the above formulation coming (I hope) as a result of my multiple emails to the editors at Columbia stating that they were in error.

In the course of my research I found that silicon nitride is used to make the crucibles to cast gallium and that selenium sulfate is used in germination. I couldn't work out why these errors should be repeated until at a metals conference I was introduced to the man who had just revised the rare earth metals section of one such encyclopedia (no, not Britannica or Columbia).

"Ah, hello then, so tell me, what sort of scientific expert are you to know so much about all of these different metals?"

"I'm a journalist, I start by reading all of the other encyclopedias."

It is at this point that Inspector Worstall is able to declare the case solved. As my gorgeous and pouting assistant, pneumatic in that very Hollywood manner, gazes on adoringly, (soon to be a major TV show folks!) I can with a folksy grin declare it to be a case of The Scribe's Problem.

We more normally see this in Biblical scholarship although it is true of all and any texts handed down to us from antiquity. Over the centuries and millennia these books have been constantly copied and the standard of the scribes doing the (handwritten) work has varied. Errors have, therefore, inevitably crept in. Indeed, I'm told by those more educated than myself that checking and correcting similar errors is at the heart of the building of a modem and thus the entire internetweb upon which we are playing at present.

Once the errors are there and then used as the basis for the next generation of scribes then what is in fact that error becomes the truth which must be promulgated. Sometimes they get caught as with the Wicked Bible of 1631 which rather got the 7th Commandment wrong:

"Thou shalt commit adultery"

And sometimes they don't get caught as with our examples above of the uses of scandium. If we switch to scientific notation for a moment, scandium is Sc, selenium is Se and silicon is Si. We can imagine at some point a weary (bleary possibly given the reputation of those in the craft) eyed journalist skimming through the information and mistaking an e or i for a c. Once this has passed the editing process and been included in a published version it then becomes the primary material from which the next generation of encyclopedia writers do their research.

What I think we end up with is that while the strictly edited version should be better, just as government by the wise and just should be, that isn't actually what we get. Not because encyclopedia editors are not just and wise, unlike their counterparts in government, but because they are human. Errors are inevitable. The very different methods used to create Wikipedia are also subject to error, but a different series of them. Just as with markets.

At this stage of the comparison between the two methods I wouldn't want to come down unequivocably on either side, unlike my known opinions on the market/bureaucracy thing.

I'll also have to admit that I'm rather with PJ O'Rourke in such matters. The last published reference book that I believe in all and every particular is the 1911 version of the Britannica.

Perhaps this will be of some comfort to Mr. McHenry though. I do very much treasure (and have safely stored away) the letter sent to me by Britannica when it was under his Editor in Chiefship. It thanks me for pointing to a small error in an entry (the use or not of hafnia in stabilizing zirconia if you'd like to know) and assures me that it will be removed in the next edition. Attempting to change Wikipedia gave me a barrage of abuse from those who could not understand how, in this day and age, someone could actually walk upright and not know how to edit a page.

So full marks for politesse to the old way of doing things, at the very least.

And a piece of advice if you are looking something up? Check with at least two different sources and make sure that they themselves are working from different primary sources.

Tim Worstall is an entrepreneur and TCS contributing writer, living in Europe.

TCS Daily Archives