I participated in a comment thread over at Megan McArdle's blog, one of my new favorite places on the web.

What they are describing is generally called a "negative externality." A negative externality is where you do not bear the full cost of your economic activity: it is borne by another individual, group of individuals, or society as a whole. Environmental issues lend themselves to easy illustrations of this, but noise pollution from a bar could also be classified as a negative externality.

The key point here for libertarians is that markets do not function efficiently in the face of negative externalities. If I'm producing widgets for $5 and selling for $10, but Joe and Bob incur $15 of damage to their property every time I produce a widget, then society is operating at a net loss. If you don't have an efficient market, then you don't have all of the much-touted benefits of capitalism.

In general, there are two ways to deal with negative externalities:

1) You can regulate them. This involves either prohibiting certain activities (NO loud music after 2 a.m., etc) or setting caps (max carbon emissions per year, etc.).

2) You can price them in. In the widget example, this would mean that the gov't imposes a tax on me of $15 to compensate Joe and Bob.

Libertarians and conservatives tend to reflexively act against anything that calls for more government involvement. The problem here is that many confuse a useful tactic (less government regulation) with the strategic goal (an efficient market). If you achieve the former at the expense of the latter, it really doesn't count as a win.