Who to Trust

An earlier post looked at the problem of how to find truth when there are no experts you can trust. This one looks at the problem of deciding what is true in a world of imperfect information and unreliable sources from a different angle, starting with the commercial version of the problem.

San Jose Camera and Video is a store near me. Some years ago I asked them what I should buy in order to convert my slides, an extensive collection of pictures of historical jewelry that I had taken in museums around the world, to digital form. The response was that they had such devices but I shouldn’t buy one; sending my slides to a service that would convert them for me would cost less and produce higher resolution images.

Nowadays I photograph medieval jewelry with the digital camera built into my cell phone. Doing so in the Irish National Museum in Dublin a month or so ago I found the optical zoom of the camera insufficient to reveal the details of the very impressive Irish pieces. Back home, I went into the same store to ask them about gadgets you could clip onto a cell phone in order to get additional optical zoom. The response was that such gadgets exist but they didn’t carry them because they did not think they were good enough to be worth buying

[Two pictures of the same piece, demonstrating the level of detail and the limitations of the camera’s optics to show it.]

.That is two cases of the firm declining to sell me something because they did not think it was in my interest to buy it. If I ever want to buy another camera — not likely, given the high and steadily improving quality of cell phone cameras — I know who I can trust to sell it to me. Or not sell it to me.

In that case I happened to come across evidence that a firm could be trusted. How could one deliberately create such evidence?

One way is by asking a question you know the answer to. Most people know quite a lot about something. If you want to know if a camera store is honest and happen to already know that sending slides out to be converted into digital images costs less and produces higher resolution images than doing it yourself, you could put my question to the store clerk and judge the honesty and competence of the store by the answer. If you have already discarded one vacuum sealer because it didn’t work reliably and replaced it by another that did,¹ you could ask a store that sold both which it would recommend.

The approach is not limited to dealing with stores. If you want to know whether someone can be trusted as a source of historical information, whether a history professor or a fellow historical recreationist, get him talking about something where you already know that the popular account is wrong. Candidates are the claim that medieval food was over spiced to hide the taste of spoiled meat, that pasta was introduced to the Mediterranean world by Marco Polo bringing it back from China, that the library of Alexandria was burned down by Caliph Umar on the grounds that everything in it was either in the Koran or false.

“I don’t know” is a legitimate answer — nobody knows everything — but confident agreement tells you that this is not a source of information you can trust.

Another approach is to find someplace where someone else will do the work for you. I used to spend a good deal of time on a Usenet forum for people interested in science fiction. It contained lots of smart, argumentative people and a wide range of political views. Any time a political controversy came up one could be pretty sure of seeing good arguments, with luck the best arguments available, for one side, followed by the best rebuttals available. Following that exchange was a lot less work than trying to research all of the relevant questions for yourself.

Even if you cannot be sure which side of a controversy is right, you can often determine with some confidence that some arguments are wrong. You can judge then sources of information by what arguments they use. A source that makes bad arguments should not be trusted. A source that refrains from bad arguments, especially from bad arguments popular with its side of the controversy, may be honest.

Two cheers for the Huffington Post

During an earlier round of the debt limit wars, President Obama warned that failure to raise the debt limit might prevent the payment of social security checks. The statement was not true. The national debt includes money owed by the treasury to the Social Security Trust fund; every dollar paid to the Trust Fund reduces the debt by a dollar, permitting an additional dollar of borrowing.²  When I pointed that out on my blog someone pointed me at an article in the Huffington Post making the same point. They published the truth even though it went against their political interest.

That reminded me of a less striking example a few years earlier, after I took up a new hobby, defending Tea Party Republican candidates against stories that exaggerated how nutty they were.³ One of the cases involved  references to "Colorado’s Ken Buck, who says he opposes the principle of separation of church and state."

I found a video of the speech by Buck that was the source for the claim on the Huffington Post. The accompanying story, unlike the one I quoted in the previous paragraph, gave a reasonably accurate account of what he said, not that he opposed the principle but that he thought it had been applied more broadly than it ought to have been.

Both incidents were a considerable while back but they, and a third where the Post turned out to be a more reliable source of information than Reason Magazine, suggest that the Huffington Post is, at least was, a more honest source of news than one might expect of an ideologically oriented publication.

Usenet still exists but its role has been largely taken over by social media, blogs with active comment threads, and web forums. The approach still works; you just have to find an online conversation with lots of smart people, a wide range of opinions, and mostly civil conversation. That has become harder with increasing polarization — Facebook, in my experience, doesn’t do it — but not impossible. For some years my best source for intelligent arguments from all sides of multiple issues was the blog Slate Star Codex; it no longer exists but its archives are still up. Neither of its successors, Astro Codex Ten or Data Secrets Lox is as good, but they do provide forums populated by people who can intelligently disagree with each other. I expect many more that I have not come across exist.

Dealing with Falsehood Online

You read a newspaper or magazine and notice that it says something that is not true. Unless the statement is libelous and you the victim your only practical recourse is to write a letter to the editor.

What about online?

Some years ago, browsing the web, I came across an article on a libertarian blog. The blog was called "Classically Liberal Student," the article signed "CLS." Its subject was the Atlas Foundation which, CLS argued, had been becoming more conservative and less libertarian under pressure from a conservative source of funding. To support the claim he offered examples of purportedly anti-libertarian organizations that Atlas had funded. One was Maxim, an organization in New Zealand that received money from Atlas. According to CLS :

Maxim was explicitly anti-free market and attacked Milton Friedman when he died. Maxim said Friedman was "simplistic" and said he ignored the "social good". They say that "the individualist view, espoused by Friedman" is just as wrong as the collectivist view mainly because it ignores the desire of theocrats like Maxim to impose Christian morality by the force of law.

For obvious reasons that caught my eye, so I clicked on their link and discovered that what CLS said was not true. The article was not an attack, it did not contain the word "simplistic" and it did not say that the individualist view is just as wrong as the collectivist view. It said not that Milton Friedman was "simplistic" but that "economists like Friedman invariably approach problems with a simplified view of the world… ." That, of course, is true — of economics and many other sciences.

The article also said nothing about the desirability of imposing Christian morality by force of law. Its central argument was that while the individual freedom Milton Friedman had worked for was a good thing it was not, by itself, sufficient to produce positive social outcomes, that doing that also required free individuals to act with a conscience concerned with the common good of society. That claim is neither implied by libertarianism nor inconsistent with it.

Checking some of the other claims CLS made, I found further misrepresentations, so posted a comment to his blog pointing them out. Eventually it appeared, as did a response by CLS. He agreed that the text from Maxim that he had linked to did not entirely fit what he said it about it and speculated that perhaps it had been changed in the two years since it was written. He explained that "when the article was originally published I wrote an article about it and quoted from it. When I mentioned it, in an article about something far bigger, I quoted my original." He provided a link to what was apparently his original article, written some two years earlier. But  the material I was complaining about was in not in it.

Additional comments by me and CLS appeared on the CLS blog, in part in response to a post I put on my blog. His blog page still exists but the text has been edited to eliminate the specific claims I complained about and the comments appear to have vanished. That is better than continuing to publish false statements, worse than admitting and retracting them.

The owner of a blog can control what appears on it but has much less control over the information reaching his readers than the editor of a magazine or newspaper. Since CLS and I were both libertarians with blogs, some of his readers were likely to  read my post — and at least one did.

I have gone into the incident in some detail because it illustrates some of the advantages of the internet for determining what is true. As long as CLS left comments on his page open, I could point out problems with what he wrote that both he and his readers could see. Both he and I could and did try to support our arguments with links to the documents we were discussing, letting readers check our claims for themselves. By editing his text and removing comments he could prevent later readers from seeing the evidence of his unreliability, but current readers had already seen it. And he had no way of preventing later readers from coming across my account of the incident on my blog and drawing their own conclusions.

If a similarly inaccurate account had appeared in print twenty years earlier I could have written the author or the journal to point out the error but unless the journal chose to publish my letter its readers would never have seen it. I might have been able to publish a correction somewhere else, but the odds that readers of the original account would see it would have been much lower in that world than in one where search engines give anyone interested access to information from thousands of sources and checking the accuracy of our accounts of what other people had written would have required access to a good library and hours of time instead of a click on a link.

It is easy to find false information on the Internet, easy to find only information supporting what you already believe. But there are ways of doing better — if you choose to.

1. The one that works is a Nesco.

2. More details in a blog post and comment thread.

3. For example … .