The have yet to Construction: A Microcomparative Account

The HYT construction poses several puzzles for syntax and the syntax-semantics interface. First, it has a negative meaning—(1) means that John has not visited his grandmother yet—and it contains a Negative Polarity Item (NPI) (yet), but there is no overt sentential negation. Second, it has the semantics of perfect aspect, but no visible perfect participle. What is possibly most interesting about the HYT construction, however, is that closer investigation reveals numerous dimensions of microvariation in its underlying syntax. For one thing, speakers seem to differ as to whether they actually treat the HYT construction as syntactically negative or affirmative—in many cases, different negativity tests give different results. For another, speakers seem to differ as to whether have must be an auxiliary or whether it can behave as a main verb. And all of this variation exists despite the construction having basically the same meaning across speakers, no matter what the underlying syntax is for those speakers. The goal of this study is to sort out the microsyntax of HYT across speakers, in the face of contradictory empirical claims and mutually incompatible proposals in the existing literature. We develop an analysis of the variation and support that analysis with data from two acceptability judgment surveys of 361 and 520 speakers each. In this paper, we explore in depth one dimension of variation: the status of have (as an auxiliary or a main verb). We find that the majority of speakers who can treat have as a main verb can also treat it as an auxiliary verb. The reverse does not hold: many speakers can treat have as an auxiliary verb, but not as a main verb. This result leads us to conclude that the difference between the main verb and auxiliary verb analysis is a relatively minor one, and we propose that it has to do with where in the structure an interpretable perfect feature [iT:PERF] is introduced. Nearly all speakers allow it in its canonical position, leading to the acceptability of the auxiliary use. Some speakers, in addition, allow it to be introduced much lower, leading to the acceptability of the main verb use. The paper is structured as follows. In section 2, we provide a discussion of previous work on HYT, and along the way provide a preview of our analysis by discussing what aspects of each proposal we do and do not adopt in our own. In section 3, we present a brief overview of our own proposal, and in section 4 we present empirical data to support it. Section 5 provides novel evidence regarding the position of yet. 6 briefly discusses a second major area of variation—the status of negation—and section 7 concludes.