6-7. Sept 2012

6. Sept 9:00-18:00

7. Sept 9:00-15:00

The meeting will be held in Paris on the two days preceding Sinn und Bedeutung. We were able to get a room through the help of Pascal Amsili and Lucia Tovena at Paris Diderot in the 13e arrondissement.

Université Paris Diderot, UFR de Linguistique, 175, rue du Chevaleret, 75013 Paris room 4C92 4th floor, “plateau” C

Closest metro station: Chevaleret, line 6.

See the following weblink for nore information (in French) on the venue and how to get there:

http://www.liafa.univ-paris-diderot.fr/web9/acces/acces_fr.php

Many thanks to Lucia Tovena (local organizer)

For members, we reimburse the travel costs to Paris and back.
We reimburse hotel costs for 3 nights:

• the 5th,
• the 6th
• and the 7th of September.

There is a price limit per night which is 100€.
If you book above this, please note that the university of Göttingen may not refund anything above 100€ per night. Typically if the difference is small, they may not object, however.

Each participant should book on his/her own.
We suggest booking in the region Place d'Italie/Les Gobelins, since that is also near the SuB site.
If you find a suitable hotel, please share the suggestion here.

Radek: I booked a room at the Neptune hotel (15, rue Godefroy, 13. Place d'Italie, Paris, 75013)
Edgar: same
Malte: same

Université Paris Diderot, UFR de Linguistique, 175, rue du Chevaleret, 75013 Paris room 4C92 4th floor, “plateau” C

Google map: http://goo.gl/maps/QZwR

Sept 6.

9 - 12 Invited talks + Discussion
12- 14 Lunch break
14- 18 Thematic Slot 1: Grammar and Discourse Structure: From Accent to Focus to Questions?

19- Dinner. suggestion: http://petitmarguery.com/fr/restaurant.html

Sept 7.

9- 12 Thematic Slot 2: Local Information Structure
12-14 Lunch break
14-15 Floris Roelofsen: Assertions and Polar Questions: the default case and beyond
15-19 Paris per pedes tour

In the second network meeting on Questions in Discourse we would like to shed more light on the relation between questions and information structure by taking a closer look at the relation between accenting/focus realization (A) and the information-structural category of focus (F=, on the one hand, and the relation between focus and the current question/question under discussion (Q) as a discourse-structuring device, on the other:

(1) A – F | F – Q

The first question to be addressed is whether the hypothesis in (2) holds in light of the empirical facts – both old and new:

(2) There is a 1:1-correlation between A and F and Q in that the nuclear pitch accent always marks the focus of the utterance and focus always indicates the current question under discussion.

Of particular relevance are apparent mismatches between A, F, and Q, which fall into two subclasses:

 i.          Unfocused accents: Instances of pitch accents that do not appear to relate directly to a question under discussion.
 ii.         Unaccented focus: Instances of question-indicating focus that are not marked by accent.

The second question to be addressed is how to integrate certain problematic instances of focus marking in embedded environments (not at-issue appositives, conditionals) into a discourse framework that is based on questions.

We will deal with each question in a separate discussion round with the following preliminary working titles:

• Grammar & Discourse Structure: From accent to Focus to Questions?
• Local Information Structure: Focus in Conditionals and Appositives

Each discussion round will be in charge of some volunteer member of the network. The discussion round will consist of a general introduction, followed by individual contributions on empirical observations and theoretical considerations (discourse modelling), followed by a general discussion.

Sept 6.

Rooth’s talk aims at bringing together Alternative Semantics and the phonology of focus. Using two binary features: F and WH, he distinguishes two types of focus: wh-focus [+F, +WH] and common focus [+F, -WH]. He showed on the basis of data from English, Chinese and Japanese, that semantic scope and phonological scope co-vary. Based on the compositional system of Shimoyama (2006) and Beck (2006), Rooth introduced a new recursive system for the projection of alternatives. In the new system, focus marking is replaced by four local operators: 01 (project alternatives from the right child), 10 (project alternatives from the left child), 11(project alternatives from both) and 00 (don’t project alternatives). Using a phonological constraint, the assignment of phonological prominence and the semantic derivation are done at the same time. The ~ operator terminates the upward projection and hence marks the scope and antecedent of contrast.

Truckenbrodt discusses whether wh-features in questions are actually F-features – his talk thus explicitly assumes F-features, in contrast to Rooth’s summarized above. He presents evidence from alternative questions, where there is a prosodic focus effect on the alternatives, and from numerous constructions (Japanese questions, echo questions, in-situ wh-phrases in multiple wh-questions in German/English) where there is focus prosody on the wh-word. He notes that English/German wh-phrases without focus prosody have alternatives without requiring a focus accent (Truckenbrodt in press-b).

Büring: we need a theory of deaccenting. Where you put the accent has nothing to do with discourse structure. e.g.

A: Let's have some french toast. B: I forgot how to MAKE french toast. → no alternative

This thematic slot discusses the relation between accent and focus, and between focus and the question under discussion.

Hartmann presented joint work with Malte Zimmermann on the Chadic languages Hausa and Bura, in which non-subject focus can remain unmarked (morpho-syntactically as well as prosodically). She suggests that this calls for a “conservative F-marking” (i.e. only F-marking of alternative inducing focus – not of new elements), via pragmatic principles (e.g. Q/A-correspondence). Cases in which focus is marked via movement and a so-called “relative aspect” are analyzed as driven by an unexpected discourse move: rejection/correction of a previous statement, mirativity, etc.

Zimmermann continued with some thoughts on how to connect the Hausa data with work on focus and the Austinian topic situation (Kratzer 2011, Schwarz 2009). He proposed that whereas in-situ focus is new information, ex-situ focus serves to indicate the QUD, and the relative marking restricts the anaphorically given or presupposed topic situation.

Onea discussed the relation between event marking and focus in Hungarian. He suggests that in Hungarian preverbal focus the event is marked as anaphoric (corresponding to that in the QUD), and every alternative is an exhaustive description of the event. In Hungarian unmarked focus a bare noun or aspectual particle indicates the event, the event is existentially closed – thus no exhaustification takes place.

Sept 7.

Onea introduced a new discourse model which incorporates the notion “potential question”. He first discussed the discourse model of Roberts (1996) and showed its limitations with respect to non-strategic discourse. Then he argued that the notion of potential question (PQ) is needed for capturing optional information without re-organizing the entire discourse. A PQ relative to a common ground (CG) is defined as a question which meets the two requirements: a) it is open relative to CG, and b) all its presuppositions are fulfilled by the CG. A proposition s triggers a PQ q in a CG iff q is not a PQ relative to the CG before adding s , but is a PQ after adding s to the CG. Onea specified a range of compositional sources of PQ e.g. indefinites trigger specificational questions, disjunction triggers decision questions etc. He argued that Conventional Implicatures which are typically not-at-issue relative to the QUD can be seen as providing answers to PQs. Then Onea introduced a new discourse model on the basis of Roberts (1996) in which the potential questions are added and the CG is eliminated.

Roelofsen presented a joint work with Donka Farkas on assertions and polar questions. The aim is to give a unified theory for assertions and polar questions which captures the similarities and differences between them with respect to their semantics and contextual effects. Integrating the inquisitive semantics framework (Groenendijk and Roelofsen 2009; Ciardelli and Roelofsen 2011 a.o) with recent works on discourse structure (Farkas and Bruce 2010 a.o), the theory is divided into a semantic component and a discourse component which are responsible for the characterization of default cases and non-default cases, respectively. First, Roelofsen showed that the differences in context effects between assertions and polar questions result from a semantic distinction: the former express a singleton set of possibilities while the latter express a proposition consisting of two mutually exclusive possibilities. Then, an account is given for the distribution of polarity particle responses by means of the notion “highlighting” and the distinction between absolute and relative polarity features. Finally, Roelofsen showed how the non-default cases such as tag questions can be dealt with using the discourse notions “source” and “dependent” (Gunlogson 2008).

In the second part of this talk presented at Konstanz University in November 2012 I explore the relation between scope marking, embedded focus accenting, and discourse structure. It is claimed that focus accents embedded in complement and conditional clauses indicate a discourse strategy (Buering 2003), which consists of a more general super-question Q', and a more specific question Q one level down which restricts the answer space of the super-question. Informativeness of Q wrt Q' is achieved not by means of the element-of relation (as in Buering's CT-constructions), but by means of the subset-relation. To give an example:

(1) Mary thinks that John kissed ANITA.

(1) simultaneously answers Q' = What does Mary think? and the embedded local question Q = Whom did John kiss (according to Mary's beliefs). I.e. Q is not interpreted independently of Q', but forms a local question in the scope of Q'. I further claim that scope marking questions are a gramamticalized way of simultaneously expressing Q' and Q within one utterance (the same can be achieved by asking two consecutive questions, as is the only option in English). The first part of the paper shows how a specific subtype of scope marking constructions in Hungarian is semantically interpreted, following the indiretc semantic accoutn of Dayal (1994, 2000).