> perverse semantics
Hmm. I'm curious as to the value of attaching such an archaic, subjective values-oriented term as "perverse" to the label here. What does "perverse" mean in this context, and is it really expected to sway opinions with what appears on its face to be rather emotional language?
> But natural language utterances actually do mean things despite their inherent ambiguity
This statement seems to work in anyone's favor, or no one's, where the argument is concerned...
> Smoking a lot produces a statistically significant increase in your risk of developing cancer
With an error rate of...zero? With a bell curve of...? With a qualitatively superior study based on the population of...? Let us have the study. Let us tear it apart. What do we have to lose by admitting that smoking may not cause cancer in this way or that one?
> They do not mean, "Smoking one cigarette is guaranteed to give you cancer"
Your assumption, but not mine. For all I know, smoking one cigarette may do the job.
> Picking at semantic nits is counterproductive.
Why? Details are important.
> There's probably even a name for it as a logical fallacy
Even logical fallacies are not fallacies when they are relevant to the discussion at hand. This is Logical Fallacy 101--see the "Fallacy Fallacy".
> I'm too lazy and frustrated at the general state of discourse in the world to bother looking it up right now.
This is discourse, right here. Bringing frustration into it is fine, but there's no guarantee it won't affect the quality of the discourse...