Gerund Zero

My initial reaction to this Language Log post about gerunds/participles was that it was imagining an ambiguity where none actually exists. I can't remember ever having been puzzled about whether a particular -ing construction was a gerund or a participle; it's generally clear, and (as some of the commenters said) it seems useful to have different names for the noun-functions and the adjective-functions of a given -ing construction.

Upon reflection, however, I've come around a little to MYL's point of view -- gerunds are not, as a rule, echt nouns. This becomes clear when you try to modify them: consider, e.g.,

(a) "Drinking continuously is a good idea" vs. (b) "Continuous drinking..."

These are both essentially idiomatic constructions to my ear; however, if drinking really were a noun, (a) should sound ridiculous. In fact there are probably instances, like "Going on endlessly about grammar will lose you friends," where you really need the adverbial form, though on a traditional parse "going" is a noun qua gerund. The fact that adverbs can modify gerunds appears quite general to constructions with gerunds in them; and -- to my mind -- offers very strong evidence that gerunds are not to be treated as true nouns. (Obviously the entire phrase is still functioning as a noun; the ambiguity is about the order of operations, as one can noun the verb either before or after adding the qualifier, and depending on what one does first the qualifier is either an adjective or an adverb.)

Some constructions w/ gerunds are of course a lot more noun-like than others -- definitely a pluralized gerund is an echt noun. But they do seem to occupy a bizarre syntactic middle ground, along with infinitives and other similar beasts.