Why does it have to be do?

Grimshaw (1997) proposed that a dummy is the use of a meaningful lexical item in a meaningless context, where it descriptive content is ignored. She further claims that this can account for why certain lexical items, e.g. do are used as dummies on the assumption that ignoring content comes at a cost and the more content that is ignored the greater the cost. Therefore dummies will tend towards those lexical items with smaller lexical content. While this goes part of the way to accounting for dummies, given that there are a number of verbs whose content could be said to be 'small', it still remains to explain why it has to be do. In this paper, we attempt to build a framework in which this question can be meaningfully addressed. Following Newson et al. (2006), we argue that do is in fact not the only dummy auxiliary, but so too are have and be. The question is therefore extended to why these are all dummies and what determines which is used in what context. The framework we choose to answer this question in is a Syntax First Alignment System (Newson 2010) which assumes late vocabulary insertion on the basis of the Superset Principle. This is well suited to accounting for Grimshaw's approach to dummies as, under these assumption, a dummy is the realisation of purely functional features by a vocabulary item whose descriptive content is treated as overspecification. We argue that have be and do are all minimal descriptive content verbs which differ from each other minimally in their association with two features: one which distinguishes verbal from non-verbal predicates and the other which distinguishes thematic from non-thematic verbs. The alignment system accounts for the distribution of these features, which, though normally realised by the verbal root, may be displaced under certain conditions so that they have to be spelled out with the functional features (tense, aspect, etc.). This is what determines which auxiliary verb is to be used as a dummy.

Grimshaw, Jane. 1997. "Projection, heads, and optimality." Linguistic Inquiry 28. 373-422.

Newson, Mark. 2010. "Syntax First, Words After: a possible consequence of doing Alignment Syntax without a lexicon." In The Even Yearbook 10, edited by László Varga. Budapest: Eötvös Loránd University, 1-46. http://seas3.elte.hu/delg/publications/even/2010.html .

Newson, Mark, Marianna Hordós, Dániel Papp, Krisztina Szécsényi, Gabriella Tóth, and Veronika Vincze. 2006. Basic English Syntax with Exercises. Budapest: Bölcsész Konzorcium.