At least since the 80s linguists have debated whether simple phrases like the cake are NPs with determiners in the specifier position or DPs with NP complements. Substantivized adjectives seem to me to provide a simple but compelling argument for the latter analysis:
- The meek and the happy must be the same type of phrase as the cake, since they are distributionally equivalent to the latter. The latter is either an NP or a DP, so the meek and the happy are also either NPs or DPs.
- It makes little sense to analyze the meek and the happy as NPs, because that would require analyzing meek and happy as noun heads. But meek and happy are adjectives rather than nouns, even when preceded by the article: note that they still can take, for example, adverbial adjuncts (the very meek, the blissfully happy). It’s far simpler to consider the (very) meek and the (blissfully) happy to be DPs of which (very) meek and (blissfully) happy are AP complements.
- Because the cake is the same type of phrase as the meek and the happy, it must also be a DP.
This seems to me both simpler and more persuasive than the arguments for the DP analysis which I’ve read in books and articles. I won’t call it airtight, since I can see ways around it—but they require adding more complexity to one’s analysis than I consider justified.
But am I mistaken? Is there an error in the argument I’ve made, or a strong objection to it? It seems to me so simple that I’m surprised I couldn’t find it anywhere. Perhaps there are e.g. problems I don’t recognize with analyzing e.g. the meek as a DP with an AP complement.