Should vs. ought to

The English modals should and ought to "seem to be largely interchangeable" (Palmer 1990;
cf. similar sentiments in Coates 1983: 69; Quirk et al. 1985: 227; Huddleston and Pullum
2002: 186). Some proposals for possible distinctions have nonetheless been made, based inter
alia on (i) subjectivity vs. objectivity (Swan 1980: 550; Gailor 1983: 348; Aarts and Wekker
1987: 193; Myhill 1996; Collins 2009: 54) (ii) absence vs. presence of an implication of nonfulfillment
(Close 1981: 121; Gailor 1983: 348-349; Westney 1996: 170) and (iii) relative
frequency vs. infrequency of epistemic reading (Coates 1983; Palmer 1987: 134; Collins
2009).
To test whether these and a few other claims made in the literature could be verified, we
conducted a corpus-based study involving over 1000 sentences with should or ought to in
contemporary spoken and written British English and coded them for 28 parameters. Raw
frequency and frequency per discourse mode could not be coded, but the data confirmed
previous observations that should is used much more frequently than ought to and that the
frequency of ought to in spoken language is higher than its frequency in written language. To
find out which of the parameters investigated contribute significantly to the distribution of
should and ought to, we fitted a logistic regression model, making sure any corruption of the
model caused by multicollinearity was excluded.
Eight parameters were shown to exert a unique and significant impact on the choice of
should vs. ought to, in order of decreasing strength: (i) inversion, (ii) a following contracted
perfect infinitive, (iii) no adverb or a following rather than preceding adverb, (iv) verb-marked
negation, (v) embedding by suggest or a similar item, (vi) reference to the non-past,
(vii) no embedding by think or a similar cognition expression and (viii) third (vs. first or
second) person subject. All other factors turned out not to be significant or were left out of the
model because they correlated with one or more parameters in the model.
Our model thus provides a parsimonious and accurate description of the should/ought to
variation. It also guides us towards a possible explanation, which we believe can be found in
the different grammatical status of should and ought to, the former being a true modal
auxiliary, the latter being a semi-auxiliary (a blend between a modal and a lexical verb). We
argue that this difference can account for the four strongest distinctions in usage we revealed.

ISBN:978-4-7589-2154-1

Publication year:2010

Keywords:deontic modality, epistemic modality, should, ought to, multivariate analysis, subjectivity