Optional OV order in Later Middle English (LME) has given rise to conflicti
ng theoretical accounts. Earlier analyses postulating movement to AgrOP or
alternative base orders are found to be inadequate to deal with the occurre
nce of OV in nonliterary LME; in a large database of 15(th) century private
familial correspondence, residual OV order is found to have been productiv
e only with negated objects. Multiple subject constructions with there expl
etives showed the same restriction. These phenomena are accounted for by po
stulating overt Neg Movement (Haegeman 1995) as a permitted option in LME.
In this framework, it is argued that LME showed a mixed typology having bot
h Neg movement and a null Neg operator. LME had three ways of satisfying th
e NEG Criterion (Haegeman 1995): Merge not in Spec NegP, coindex [OP](i)...
[XP(Neg)](i), and Move XP(Neg) to Spec NegP. Modern English has only the fi
rst two. The distribution in this period of negative concord with not is sh
own to support our analysis.