August 12, 2006

Israelis killed, Lebanese die

Dr. James Eitel wrote to the Oakland Tribune to point out that an Aug. 7 front page story headline in the paper had said, in large print, Rocket attack kills 12 Israelis, and then in smaller print, At least 16 in Lebanon die from airstrikes. Dr Eitel remarked: "This has an immediate bias that Israeli lives are more valuable than Lebanese lives." Does it? It's certainly not immediately clear.

To say that an event kills people is to say (roughly) that the event was the immediate cause of their their death in a direct way. To say that people died from an event is to say (roughly) the same.

I think it might be more common to use "die from" with causes of death like diseases or effects of injuries that might be spread over time (The man had died from exposure/burns); and "kill" is more common with sudden events (The fall/gunshot killed him instantly), but I don't have statistical evidence for that, and one could hardly be surprised to read The man had died from a fall, or The cancer had slowly killed him. And anyway, the difference would hardly support a claim of implying greater value for the lives lost from events described one way rather than the other. Dr Eitel is right to be looking for evidence of bias one way or the other, but the linguistic analysis involved is subtle, and must be done with real care. (Recall this case where some claims about bias in the opposite direction — anti-Israeli — involved claims about passive clauses but the analysts could only identify passives correctly one time in three.)

Dr Eitel goes on to say that "in the text of the article, there were four paragraphs describing the location, circumstances, and human consequences of the rocket attack on Israel. There were exactly zero words describing the location, circumstances and human consequences of the Lebanese killed by Israeli weapons. Are Israeli lives more valuable than Lebanese?" They are not, of course. But I bet the exact locations and human consequences of an air strike north of the Lebanese border in a bombing zone are harder for a journalist in Israel to find out about than the exact locations and human consequences of a rocket hitting a kibbutz in the relative safety of Israel. One has to think about that too: reporters are located in specific places, and have just hours to find out interesting facts and file a story, and often don't want to just ride a motorcycle into a war zone and die for their paper. We can ask for some editorial judgment back home where the paper is composed, but we can't insist that nothing be printed until the reporters have defied death to get exactly the same amount of it regarding the deaths of the Lebanese victims as they have regarding the deaths of the Israeli victims.

[Update: A correspondent points out to me that Dr Eitel might have been talking solely about the size of the letters in the two headlines as suggesting differential importance. If that were so, the above might be just an irrelevance, stimulated by a misreading. I note, though, that "Rocket attack kills 12 Israelis" has only 31 characters (including space). "At least 16 in Lebanon die from airstrikes" has 42. I don't know whether that length difference alone would have required the smaller type, and I also don't know whether one could argue that size of type in headlines symbolizes a newspaper's feeling of differential importance of the facts stated. Maybe. Again, this should be an empirical matter: it ought to be checkable by comparing (I suppose) editors' opinions with lengths of headlines and subheaders suggested. I wouldn't bet on what investigations of this topic might turn up.]

Posted by Geoffrey K. Pullum at August 12, 2006 10:51 AM