NEW YORK (FORTUNE) - One of the most important lessons you can ever learn about markets is also one of the easiest to forget: Just because prices are more reasonable than they were doesn't mean they're reasonable. I'm sorry to report that it's absolutely the lesson to keep in mind now that the Dow has hit 42-year highs and crept back up near 11,000.

The preeminent teacher of that lesson is Robert Shiller, a Yale professor with a strong record of thinking independently and being right. His book "Irrational Exuberance," arguing that stock prices were insanely high, appeared almost precisely at their peak in March 2000. Now he has updated the book to reflect 2005 valuations and concludes that, believe it or not, the market is still irrationally exuberant.

How does he come to this conclusion? After all, stocks are generally lower than back in the bubble days, and we've had four years of economic growth to rehabilitate corporate profits. His answer is simple. As he told me the other day, all the competing theories boil down to one easy-to-understand calculation: "The trailing P/E ratio for the S&P composite is still around 25, vs. a long-term average of 15."

That's a huge difference, much greater than what you read about in the newspapers. The commonly cited figures -- a current market multiple of 17, vs. a historical average of 15.2 -- are based on the previous 12 months' earnings. But, as Shiller points out, that's foolish: "Twelve months is kind of short, only a fraction of one business cycle."

So he uses a ten-year earnings average, an approach advocated by Graham and Dodd in Security Analysis, the value investor's bible. And while prices are clearly above the long-term trend any way you cut it, by that measure they are still mountainously beyond normal.

For some people -- I don't want to mention any names, but cast a glance at "Nope -- We're Too Gloomy" -- that conclusion is impossible to accept. So they contort the numbers and cook up theories about why today's prices aren't really as high as they appear. The most significant theory, which surveys show is believed by vast armies of investors, is that stocks aren't as risky as we used to think they were, so they're actually worth more than investors have historically been willing to pay. In other words, we were simply wrong for the past several decades but at last have seen the light, and in that light today's overall market valuations make sense.

Shiller would hoot at that one if hooting were his style. Instead he just mentions that this is "the Dow 36,000 theory." That 1999 book by investing columnist James Glassman and former Fed economist Kevin Hassett, you'll recall, argued that prices would rocket as the populace realized that in the long run, stocks always beat other investments, so they're really safer than conventionally thought.

We must all thank Shiller for reminding us of this prediction from the book: "... a sensible target date for Dow 36,000 is early 2005, but it could be reached much earlier." Or not.

It's easy to make fun of Dow 36,000, but it's more important to recognize that the theory behind it is still at work, and it still doesn't add up. As Shiller points out with voluminous support, it just isn't true that stocks always outperform other investments over long-term periods, and, he says, "there is certainly no reason to think they must in the future." If that's true, then stocks would appear to be just as risky as ever. We are not in a "new era." Math still works the same way. And today's valuations are too high.

No one wants to hear that. It's almost irresistible to believe that after all we investors have endured -- the hellish bear market, the recession, the scandals -- we've emerged from the crucible sadder but wiser, finally willing to face the truth about stock values. But it isn't so. The amazing reality is that we haven't learned our lesson even yet.

Find out why FORTUNE's Andrew Serwer says we're still too gloomy.