Researcher: Marijke De Belder
> Download the project description (pdf)
It seems obvious why the sentence the cat played with the mouse is considerably better than the sentence the play catted with the mouse. Cat is a noun and play is a verb and we should use them as such. Using play as a noun and cat as a verb simply does not seem to work. Put differently, words seem to belong to a specific category and this category determines how the word is used. This view is commonly held in linguistics because it is straightforward: nouns are used as nouns, verbs are used as verbs, etc. More generally, the category of a word matches its use in the sentence. I will refer to it as the matching principle.
One expects that if a principle is commonly held in a field that counter-examples are hard to find. This is not the case for the matching principle, however. Consider the following example. The words Britney and Timberlake are proper names. According to the matching principle, we expect that one can only use them as such, as in the examples below.
(1) Britney is pregnant.
(2) Timberlake has got a new song.
In the examples above Britney and Timberlake are used as proper names. However, they can be used as adjectives too:
(3) That’s so Britney!
(4) That’s so Timberlake!
These data are not accounted for by the matching principle. To include such examples in the theory, I will propose an alternative to the matching principle. I hypothesize that words mainly get their category from the syntactic structure.