上外研究生学术训练营(语言学)
The status of tense in non-finite clauses in English is controversial; the issues in dispute are partly terminological (people disagree on what counts as a “tense”) and partly empirical (non-finite clauses differ in some respects from finite clauses vis à vis temporal interpretation but resemble finite clauses in other respects.) In Stowell (1981, 1982), I suggested that a subset of infinitival clauses (control complements and for-infinitives) contain a future-shifting “unrealized” tense. Hackl and Nissenbaum (2012) argue that the putative future-shifting tense is, at least in infinitival relative clauses, really a necessity or possibility modal. Wurmbrand (2014) adopts a similar position, arguing that infinitives may contain a counterpart to the future-shifting modal woll (Abusch 1997) but not true tense. Wurmbrand also discusses cases of infinitival complements of attitude verbs with simultaneous interpretations and argues against the presence of present tense on the basis of the fact that finite sequence-of-tense effects can operate across them.
Setting aside Wurmbrand’s arguments (which, I claim, rest on a flawed conception of ‘present tense,’) I argue for the existence of true past tense in non-finite clauses—both in infinitives and in “bare VP” complements of modals, in the form of the non-finite perfect. The argument is based partly on observations of Hofmann (1966), and partly on independent observations concerning non-finite analogs of sequence-of-tense with the non-finite perfect, as well as what appear to be counterfactual ‘fake past’ interpretations.
The debate about non-finite tense in English is in some respects parallel to the debate about whether tense exists in languages like Chinese and St’át’imcets, a Salishan language of British Columbia. A widely held traditional view is that such languages lack true tenses in the traditional sense, and that they rely on a combination of verbal aspect, time-denoting adverbs, and/or pragmatic reasoning based on contextual information to draw inferences about the temporal location of events and situations. More recently, Matthewson (2002, 2006, 2017), Sybesma (2007), Sun (2014), and Demirdache and Sun (2017) have challenged this view, arguing for the existence of a phonetically null tense in St’át’imcets and Chinese. Demirdache and Sun’s arguments are particularly compelling, since they show that the full array of constraints on tense construal operative in English finite clauses have direct counterparts in Chinese.
I will draw comparisons between these two debates over the status of tense in non-finite clauses in English and ‘tenseless’ languages like Chinese), and argue in favor of the null tense approach.