Focus in German Towards a Corpus-Based Study Motivation ...

Focus in German: Focus in German:

Towards a Motivation Towards a

Corpus-Based Study Corpus-Based Study

Kordula De Kuthy Kordula De Kuthy

Detmar Meurers Detmar Meurers

◮ The nature of the integration of a sentence into discourse

Motivation Motivation

Background

can provide an explanation for constraints stipulated in Background

Focus in German: Towards a Expressing inf. struc

Characterizing intonation

syntax (De Kuthy 2002; De Kuthy & Meurers 2003). Expressing inf. struc

Characterizing intonation





Corpus-Based Study Autosegmental-metricalappr.

Disco. fct. of pitch accents

◮ To explore this line of research, we need an explicit Autosegmental-metricalappr.

Disco. fct. of pitch accents

Relating inton. & interpretation

Focus projection

representation of information structure and its interaction. Relating inton. & interpretation

Focus projection

Constraining projection

An open issue

◮ German and English are intonation languages: Constraining projection

An open issue







Kordula De Kuthy Detmar Meurers

A pilot study ◮ Prosody plays an important role in constraining the A pilot study

IMS Radionews Corpus IMS Radionews Corpus



¨ ¨

Universitat Tubingen Syntactic corpus annotation possible integration of a sentence into the discourse. Syntactic corpus annotation

Examples Examples

PP with one accent ◮ Research relating syntax, information structure, and PP with one accent

NP with multiple accents NP with multiple accents

Accents on every part intonation predominantly theoretically driven Accents on every part

An unexpected accent An unexpected accent

Cogeti Workshop Multiple unexp. accents ◮ partly contradictory empirical assumptions Multiple unexp. accents

Examples occurring twice Examples occurring twice

¨

Tubingen. October 9, 2009

Conclusion ◮ Investigate this empirically by looking at the intonation Conclusion

Appendix

ToBi

of naturally occurring sentences in a larger discourse. Appendix

ToBi

Challenging focus projection

⇒ A pilot study based on intonationally annotated corpora

Challenging focus projection

Published empirical evidence Published empirical evidence









1 / 40 2 / 40



Focus in German: Focus in German:

Expressing information structure Towards a Characterizing intonation Towards a

Corpus-Based Study Corpus-Based Study

Kordula De Kuthy Kordula De Kuthy

Detmar Meurers Detmar Meurers





Motivation

◮ Intonation patterns consist of intonation features or Motivation



◮

subsystems of several kinds:

Languages differ with respect to how the information Background

Expressing inf. struc

Background

Expressing inf. struc

◮ intonational contour (tune)

structure of an utterance is represented. Characterizing intonation Characterizing intonation

Autosegmental-metricalappr.

Disco. fct. of pitch accents

◮ prominence (stress) Autosegmental-metricalappr.

Disco. fct. of pitch accents

◮ Linguistic means of marking information structure include: Relating inton. & interpretation ◮ intonational phrasing Relating inton. & interpretation

Focus projection Focus projection

◮ word order Constraining projection

◮ pitch range Constraining projection



◮ morphology An open issue

◮ The contour indicates the movement of pitch.

An open issue



A pilot study A pilot study

◮ prosody IMS Radionews Corpus ◮ For example, the intonation pattern of an assertion has IMS Radionews Corpus

Syntactic corpus annotation Syntactic corpus annotation

◮ English and German are so-called intonation languages Examples a distinct contour from that of a question. Examples

PP with one accent PP with one accent

◮ Information structuring is signaled by the intonation NP with multiple accents ◮ Intonational phrasing divides the sequence of words NP with multiple accents

Accents on every part Accents on every part

(contour) of an utterance, including pitch accents. An unexpected accent into intonational units, the intonational (prosodic) phrases. An unexpected accent

Multiple unexp. accents Multiple unexp. accents

◮ The absence or presence of an accent is an indicator of Examples occurring twice

◮ Phrase boundaries are marked by pauses, boundary Examples occurring twice





the discourse function of a constituent in a sentence. Conclusion

tones and duration patterns.

Conclusion



Appendix Appendix

ToBi

Challenging focus projection

◮ Pitch range controls limits in which contours are realized. ToBi

Challenging focus projection

Published empirical evidence Published empirical evidence









3 / 40 4 / 40

Focus in German: Focus in German:

Autosegmental-metrical approach to intonation Towards a Tones and Break Indices (ToBI) Towards a

Corpus-Based Study Corpus-Based Study



◮ Pierrehumbert (1980) proposes a description of intonation: Kordula De Kuthy

Detmar Meurers

Kordula De Kuthy

Detmar Meurers



◮ the grammar of phrasal tunes, consisting of L and H tones: Motivation Motivation

◮ pitch accents Background

◮ ToBI is a system for transcribing the intonation patterns Background

◮ phrase accents Expressing inf. struc

Characterizing intonation

and other aspects of the prosody of English utterances. Expressing inf. struc

Characterizing intonation

◮ boundary tones Autosegmental-metricalappr. Autosegmental-metricalappr.

Disco. fct. of pitch accents

◮ Based on autosegmental-metrical approach to intonation. Disco. fct. of pitch accents

◮ the metrical representation of the text Relating inton. & interpretation Relating inton. & interpretation



◮ rules for lining up the tune with the text

Focus projection

Constraining projection

◮ The perceived intonation contour is described in terms Focus projection

Constraining projection





◮ Phonological tones

An open issue

of high (H) and low (L) targets in the local pitch range. An open issue



A pilot study A pilot study

◮ Each phrase requires at least one pitch accent IMS Radionews Corpus ◮ ToBI transcription marks: IMS Radionews Corpus

Syntactic corpus annotation Syntactic corpus annotation

◮ English: H*,L*, or bitonal: H*+L, H+L*, L*+H, L+H*, H*+H Examples ◮ accented syllables, for English: Examples

PP with one accent PP with one accent

◮ Each phrase receives a phrase accent at the end of the NP with multiple accents ◮ H*,L*, or bitonal: H*+L, H+L*, L*+H, L+H*, H*+H NP with multiple accents

Accents on every part Accents on every part

word associated with the last pitch accent: An unexpected accent

◮ The * marks the tone on the accented syllable. An unexpected accent

− −

◮ H ,L Multiple unexp. accents

Examples occurring twice

◮ intonational boundaries (with break values from 0–4): Multiple unexp. accents

Examples occurring twice



◮ Each phrase ends with a boundary tone: Conclusion

◮ intermediate boundary (0–3): H− , L− Conclusion

◮ ◮ full boundary (4): L% or H%

H%, L% Appendix Appendix

ToBi ToBi



◮ The approach has been modified and further developed Challenging focus projection Challenging focus projection

Published empirical evidence Published empirical evidence



◮ e.g., in Beckman & Pierrehumbert (1986)

5 / 40 6 / 40



Focus in German: Focus in German:

An example Towards a Discourse function of pitch accents Towards a

Corpus-Based Study Corpus-Based Study

Kordula De Kuthy Kordula De Kuthy

Detmar Meurers Detmar Meurers





Motivation Motivation



Background ◮ Pitch accents can have different shapes allowing them Background

Expressing inf. struc Expressing inf. struc

VROOT

Characterizing intonation to signal different functions in the discourse. Characterizing intonation

Autosegmental-metricalappr. Autosegmental-metricalappr.





S

Disco. fct. of pitch accents

Relating inton. & interpretation

◮ The H∗ pitch accents are often assumed to signal focus, Disco. fct. of pitch accents

Relating inton. & interpretation

Focus projection

Constraining projection

i.e., new material. Focus projection

Constraining projection



NP PP PP

An open issue ◮ A more precise mapping of each possible pitch accent An open issue



A pilot study A pilot study

IMS Radionews Corpus

to its information structure role is subject of debate. IMS Radionews Corpus



Das Gesetz tritt am kommenden Montag in Kraft . Syntactic corpus annotation Syntactic corpus annotation



das1 g@0 zEts1 (L*H) tRIt1 am1 kOm1 m@n0 d@n0 mo:n1 (!H*L?) ta:k0 In1 kRaft1 (L%) Examples ◮ Autosegmental-metrical approach to German intonation Examples

PP with one accent PP with one accent

ART

1

NN

1

VVFIN

1

APPRART

1

ADJA

1

NN

1

APPR

1

NN

4

$.

0

NP with multiple accents ´

(Uhmann 1991; Fery 1993) assumes: NP with multiple accents

Accents on every part Accents on every part

L*H !H*L? L% An unexpected accent

Multiple unexp. accents

◮ H* or H*+L represent focus accents An unexpected accent

Multiple unexp. accents

Examples occurring twice ◮ L* highlights background constituents Examples occurring twice



Conclusion ◮ L*+H is a topic marker Conclusion



Appendix Appendix

ToBi ToBi

Challenging focus projection Challenging focus projection

Published empirical evidence Published empirical evidence









7 / 40 8 / 40

Focus in German: Focus in German:

Relating intonation and interpretation Towards a From pitch accent to projected focus: Example Towards a

Corpus-Based Study Corpus-Based Study

Focus projection Kordula De Kuthy

Detmar Meurers

Kordula De Kuthy

Detmar Meurers





Motivation Motivation

◮ The word marked by a pitch accent and the extension of Background

the focus are related by rules of focus projection. Expressing inf. struc

[[S]]F Background

Expressing inf. struc

Characterizing intonation Characterizing intonation

Autosegmental-metricalappr. Autosegmental-metricalappr.

(1) Marius hat ein Buch mit BAGGERN bekommen. Disco. fct. of pitch accents Disco. fct. of pitch accents

Relating inton. & interpretation NP [[VP]]F Relating inton. & interpretation

Focus projection Focus projection

¨

(2) Was fur ein Buch hat Marius bekommen? Constraining projection Constraining projection

An open issue An open issue

Marius hat ein Buch [[mit BAGGERN]]F bekommen. V [[NP]]F V

A pilot study

IMS Radionews Corpus

Marius A pilot study

IMS Radionews Corpus



(3) Was hat Marius bekommen? Syntactic corpus annotation

Examples

Syntactic corpus annotation

Examples



Marius hat [[ein Buch mit BAGGERN]]F bekommen. PP with one accent

NP with multiple accents hat Det N [[PP]]F bekommen

PP with one accent

NP with multiple accents

Accents on every part Accents on every part

An unexpected accent An unexpected accent

¨

(4) Wie war Weihnachten fur Marius? Multiple unexp. accents Multiple unexp. accents





Marius [[hat ein Buch mit BAGGERN bekommen]]F . Examples occurring twice

ein Buch mit BAGGERN Examples occurring twice



Conclusion Conclusion



Appendix Appendix

(5) Was war los? ToBi ToBi



[[Marius hat ein Buch mit BAGGERN bekommen]]F . Challenging focus projection

Published empirical evidence

Challenging focus projection

Published empirical evidence









9 / 40 10 / 40



Focus in German: Focus in German:

Constraining focus projection Towards a An open issue Towards a

Corpus-Based Study Corpus-Based Study



◮ Focus projection rules determine the focus projection Kordula De Kuthy

Detmar Meurers

Kordula De Kuthy

Detmar Meurers





potential of a pitch accent depending on syntactic structure Motivation Motivation



Background Background

◮ Strongest accent is generally constrained to fall on the ◮

Expressing inf. struc Popular syntax-driven approach to focus projection: Expressing inf. struc





last element (e.g., Nuclear Stress Rule of Chomsky & Characterizing intonation

Autosegmental-metricalappr. Selkirk (1995) and many variants

Characterizing intonation

Autosegmental-metricalappr.



Halle 1968 for English; Jacobs 1988, p. 124 for German) Disco. fct. of pitch accents

Relating inton. & interpretation

Disco. fct. of pitch accents

Relating inton. & interpretation

Focus projection

◮ Recently, some authors have questioned whether there Focus projection



◮ Narrow focus examples exemplifying this: Constraining projection Constraining projection

An open issue is an interesting constraining relation between syntax, An open issue





(6) Wer hat ein Buch mit Baggern bekommen? A pilot study intonation, and information structure: A pilot study

IMS Radionews Corpus IMS Radionews Corpus

[[MARIUS]]F hat ein Buch mit Baggern bekommen. Syntactic corpus annotation

◮ ¨

Buring (2006) any accent within a phrase can project Syntactic corpus annotation

Examples Examples



(7) Was mit Baggern hat Marius bekommen? PP with one accent focus to the phrase, i.e., focus can always project PP with one accent

NP with multiple accents NP with multiple accents



Marius hat ein [[BUCH]]F mit Baggern bekommen. Accents on every part ◮ Kadmon (2006) and Roberts (2006): focus never Accents on every part

An unexpected accent An unexpected accent

Multiple unexp. accents

projects (new elements must be accented; unaccented Multiple unexp. accents





These are not felicitous answers to questions such as:

Examples occurring twice

focused elements are given/retrievable/expectable) Examples occurring twice



Conclusion Conclusion



(8) a. Was hat Marius bekommen? Appendix Appendix

ToBi ToBi



¨

b. Wie war Weihnachten fur Marius? Challenging focus projection Challenging focus projection

Published empirical evidence Published empirical evidence



c. Was war los?

11 / 40 12 / 40

Focus in German: Focus in German:

An open issue (cont.) Towards a The IMS Radionews Corpus (Rapp 1998) Towards a

Corpus-Based Study Corpus-Based Study

Kordula De Kuthy Kordula De Kuthy

Detmar Meurers Detmar Meurers





Motivation ◮ Recordings of radio broadcasts on Deutschlandfunk: Motivation

◮ There is very little published empirical research on Background Background

Expressing inf. struc

◮ 4 consecutive news broadcasts (28.07.1995) Expressing inf. struc

focus projection (Gussenhoven 1983; Birch & Clifton Characterizing intonation Characterizing intonation

Autosegmental-metricalappr. ◮ 17 news stories, 3 weather forecasts (617.272 s), female Autosegmental-metricalappr.

´

1995; Welby 2003; Fery 1993). Disco. fct. of pitch accents Disco. fct. of pitch accents

◮ 11 news stories, 1 weather forecast (476.553 s), male

Relating inton. & interpretation Relating inton. & interpretation

◮ substantial evidence for the existence of some focus Focus projection Focus projection

Constraining projection ◮ 14 consecutive news broadcasts (21.11.1995) Constraining projection

projection, i.e., for certain pairs of verbs with object-NP An open issue An open issue

◮ 13 news stories, 3 weather forecasts (541.659 s), female

arguments in English A pilot study

◮

A pilot study

IMS Radionews Corpus

94 news stories, 12 weather forecasts (3551.8 s), male IMS Radionews Corpus

◮ but more evidence is needed, specifically: Syntactic corpus annotation Syntactic corpus annotation



◮ In which constructions can what kind of elements be

Examples

PP with one accent

Total length: 1 h 26 min, 514 sentences (≈ 10 sec/sent) Examples

PP with one accent



accented (with which type of accents) and project focus NP with multiple accents NP with multiple accents

Accents on every part ◮ Corpus preparation involved: Accents on every part

how far? An unexpected accent An unexpected accent

Multiple unexp. accents

◮ manual segmentation into news stories Multiple unexp. accents



◮ Can we obtain more robust empirical evidence for focus Examples occurring twice Examples occurring twice



Conclusion

◮ orthographic transliteration Conclusion

projection from authentic language data? ◮ automatically word alignment

Appendix Appendix

ToBi ◮ manual prosodic labeling with ToBI ToBi

Challenging focus projection Challenging focus projection

Published empirical evidence Published empirical evidence









13 / 40 14 / 40



Focus in German:

Syntactic annotation of the corpus Towards a PP with one accent

Corpus-Based Study

Kordula De Kuthy

Detmar Meurers





Motivation



Background VROOT



◮ IMS Radionewscorpus is not syntactically annotated. Expressing inf. struc

Characterizing intonation S





◮ Parsed corpus with the Berkley parser (Petrov & Klein Autosegmental-metricalappr.

Disco. fct. of pitch accents VP





2007) to obtain syntactic analysis (thanks to Adriane Boyd). Relating inton. & interpretation

Focus projection

PP

Constraining projection

◮ Converted corpus including An open issue

NP



A pilot study

◮ orthographic transcription IMS Radionews Corpus NP PP



◮ phonetic transcription Syntactic corpus annotation

Examples Bundesinnenminister Kanther hat sich gegen die Aufnahme weiterer Fl¸chtlinge aus Bosnien ausgesprochen .

◮ ToBI annotation PP with one accent bUn0 (L*H) d@s0 In1 n@n0 mi| kan1 (L*H) t@R0 hat1 zIC1 ge:1 (L*H) g@n0 di:1 aUf1 na:0 m@0 vaI1 t@0 R@R0 flYCt1 lIN0 N@0 aUs1 bOs1 (H*L) ni:0 @n0 aUs1 g@0 SpROx0 x@n0 (L%)|

NN NE VAFIN PRF APPR ART NN ADJA NN APPR NE VVPP $.

◮ syntactic analysis NP with multiple accents

Accents on every part

1

L*H

1

L*H

1 1 1

L*H

1 1 1 1 1 2

H*L

4

L%

0





An unexpected accent

into TiGer-XML so that it can be browsed and searched Multiple unexp. accents

Examples occurring twice

using TiGerSearch tool (Lezius 2002).

Conclusion



Appendix

ToBi

Challenging focus projection

Published empirical evidence









15 / 40

Focus in German:

NP with multiple accents Towards a Accents on every part of an NP

Corpus-Based Study

Kordula De Kuthy

PP

Detmar Meurers





Motivation

PP



Background

Expressing inf. struc

NP von den Versuchen auf dem Mururoa-Atoll

Characterizing intonation

fOn1 de:n1 fER0 zu:1 (L*H) x@n0 (-) aUf1 de:m1 mu:0 Ru:0 Ro:1 a:0 (L*H) a"t0 t|

Autosegmental-metricalappr.

APPR ART NN APPR ART NN

Disco. fct. of pitch accents

Der nordrhein-westf‰lische Ministerpr‰sident Rau 1 1 3 1 1 4

Relating inton. & interpretation

de:R1 nORt1 (H*L) RaIn0 vEst0 fE:1 lI| mi:0 nIs1 (!H*L) t@R0 pREz0 zi| RaU1 (L*HL) (%) L*H - L*H %

Focus projection

ART ADJA NN NE VAFIN

Constraining projection

1 1 1 2 S

An open issue

H*L !H*L L*HL %

A pilot study

IMS Radionews Corpus VP

VROOT Syntactic corpus annotation

Examples

NP NP

PP with one accent

S NP with multiple accents

Accents on every part

werde keinerlei Gef‰hrdung der Umwelt ausgehen

An unexpected accent

VP mu:0 Ru:0 Ro:1 a:0 (L*H) a"t0 t| ve:R1 d@0 kaI1 (H*L) n@R0 laI0 g@0 fE:R1 (H*L) dUN0 de:R1 Um1 (H*L) vElt0 aUs1 ge:0 @n0 (L%)

Multiple unexp. accents

VAFIN PIAT NN ART NN VVINF

Examples occurring twice

2p 1 1 1 1 4

NP PP Conclusion H*L H*L H*L L%



Appendix

hat den F¸hrungsstreit bei den Sozialdemokraten kritisiert . ToBi

¨ ¨

Die franzosische Regierung hat in einem offenen Brief an die Bevolkerung Australiens um Verstand-

¨

HL) (%) hat1 de:n1 fy:1 (H*L) RUNs0 StRaIt0 (-) baI1 de:n1 zo:0 tsi:0 a:l1 (H*L) de:0 mo:0 k| kRi:0 ti:0 zi:Rt1 (*?) (L%) Challenging focus projection

VAFIN ART NN APPR ART NN VVPP $. Published empirical evidence

¨

nis fur die geplanten Atomtests im Sudpazifik geworben. Außenminister [”d9[S]a[R]Et] versicherte in

¨

1 1 1 1 1 1 4 0 dem heute von der Zeitung Sydney Morning Herald veroffentlichten Schreiben, von den Versuchen

¨

H*L - H*L *? L% ¨

auf dem Mururoa-Atoll werde keinerlei Gefohrdung der Umwelt ausgehen.

17 / 40





An unexpected accented in an NP Multiple unexpected accents in a PP



S S







PP





VP

PP







NP PP

PP



Der SPD-Finanzexperte Poss setzt nach dem Steuerkompromi# mit der Regierungskoalition in Bonn jetzt

de:R1 Es1 (H*) pe:1 de:1 (L*H) fi:0 na| pOs1 (H*L) (%) zEtst1 na:x1 de:m1 StOY1 @R0 (L*H) kOm0 pRo:0| mIt1 de:R1 Re:0 gi:1 (L*H?) RUNs0 ko:0 al| In1 bOn1 (L*H) (H%) jEtst1 (L*H) (-)

NP ART NN NN VVFIN APPR ART NN APPR ART NN APPR NE ADV

1 1 4 1 1 1 3 1 1 1 1 4 3

H* L*H *? H*L % L*H - L*H? L*H H% L*H -





, es d¸rfe nicht zu einer Verunsicherung der Menschen kommen

Es1 dYR1 f@0 nICt1 (*?) tsu:1 aI1 n@R0 fER0 Un1 (H*L) zIx0 C@0 RUN| de:R1 mEn1 S@n0 kOm1 m@n0 (L%)

$, PPER VMFIN PTKNEG APPR ART NN ART NN VVINF

1 1 1 1 1 1 1 1 4

*? H*L L%

PP







NP PP

Der deutsche Sparkassen- und Giroverband hat davor gewarnt, die psychologischen und praktis-

¨ ¨ ¨

chen Probleme bei der Einfuhrung einer gemeinsamen europaischen Wahrung zu unterschatzen.¨

auf ein Ende der internen Streitigkeiten in seiner Partei

¨ ¨

Die Konvergenzkriterien mußten unbedingt eingehalten werden, betonte Kohler in einem Interview. jEtst1 (L*H) (-) aUf1 aIn1 En1 (H*L) d@0 (-) de:R1 In0 tER1 (H*L) n@n0 StRaIt1 tIC0 kaI0 t@n0 In1 zaI1 n@R0 paR0 (HH*L?) taI1 (L%)

¨

Bloße Tendenzen reichten dabei nicht aus, es durfe nicht zu einer Verunsicherung der Menschen APPR ART NN ART ADJA NN APPR PPOSAT NN

1 1 1 1 1 1 1 1 4

kommen. H*L - H*L HH*L? L%

Examples occurring twice with different accents (1,2) Examples occurring twice with different accents (1,2)

S







VP









NP PP









NP PP NP









Der Verband s¸dostasiatischer Staaten , [aI][si:][@n] , hat heute auf seiner Jahrestagung im Sultanat Brunei Vietnam aufgenommen

de:R1 fER0 bant1 (L*H?) zy:t0 Ost1 az0 zi:0 a:0 tIS0 S@| Sta:1 (L*!H) t@n0 (-) aI0 (L*H) si:0 @n0 (H%) hat1 hOY1 (H*L?) t@0 aUf1 zaI1 n@R0 ja:1 (L*!H) R@s0 ta:0 gUN0 (-) Im1 zUl0 ta:0 na:t1 bRu:1 naI0 (L*H) (-) vi:0 Et0 nam1 (H*L) aUf1 g@0 nOm0 m@n0 (L%)

ART NN ADJA NN $, NE $, VAFIN ADV APPR PPOSAT NN APPRART NN ADV NE VVPP

1 3 1 3 0 4 0 1 3 1 1 1 1 1 4 1 4

L*H? L*!H - L*H H% H*L? L*!H - L*H - H*L L%









S









VP







PP



NP



PP NP



NP





hat heute auf seiner Jahrestagung im Sultanat Brunei Vietnam aufgenommen

Der Verband s¸dostasiatischer Staaten , [a:][si:][a:n] , hat1 hOY1 (L*H) t@0 (-) aUf1 zaI1 n@R0 ja:1 (L*H) R@s0 ta:0 gUN0 Im1 zUl0 ta:0 na:t1 bRu:1 naI0 (L*H) (%) vi:0 Et0 nam1 (H*L) aUf1 g@0 nOm0 m@n0 (%)

de:R1 fER0 bant1 (L*H) zy:t0 Ost1 (H*L) az0 zi:0 a:0 tIS| Sta:1 (L*H) t@n0 (-) a:0 (L*H) si:0 a:n0 (%) VAFIN ADV APPR PPOSAT NN APPRART NN ADV NE VVPP

ART NN ADJA NN $, NE $, 1 3 1 1 1 1 1 3 1 4

2p 1 1 3 0 4 0 L*H - L*H L*H % H*L %

L*H H*L L*H - L*H %









Conclusion

Focus in German:

Towards a References Focus in German:

Towards a

Corpus-Based Study Corpus-Based Study

◮ In light of conflicting theory-driven proposals on the Kordula De Kuthy

Beckman, M. & J. Pierrehumbert (1986). Intonational Structure in Japanese and Kordula De Kuthy

Detmar Meurers Detmar Meurers



interaction of syntax, intonation, and information Motivation

English. Phonology Yearbook 3, 255–309.

Motivation

structure, it is crucial to strengthen the empirical basis. Birch, S. & C. Clifton, Jr (1995). Focus, Accent, and Argument Structure: Effects on

Background Background

Expressing inf. struc

Language Comprehension. Language and Speech 38(4), 365–391. Expressing inf. struc

◮ Spoken corpora are attractive source of authentic data Characterizing intonation

¨ ´

Buring, D. (2006). Focus projection and default prominence. In V. Molnar & Characterizing intonation

Autosegmental-metricalappr. Autosegmental-metricalappr.

◮ Targeted exploration requires annotated corpora. Disco. fct. of pitch accents

S. Winkler (eds.), The Architecture of Focus , Berlin: Mouton De Gruyter, Disco. fct. of pitch accents

Relating inton. & interpretation vol. 82 of Studies in Generative Grammar , pp. 321–346. Relating inton. & interpretation

◮ Where treebanks are not available, parsing is a viable Focus projection Focus projection

Chomsky, N. & M. Halle (1968). The Sound Pattern of English . New York, NY:

option for supporting syntactic corpus queries. Constraining projection

An open issue Harper and Row.

Constraining projection

An open issue





◮ Reported first observations on IMS Radionews Corpus A pilot study De Kuthy, K. (2002). Discontinuous NPs in German — A Case Study of the A pilot study

IMS Radionews Corpus IMS Radionews Corpus

◮ One finds more accents than assumed by traditional Syntactic corpus annotation

Interaction of Syntax, Semantics and Pragmatics . Stanford, CA: CSLI Syntactic corpus annotation

Examples Publications. URL http://ling.osu.edu/∼kdk/papers/dekuthy00.html. Examples

theories of focus projection. PP with one accent PP with one accent

NP with multiple accents De Kuthy, K. & W. D. Meurers (2003). The secret life of focus exponents, and what NP with multiple accents

◮ A number of accents occur in positions unexpected for Accents on every part it tells us about fronted verbal projections. In S. Muller (ed.), Proceedings of

¨ Accents on every part



standard theories of focus projection. An unexpected accent

Multiple unexp. accents

the Tenth Int. Conference on HPSG . Stanford, CA: CSLI Publications, pp. An unexpected accent

Multiple unexp. accents

97–110. URL http://purl.org/dm/papers/dekuthy-meurers-hpsg03.html.

◮ There is significant variation, even when the same Examples occurring twice Examples occurring twice





Conclusion Fery, C. (1993). German Intonational Patterns . No. 285 in Linguistische Arbeiten.

´ Conclusion

information in the same context is reported. ¨

Tubingen: Max Niemeyer Verlag.

Appendix Appendix

◮ Larger intonationally annotated corpora of spoken ToBi

Gussenhoven, C. (1983). Testing the reality of focus domains. Language and ToBi

Challenging focus projection Challenging focus projection

Speech 26, 61–80.

language in context are needed to test theories of the Published empirical evidence Published empirical evidence



Jacobs, J. (1988). Fokus-Hintergrund-Gliederung und Grammatik. In H. Altmann

interaction of information structure, syntax, and intonation. (ed.), Intonationsforschungen , Tubingen: Max Niemeyer Verlag, pp. 89–134.

¨

23 / 40 23 / 40

Kadmon, N. (2006). Some Theories of the Interpretation of Accent Placement. Focus in German: Focus in German:

Handout, colloquium at OSU, October 19, 2006. Towards a Phrasal Tones Towards a

Corpus-Based Study Corpus-Based Study

¨

Lezius, W. (2002). Ein Suchwerkzeug fur syntaktisch annotierte Textkorpora. Ph.D. Kordula De Kuthy Kordula De Kuthy

¨

thesis, IMS, University of Stuttgart. Arbeitspapiere des Instituts fur Maschinelle Detmar Meurers Detmar Meurers



Sprachverarbeitung (AIMS), vol 8, nr 4. Motivation Motivation

Petrov, S. & D. Klein (2007). Improved Inference for Unlexicalized Parsing. In Background Background

HLT/ACL 2007 . pp. 404–411. URL Expressing inf. struc Expressing inf. struc



http://www.aclweb.org/anthology/N/N07/N07-1051. Characterizing intonation Characterizing intonation





Pierrehumbert, J. (1980). The Phonetics and Phonology of English Intonation.

Autosegmental-metricalappr.

Disco. fct. of pitch accents

Phrasal tones will be assigned at every intermediate or Autosegmental-metricalappr.

Disco. fct. of pitch accents



Ph.d., MIT. Relating inton. & interpretation

Focus projection

intonation phrase: Relating inton. & interpretation

Focus projection

¨

Rapp, S. (1998). Automatisierte Erstellung von Korpora fur die Prosodieforschung. Constraining projection Constraining projection



Ph.D. thesis, University of Stuttgart. An open issue ◮ L- or H- phrase accent, which occurs at an intermediate An open issue



A pilot study A pilot study

Roberts, C. (2006). Resolving Focus. Conference abstract. Sinn und Bedeutung

IMS Radionews Corpus

phrase boundary (level 3 and above); IMS Radionews Corpus

11. Barcelona, Spain. URL http://www.upf.edu/dtf/sub11/abstracts/roberts.pdf. Syntactic corpus annotation Syntactic corpus annotation

Examples

◮ L% or H% (final) boundary tone, which occurs at every Examples

Schwarzschild, R. (1999). GIVENness, Avoid F and other Constraints on the PP with one accent PP with one accent

Placement of Focus. Natural Language Semantics 7(2), 141–177. URL NP with multiple accents full intonation phrase boundary (level 4) NP with multiple accents



http://www.rci.rutgers.edu/∼tapuz/nls98.pdf. Accents on every part Accents on every part

An unexpected accent An unexpected accent

Selkirk, E. (1995). Sentence Prosody: Intonation, Stress, and Phrasing. In J. A. Multiple unexp. accents Multiple unexp. accents



Goldsmith (ed.), The Handbook of Phonological Theory , Oxford: Basil Examples occurring twice Examples occurring twice





Blackwell, no. 1 in Blackwell handbooks in linguistics, chap. 16, pp. 550–569. Conclusion Conclusion



Uhmann, S. (1991). Fokusphonologie - Eine Analyse deutscher Appendix Appendix

ToBi ToBi

Intonationskonturen im Rahmen der nicht-linearen Phonologie , vol. 252 of

Challenging focus projection Challenging focus projection

Linguistische Arbeiten . Tubingen: Max Niemeyer Verlag.

¨ Published empirical evidence Published empirical evidence



Welby, P. (2003). Effects of Pitch Accent Position, Type, and Status on Focus

Projection. Language and Speech 46(1), 53–81.

24 / 40 24 / 40



Focus in German: Focus in German:

Pitch accents Towards a Break Index Values Towards a

Corpus-Based Study Corpus-Based Study

◮ Pitch accent tones are marked at every accented syllable. Kordula De Kuthy Kordula De Kuthy

Detmar Meurers Detmar Meurers

◮ Lack of pitch accent assignment for a syllable is

Motivation ◮ Values for the break index are: Motivation

interpreted as meaning that the syllable is not accented.

Background Background

◮ H* ‘peak accent’ Expressing inf. struc 0 for cases of clear phonetic marks of clitic groups; e.g. Expressing inf. struc

◮ an apparent tone target on the accented syllable which Characterizing intonation Characterizing intonation

Autosegmental-metricalappr.

the medial affricate in contractions of ‘did you’ or a flap Autosegmental-metricalappr.

is in the upper part of the speaker’s pitch range for the Disco. fct. of pitch accents

as in ‘got it’. Disco. fct. of pitch accents



phrase. Relating inton. & interpretation Relating inton. & interpretation

Focus projection 1 most phrase-medial word boundaries. Focus projection

◮ L* ‘low accent’ Constraining projection Constraining projection



◮

2 a strong disjuncture marked by a pause or virtual

an apparent tone target on the accented syllable which An open issue An open issue





is in the lowest part of the speaker’s pitch range. A pilot study pause, but with no tonal marks; i.e. a well-formed tune A pilot study



◮ L*+H ‘scooped accent’

IMS Radionews Corpus

Syntactic corpus annotation

continues across the juncture. OR a disjuncture that is IMS Radionews Corpus

Syntactic corpus annotation



◮ a low tone target on the accented syllable which is Examples weaker than expected at what is tonally a clear Examples

PP with one accent PP with one accent

immediately followed by relatively sharp rise to a peak in NP with multiple accents intermediate or full intonation phrase boundary. NP with multiple accents

Accents on every part Accents on every part

the upper part of the speaker’s pitch range. An unexpected accent

3 intermediate intonation phrase boundary; i.e. marked An unexpected accent

◮ L+H* ‘rising peak accent’ Multiple unexp. accents

by a single phrase tone affecting the region from the last Multiple unexp. accents

Examples occurring twice Examples occurring twice

◮ a high peak target on the accented syllable which is

Conclusion

pitch accent to the boundary. Conclusion

immediately preceded by relatively sharp rise from a 4 full intonation phrase boundary; i.e. marked by a final

Appendix Appendix

valley in the lowest part of the speaker’s pitch range. ToBi boundary tone after the last phrase tone. ToBi

◮ H+!H* Challenging focus projection Challenging focus projection

Published empirical evidence Published empirical evidence

◮ clear step down onto the accented syllable from a high

pitch which itself cannot be accounted for by a H phrasal

tone ending the preceding phrase or by a preceding H 25 / 40 26 / 40

Focus in German: Focus in German:

Recent proposals Towards a ¨

Buring: Focus projection and default prominence Towards a

Corpus-Based Study Corpus-Based Study

Questioning focus projection rules Kordula De Kuthy

Detmar Meurers

◮ Starting point: Selkirk (1995) Kordula De Kuthy

Detmar Meurers





Motivation

◮ F-marking: Motivation



Background ◮ An accented word is F-marked. Background

Expressing inf. struc ◮ Vertical Focus Projection: Expressing inf. struc

Characterizing intonation Characterizing intonation

Autosegmental-metricalappr. F-marking of head of phrase licenses F-marking of phrase. Autosegmental-metricalappr.



◮ ¨

Buring (2006): Focus Projection and Default Prominence Disco. fct. of pitch accents

◮ Horizontal Focus Projection: Disco. fct. of pitch accents

Relating inton. & interpretation Relating inton. & interpretation



[recent article] Focus projection F-marking of internal argument licenses F-marking of head. Focus projection

Constraining projection Constraining projection

An open issue

◮ Focus of the sentence (FOC): An open issue



◮ F-marked node not dominated by another F-marked node.

◮ Roberts (2006): Resolving Focus A pilot study A pilot study

IMS Radionews Corpus IMS Radionews Corpus



[Sinn & Bedeutung abstract] Syntactic corpus annotation

◮ F Interpretation: Syntactic corpus annotation

Examples ◮ constituent F-marked (but not FOC): New Examples

PP with one accent PP with one accent

◮ constituent not F-marked: Given

◮

NP with multiple accents NP with multiple accents

Kadmon (2006): Some Theories of the Interpretation of Accents on every part Accents on every part





Accent Placement [OSU talk handout] An unexpected accent ◮ ¨

Buring (2006): eliminates focus projection rules An unexpected accent

Multiple unexp. accents Multiple unexp. accents

Examples occurring twice Examples occurring twice

◮ No restrictions on vertical focus projection, i.e., any

Conclusion Conclusion

accent within a phrase can project focus to the phrase.

Appendix Appendix

ToBi ◮ Horizontal focus projection is the consequence of ToBi

Challenging focus projection Challenging focus projection

Published empirical evidence default prominence assignment, not focus projection. Published empirical evidence





⇒ No syntactic constraints on focus projection are needed.

27 / 40 28 / 40



Focus in German: Focus in German:

¨

Buring: Vertical focus projection Towards a ¨

Buring: Horizontal focus projection Towards a

Corpus-Based Study Corpus-Based Study

Kordula De Kuthy Kordula De Kuthy



◮ ¨

Buring (2006, based on Schwarzschild 1999) uses Detmar Meurers Detmar Meurers





Motivation Motivation

de-accenting examples to argue: focus can always project.

Background Background

Expressing inf. struc

◮ ¨

Horizontal Focus Projection (Buring 2006) Expressing inf. struc



(9) I know that John drove Mary’s red convertible. Characterizing intonation

Autosegmental-metricalappr.

◮ In configuration [A B ], one of A, B can be unaccented,

Characterizing intonation

Autosegmental-metricalappr.



But what did Bill drive? Disco. fct. of pitch accents

even though it is interpreted as F-marked. Disco. fct. of pitch accents

Relating inton. & interpretation Relating inton. & interpretation





(10) a. He drove [[his motorcycle]]F . Focus projection

Constraining projection

◮ ¨

Buring (2006) wants to derive this effect based on a

Focus projection

Constraining projection





He drove [[her blue convertible]]F .

An open issue An open issue

b. theory of default prominence.

A pilot study A pilot study



⇒ focus can project from adjectives/adjuncts

IMS Radionews Corpus

◮ idea of default prominence: default accent placement, IMS Radionews Corpus

Syntactic corpus annotation Syntactic corpus annotation

Examples

PP with one accent

independent of whether focus or background Examples

PP with one accent

◮ Similar de-accenting examples support focus projection NP with multiple accents NP with multiple accents

Accents on every part

◮ The idea is only sketched, not worked out. A default Accents on every part



from transitive subjects, indirect objects, adverbs, minor An unexpected accent

pattern suggested for English states that predicates An unexpected accent

Multiple unexp. accents Multiple unexp. accents

categories and headless structures—all of which had Examples occurring twice

don’t receive a pitch accent if an argument does. Examples occurring twice





been claimed to not support focus projection. Conclusion Conclusion



Appendix Appendix

◮ ¨

Vertical focus projection (Buring 2006): ToBi ToBi

Challenging focus projection Challenging focus projection



◮ F-marking of any daughter licenses F-marking of mother. Published empirical evidence Published empirical evidence









29 / 40 30 / 40

Focus in German: Focus in German:

Roberts (2006): Resolving Focus Towards a Kadmon (2006): Some theories of the Towards a

Corpus-Based Study Corpus-Based Study

Kordula De Kuthy

Detmar Meurers

interpretation of accent placement Kordula De Kuthy

Detmar Meurers

◮ Roberts (2006) eliminates focus projection entirely and

Motivation Motivation

instead proposes to relate accent placement to ◮ Parallel to Roberts (2006), Kadmon eliminates focus

Background Background

interpretation using the notion of retrievability. Expressing inf. struc projection entirely and instead relies on the notion of Expressing inf. struc

Characterizing intonation Characterizing intonation

◮ Core components: Autosegmental-metricalappr. expectable to relate accent placement and interpretation. Autosegmental-metricalappr.

Disco. fct. of pitch accents Disco. fct. of pitch accents

◮ Accentuation: Freely align pitch accents (in an Relating inton. & interpretation ◮ Core components: Relating inton. & interpretation

Focus projection Focus projection

independently generated prosodic structure) with words Constraining projection ◮ Interpretation of pitch accent placement: A word is Constraining projection



(in an independently generated syntactic structure). An open issue

interpreted as expectable iff it is unaccented.

An open issue



A pilot study A pilot study

◮ Retrievability presupposition: If a contentful constituent IMS Radionews Corpus

◮ An expression B is expectable in an utterance U iff the IMS Radionews Corpus





bears no accent, then its denotation is conventionally Syntactic corpus annotation

Examples

following holds: Syntactic corpus annotation

Examples



implicated to be retrievable. PP with one accent

NP with multiple accents

◮ Presented with the result of replacing B in U with a PP with one accent

NP with multiple accents



◮ Novelty Implicature of Focus: If a constituent bears an Accents on every part variable, it would be possible for the hearer to infer on Accents on every part

An unexpected accent

the basis of prior context that in the actual utterance, the An unexpected accent



accent, then its denotation is irretrievable. Multiple unexp. accents Multiple unexp. accents

Examples occurring twice position of that variable should be occupied by B . Examples occurring twice



◮ Elements without nuclear pitch accent which traditionally Conclusion Conclusion

◮ Elements without nuclear pitch accent which traditionally

were analyzed as part of a projected focus must be Appendix Appendix

ToBi

were analyzed as part of a projected focus must be ToBi

◮ retrievable, or Challenging focus projection Challenging focus projection

◮ expectable, or

◮ accented after all Published empirical evidence Published empirical evidence



◮ accented after all

31 / 40 32 / 40



Focus in German: Focus in German:

Roberts (2006)/Kadmon (2006): An Example Towards a Empirical Evidence: Gussenhoven (1983) [I] Towards a

Corpus-Based Study Corpus-Based Study

Kordula De Kuthy Kordula De Kuthy

Detmar Meurers Detmar Meurers

◮ Context-retrievability experiment: judge whether a question

Motivation Motivation



Background

and an answer are from the same or a different dialogue. Background

Expressing inf. struc Expressing inf. struc

Characterizing intonation

◮ Experiment includes two types of questions: Characterizing intonation

Autosegmental-metricalappr. Autosegmental-metricalappr.



(11) What did you do? (traditionally: wide, VP focus) Disco. fct. of pitch accents

Relating inton. & interpretation

(13) What do you do? (wide, VP focus) Disco. fct. of pitch accents

Relating inton. & interpretation





(12) a. I invited bill. (R/K claim: good without further context) Focus projection

Constraining projection (14) What do you teach? (narrow, NP focus)

Focus projection

Constraining projection





b. I invited bill. (R/K claim: bad without further context, An open issue An open issue



A pilot study A pilot study

good if party context) ◮ Two types of answers:

IMS Radionews Corpus IMS Radionews Corpus

Syntactic corpus annotation

Examples

(15) I teach linguistics. (accents on verb and NP) Syntactic corpus annotation

Examples



(How is it ensured that the focus must contain some accent, PP with one accent PP with one accent





even if Bill is also retrievable?)

NP with multiple accents

Accents on every part

(16) I teach linguistics. (accent on NP only) NP with multiple accents

Accents on every part

An unexpected accent An unexpected accent

Multiple unexp. accents ◮ Results Multiple unexp. accents

Examples occurring twice Examples occurring twice



Conclusion

◮ Listeners performed no better than chance in judging Conclusion



Appendix

whether questions and answers were matched. Appendix

ToBi ◮ This finding supports focus projection: To focus the VP, ToBi

Challenging focus projection Challenging focus projection

Published empirical evidence it is sufficient to accent the object NP. Published empirical evidence









33 / 40 34 / 40

Focus in German: Focus in German:

Empirical Evidence: Gussenhoven (1983) [II] Towards a Empirical Evidence: Birch & Clifton (1995) [I] Towards a

Corpus-Based Study Corpus-Based Study

Kordula De Kuthy ◮ Two sets of experiments, each consisting of: Kordula De Kuthy

◮ Second experiment tested dialogues with questions: Detmar Meurers Detmar Meurers

◮ make-sense judgment task (appropriateness of dialogue):

Motivation Motivation

(17) Please tell me what happened that night? response times to make yes/no-judgments measured.

Background Background

(wide, VP focus) Expressing inf. struc

◮ linguistic judgment task: rate prosodic appropriateness Expressing inf. struc

Characterizing intonation

on Likert scale (1–5) Characterizing intonation



(18) What do you remember from the last lesson? Autosegmental-metricalappr. Autosegmental-metricalappr.

Disco. fct. of pitch accents

◮ Experiment 1: Questions and answers used in dialogues: Disco. fct. of pitch accents



(narrow, NP focus) Relating inton. & interpretation

Focus projection

Relating inton. & interpretation

Focus projection

Constraining projection (21) a. Isn’t Kerry pretty smart? (wide, VP focus) Constraining projection



◮ The answers contain “non-merging predicates”: An open issue

b. Isn’t Kerry good at math? (V focus) An open issue



A pilot study A pilot study



(19) I remember nothing. (accent on V and NP)

IMS Radionews Corpus

(22) a. Yes, she teaches math. IMS Radionews Corpus





b. Yes, she teaches math.

Syntactic corpus annotation Syntactic corpus annotation

Examples Examples



(20) I remember nothing. (accent on NP only) PP with one accent

NP with multiple accents

c. Yes, she teaches math. PP with one accent

NP with multiple accents

Accents on every part Accents on every part

◮ Results: Listeners matched An unexpected accent ◮ Results for wide, VP focus question (21a): An unexpected accent

Multiple unexp. accents Multiple unexp. accents

◮ narrow focus questions (18) with answers accenting Examples occurring twice

◮ make-sense judgment: same reaction times for answers Examples occurring twice





only the NP (20) Conclusion with accent on both V and NP (22a) or only on NP (22b) Conclusion

◮ linguistic judgment: subjects preferred answers with

◮ wide focus questions (17) with answers accenting both Appendix Appendix

ToBi ToBi

the verb and the NP (19) Challenging focus projection

accent on both V and NP (22a) over only on NP (22b) Challenging focus projection

Published empirical evidence

⇒ Birch & Clifton (1995) interpret this a saying that Published empirical evidence



⇒ Non-merging predicates do not allow focus projection. accenting the verb of a focused VP is optional.

35 / 40 36 / 40



Focus in German: Focus in German:

Empirical Evidence: Birch & Clifton (1995) [II] Towards a Empirical Evidence: Welby (2003) Towards a

Corpus-Based Study Corpus-Based Study

Kordula De Kuthy ◮ Welby (2003) investigates the influence of prosodic Kordula De Kuthy



◮

Detmar Meurers Detmar Meurers

Experiment 2 used questions supporting VP focus and phrasing in the Birch & Clifton (1995) setup.

Motivation Motivation

answers with “non-lexical” argument NPs (quantifiers):

Background Background

◮ Questions: VP focus (25a) or object-NP focus (25b):

(23) What can you tell me about the math program at Expressing inf. struc

Characterizing intonation

Expressing inf. struc

Characterizing intonation



Cornell this year? Autosegmental-metricalappr. (25) a. What’s that terrible smell coming from the Autosegmental-metricalappr.

Disco. fct. of pitch accents Disco. fct. of pitch accents

neighbors’ yard?

(24) a. They accepted no one. Relating inton. & interpretation

Focus projection

Relating inton. & interpretation

Focus projection

Constraining projection b. There’s a terrible smell coming form the Constraining projection



b. They accepted no one. An open issue

neighbors’ yard. What are they burning?

An open issue



A pilot study A pilot study

◮ Results: IMS Radionews Corpus IMS Radionews Corpus

Syntactic corpus annotation ◮ There were four possible answer types: Syntactic corpus annotation



◮ make-sense judgment: faster response times for answers Examples Examples





(26) a. They’re burning

PP with one accent PP with one accent

with accent on both V and NP (24a) than for NP only (24b). NP with multiple accents their garbage. (verb) NP with multiple accents





◮linguistic judgment: no preference for answers with

Accents on every part

An unexpected accent

H* L-L% Accents on every part

An unexpected accent





their garbage.

Multiple unexp. accents Multiple unexp. accents

accents on V and NP (24a) over only on NP (24b) Examples occurring twice b. They’re burning (obj-NP) Examples occurring twice





⇒ Conclusion: Accented lexically filled argument NPs Conclusion H* L-L% Conclusion



project focus, while non-lexically filled ones do not. Appendix

ToBi

c. They’re burning their garbage. (“hat”) Appendix

ToBi

◮ Parallel to Gussenhoven’s “non-merging predicates” results Challenging focus projection H* H* L-L% Challenging focus projection



◮ Reverse linguistic judgment results remain as a puzzle. Published empirical evidence Published empirical evidence



d. They’re burning their garbage. (two peak)

H* L- H* L-L%

37 / 40 38 / 40

Focus in German: Focus in German:

Empirical Evidence: Welby (2003) Results Towards a ´

Empirical Evidence: Fery (1993) Towards a

Corpus-Based Study Corpus-Based Study

Kordula De Kuthy ◮ Context-retrievability experiment: judge whether a question Kordula De Kuthy

Detmar Meurers Detmar Meurers





Motivation

and an answer are from the same or a different dialogue. Motivation



Background

◮ Experiment includes two types of questions: Background

◮ The experiment used the linguistic judgment task of Expressing inf. struc Expressing inf. struc



Birch & Clifton (1995), rating using Likert scale. Characterizing intonation (27) Wer ist verhaftet worden? Characterizing intonation

Autosegmental-metricalappr. Autosegmental-metricalappr.

Disco. fct. of pitch accents who has arrested been Disco. fct. of pitch accents



◮ Results for questions supporting VP-focus and for those Relating inton. & interpretation Relating inton. & interpretation

Focus projection (narrow, NP focus) Focus projection



supporting object-NP-focus were identical. Constraining projection

An open issue

Constraining projection

An open issue





◮ “hat” pattern and the object-NP-only accent pattern A pilot study A pilot study



were rated as equally appropriate.

IMS Radionews Corpus ¨

(28) Hast Du heute die Nachrichten gehort? IMS Radionews Corpus

Syntactic corpus annotation Syntactic corpus annotation



◮ Appropriateness of hat pattern → prenuclear pitch Examples have you today the news heard Examples

PP with one accent PP with one accent



accent does not affect focus structure interpretation. NP with multiple accents (wide focus) NP with multiple accents

Accents on every part Accents on every part

◮ Two-peak pattern was less acceptable An unexpected accent An unexpected accent

Multiple unexp. accents Multiple unexp. accents

→ Two-peak pattern is disfavored for single focus Examples occurring twice Examples occurring twice



interpretation (favors double focus interpretation). ◮ Answer recorded twice, once for each question:

Conclusion Conclusion



Appendix

ToBi

(29) gorbatschov ist verhaftet worden. Appendix

ToBi

Challenging focus projection

Published empirical evidence

Gorbachev has arrested been Challenging focus projection

Published empirical evidence







◮ Result: Listeners decided at random

39 / 40 40 / 40