conf. Demirdache & Martin 19/5

[Bridget Copley bridget.copley@sfl.cnrs.fr [parislinguists]]

Le projet « La causalité dans le langage et la cognition »  de la Fédération «Typologie et universaux linguistiques » (CNRS FR 2559) a le plaisir d’annoncer un exposé :  


On how Not to culminate

Hamida Demirdache & Fabienne Martin
Université de Nantes LLING (EA3827) / Universität Stuttgart SFB 732


lundi 13 mai 2014 - 14:30 - 16:30
CNRS site Pouchet: 59 rue Pouchet, Paris 75017, salle 159
métro ligne 13 Guy Môquet ou Brochant, RER C Porte de Clichy (sortie La Jonquière)




résumé : In Romance and Germanic languages, a perfective sentence with an accomplishment predicate is taken to describe an event that has culminated – that is to say, an event that has reached its telos or inherent, natural, endpoint.

(1) Pierre a tué son chat, #mais il n'est pas (encore) mort. 
(2)  Peter killed his cat, #but it is not dead (yet).

In many languages, however, perfective accomplishments are known to allow for non­culminating readings. There are (at least) two different ways in which culmination can be cancelled across predicate types and languages. In Mandarin (3) with causative ‘burn’, non culmination involves non completion of the expected change of state since the book burns partially, but not completely. Mandarin ‘burn’, however, can also fail to culminate via the absence of the expected change of state, as in (4) where the book fails to undergo any burning change whatsoever. These two non culminating construals correspond roughly to the distinction found in the litterature between failed attempt readings and partial success readings (Tatevosov 2008, Tatevosov & Ivanov 2009, Lyutikova & Tatevosov 2009). We label them partial result  and zero result non-culminating readings.

(3) Partial result non-culminating reading

Yuēhàn    shāo le       tā­de shu,        dàn      méi      quán shāo­huǐ          (Demirdache & Sun 2014)
Yuēhàn    burn perf     3sg­DE book     but       neg      completely burn­destroy
‘Yuēhàn burned his book, but it didn’t burn completely.’

 (4) Zero result non-culminating reading

Yuēhàn    shāo le       tā­de shu,        dàn      méi shāo­zháo                                                        (id.)
Yuēhàn    burn perf     3sg­DE book     but       neg burn­touch
‘Yuēhàn burned his book, but it didn’t get burnt at all.’

This talk seeks to probe a correlation gone to a large extent unnoticed in the literature, namely, that the availability of non­culminating construals for accomplishments correlates with the control of the agent over the described event: whenever an accomplishment admits a non-culminating construal, this is the case only if we can ascribe agenthood to the subject; if the subject of the very same verb is a (pure) causer, culmination cannot be cancelled, compare e.g. (3) and (5):

 (5) # Huǒ shāo le   tā­de shu,    dàn   méi    shāo­zháo                             (Demirdache & Sun 2014) 
         fire burn perf 3sg­DE book but    neg    burn­touch
Intended: ‘The fire burned his book, but it didn’t get burnt at all.’ 

Demirdache & Martin (2013) refer to this correlation as the Agent Control Hypothesis (ACH). We explore here the scope of the ACH relative to each of the different ways of cancelling culmination identified above. The question we thus ask is  to what extent the ACH holds crosslinguistically across the subtypes of non-culminating construals illustrated with Mandarin in (3) vs. (4): non completion of the expected change of state vs. the (total) absence of the expected change of state.
This will lead us to consider two versions of the ACH:

 (7)    a.    S­ACH (Strong version) 

         Zero result and partial result NC construals require the predicate's external argument to be associated with 'agenthood' properties.

b.    W­ACH (Weak version)


Crosslinguistically, at least zero result non­culminating construals require the predicate's external argument to be associated with 'agenthood' properties.

On the strong version, cancelling culmination requires agent control whether what is being denied is the occurence in w0 of any result (‘zero result' construal), or merely that the result state satisfies the maximal value of the relevant scale (‘partial result' construal). The question is whether there are languages (or predicate classes across languages) of which the ACH holds in its strong version.

On the weak version, the prediction is that, across languages and predicate types, agenthood is required to licence zero result construals, while partial result construals will be, in some languages and with some predicates, licensed with causer (non­agentive) subjects. 





Nous vous accueillons également à notre séance à venir : 


2 juin (lundi) - 14:30-16:30
Hilary Chappell (CRLAO, EHESS) : Réanalyse des constructions causatives et passives périphrastiques dans les langues sinitiques ayant pour source des marqueurs issus de verbes “donner” et “attendre"


Groupe sfl-cause
Agenda du laboratoire Structures Formelles du Langage 
Fédération Typologie et Universaux Linguistiques
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://listserv.linguistlist.org/pipermail/parislinguists/attachments/20140513/5f4fef5a/attachment.htm>