Gricean Maxims, and why I have trouble with them...


Categories: Personal Politics

A while back, I commented on Romney’s “Keep America American”. At least one person told me that I had given the impression that I thought that the people who liked this slogan were racists; this confused the heck out of me, because I thought what I said was that racists liked the slogan, and that I believe it was constructed the way it was in order to appeal to them.

To me, these statements are virtually unrelated, and it is confusing that anyone would read one as the other. But let’s look at more cases.

In a recent discussion pertaining to various claims made by politicians, I saw a claim in passing which struck me as incorrect, so I responded to point this out. Since it was on tumblr, I left the entire quoted post (because that’s tumblr-custom) even though I was only responding to part of it. But the thing is, I responded to a thing which is… well, not what most people would have responded to.

There’s a set of rules, called Gricean maxims, which are basic principles of interpreting communication. They are instinctive in humans… no, wait. They are instinctive in non-autistic humans. They’re not instinctive for me. I have to think about them. As a result, I can forget to, and this often results in me communicating things wildly unlike what I thought I was saying.

The particular thing that gets me with corrections is an order-of-operations problem. For most people, the processing order is:

  1. See statement.
  2. Interpret statement according to basic rules.
  3. Evaluate statement for things like truth or falsehood.

For me, it’s:

  1. See statement.
  2. Evaluate statement for things like truth or falsehood.
  3. If statement is true, interpret statement according to basic rules.

Now, here’s where it gets funky. If I am arguing with a claim because the statement pre-interpretation was false, people who aren’t autistic are very likely to end up perceiving me as disagreeing with the post-interpretation meaning. Which can be a source of conflict, especially in the fairly common case where I happen to accept the intended communication as true but uninteresting, so I’m arguing entirely about a related point. e.g., when someone posted a list of accomplishments by Muslims (in response to a picture of the WTC buildings captioned “imagine a world withous muslims”), I jumped on a couple of them as obviously bogus, without a moment’s thought to the question of whether the underlying implied arguments had any merit. I wasn’t going to think about the logical implications of the statements until I’d resolved the question of why some of them were false.

That said, these really are lovely and useful principles. As always, Wikipedia has a good quick summary. And there’s an interesting variance; I can forget to apply these because I’m in a hurry. People who apply them only unconsciously, however, run the risk of being unaware of whether or not they are applying these rules, and are quite likely to apply them selectively to improve the perceived performance of people they like, or disimprove the perceived communications of people they dislike. (This is a failure to apply the “principle of charity”, another wonderful communicative tool.)

For an interesting example, consider the claim from tonight’s debate that Obama did, or did not, call the attack in Libya an “act of terror” the day after it occurred. (Language log has now written about this too, and gives better examples.)

The actual words from the president’s speech:

No acts of terror will ever shake the resolve of this great nation, alter that character, or eclipse the light of the values that we stand for. Today we mourn four more Americans who represent the very best of the United States of America. We will not waver in our commitment to see that justice is done for this terrible act. And make no mistake, justice will be done.

A number of conservative pundits have rushed to point out that this does not actually state that the attack in question was an act of terror. This is… well, if they’re autistic it’s careless, if they’re not it’s dishonest. This is a speech about the attack on the embassy. Everything in this speech is presumptively related to or about that attack. Which is to say: Either this constitutes an (indirect, but clear) assertion that the attack in question was an “act of terror”, or it is a lone paragraph on a completely unrelated topic which is inexplicably sandwiched between two paragraphs of talking about the attack.

Only one of these interpretations is remotely plausible.

I don’t dispute that, in general, politicians lie. If someone were to assert that Obama chose that phrasing so that, if he were later found to be wrong about the attack, he could pretend that he never meant to imply that it was an act of terror, I would consider that an untestable claim, but certainly a highly plausible one. The man’s not an idiot; he had insufficient information, and probably acted to hedge his bets. But to jump around yelling that he didn’t actually specifically identify the subject of his speech as the subject of a paragraph within his speech is idiotic.

Had these commentators been asked, without awareness of the speaker, whether a paragraph like this in a speech about the attack in question could be reasonably construed as asserting that the attack was an act of terror, I do not doubt for a minute that they would have said yes. But they weren’t asked about that. They were asked, after the candidate who is their sole hope of defeating a hated opponent had asserted that Obama had said nothing of the sort until two weeks later, whether these remarks proved their hero to be a liar.

Of course, both candidates lie. They lie at an astounding pace. And yet, for some reason, the vast majority of people in the US are thoroughly convinced that their candidate doesn’t lie, and the other one does. A moment’s thought dispels this notion. Which of the following is more likely?

  1. 47-48% or so of the American public are completely and totally incapable of critical thinking, and unhesitatingly believe good things about their preferred candidate and bad things about his opponent, without regard for the evidence. The other 52-53% are fine.
  2. 48-49% or so of the American public are completely and totally incapable of critical thinking, and unhesitatingly believe good things about their preferred candidate and bad things about his opponent, without regard for the evidence. The other 51-52% are fine.
  3. The entire populace is systematically biased in exactly the way that has been confirmed by basically all of the research we have ever done on how humans evaluate truth claims and how halo effects work.

I’m betting on that last one.