Re: RT list: the projection problem for presuppositions

Dear Hanno and all,

Hanno asked:
> the "projection problem for presuppositions." ... have Relevance Theory researchers weighed in on this topic?

One D. Wilson's PhD thesis was on the subject: Wilson, D. (1975). Presuppositions and Non-Truth-Conditional Semantics. New York: Academic Press.
See also Wilson, D. (1975). Presupposition, assertion, and lexical items. Linguistic Inquiry, 95-114.

The main thesis (if I have understood correctly: this was before my time) was that 'presuppositions' are not a matter for the grammar, and that 'presuppositional' phenomena are best understood pragmatically.

Ruth Kempson (who was close at that time to the work that became relevance theory) also gave arguments for this conclusion. See:
Kempson, R. M. (1975). Presupposition and the Delimitation of Semantics. Cambridge: Cambridge University Press.

Around this time presuppositional theories became much more sophisticated (e.g. Stalnaker, Gazdar and a little later, Heim) partly in response to objections that the accounts of the early 70s were empirically inadequate in their predictions about 'projection'. Most of the huge literature that Hanno refers to has developed along lines set down at that time -- with increasingly complex formal theories to save the phenomena (e.g. van der Sandt, Beaver).

The idea that presuppositions are pragmatic in the sense that, strictly, they are properties of utterances rather than sentences is widely accepted (mainly following Stalnaker). But the more radically pragmatic approach that suggests that it is a mistake to write properties of a speech act, assertion, into the lexical entries of certain words or constructions ('presupposition triggers' in the now-standard terminology) has mostly been forgotten or ignored. (Mandy Simons' work is an exception, I think.)

However, Wilson and Kempson's arguments, where they are applicable to the more recent theories, still look compelling (to me at least, and perhaps to other relevance theorists, hence, perhaps, the relative silence -- but not total silence: note Carston's work in the 1990s on 'presupposition' and metalinguistic negation).
In fact some recent work seems to succumb rather directly since it gives up a left-to-right asymmetry in projectional behaviour.

Phenomena analysed as presuppositional can be dealt with in relevance theory by a divide-and-conquer strategy. a) Many 'presuppositions' are simply entailments/components of the proposition expressed. b) Some may be implicated premises. c) For lexical items like 'too' and 'even', and for pronouns, in relevance theory it is natural to look for accounts in terms of procedural meaning. d) Some alleged presuppositions may not, in typical uses, be communicated at all -- e.g. the alleged existence presupposition of proper names -- and thus fall outside of pragmatics as well as semantics.

Having said all that, there seems to be some common ground with the move in the last few years, particularly by Craige Roberts, David Beaver, Judith Tonhauser and Mandy Simons, to find a taxonomy of the area. They argue (again, if I have correctly understood) that there seem to be several categories of item involved, with no one feature shared between all of them, and that it is time to try to understand the data that have been thrown up by the last several decades of work in a more sophisticated framework. They might be happy with the idea that many 'presuppositions' are entailed, but would like to know (among other things) what makes that entailment (typically) backgrounded.

One remaining point of disagreement might come over the status of this backgrounding. Is it lexically specified? This might come down to familiar arguments over the status of commonly-made pragmatic inferences: are they defaults in some meaningful sense?

Of course a thoroughly pragmatic theory (like other theories of the phenomena) ought to be able to explain 'projection' in the clear cases. A caveat: I write 'in the clear cases', because it seems to me that, for quite a lot of the data which has been evolved to distinguish between existing theories, intuitions are unclear, and probably rather context sensitive.

Given that several people whose work I have referred to are subscribers to this list, I hope to be corrected where this account errs.

Best,
Nick

Nicholas Allott
Postdoctoral research fellow
CSMN
University of Oslo

n.e.allott@csmn.uio.no
nicholas.allott@gmail.com
Received on Mon Feb 28 14:43:36 2011