A lot of times, people put forward very complicated models of political situations.  While an accurate model is likely to be complicated, it doesn't work in reverse: the complicated models can be very, very inaccurate, and can serve more to (sometimes intentionally) confuse things or divert attention from important facts than to illuminate the situation.

Chomsky will occasionally toss a "naive" model (as in, "How would we model this if we didn't know anything else about the situation?") into the argument, not necessarily because he believes it's accurate, but to clarify how people are thinking about things.  One example would be where he explains the closeness of the 2000 US Presidential election with a random model: if all the voters voted at random, you'd expect results like that.

Another is when he talks about his model of the media, which is basically that it's corporate-owned and staffed largely by people of one class.  The first-order predictions are more or less that the media won't rock the boat, will avoid deep issues, and will represent a narrow range of views.  There are second- and third-order effects, too, but these are less important (describe less of the media's behavior); the simple model works really, really well, even if it isn't a complete or completely accurate description.  It's analogous to how the moon isn't really a sphere, but if you model it as one, you won't go far wrong.

This is a really crummy ''persuasive'' technique, by the way -- unless you're debating with scientists, and scientists who aren't already wedded to a complicated theory.  People tend to interpret simple theories as not being serious, possibly because they are threatening, and therefore get tangled up in the humor response.

An example would be the article by Chomsky that I read in an International Relations class, in which he pointed out that the simple model of American foreign intervention was that American intervention made people ''less'' free in nearly every instance.  This really put me off, to the point where I suspected the professor included it in the readings solely to discredit Chomsky.  I ended up not reading anything more by him for a few years (then I plowed through ''Deterring Democracy'', ''Necessary Illusions'', ''Manufacturing Consent'', and a few others, and things started to click).
--GeorgePaci

You might be interested in what Micheal Albert's testimony at the WSF (http://www.zmag.org/albertpa.htm). He makes the point that most people already know capitalism is evil but think it is inevitable, so talking about the evils of capitalism is like talking about the evils of old age; painful and useless. Perhaps, by giving a 'naive model' that cuts across the bullshit, Chomsky just hastens the point at which people demand an alternative vision. That doesn't seem to fit your own experience though. -- RichardKulisz