Abstract
In this paper I shall argue that not is neither a head nor a specifier of any maximal projection, but an adverbial modifier adjoining to I directly. The paper pursues this thesis from several pointsof view (VP ellipsis, HMC, I-to-C movement, and the strict adjacency of not to do/to), before briefly examining the historical changes of not. The examined data include negative subjunctive that-clauses (Section 2)and some archaic negative sentences (Section 3). By comparing various analyses of not and looking into the history of English negation, this paper defends the adverbial status of not in I against the dominant NegP hypothesis with not in its head. With a certain strategy, this status of adjoining not does not conflict with Potsdam's (1997) argument for VP ellipsis. Furthermore, Jespersen's (1917)historical cycle of negation can be naturally captured in the NegP-less clausal structure.
