Unravelling intentions: Speaker attitudes in new quotatives

Stef Spronck

Amsterdam, The Netherlands

WS146: New Perspectives on New (and old) Quotatives

Reports take many different forms in the languages of the world, cf. (1)-(3)

(1) [John said [that it was going to rain]]

(2) [palle na?~ tèl?,] [kard?e mel? mel?]

tiedendof_sari not ½-paisa do.3f.PRES fair-REDUP

‘She hasn't a paisa in her pocket but she says, “I want to go to the fair.”’ (Bashir, 2005)

(3) [Tuumasi- n- nguuq qilalugaq pisar-aa]

T. REL- QUOT beluga catch-3sg/3sg.INDIC

‘Tuumasi caught a beluga (they say)’ (Fortescue, 2003: 295)

The English (1) is traditionally analysed as a two clause sentence, while for the Panjabi (2) such an analysis is less apparent and in the West-Greenlandic (3) clearly false. Semantically, however, (1)-(3) are comparable in that they convey reported messages and cross-linguistically grammaticalised quotatives are not inherently different from periphrastic structures (pace Feuillet, 1997; McGregor, 1994; Spronck 2006b, fc.).

In order to surpass superficial structural differences, and thus bare the relevant cross-linguistic semantic and pragmatic differences between reporting utterances, I present a functionalist construction grammar account of quotatives. In construction grammar approaches to syntax syntactic constructions are analysed as conventional pairings of meaning and form, irrespective of the morphology in which a particular meaning is encoded in a language L (cf. Croft, 2001; Goldberg, 1995; 2006, Östman & Fried, 2005). I propose the generalised semantic structure in (4) as a description of the basic meaning reporting utterances in the languages of the world combine.

(14) [SOURCE (speaker hearer)

MESSAGE (speaker attitude-message)]evidential

In words: reporting utterances consist of an (implied) source and a reported message, jointly functioning as an evidential construction (cf. Aikhenvald, 2004). The italic variables in (4) may be merely inferable in a language L. The variable ‘speaker attitude’ may be explicitly encoded in e.g Lele, cf (5) or implied in the Tariana (6a) where the speaker ‘does not vouch for the information reported’, unlike in (6b) (Aikhenvald, 2004: 139).

(5) cànìgé ná-y no go lele sá?


‘Canige said that he is a Lele man, but I have my doubts’ (Frajzyngier, 2002: 174)

(6a) 'o-he: gi-ba na-tú


‘“I’m sick”, he said’

(6b) 'o-he: gi na-tú


‘He said that he is sick’ (Aikhenvald, 2004: 139)

In Spronck (2006b, fc.) I argue that the scale in (6) characterises the evidential values for reporting constructions and the one in (7) the speaker attitude.

(6) direct evidence > indirect evidence > inferential evidence

(7) doxic > dubitative > volitive

In this talk I apply the proposed model to new quotatives in Dutch, English and Russian and compare these to quotative marker constructions in a random sample of non-Indo-European languages. I argue that in every language new quotatives serve specific semantic and pragmatic functions but that there are universal patterns to be found in the expression of evidentiality and speaker attitude. The aim is to place the speaker attitudes found in the respective new quotatives in a cross-linguistic perspective and in so doing indicating which functions are language-specific and which are cross-linguistically relevant for quotatives.

Session: Workshop (part 1)
New Perspectives on New (and old) Quotatives
Friday, April 4, 2008, 10:30-12:00
room: 02