If one plus one is one
I am about to do something unfair. But don’t worry, names will be changed to protect the guilty. The following is a passage from a recent dissertation from a well-known linguistics department. I’m sure if you do a little work, you’ll be able to figure out whose it is. So just don’t do the work, okay!
In order to get the case where the type-3 adverb follows the article, something else must happen. I will assume the existence of a position higher than “MaxP” to which adverbs may raise. Note that, at least for type-3 adverbs, it must be a position to which they raise, since they are merged close enough to VP to select V0. I label this position, “ArbP”, (following Smith 2002) a so-called Q-position where adverbs can occur in several Indo-European languages. When MaxP raises to SpecOP, the adverb is stranded in ArbP and therefore follows the determiner: [figure omitted] Thus, the distribution of adverbs provides strong independent evidence that MaxP raising in the derivation of clefts in this particular language. If this were not so, it would be quite difficult to explain how type-3 adverbs, for instance, end up following the definite determiner. If there is a landing spot in the vicinity of the Q-position, namely ArbP, to which adverbs can raise before MaxP fronting, the distribution of adverbs falls out without stipulation.
You see, all you have to do is “assume the existence of a position higher than MaxP to which the adverbs may raise,” and then “the distribution of adverbs falls out without stipulation”! It’s a miracle.
Like I said, it’s not really fair. After all, I’m sure the linguistic argumentation is entirely internally consistent.
[This was inspired by this entry to marginaling.]