The Psychological Foundations of the Grammar 79
notably the failure to account for more than one accent per utterance.^30
S&W’s lack of treatment of prosody confl icts with the idea that hearers form
anticipatory hypotheses. S&W’s claim that only the fi nal accented syllable
serves to signal logical entailments means that the hearer has to wait until
the utterance is complete before being able to make a judgement as to which
entailment is relevant. Levinson (2000: 5) notes that the preponderance
of evidence available from the psycholinguistic literature indicates that
‘hypotheses about meaning are entertained incrementally – as the words
come in, as it were’, and concludes that S&W’s concept of presuppositions
is not psycholinguistically plausible.
A further problem with S&W’s theory is that as they do not study corpus
data, many of their example sentences appear too long to be normally
spoken in one tone unit. An example with nuclear accent on France is ‘The
exhibition was visited by the king of FRANCE’ (ibid. 214). In discourse a speaker
has the option to utter it as two tone units, e.g.
(39) // the exhiBItion was VISited // by the KING of FRANCE //and so it is not at all clear what presupposition a hearer should or could
infer. Bolinger (1989: 357) argues for the removal of what he labels ‘the
dead hand of transformational prosody’ and the recognition that the key
factor guiding S&W’s Principle of Relevance is speakers’ assumptions of
the state of convergence existing between themselves and their hearers.
Inferences and entailments, according to Bollinger, are generated by the
occurrence of words in context. Hearers are not required to wait until the
fi nal accent to form logical entailments. Pre-nuclear accents help hearers
form anticipatory incremental hypotheses which enable them to under-
stand speakers’ messages. If one follows Bolinger and removes ‘the dead
hand of transformational prosody’, S&W’s principle of relevance is a
valuable and incisive method of explicating discoursal meaning.
A fi nal problem with S&W’s theory was alluded to in Section 2 of this
chapter. They do not overtly take into account the interactive tension inher-
ent between the language and social systems which both shapes the social
systems as well as concomitantly forcing the speaker to adopt a register
appropriate to the discourse setting, the speaker’s communicative goals;
and the relative statuses of the interlocutors.^31 However, S&W’s theory states
that a communicator’s cognitive environment includes all the facts that
are manifest to him/her, plus any resulting inferences arising from these
facts (1995: 38–9). Facts that are manifest to individuals include those
which arise from their perceptual abilities, their previous experiences and