WASHINGTON –

Economist Robert Pindyck of the Massachusetts Institute of Technology recently examined the computer models that estimate the effects and costs of climate change — and didn't like what he found. The models reflect two gaping uncertainties, he says. First, we don't know how much increased atmospheric carbon dioxide will raise global temperatures. "There are feedback loops" — interactions between greenhouse gases and weather — "that aren't easy to measure." The models make assumptions. Next, he says, we don't know what economic losses will result from higher temperatures. More assumptions. The "damage functions" in the models, he says, "are completely made-up."

Pindyck sounds like a "global warming denier." He isn't. True, he thinks climate change and its adverse economic consequences could be wildly overstated. He also thinks they could be wildly understated. The effects might ultimately be catastrophic. We simply don't know. Ignorance reigns. The best course, he says, would be to adopt a modest carbon tax — because there are certainly some ill effects — and adjust it as we learn more. Meanwhile, we shouldn't assume that computer models convey scientific truth. "The models create an illusion of knowledge," he says. "For me, the issue is being honest."

I'd call Pindyck a global warming pragmatist, and it's a middle path that I find appealing. It acknowledges warming's uncertainties but doesn't use them as an excuse for inaction. For years, I've advocated an energy tax — my preference now is a carbon tax — because it could advance other national goals. It could reduce budget deficits and enhance energy security by pushing consumers toward more efficient cars and trucks. That's my standard: Support policies that, though they might address climate change, can be justified on other grounds.