Contrast, Information Structure and Intonation

Workshop October 28 -29th, 2005
Computational Linguistics Department, Stockholm, Sweden

Local Organizer:

Jennifer Spenader (j.spenader(a)gmail.com)

Confirmed speakers

Day and time
Speaker
Title

Friday, October 28th



14:00 - 14:10
Welcome


14:10-14:55
Kjell-Johan Saebo (University of Oslo) homepage Intensifying "selv" in Germanic: a Reanalysis
14:55-15:40
Lotte Hogeweg (Utrecht/Nijmegen)

On the Dutch particle "wel" and its contrastive power


15:40-16:00
Coffee break


16:00-16:45
Sofia Gustafson-Capkova (Stockholm University) homepage Contrast as a principle governing the perceived discourse structure
16:45-17:30
Valeria Molnar (University of Lund, Sweden) On different kinds of CONTRAST
Saturday, October 29th



9:45-10:30
Carla Umbach (University of Osnabruck) homepage Contrast chains
10:30-10:45
Short break

10:45-11:30
Peter de Swart (Radboud University, Nijmegen) homepage And now for something completely different: the use of "anders" in Dutch
11:30-12:15
Henk Zeevat (ILLC Amsterdam) homepage
& Elena Karagjosova

The history and grammaticalization of "doch"
12:15-13:15
Lunch


13:15-1400
Sophie Repp (Humboldt University Berlin) homepage "Only"and "Even" in contrastive coordinations
14:00-14:45
Petra Hendriks (University of Groningen) homepage Asymmetries in the acquisition of contrastive stress
14:45-15:00
Break

15:00-15:45
Beata Gyuris (Budapest) homepage Contrastive topics in Hungarian: a prosodic distinction with interpretational correlates
15:45-16:30
Helen de Hoop (Radboud University, Nijmegen) homepage Escape from stress: the position of pronouns in Dutch
16:30-16:45
Break

16:45-17:30
Jennifer Spenader (Stockholm University) homepage Contrast in Context
17:30 +
Free discussion, continuing at resturant


Workshop Purpose

Contrast has been approached from many different perspectives. People working in computational theories of discourse have seen it a basic discourse relation. Researchers in Pyscholinguistics and text interpretation see it as a basic organizing principle in interpretation. Within alternative semantics and information structure work, contrast is also given a key role, and in this work its interaction with intonation has been studied. In short, the concept of contrast interfaces with a variety of linguistic phenomena, and the effects of contrast, and its twin, parallelism, are pervasive in natural language.

Yet there is still little consensus about exactly how this interaction plays out, and how the different perspectives on contrast relate. Even formal characteristics of such well-used concepts as ``contrastive topics'' are blurry, and general, formal procedures for their identification lacking. Much recent work has successfully used a more formal treatment of contrast to explain contraints on the inteprtation of other linguistic phenomena (e.g. Saebo 2004 on additive particles, and Hendriks 2004 with ellipsis), and problems similar to these are obvious applications for a better understanding of contrast. Also, surveys such as Umbach (2004) and Molnar (2001), have tried to clarify how the different perspectives of contrast related. Contrast is a useful concept only insofar as it is formally delimited, and where its definition in one application is comparable to its use in another. The main purpose of the workshop will be to give a more unified formal semantic/pragmatic account of contrast and how it relates to information structure, discourse structure and intonation. By organizing the workshop we hope to encourage research related to :


This workshop is a follow-up of the workshop organized in April of 2003 Helen de Hoop and Peter de Swart organized a workshop on Contrast in Discourse (see link left) at Radboud University (formally the University of Nijmegen) in Nijmegen, the Netherlands. Work presented at the original workshop was stimulating and of such high quality that it resulted in a special issue of Journal of Semantics. We hope to continue to encourage work on this important topic by once again bringing together top researchers in the area. There is also the intention that a selection of the papers should appear in an edited volume.

Submission details:

Authors are invited to submit a proposal for their talk, or an abstract to be made available before the worksho. Submissions should not exceed 8 page, but may of course be less. The following formats are acceptable: PDF, PS, ASCII text. Please send your submission by email to j.spenader(a)gmail.com by September 15.

Important Dates

Submissions: September 15th, 2005

Where to stay:

Here are my two hotell suggestions. If you prefer something different, need help booking, or want some more information, just send me an email. I am really very happy to help!

Gustav Vasa Hotell:
This is a beautiful hotell located near Odenplan in Stockholm. The entire hotel is architecturally marked. The rooms are all a bit different, due to the age of the building, but it has a lot of charm. Actually I booked this hotell for my father and for good friends when they have come to visit and it wasn't possible to stay with me, so I really recommend it. It's near a subway line that goes to the central station (3 stops) and there is a bus that stops almost in front that goes to the University campus. Gustav Vasa Hotell info in English
A single is about 70 Euros a night, a double, 90 Euros. Gustav Vasa Hotel's Swedish Homepage


Hotel Arcadia:
This is a more normal hotel (no odd shaped rooms...but also less "atmosphere") located very near to the subway line that goes directly to the university campus, and is only one stop away! University guests often stay here. It's also right in downtown Vasastan, and only tow stops away to the Central Station.
A double room seems to be about 90 Euros
Hotell Arcadia

THE WORKSHOP LOCATION
The works will be held at the department of linguistics at Stockholm
University. The red subway line stops at the university, direction
Mörby Centrum. It is one stop after Tekniska Högskolan. From Hotell
Arcadia you just have to get on the subway and take it one more stop.
If you are coming from Gustav Vasa Hotel, you have two possibilites.
You can walk down Valhallavägen, with the church at your back, and see
beautiful Vasa-stan (Vasa-town, a part of Stockholm). After about 10
minutes you'll have reached the subway station Tekniska Högskolan and
you can take the same subway one station as the people staying at the
Arcadia.

Or you can take the bus from Odenplan. Bus 70 or Bus 40 go to the
University subway station. The trip takes about 10 minutes. The hotell
staff should be able to point you in the direction of the bus stop.

 After arriving at the university campus you need to walk towars a row
of tall, light blue buildings. These buildings are collectively known
as "Södra huset" (South building) and the department of linguistics is
in HUS C, on the third floor. Hus C is also color-coded green inside.
We will be having the conference in room C307. I will put up signs
inside the building to guide you there. Otherwise, try asking someone.
:)

Abstracts


Helen de Hoop

Escape from stress: the position of pronouns in Dutch
Pronouns tend to scramble to the left in Dutch sentences. For example, pronominal objects almost obligatorily occur to the left of a sentential adverb (i.e., in scrambled position). Exceptions to this generalization are pronominal objects which receive a deictic or a contrastive reading, as for example in (1):

(1) “Afgelopen”, zei oom schor, en Kees hoorde hem snikken. “Ach God”, zei tante. En Kees hoorde ook haar snikken.
That's the end, said uncle hoarse, and Kees heard him sob. O God, said aunt. And Kees heard also her sob.

The pronoun in (1) gets a contrastive interpretation, 'Kees heard her sob too (after he heard his uncle sob). I will show how a bidirectional Optimality Theoretic analysis naturally explains the fact that unscrambled pronouns get a deictic (non-anaphoric) or contrastive (anaphoric) interpretation, both of which come with stress on the pronoun. One observation that is not yet explained by this analysis, is that pronominal objects do not have to scramble when the sentential adverb is stressed, witness (2):

(2) Maar deze mevrouw was een edele dame, en als ze ooit hem nodig had…
But this madam was a noble lady, and if she ever him needed

In those exceptional cases, the pronoun obtains a non-contrastive reading in situ and it remains unstressed although it is in unscrambled position (cf. van Balen and de Hoop, 2005). I will argue that this can be explained in a bidirectional OT analysis as well when we assume that the constraint that pronouns should not receive stress is ranked higher than the constraint that pronouns scramble. Most of the time, the latter constraint is satisfied simply because it goes hand in hand with the former constraint, but if the former constraint can be satisfied independently of the latter one, then we sometimes see as an effect the emergence of an unstressed pronoun in unscrambled position.

Petra Hendriks

Asymmetries in the acquisition of contrastive stress
A well-known property of focus-sensitive operators such as only is that the truth conditions they give rise to can vary with the placement of main sentence stress. A different placement of main sentence stress can result in a different interpretation of the bound focus with which the focus-sensitive operator associates. Whereas children seem to correctly interpret sentences with bound focus in which stress falls on the default position (i), they appear to experience problems with the comprehension of sentences in which (contrastive) stress falls on a non-default position (ii) (e.g., Gennari et al., 2001; Gualmini et al., 2002; Halbert et al., 1995; Szendröi, 2003; but see Miller et al. 2005, who argue that children do not have difficulty using contrastive stress in comprehension).

(i) Tigger only threw a chair to PIGLET.
(ii) Tigger only threw a CHAIR to Piglet.

In production, on the other hand, there is no evidence that children have problems with the correct placement of contrastive stress (Baltaxe, 1984; Cutler & Swinney, 1987; Hornby & Hass, 1970; Nederstigt, 2001).
This remarkable asymmetry between production and comprehension, with correct comprehension lagging behind correct production as much as three years, is given an explanation in terms of bidirectional Optimality Theory (Blutner, 2000). In particular, it is argued that children until the age of 6 or 7 optimize in one direction only: from form to meaning in comprehension, and from meaning to form in production. As a result, sentences with contrastive stress on a non-default position are ambiguous as well, as are sentences with stress on the default position, and can have a narrow focus interpretation or a wide focus (VP focus) interpretation. This ambiguity is resolved by taking into account speaker's alternatives in comprehension, that is, by optimizing bidirectionally. Hearers will then realize that a VP focus interpretation is better expressed by a sentence with default stress. Only after the age of 6 or 7 are children able to resolve the ambiguity of (ii) through this bidirectional reasoning, and block the VP focus interpretation for this sentence.
Children's difficulties in determining the focus of the focus particle on the basis of the placement of contrastive stress resemble their difficulties with the interpretation of indefinite subjects and objects (de Hoop & Krämer, to appear), pronouns (Hendriks & Spenader, 2004), and scalar implicatures, but seem to contrast with the ease with which children evoke a set of alternatives on the basis of the placement of contrastive stress (cf. Miller et al., 2005).

References


Baltaxe, Christiane A.M. (1984). Use of contrastive stress in normal, aphasic, and autistic children. Journal of Speech and Hearing Research 27, 97-105.
Blutner, Reinhard (2000). Some aspects of optimality in natural language interpretation. Journal of Semantics 17.3, 189-216.
Cutler, Anne, & David A. Swinney (1987). Prosody and the development of comprehension. Journal of Child Language 14, 145-167.
Gennari, S., Andrea Gualmini, Stephen Crain, L. Meroni & Simona Maciukaite (2001). How adults and children manage stress in ambiguous contexts. Proceedings of 1st Workshop on Cognitive Models of Semantic Processing, University of Edinburgh, Edinburgh.
Gualmini, Andrea, Simona Maciukaite & Stephen Crain (2002). Children's insensitivity with contrastive stress with ONLY. Penn Working Papers in Linguistics 9.1.
Halbert, Anne, Stephen Crain, Donald Shankweiler, & Elaine Woodams (1995). Children's Interpretive Use of Emphatic Stress. Paper presented on the 8th Annual CUNY Conference on Human Sentence Processing, Tucson, AZ.
Hendriks, Petra, & Jennifer Spenader (2004). A bidirectional explanation of the pronoun interpretation problem. In: P. Schlenker en E. Keenan (eds.), Proceedings of the ESSLLI '04 Workshop on Semantic Approaches to Binding Theory, Nancy, France.
Hoop, Helen de, & Irene Krämer (to appear). Children's Optimal Interpretations of Indefinite Subjects and Object. Language Acquisition.
Hornby, P. A., & W. A. Hass (1970). Use of contrastive stress by preschool children. Journal of Speech and Hearing Research 13, 395-399.
Miller, Karen, Cristina Schmitt, Hsiang-Hua Chang & Alan Munn (2005). Young children understand some implicatures. In: A. Brugos, M.R. Clark-Cotton, and Seungwan Ha (eds.), BUCLD 29 Proceedings, Sommerville, MA: Cascadilla Press, pp. 389-400.
Nederstigt, Ulrike (2001). The acquisition of additive “focus particles” in German. A. H.-J. Do et al., BUCLD 25 Proceedings, Sommerville, MA: Cascadilla Press, pp. 554-565.
Szendröi, Kriszta (2003). Acquisition Evidence for an Interface Theory of Focus. In: J. van Kampen & S. Baauw (eds.), Proceedings of Generative Approaches to Language Acquisition 2003.

Lotte Hogeweg

On the Dutch particle wel and its contrastive power

On a forum on the internet I saw a discussion about an article announcing the death of a famous Dutch person. The article ended with the following line:

(1) Willem Oltmans zal wel in stilte begraven worden
  Willem Oltmans will PARTICLE in silence buried be
  'Willem Oltmans will be buried in silence'

The person who placed (part of) the article on the forum finds the use of the word wel very inappropriate. 'Was he such a noisy man?' he wonders. Other forum members speculate about why the word wel is used here too. One discussant puts forward the idea that the word indicates that Willem Oltmans will nót be put on the centre spot of the Arena stadium (unlike the famous Dutch singer Andre Hazes who died just before the person discussed here). Another person suggests that the way he will be buried is put in contrast with the image of his rather turbulent life that rises from the rest of the article. Finally a person suggests that the word wel indicates a contrast between the fact that he will be buried in silence and the fact mentioned in the previous sentence (not cited on the forum) that a public website has been created where people can offer their condolences.

Wel could be called the positive counterpart of niet 'not'. When children disagree about a certain fact they often use those two words as recurring arguments: wel(les), niet(es), wel(les), niet(es) 'it is, it is not, it is, it is not'. Because the affirmative meaning of a sentence is the unmarked one, adding the particle wel has to have another reason then just creating a positive meaning. Often this reason is stressing the affirmative nature of the sentence, and by doing so creating contrast, like in the first example. In other situations this contrastive meaning is not that clear, or even absent:

(2) Ik heb vandaag wel honderd boten geteld!
I have today PARTICLE hundred boats counted
'I have counted no less than a hundred boats today!'

Wel is connected to the quantifier hundred in (2) and the speaker indicates by using the word that he thinks hundred boats is a lot. Another different function of wel is illustrated in (3):

(3) Het feestje afgelopen zaterdag was wel leuk
The party last Saturday was PARTICLE nice
'The party last Saturday was OK'

Here the speaker says that the party was OK, not good not bad. Wel functions as a moderator to the predicate leuk (nice), and weakens its meaning. And what about the, for non-native speakers of Dutch quite confusing use of both wel and niet right next to each other. This use is only possible in questions or other wh-phrases. In (4) the speaker expresses his surprise over a situation in which it appears that the hearer has eaten a lot of cake.
(4) Grote grutten, hoeveel taart heb je wel niet
Great groats, how much cake have you PARTICLE not gegeten?!
eaten?! 'My God, how much cake did you eat?!'

Wel has not often been included in research on Dutch particles. Not much has been written about it to my knowledge. While in the Spoken Dutch Corpus (Corpus Gesproken Nederlands) it takes in the twentieth position in the frequency list containing all words! The only word considered to be a particle in a higher position is maar 'but'. And as we have seen a lot can been said about the word as well.
I have been calling wel a particle up till now. But what does it mean to be a particle? Foolen (1993) considers the class of particles to include every element that doesn't add to the propositional meaning of a sentence. He subdivides them into the following classes: interjections, conjunctions, modal adverbs, focus particles and modal particles. I have only mentioned few of the possible functions of wel. They are quite diverse and function in different word classes. In this talk I will make an inventory of the different meanings or functions wel expresses or fulfils and argue they share a core meaning based on denial of an explicit or implicit previous denial.

Sometimes the different readings are in conflict. In (1) the forum members are discussing the meaning of wel. They interpret it as creating a contrastive reading and try to find alternatives for elements of the sentence. But there are actually several other possible interpretations that the discussants don't take into consideration.
In contrast to focus particles wel has scope over the whole sentence. Wel doesn't create a set of alternatives on its own. That actually makes sense, because the only alternative for a situation that ís the case, is a situation that isn't; the only opposite of wel is niet 'not'. Use of the word wel with main stress, in other words, implies a previous state of knowledge in which the proposition of the sentence was hold not true. In that case the utterance containing wel is correcting a previous state of knowledge.
Only when phonetic focus is added to one of the elements of the proposition another set of alternatives is created. Additional focus on the subject for example implies a previous state of knowledge in which for (a set of) alternative(s) for the subject the negation of the proposition held. Or when the verb is in focus, (a set of) alternative(s) for the activity is evoked.
This is in fact how the forum members interpret (1). They are asking the question “what the word wel indicates”. Well, the word wel indicates not. What they should actually ask is what element is meant to have additional focus. The person talking about the centre spot of the Arena stadium interprets the sentence as having focus on Willem Oltmans. By that a set of alternatives for that word is activated. Because of the recent death of Andre Hazes and the hype created around it, the saliency of his name is “in the air”, so to speak. The other two suggestions both need the sentence to have focus on begraven (buried). One person abstracts from the article in general the rather turbulent live of Willem Oltmans as the alternative for begraven. The other discussant abstract the fact that a public website has been created as mentioned in the previous sentence, as an alternative. There is also another non-contrastive possible interpretation of wel in (1). In this interpretation wel functions as a modal particle. It takes scope over the whole sentence and indicates that the speaker is not a hundred percent sure about the proposition. This would mean for sentence (1) that the writer of the article suspects the person will be buried in silence but is not sure about it. In this reading wel will have no phonetic focus.

Since prosody has not been marked in this case, it gives us no clue about how this sentence should be interpreted. Nonetheless people come to an interpretation. In the talk I will argue a hierarchy of the different readings plays a role in interpreting the particle wel, meaning some functions are stronger than others. If for example in the article containing (1) a previous utterance stated that Willem Oltmans wouldn't be buried in silence, (1) would for sure have been interpreted as correcting that statement. And although people have trouble finding a plausible alternative for the contrastive reading, they still prefer it over the possibility of wel being a modal particle. A combination of such a hierarchy and constraints concerning prosodic intonation put in an Optimality Theoretic framework can account for the diversity in production and interpretation of the particle wel.

Since prosody has not been marked in this case, it gives us no clue about how this sentence should be interpreted. Nonetheless people come to an interpretation. In the talk I will argue a hierarchy of the different readings plays a role in interpreting the particle wel, meaning some functions are stronger than others. If for example in the article containing (1) a previous utterance stated that Willem Oltmans wouldn't be buried in silence, (1) would for sure have been interpreted as correcting that statement. And although people have trouble finding a plausible alternative for the contrastive reading, they still prefer it over the possibility of wel being a modal particle. A combination of such a hierarchy and constraints concerning prosodic intonation put in an Optimality Theoretic framework can account for the diversity in production and interpretation of the particle wel.

Kjell-Johan Saebo

Intensifying selv in Germanic: A Reanalysis
Regine Eckardt (2001) has proposed an analysis of "intensifying" selbst which can account for many of the relevant facts. As it stands, however, it cannot quite cope with adverbal cases where the "associate" is apparently itself in focus, and it does not answer the question what makes intensifying selbst necessary in many cases.

In this paper, I elaborate on that analysis to say that the associate can be a type e variable and that the selbst focus presupposition can be interpreted at the level of the VP. I show that in this way, intensifying selbst enables one and the same argument to be both a continuing theme and a contrastive theme or both a theme focus and a rheme focus, and that it allows for focus presuppositions involving alternatives to alternatives.

Sophie Repp

'Only' and 'even' in contrastive coordination

.pdf version of abstract

Jennifer Spenader

Contrast in Context

Why do speakers use contrastive relationships in dialogue, and why do writer's use it in texts? Traditionally approaches to contrast have focussed on what licenses the contrastive relationship, to a great degree ignoring its function in the discourse. Different theories have given potential answers to this question. I'll look at each one, discuss the shortfalls of the theory, and then present examples from a corpus study that illustrate what types of relationships seem to be relevant for identifying the function of a contrastive relationship, understanding it, and for even generating similar discourse relations. These three subtasks are necessary to be able to correctly deal with contrastive relationships in any language technology system, and they all hinge on an ability to identify what contrasts function is.


Henk Zeevat & Elena Karagjosova

The History and Grammaticalisation of "Doch"

.pdf version of abstract

Beata Gyuris

Contrastive topics in Hungarian: a prosodic distinction with interpretational correlates 

Since Szabolcsi 1981, the term contrastive topic has been used in the study of Hungarian to refer to  certain constituents that are situated in the so-called topic position of the sentence in the syntactic framework proposd by Katalin É. Kiss (cf. É. Kiss 1981, 1987, 1994, 2002, among others), the highest among the operator positions on the left periphery of the hierarchically structured preverbal field of the Hungarian sentence. According to É. Kiss, when the constituents occupying this position are pronounced with a falling intonation contour, they must be noun phrases with a referential and specific interpretation. Szabolcsi 1981 points out that a wide range of constituent types (e.g. non-referential noun phrases, verb, adjectives, etc.) becomes available for occupying the above syntactic position (testable by syntactic means), when pronounced with a rising contour. She argues that the rising intonation is associated with a kind of interpretational surplus: it is suggested, on the one hand, that there are alternatives to the denotation of the contrastive topic about which „the same question could be asked”, and, on the other hand, that the answer to these questions would differ from the one given in the original sentence containing the contrastive topic. For example, (1) below evokes the contrast that there are relevant individuals other than John about whom the question when they arrived could be raised, and it suggests that the answers to these questions wouold specify a day different from Tuesday:

 

(1)       [CT ´János] [F                      `kedden]    érkezett.
                 John          Tuesday-on   arrived
            ‘As for John, he arrived on Tuesday.’   

A closer examination reveals, however, that the two aspects of this double contrast (cf. Molnár 1998) do not have the same status. Whereas evoking a set of alternatives seems to be a necessary component in the interpretation of contrastive topics across languages, it is not necessary that in sentences where alternatives are substituted for the original contrastive topic, the foci also be different (cf. Eckardt 2002, van Hoof 2003). Consequently, in Gyuris (to appear), the first type of contrast is assigned the status of  presupposition, whereas the second one that of an implicature. (A similar view about German and English is reflected in Büring's 2003 theory.)

The intonation pattern of the contrastive topic in Hungarian has in most works been described as a rise (e.g. Szabolcsi 1981, Kálmán and Nádasdy 1994), in others it is referred to as a (fall)-rise (cf. Molnár 1998, É.Kiss and Gyuris 2003). It is not discussed in these works, however, under what circumstances the rising contour can or must be preceded by a fall.

It will be shown in the paper that the choice between the alternative prosodic realizations is not arbitrary, but depends strongly on which of the prototypical utterance-types that tend to precede contrastive topics cross-linguistically appears in the actual discourse. The types of utterances that can precede contrastive topics include, among others  i) the single wh-question to which the declarative with the contrastive topic provides a congruent answer (von Stechow 1990), ii) the question to which it provides a partial answer (the root question in Büring's 2003 theory), and iii) a question to which the declarative with the contrastive topic does not provide an answer at all, since it is asking for information about an alternative of the contrastive topic denotation (sister-question of question type i) in Büring's 2003 theory). The English versions of the relevant question-types for (1) are illustrated in (2)-(4):

 

(2)        When did John arrive?

(3)        Who arrived when?

(4)        When did Mary arrive?

Whereas (1) can function as an answer to (2) either when the subject noun phrase is pronounced with a rise or with a fall-rise, it can only function as an answer to (3) when the latter is pronounced with a rise, and to (4) when it is pronounced with a fall. 

            It is proposed that the unequal status of the two types contrast associated with contrastive topics in Hungarian, as well as their incompatibility with some of the questions of the type in (2)-(4) on any of their actual utternaces can be accounted for if the two facets of the double contrast are attributed to the two different aspects of the intonation pattern characterizing these constituents. According to this, the rising contour common to all instances of contrastive topics should be responsible for introducing the presupposition that there are relevant alternatives to the contrastive topic denotation, and the fall preceding the rise for the implicature that these alternatives have to be associated with a value different from that specified by the focus in the original sentence.

            The paper discusses the cross-linguistic implications of the parallels between the interpretation and preferred discourse-environments for the two intonational variants of Hungarian contrastive topics pointed out here and those between the interpretation and preferred discourse environments for I-topics vs. multiple foci in German, pointed out by Jacobs 1997.

References

Büring, Daniel. 1997. The Meaning of Topic and Focus. The 59th Street Bridge Accent. London, New York: Routledge.

Büring, Daniel. 2003. On D-Trees, Beans, and B-Accents. Linguistics and Philosophy 26:
511-545.

Eckardt, Regine. 2002. Referential and denotational topics. Talk given at the Conference on Contrast at the institute of Linguistics, University of Leipzig, February 2002.

É. Kiss, Katalin 1981. Structural relations in Hungarian, a free word order language. Linguistic Inquiry 12: 185–213.

É. Kiss, Katalin 1987. Configurationality in Hungarian. The Hague: Mouton.

É. Kiss, Katalin. 1994. Sentence structure and word order. In: Kiefer, Ferenc and Katalin É. Kiss (eds.) The Syntactic Structure of Hungarian [Syntax and Semantics 27]. San Diego, London: Academic Press.

É. Kiss, Katalin 2002. The Syntax of Hungarian. Cambridge University Press.

É. Kiss, Katalin and Beáta Gyuris 2003. Apparent scope inversion under the rise fall contour. Acta Linguistica Hungarica 50: 371-404.

Gyuris, Beáta to appear. A new approach to the scope of contrastive topics. In Benjamin Shaer, Philippa Cook, Werner Frey, Claudia Maienborn (eds.) Dislocation: Syntactic, semantic and discourse perspectives. London, New York: Routledge.

van Hoof, Hanneke. 2003. The Rise in the Rise-Fall Contour: Does it Evoke a Contrastive Topic or a Contrastive Focus? Linguistics 41:515-563.

Kálmán, László and Ádám Nádasdy 1994. A hangsúly [Stress]. In: Kiefer F. (ed.) Strukturális magyar nyelvtan 2. Fonológia. [A Structural Grammar of Hungarian 2: Phonology], 393–467. Budapest: Akadémiai Kiadó.

Molnár, Valéria 1998. Topic in focus: the syntax, phonology, semantics, and pragmatics of the so-called “contrastive topic” in Hungarian and German. Acta Linguistic Hungarica  45: 389-466.

Szabolcsi, Anna 1981. Compositionality in Focus. Folia Linguistica 15: 141–61.

Valeria Molnar

On different kinds of CONTRAST

The talk aims at a clarification of the notion of contrast and its impact
on linguistic structure. Despite the agreement in the literature that the
definition of contrast includes highlighting and operation on alternatives,
the linguistic status of contrast and its relation to the two key concepts
of information structure 'focus' and 'topic' is still a controversial
issue. Whereas in the contrast-focus-relation the closeness of these two
concepts and the difficulties in their formal distinction often leads to
confusion of these two statuses, the problematic issue in the
contrast-topic relation is the compatibility of these two concepts. The
main claim of the talk is that contrast is a linguistically relevant
phenomenon compatible with both focusing and topicality and cannot merely
be regarded as "the result of general cognitive processes referred to as
'conversational implicatures'" (Lambrecht 1994: 291). The linguistic
relevance of contrast must, however, be differentiated depending on which
languages and which types of constructions are taken into consideration.
Contrast can not only serve as a feature for further distinctions within
focus and topic, but it can also be superimposed upon these two relevant
discourse notions. It is argued for two main types of contrast superseding
the functions of topic and focus: (i) the intonationally marked type,
called "I-contrast", obligatorily realized by (fall)-rise and connected
with a special semantic/pragmatic interpretation (non-exhaustive
identification), and (ii) the syntactically marked type, termed
"S-contrast", in which case besides the prosodic highlighting the
left-peripheral position is essential for securing the discourse connection
with competing (explicitly mentioned or contextually present) alternatives
of the relevant set.


TBA

14 March 2005
Contents by Jennifer Spenader <j.k.spenader(at)let.rug.nl>