• On Wh-Movement from Subject Position

      Tsujimura, Natsuko (University of Arizona Linguistics Circle, 1983)
      Among the current approaches to universal grammar, those within the framework of the Extended Standard Theory have been considered as promising by a number of transformationalists. In this theory, the transformational component "Move " maps D- structures onto S- structures leaving behind traces which are coindexed with the roved elements. "Move a" represents both "Move NP" and "Move wh-phrase ". In this framework, a large number of phenomena involving WH-Movement and traces have been accounted for. One interesting case is the to contraction phenomenon in English. A familiar example of the phenomenon can be seen in the contraction of "want + to → wanna", and the following are, according to Chomsky and Lasnik (1977), the cases which involve traces (indicated as t) left behind by WH- Movement (i.e., (3) , (4)) : (1) I want to meet John. (2) I want Mary to meet John. (3) Who do you want to meet t? (4) Who do you want t to meet John? (5) Who do you wanna meet t? (6) *Who do you wanna meet John? where sentences of the type in (3) and (4) are derived from sentences like (1) and (2), respectively, (ignoring the difference in subjects) by WH- Movement. The contraction in question is observed in (5) and (6), which are the contracted versions of (3) and (4), respectively. WH- Movement leaves traces in (3) and (4) as shown in (7) and (8): (7) [who [do you want [to meet t] ] ] (8) [who [do you want [t to meet John]]] Chomsky and Lasnik claim that the trace left by WH- Movemrent in (8) intervenes between want and to, which makes the contraction impossible. On the other hand, the trace in (7) does not cone between want and to, hence, contraction may take place (cf. (5)). 142 Jaeggli (1980) refines this account for to contraction phenomena by distinguishing between two kinds of traces, i.e., Case-marked versus non-Case-marked trace. Only Case-marked traces, i.e., traces of WH- Movement, prevent contraction, whereas traces which are not Case- marked allow the contraction. The crucial cases of the latter involve contraction in semiauxiliaries, where the trace is left by Raising.' At this point, we should ask the question: can contraction phenomena be generalizable within the framework of trace theory? In other words, can trace theory be extended to account for contraction phenomena in general? If we assume that contraction phenomena cannot be generalized, the to constraction phenomenon would be an idiosyncratic feature, and could be treated in the lexical domain by considering the contracted forms as independent entries in the lexicon. On the other hand, if we assume that the phenomena can be generalized, trace theory should account for other contractions such an auxiliary contraction as well as the to contraction phenomenon. Although the choice between the two assumptions seems to be theory -dependent, we will take the latter assumption in this paper; that is, contraction phenomena can be generalized. Given this assumption, we encounter a problem with auxiliary contraction phenomena. Consider the following: (9) a. Who t has seen John? b. Who's seen John? where t in (9a) is the trace of WH-Movement, and (9b) is the contracted version of (9a). While trace theory accounts for the contraction phenomenon in (5)-(6) on the one hand, it would wrongly predict that the trace in (9a), which is left behind by WH-Movement, prevents the contraction between who and has on the other hand, since the trace in (9a) is Case - marked, and is supposed to prevent the contraction. Provided that we continue assuming trace theory can account for contraction phenomena in general, we might posit the contraction revealed in (9) is attributed to other assumptions underlying the derivation of the sentences of (9). One such assumption I would like to examine in this paper has to do with WH-Movement from subject position. Thus far, when we assume "Move wh-phrase" in core grammar for English, we have also assumed movement from subject position as in (9) as well as from object position (e.g. (3)). On the other hand, if we suppose that WH- Movement does not apply to subject position in a root sentence, preserving the generalization of contraction phenomena within the framework of trace theory, the problem with regard to the auxiliary contraction as seen in (9) does not arise, since no trace intervenes between the two elements to be contracted. Besides the problem stated above, WH- Movement from this position raises other problems, which I will discuss in Section 3. Given the above outline of the discussion, in Section 2, I would like to propose the hypothesis that WH- Movement does not apply to subject position in a root sentence (call it the No WH-Movement Hypothesis). The rest of the discussion consists of the following two sections: in Section 1, a set of counterexamples are presented to the trace theory account for contraction phenomena, assuming the phenomena are generalizable, and some attempts are made to solve the problem with respect to the auxiliary contraction as in (9), under the assumption that WH-Movement takes place in subject position; in Section 3, I will formulate a constraint on "Move wh-phrase" in order to prevent WH-Movement from subject position. Throughout this paper, I mean by subject position, the position in a root sentence.