Cross-sentential reference assignment and the grammar/inference distinction

​

Roberto B. Sileo

University of Cambridge

​

Abstract

In an attempt to show that utterance interpretation is not as pragmatics-driven as originally envisaged by Grice (1989), Lepore and Stone (2015) purport that the language system itself contains a specific rule for cross-sentential (anaphoric) reference assignment. In their view, the referent of a subject of an English sentence is preferably resolved by the immediately preceding subject. They argue, in this respect, that in the absence of contextual assumptions to the contrary “He” in (1) and (2) refers, respectively, to “the president” and to “Jones”:

 

(1)          The president nominated Jones. He expected a quick confirmation.

(2)          Jones was nominated by the President. He expected a quick confirmation.

(Lepore and Stone 2015:91)

 

Lepore and Stone’s observations appear to depart from van der Sandt’s (1992, 2012) now long-standing idea that cross-sentential pragmatic binding normally takes place at the nearest accessible position. According to van der Sandt, “he”, for example, in (3), would preferably refer to “Harry’s child” even though the presuppositional inference disappears:

 

(3)          Harry does not have a child. So he cannot be on holiday.

                (van der Sandt 1992:335)

 

In this paper, I assess the debate and I support a post-Gricean rather than a non-Gricean explanation. My approach to cross-sentential reference assignment stands, first, on the all-pervasive role of intentions in natural language interpretation and, second, on the theoretical claim that semantic and pragmatic sources of information, on the one hand, and automatic and inferential processing mechanisms, on the other, can all be taken to be involved in anaphoric reference resolution. I propose that cross-sentential reference assignment can be comprehensively and adequately accounted for by Jaszczolt’s (2010, 2016a, 2016b, 2017) Default Semantics framework. I demonstrate, in a nutshell, that (i) reference assignment relies on both linguistic and non-linguistic sources of information (lexicon, situations of discourse, world knowledge, the human inferential system, and/or social and cultural assumptions), (ii) that reference is resolved via distinct processing mechanisms in different contexts, and (iii) that reference is not necessarily assigned by following grammatical constraints. While the language system itself is, indeed, a valid object of philosophical and/or linguistic analysis, there are no sound reasons to force pragmatic effects into our grammar.

​

Reference

Grice, P. (1989). Studies in the Way of Words.  Cambridge, MA: Harvard University Press.

Jaszczolt, K. M. (2010).  Default Semantics.  In Heine, B. and Narrog, H. (eds.)  The Oxford Handbook of Linguistic Analysis.  Oxford: Oxford University Press, 193-221.

Jaszczolt, K.M. (2016a). Meaning in Linguistic Interaction: Semantics, Metasemantics, Philosophy of Language.  Oxford: Oxford University Press.

Jaszczolt, K.M. (2016b). On Unimaginative and Conventional Conventions: Response to Lepore and Stone. Polish Journal of Philosophy X (1), 89-98.

Jaszczolt, K. M. (2017). Pragmatics and Philosophy: In Search of a Paradigm. In press. Intercultural Pragmatics. Special Issue ‘Pragmatics and Philosophy’.

Lepore, E. and Stone, M. (2015). Imagination and Convention: Distinguishing Grammar & Inference in Language.  Oxford: Oxford University Press.

Sandt, R.A. van der (1992).  Presupposition projection as anaphora resolution.  Journal of Semantics 9, 333-377.

Sandt, R.A. van der (2012).  Presupposition and accommodation discourse.  In Allan, K. and Jaszczolt, K.M. (eds.).  The Cambridge Handbook of Pragmatics.  Cambridge: Cambridge University Press, 329-350.