Verb second as a function of Merge

I hypothesize that Merge as defined in (1) automatically creates a dependency relation S (for sisterhood) where x is invariably the antecedent (or nondependent) and y the dependent. I suggest furthermore that S can be (and perhaps universally is) marked on y, and spelled out on one of the terms of y. The proposal of this paper is that this dependency marking may be realized in two ways: by inflectional morphology (tense, agreement marking) or by position. In particular, the proposal is that V2 is positional marking of the relation between a fronted element and its sister, to the effect that the term of y spelling out S is realized as the leftmost element in y. It can be seen that V2, on this proposal, is really a verb-first requirement applied to the domain of the dependent in a dependency relation. This eliminates a property that has always been commented on as strange, namely reference to an arbitrary number (two) in the description of the pattern. It raises the question, though, why no V-last counterpart to V2 seems to exist (i.e. no pattern that requires fronting of x, creating , to be accompanied by realization of the verb on the right edge of the dependent y). I suggest that the verb in V2-constructions is a member of a larger class of elements described as ‘linkers’, appearing more generally in constructions of predication or modification. We define linker as in (2):