The 1st Amendment, which guarantees freedom of speech and freedom of the press, takes the view that the people should dictate to the government, not the other way around. But no one told a group of 32 state attorneys general, who have taken it upon themselves to instruct the film industry on the appropriate content of movies.

This time, the cause is not raunchy sex, foul language or blood-spattering violence. It's cigarettes. Many experts think that when actors puff away, they cause teenagers to do likewise. One study went so far as to say that 38 percent of all the kids who acquire the habit do so because of the influence of films. So all these state government officials want filmmakers to stop depicting tobacco use.

They evidently have had an effect. Not long after the attorneys general sent a letter requesting action, the Motion Picture Association of America (MPAA) agreed to use smoking in determining each film's rating. "Depictions that glamorize smoking or movies that feature pervasive smoking outside of a historic or other mitigating context" would run afoul of the ratings board. Apparently it would be OK to show an unwashed lowlife taking a drag just before he drops dead of a heart attack.

The MPAA didn't go as far as demanded by some anti-tobacco groups that want to slap an R rating on just about every film in which actors light up. But it accepted the basic principle that public health lobbyists and politicians should have a big role in deciding what people will see, instead of letting the industry merely cater to its audience.

It's hard to fully credit the notion that kids start smoking just because they see Scarlett Johansson doing it. Steven Milloy, publisher of the website JunkScience.com, points out that adolescent smoking has declined even as onscreen smoking has increased. If movies exert such a mammoth influence on impressionable youngsters, shouldn't teen tobacco use be on the rise?

The studies themselves are not as damning as they purport to be. They indicate that kids who watch more movies with smoking are more likely to smoke. But a correlation does not necessarily show a cause: Just because there is lots of beer drinking at baseball games doesn't mean beer drinking causes baseball.

It may be that kids see a star light up and rush out to imitate him. Or it may be that teens who are inclined to smoke anyway are also inclined to see the sort of movies that feature smoking. Continued...