Stative passive: A window on some smaller syntactic structures
An idea that has been investigated in some syntactic approaches to morphology is that certain elements can enter the syntax by being "directly attached" to other heads, with the output of this operation being itself a head, not a phrase. Recent applications of this idea are found in discussions of (certain) compounds, and of different types of derivational morphology (e.g. Embick 2016, Wood 2018). Stative passives (a.k.a. "adjectival passives") like opened in The door is opened, or hammered in The metal is hammered provide an interesting testing ground for syntactic approaches to word formation. On the one hand, they have certain properties that might make this "direct attachment" or "small" analysis look appealing. On the other hand, they appear to scope over resultative secondary predicates and other vP internal material (The metal is hammered flat); this and other facts have led different researchers (Kratzer 2001, Embick 2004, and others in that vein) to propose that stative passives are (sometimes/always) phrasal.
A reconsideration of some of the facts about stative passives, along with a new look at how un-prefixation works, suggests that there are serious problems with a phrasal analysis. I show how these difficulties can be overcome with "direct attachment" analyses of both the head realized as un- and certain stative passives, and point to the broader possibility that all stative passives in English might be "small". The general conclusions provide further reasons to think that Roots are introduced into the syntax in ways that are different from superficially similar full phrases, providing evidence for the general outlook on this question advanced in Marantz (2009).