When is a Verb not a Verb?

Carson Schütze

Full Text:


DOI: http://dx.doi.org/10.7557/12.11


Auxiliary; Modal; <em>Do</em>-support; <em>Be</em>; Tense; Mood; Participle


Changes are proposed to the categorial status traditionally accorded to Aux-related and verbal elements in the clause, and the new taxonomy is applied in implementing the old insight that be should be analyzed as the default, semantically empty verb. The central issue is when a verb-like element does (not) count as categorially a V for distributional purposes. The major proposals are: 1) to remove be and have from the category Aux and treat them as Vs; 2) to separate out participles from genuine tensed and bare verbs; 3) to group do with modals, rather than with have and be, into a category Mood that also includes a null indicative morpheme.
This scheme is used to account for the entire distribution of the forms of be just by treating it as V with no properties. Be fulfills two requirements that cannot always be met by contentful verbs: first, it satisfies the syntactico-semantic need for Tense to c-command a clause-mate V (the “V Requirement”); second, it satisfies the morphosyntactic need for participial affixes ( -ing, -en) to have hosts. It is shown how the former requirement derives the exceptionally high position of finite be by base-generating it above negation etc., rather than raising it across. VP-ellipsis data provide independent support for this treatment. Finally, some tentative suggestions are offered for how the V Requirement might be derived from deeper principles, while still allowing for the fact that it is apparently not fully enforced in languages with null copulas.

Copyright (c) 2003 Carson Schütze