Japanese <i>wh</i>-questions always exhibit focus intonation (FI). Furthermore, the domain of FI exhibits a correspondence to the <i>wh</i>-scope. I propose that this phonology-semantics correspondence is a result of the cyclic computation of FI, which is explained under the notion of <i>Multiple Spell-Out</i> in the recent Minimalist framework. The proposed analysis makes two predictions: (1) embedding of an FI into another is possible; (2) (overt) movement of a <i>wh</i>-phrase to a phase edge position causes a mismatch between FI and <i>wh</i>-scope. Both predictions are tested experimentally, and shown to be borne out.

The 10th volume of the working paper series contains two papers contributed by SFB-members. The first paper “Single prosodic phrase sentences” by Caroline Féry (A1) and Heiner Drenhaus (C6, University of Potsdam) investigates the prosody of Wide Focus Partial Fronting in a series of production and perception experiments. The second paper “Focus Asymmetries in Bura” by Katharina Hartmann, Peggy Jacob (B2, Humboldt University Berlin) and Malte Zimmermann (A5, University of Potsdam) explores the strategies of marking focus in Bura (Chadic).

This paper surveys a range of constructions in which prosody affects discourse function and discourse structure.We discuss English tag questions, negative polar questions, and what we call “focus” questions. We postulate that these question types are complex speech acts and outline an analysis in Segmented Discourse Representation Theory (SDRT) to account for the interactions between prosody and discourse.

New evidence is provided for a grammatical principle that singles out contrastive focus (Rooth 1996; Truckenbrodt 1995) and distinguishes it from discourse-new “informational” focus. Since the prosody of discourse-given constituents may also be distinguished from discourse-new, a three-way distinction in representation is motivated. It is assumed that an F-feature marks just contrastive focus (Jackendoff 1972, Rooth 1992), and that a G-feature marks discoursegiven constituents (Féry and Samek-Lodovici 2006), while discoursenew is unmarked. A crucial argument for G-marking comes from second occurrence focus (SOF) prosody, which arguably derives from a syntactic representation where SOF is both F-marked and G-marked. This analysis relies on a new G-Marking Condition specifying that a contrastive focus may be G-marked only if the focus semantic value of its scope is discourse-given, i.e. only if the contrast itself is given.

Prosodi kan sammanfattas som talets rytmiska, dynamiska och melodiska aspekter. Utan prosodi skulle talet förefalla monotont och kommunikationen kunna kompliceras. Syftet med föreliggande studie var att undersöka prosodiska aspekter av nonordsproduktion hos barn med cochleaimplantat och barn med språkstörning. Tidigare studier av prosodi hos nämnda grupper har inte i detalj beskrivit vilka typer av fel som förekommer varför detta var intressant att undersöka. Föreliggande studie baseras på redan insamlat material av nonordsrepetition av sammanlagt 41 barn vilka deltagit i tidigare studier. Av dessa var 27 stycken barn med språkstörning i åldrarna 4:6-7:6 år och fjorton var barn med CI i åldrarna 3:0-13:4. Det inspelade materialet transkriberades och felanalyserades gruppvis och medelvärden för grupperna räknades ut. Barnen med CI uppvisade större prosodiska problem vid nonordsrepetition än barnen med språkstörning. Såväl barnen med språkstörning som barnen med CI uppvisade svårigheter med stavelseproduktion, betoningsplacering, ordaccent och vokalkvantitet. Dock gjorde barnen med CI generellt större antal fel samt fler olika typer av fel jämfört med barnen med språkstörning. Alla barnen med CI gjorde någon typ av prosodiskt fel medan 11 av 27 de barnen med språkstörning producerade korrekt prosodi på samtliga nonord. Barn som hade en högre ålder vid implantation uppvisade större prosodiska svårigheter än barn som implanterats vid en lägre ålder / Prosody can be defined as the rythmic, dynamic and melodic aspects of speech. Without prosody, speech would sound monotonous and communication could be obstructed. The aim of the present study was to examine prosodic aspects of nonword repetition by children with cochlear implants and children with language impairment. Previous studies of prosody in these groups have not in detail described what kinds of errors that occur and therefore this is interesting to investigate. The present study is based on previously collected data of nonword repetition among a total of 41 children, all participating in previous studies. Of these children, 27 were children with language impairment aged between 4:6-7:6 years. Fourteen were children with CI aged between 3:0-13:4. The recorded data was transcribed and analyzed group wise and the mean value of the groups were calculated. Children with CI showed prosodic problems in nonword repetition to a greater extent than children with language impairment did. Both children with language impairment and children with CI had difficulties with number of syllables, stress, tonal word accent and quantity of vowel. However, the children with CI generally made errors to a greater extent as well as more types of errors, compared to the children with language impairment. All children with CI made some type of prosodic error whilst 11 out of 27 of the language impaired children produced all the words prosodically correct. Children who were implanted at an older age showed greater difficulties with prosody than children who were implanted at a younger age.

In a production experiment and two follow-up perception experiments on read German we investigated the (de-)coding of discourse-new, inferentially and textually accessible and given discourse referents by prosodic means. Results reveal that a decrease in the referent’s level of givenness is reflected by an increase in its prosodic prominence (expressed by differences in the status and type of accent used) providing evidence for the relevance of different intermediate types of information status between the poles given and new. Furthermore, perception data indicate that the degree of prosodic prominence can serve as the decisive cue for decoding a referent’s level of givenness.

Languages differ in the means they avail themselves of for the structural realization of information structure, where available options are word order, prosody and morphology. Turkish has long been characterized as predominantly using word order and its variation in realizing information structure, where certain positions in a sentence are associated with certain pragmatic functions related to information structure. Prosody has been proposed to play only a secondary role interacting with word order. Contrary to this widely established view, the thesis argues that the notion that sentential positions have pragmatic functions and word order variation is a syntactic means to realize these functions can be abandoned, without any loss of explanatory power, in favor of a tune-based perspective where prosody is the sole structural determinant of information structure. In this setting word order variations are argued to be prosodically motivated, in that Turkish phonology imposes some precedence constraints on intonational contours. Word order variation then turns out to be just a consequence as opposed to being a determinant in attaining the right information structure required by the discourse context. To substantiate these claims a tune-based account, based on Steedman&#039 / s account of English information structure, is proposed for the structural realization of information structure in Turkish, whereby information structural units are directly associated with prosodic phrases intonationally marked in certain ways. Validity of the account is tried to be established by intonational analysis of recorded speech data. As for the explanatory value, the information structure phenomena that has received positional explanation in the relevant literature, are tried to be captured only in prosodic terms, without committing to positions, syntactic strategies and such.

For the last several years, speaker diarization has been attracting substantial research attention as one of the spoken language technologies applied for the improvement, or enrichment, of recording transcriptions. Recordings of meetings, compared to other domains, exhibit an increased complexity due to the spontaneity of speech, reverberation effects, and also due to the presence of overlapping speech. Overlapping speech refers to situations when two or more speakers are speaking simultaneously. In meeting data, a substantial portion of errors of the conventional speaker diarization systems can be ascribed to speaker overlaps, since usually only one speaker label is assigned per segment. Furthermore, simultaneous speech included in training data can eventually lead to corrupt single-speaker models and thus to a worse segmentation. This thesis concerns the detection of overlapping speech segments and its further application for the improvement of speaker diarization performance. We propose the use of three spatial cross-correlationbased parameters for overlap detection on distant microphone channel data. Spatial features from different microphone pairs are fused by means of principal component analysis, linear discriminant analysis, or by a multi-layer perceptron. In addition, we also investigate the possibility of employing longterm prosodic information. The most suitable subset from a set of candidate prosodic features is determined in two steps. Firstly, a ranking according to mRMR criterion is obtained, and then, a standard hill-climbing wrapper approach is applied in order to determine the optimal number of features. The novel spatial as well as prosodic parameters are used in combination with spectral-based features suggested previously in the literature. In experiments conducted on AMI meeting data, we show that the newly proposed features do contribute to the detection of overlapping speech, especially on data originating from a single recording site. In speaker diarization, for segments including detected speaker overlap, a second speaker label is picked, and such segments are also discarded from the model training. The proposed overlap labeling technique is integrated in Viterbi decoding, a part of the diarization algorithm. During the system development it was discovered that it is favorable to do an independent optimization of overlap exclusion and labeling with respect to the overlap detection system. We report improvements over the baseline diarization system on both single- and multi-site AMI data. Preliminary experiments with NIST RT data show DER improvement on the RT ¿09 meeting recordings as well. The addition of beamforming and TDOA feature stream into the baseline diarization system, which was aimed at improving the clustering process, results in a bit higher effectiveness of the overlap labeling algorithm. A more detailed analysis on the overlap exclusion behavior reveals big improvement contrasts between individual meeting recordings as well as between various settings of the overlap detection operation point. However, a high performance variability across different recordings is also typical of the baseline diarization system, without any overlap handling.

Prosody, the structures governing the pitch and rhythm of speech, is essential to the correct and authentic use of a language. Unfortunately, many students of a second language find it difficult to learn these patterns, particularly when they differ significantly from their mother tongue. Several pedagogical models are now widely used, but only within the confines of a conventional classroom environment. Dramatisation, the process of preparing a full theatrical performance, has already shown itself to be useful in the acquisition of second language grammar and vocabulary. This thesis demonstrates by means of a qualitative case study that this approach to language teaching also results in a statistically significant improvement in the area of prosody. The results were obtained by analysing the digitised speech of four students who participated in a course employing dramatisation in order to teach French as a second language.

<p>In this corpus-based study, the two verb pairs arise and rise and awake and wake are investigated. The paper focuses on seven research questions that are related to the meanings of the verbs in question, the semantic specialisations of those verbs, and the semantic relation</p><p>of the specific verb pair constituents. Furthermore, tendencies of language change are investigated, and an attempt is made to generalise over the influence of the prefix a- on those</p><p>tendencies.</p><p>The results of the quantitative and qualitative analysis show that the verbs awake and wake are more synonymous than the verbs arise and rise. It seems as if due to this difference, the two verbs arise and awake are subject to different processes of language change that take</p><p>their development into different directions. The observations made about the characteristics of the prefix a- that is involved in the two verbs arise and awake are often ambiguous and inconclusive. Therefore, the influence of this prefix on the processes of language change needs to be analysed further by investigating more word pairs distinguished by the absence or presence of the prefix a-.</p>