Research in theoretical linguistics is often driven by a goal of a reduction of theoretical postulates and the simplification of the overall system. Apparently disparate accounts of disparate phenomena may be shown to all full under one meta-principle, or the effects of some mechanism may be shown to follow from other independent components of the grammar. Specifically, I will only be concerned here with the part of X`-theory that distinguishes words from phrases-considering in what ways the syntax needs to make reference to the notions of minimal and maximal projection. We will see that non-saturated elements can be selected as X^0s, and that saturated elements may have special distribution of X^0, not XP. Adverbs, particles and pronouns lack valence or selectional properties, yet need not, and in some cases must not, behave as phrases. Consequently, the attempt to reduce phrasality to saturatedness seems misguided.

• Peter Sells