Juan Non-Volokh said that Joe Lieberman said something false on the weekend:

For example, Lieberman stated that the Bush Administration’s “Clear Skies” proposal to reform the Clean Air Act “actually would increase pollution” … He’s wrong … and should know better as a member of the Senate Environment Committee.

First, the proposed “Clear Skies” legislation will reduce utility emissions of NOx and SOx by around 70 percent. As I have noted before, the worst that can be said of “Clear Skies” is that it will reduce utility emissions marginally less than they might be reduced under current law – I say “might” because current projections presume that the current regulatory process will stay on schedule, and this is unlikely. Either way, this is not a policy that “actually would increase pollution.”

My first thought was that there’s a meaning for ‘increase’ that Lieberman could be using here. On second thoughts, I’m not so sure, but the semantic question is pretty interesting I think, at least if you’re a semi-professional semanticist.

I’d have thought that the meaning of ‘increase’ would be in the ballpark of these two concepts, which I define stipulatively.

X increases_t Q if X causes Q to be larger than it previously was. (The t subscript is to indicate that this is a temporal concept – what matters is that Q grows over time as a consequence of X.)

X increases_c Q if X causes Q to be larger than it othewise would have been. (The c subscript is to indicate that this is a counterfactual concept – what matters is that Q is larger than it is in some relevant counterfactual situation.)

What is the relation between the meaning of ‘increase’ and my concepts increase_t and increase_c? If it means the disjunction of the two, or it is ambiguous between the two, then Lieberman is (arguably) off the hook, since Juan is conceding (at least for the sake of the argument) that the Clear Skies legislation increases_c pollution. Unfortunately, it isn’t obvious that either of these claims holds. Here are a couple of cases to test intuitions.

Merlin
We have here a pile of rocks. The pile is a little unstable, and some rocks are rolling down it, and if left untouched they will soon roll off the pile. Merlin casts a spell that halts one of the rocks, and moves it to a stable point. While this spell is being cast, some other rocks roll off the pile. Did Merlin increase the size of the rock pile? My intuition is that he did not, even though he did
increase_c it.

RHAWP
The Red-Haired Australian Welfare Plan (RHAWP) calls for all red-haired Australians to be given a one-time payment of $100,000. Let’s imagine (contrary to fact) that we are in a scenario where the deficit is falling fast enough that even with the RHAWP in place, it will still fall next year. In that circumstance, would the RHAWP increase the deficit? It would increase_c the deficit, but not increase_t it. In this circumstance I’m a little more conflicted – I’m certainly more inclined to say that the RHAWP will increase the deficit than that Merlin increased the size of the rockpile. That’s good news for Lieberman, since arguably the Clear Skies legislation will have the same effect on pollution as the RHAWP has on the deficit.

Here’s my best guess as to what is going on. ‘Increase’ really does mean increase_t, and Juan is right than what Lieberman said is false. But we are happy to use it of people who exacerbate, that is increase_c, something we perceive to be a problem. To test this prediction, imagine what would happen if we thought the rock pile was getting in the way of something we want to do. For example, assume it covers up a mine shaft down which a young child is trapped. I think in that case I’d be more prepared to say that Merlin increased the rock pile. And I think (very tentatively) the best analysis of the case is that it’s a case where one can appropriately say something not true because it’s a useful way of communicating something that is true.

One final question. Juan says that

A “lie” is a deliberately false statement, typically made with an intent to deceive. Not just any false statement, or bit of spin, will do. Intentional deception is key.

Hmmm, is intent ‘typical’ or ‘key’? Anyway, we might wonder whether the following situation constitutes lying. X utters S (deliberately), S means that p, X knows p is false, but X thinks S means q, and X believes q is true. I’d say that’s not a lie. And in Lieberman’s defence, that might be what happened here. Before thinking through the cases, I thought ‘increase’ might mean increase_c. I now think that’s wrong, but I think it’s a mistake a competent speaker could make. (I don’t think I was incompetent before I thought about the Merlin case.) Juan makes a point of not accusing Lieberman of lying, and I think that was correct given these considerations.