Refine
Year of publication
Document Type
- Conference Proceeding (411)
- Part of a Book (332)
- Article (253)
- Book (27)
- Doctoral Thesis (19)
- Working Paper (14)
- Other (7)
- Contribution to a Periodical (6)
- Review (3)
- Part of Periodical (2)
Language
- English (1078) (remove)
Is part of the Bibliography
- no (1078) (remove)
Keywords
- Deutsch (225)
- Korpus <Linguistik> (213)
- Computerlinguistik (118)
- Englisch (78)
- Konversationsanalyse (69)
- Annotation (59)
- Automatische Sprachanalyse (50)
- Interaktion (42)
- Semantik (42)
- Gesprochene Sprache (41)
Publicationstate
- Veröffentlichungsversion (559)
- Postprint (148)
- Zweitveröffentlichung (112)
- Preprint (2)
- Ahead of Print (1)
- Erstveröffentlichung (1)
Reviewstate
- Peer-Review (397)
- (Verlags)-Lektorat (303)
- Qualifikationsarbeit (Dissertation, Habilitationsschrift) (16)
- Peer-review (14)
- Verlags-Lektorat (9)
- Review-Status-unbekannt (5)
- Peer-Revied (3)
- Abschlussarbeit (Bachelor, Master, Diplom, Magister) (Bachelor, Master, Diss.) (2)
- Zweitveröffentlichung (2)
- (Verlags-)Lektorat (1)
Publisher
- de Gruyter (60)
- Benjamins (50)
- IDS-Verlag (47)
- Springer (42)
- Association for Computational Linguistics (35)
- European Language Resources Association (ELRA) (32)
- Institut für Deutsche Sprache (24)
- European Language Resources Association (22)
- Elsevier (21)
- Niemeyer (20)
Ungoliant: An optimized pipeline for the generation of a very large-scale multilingual web corpus
(2021)
Since the introduction of large language models in Natural Language Processing, large raw corpora have played a crucial role in Computational Linguistics. However, most of these large raw corpora are either available only for English or not available to the general public due to copyright issues. Nevertheless, there are some examples of freely available multilingual corpora for training Deep Learning NLP models, such as the OSCAR and Paracrawl corpora. However, they have quality issues, especially for low-resource languages. Moreover, recreating or updating these corpora is very complex. In this work, we try to reproduce and improve the goclassy pipeline used to create the OSCAR corpus. We propose a new pipeline that is faster, modular, parameterizable, and well documented. We use it to create a corpus similar to OSCAR but larger and based on recent data. Also, unlike OSCAR, the metadata information is at the document level. We release our pipeline under an open source license and publish the corpus under a research-only license.
The aim of this work is to describe criteria used in the process of inclusion and treatment of neologisms in dictionaries of Spanish within the framework of pandemic instability. Our starting point will be data obtained by the Antenas Neológicas Network (https://www.upf.edu/web/antenas), whose representation in three different lexicographic tools will be analyzed with the purpose of identifying problems in the methodology used to dictionarize – that is, how and what words were selected to be included in dictionaries and how they were represented in their entries – neologisms during the COVID-19 pandemic (sources and corpora of analysis, selection criteria, types of definition, among other aspects). Two of them are monolingual and COVID-19 lexical units were included as part of their updates: the Antenario, a dictionary of neologisms of Spanish varieties, and the Diccionario de la Lengua Española [DLE], a dictionary of general Spanish, published by the Real Academia Española [RAE], Spanish Royal Academy). The other is a bilingual unidirectional English-Spanish dictionary first published as a glossary, Diccionario de COVID-19 EN-ES [TREMEDICA], entirely made up of neological and non-neological lexical units related to the virus and the pandemic. Thus, the target lexis was either included in existing works or makes up the whole of a new tool located in a portal together with other lexicographic tools. Unlike other collections of COVID-19 vocabulary that kept cropping up as the pandemic unfolded, all three have been designed and written according to well-established lexicographic practices.
Our working hypothesis is that the need to record and define words which were recently created impacts the criteria for inclusion and treatment of neologisms in dictionaries about Spanish, including a certain degree of overlap of some features which are traditionally thought to be specific to each type of dictionary.
This paper outlines the generation process of a specifi computational linguistic representation termed the Multilingual Time Map, conceptually a multi-tape finit state transducer encoding linguistic data at different levels of granularity. The fi st component acquires phonological data from syllable labeled speech data, the second component define feature profiles the third component generates feature hierarchies and augments the acquired data with the define feature profiles and the fourth component displays the Multilingual Time Map as a graph.
Although the N400 was originally discovered in a paradigm designed to elicit a P300 (Kutas and Hillyard, 1980), its relationship with the P300 and how both overlapping event-related potentials (ERPs) determine behavioral profiles is still elusive. Here we conducted an ERP (N = 20) and a multiple-response speed-accuracy tradeoff (SAT) experiment (N = 16) on distinct participant samples using an antonym paradigm (The opposite of black is white/nice/yellow with acceptability judgment). We hypothesized that SAT profiles incorporate processes of task-related decision-making (P300) and stimulus-related expectation violation (N400). We replicated previous ERP results (Roehm et al., 2007): in the correct condition (white), the expected target elicits a P300, while both expectation violations engender an N400 [reduced for related (yellow) vs. unrelated targets (nice)]. Using multivariate Bayesian mixed-effects models, we modeled the P300 and N400 responses simultaneously and found that correlation between residuals and subject-level random effects of each response window was minimal, suggesting that the components are largely independent. For the SAT data, we found that antonyms and unrelated targets had a similar slope (rate of increase in accuracy over time) and an asymptote at ceiling, while related targets showed both a lower slope and a lower asymptote, reaching only approximately 80% accuracy. Using a GLMM-based approach (Davidson and Martin, 2013), we modeled these dynamics using response time and condition as predictors. Replacing the predictor for condition with the averaged P300 and N400 amplitudes from the ERP experiment, we achieved identical model performance. We then examined the piecewise contribution of the P300 and N400 amplitudes with partial effects (see Hohenstein and Kliegl, 2015). Unsurprisingly, the P300 amplitude was the strongest contributor to the SAT-curve in the antonym condition and the N400 was the strongest contributor in the unrelated condition. In brief, this is the first demonstration of how overlapping ERP responses in one sample of participants predict behavioral SAT profiles of another sample. The P300 and N400 reflect two independent but interacting processes and the competition between these processes is reflected differently in behavioral parameters of speed and accuracy.
Preface
(2019)
Preface
(2020)
Physicists look at language
(2006)
This paper aims at verifying if the most important online Brazilian Portuguese dictionaries include some of the neologisms identified in texts published in the 1990s to 2000s, formed with the elements ciber-, e-, bio-, eco- and narco, which we refer to as fractomorphemes / fracto-morphèmes. Three online dictionaries were analyzed (Aulete, Houaiss and Michaelis), as well as Vocabulário Ortográfico da Língua Portuguesa (VOLP). We were able to conclude that all three dictionaries and VOLP include neologisms with these elements; Michaelis and VOLP do not include separate entries for bound morphemes, whereas Houaiss includes entries for all of them and Aulete includes entries for bio-, eco- and narco-. Aulete also describes the neological meaning of eco- and narco-, whereas Houaiss does not.
This study investigates cross-language differences in pitch range and variation in four languages from two language groups: English and German (Germanic) and Bulgarian and Polish (Slavic). The analysis is based on large multi-speaker corpora (48 speakers for Polish, 60 for each of the other three languages). Linear mixed models were computed that include various distributional measures of pitch level, span and variation, revealing characteristic differences across languages and between language groups. A classification experiment based on the relevant parameter measures (span, kurtosis and skewness values for pitch distributions for each speaker) succeeded in separating the language groups.
This study presents the results of a large-scale comparison of various measures of pitch range and pitch variation in two Slavic (Bulgarian and Polish) and two Germanic (German and British English) languages. The productions of twenty-two speakers per language (eleven male and eleven female) in two different tasks (read passages and number sets) are compared. Significant differences between the language groups are found: German and English speakers use lower pitch maxima, narrower pitch span, and generally less variable pitch than Bulgarian and Polish speakers. These findings support the hypothesis that inguistic communities tend to be characterized by particular pitch profiles.
Based on specific linguistic landmarks in the speech signal, this study investigates pitch level and pitch span differences in English, German, Bulgarian and Polish. The analysis is based on 22 speakers per language (11 males and 11 females). Linear mixed models were computed that include various linguistic measures of pitch level and span, revealing characteristic differences across languages and between language groups. Pitch level appeared to have significantly higher values for the female speakers in the Slavic than the Germanic group. The male speakers showed slightly different results, with only the Polish speakers displaying significantly higher mean values for pitch level than the German males. Overall, the results show that the Slavic speakers tend to have a wider pitch span than the German speakers. But for the linguistic measure, namely for span between the initial peaks and the non-prominent valleys, we only find the difference between Polish and German speakers. We found a flatter intonation contour in German than in Polish, Bulgarian and English male and female speakers and differences in the frequency of the landmarks between languages. Concerning “speaker liveliness” we found that the speakers from the Slavic group are significantly livelier than the speakers from the Germanic group.
New KARL (Knowledge Acquisition and Representation Language) allows to specify all parts of a problem-solving method (PSM). It is a formal language with a well-defined semantics and thus allows to represent PSMs precisely and unambiguously yet abstracting from implementation detail. In this paper it is shown how the language KARL has been modified and extended to New KARL to better meet the needs for the representation of PSMs. Based on a conceptual structure of PSMs new language primitives are introduced for KARL to specify such a conceptual structure and to support the configuration of methods. An important goal for this extension was to preserve three important properties of KARL: to be (i) a conceptual, (ii) a formal, and (iii) an executable language.
An ongoing academic and research program, the “Vocabula Grammatica” lexicon, implemented by the Centre for the Greek Language (Thessaloniki, Greece), aims at lemmatizing all the philological, grammatical, rhetorical, and metrical terms in the written texts of scholars (philologists and scholiasts) who curated the ancient Greek literature from the beginning of the Hellenistic period (4th/3rd c. BC) until the end of the Byzantine era (15th c. AD). In particular, it aspires to fill serious gaps (a) in the study of ancient Greek scholarship and (b) in the lexicography of the ancient Greek language and literature. By providing specific examples, we will highlight the typical and methodological features of the forthcoming dictionary.
Prominence has been widely studied on the word level and the syllable level. An extensive study comparing the two approaches is missing in the literature. This study investigates how word and syllable prominence relate to each other in German. We find that perceptual ratings based on the word level are more extreme than those based on the syllable level. The correlations between word prominence and acoustic features are greater than the correlations between syllable prominence and acoustic features.
The current paper presents a corpus containing 35 dialogues of spontaneously spoken southern German, including half an hour of articulography for 13 of the speakers. Speakers were seated in separate recording chambers, mimicking a telephone call, and recorded on individual audio channels. The corpus provides manually corrected word boundaries and automatically aligned segment boundaries. Annotations are provided in the Praat format. In addition to audio recordings, speakers filled out a detailed questionnaire, assessing among others their audio-visual consumption habits.
In our study we use the experimental framework of priming to manipulate our subjects’ expectations of syllable prominence in sentences with a well-defined syntactic and phonological structure. It shows that it is possible to prime prominence patterns and that priming leads to significant differences in the judgment of syllable prominence.
The perception of prosodic prominence is influenced by different sources like different acoustic cues, linguistic expectations and context. We use a generalized additive model and a random forest to model the perceived prominence on a corpus of spoken German. Both models are able to explain over 80% of the variance. While the random forests give us some insights on the relative importance of the cues, the general additive model gives us insights on the interaction between different cues to prominence.
A frequently replicated finding is that higher frequency words tend to be shorter and contain more strongly reduced vowels. However, little is known about potential differences in the articulatory gestures for high vs. low frequency words. The present study made use of electromagnetic articulography to investigate the production of two German vowels, [i] and [a], embedded in high and low frequency words. We found that word frequency differently affected the production of [i] and [a] at the temporal as well as the gestural level. Higher frequency of use predicted greater acoustic durations for long vowels; reduced durations for short vowels; articulatory trajectories with greater tongue height for [i] and more pronounced downward articulatory trajectories for [a]. These results show that the phonological contrast between short and long vowels is learned better with experience, and challenge both the Smooth Signal Redundancy Hypothesis and current theories of German phonology.
Streefkerk defines prominence as the perceptually outstanding parts in spoken language. An optimal rating scale for syllable prominence has not been found yet. This paper evaluates a 4-point, an 11-point, a 31-point, and a continuous scale for the rating of syllable prominence and gives support for scales using a higher number of levels. Priming effects found by Arnold, et al., could only be replicated using the 31-point scale.
In previous research we showed that the priming paradigm can be used to significantly alter the prominence ratings of subjects. In that study we only looked at the changes in the subjects’ ratings. In the present study, we analyzed the acoustic parameters of the stimuli used in the priming study and investigated the correlation between prominence ratings and acoustic parameters. The results show that priming has a significant effect on these correlations. The contribution of acoustic features on perceived prominence was found to depend on the prominence pattern. If a dominantly prominent syllable is present in a given utterance, f0 and intensity contribute most to the perceived prominence, while duration contributes most when no syllable is dominantly prominent.
A trainable prosodic model called SFC (Superposition of Functional Contours), proposed by Holm and Bailly, is here confronted to German intonation. Training material is the publicly available Siemens Synthesis Corpus that provides spoken utterances for high-quality speech synthesis. We describe the labeling framework and first evaluation results that compares the original prosody of test sentences of this corpus with their prosodic rendering by the proposed model and state-of-the-art systems available on-line on the web.
The classification of verbs in Levin's (1993) English Verb Classes and Alternations: A preliminary Investigation, on the basis of both intuitive semantic grouping and their participation in valence alternations, is often used by the NLP community as evidence of the semantic similarity of verbs (Jing & McKeown 1998; Lapata & Brew 1999; Kohl et al. 1998). In this paper, we compare the Levin classification with the work of the FrameNet project (Fillmore & Baker 2001), where words (not just verbs) are grouped according to the conceptual structures (frames) that underlie them and their combinatorial patterns are inductively derived from corpus evidence. This means that verbs grouped together in FrameNet (FN) might be semantically similar but have different (or no) alternations, and that verbs which share the same alternation might be represented in two different semantic frames.
Playing videogames is a popular social activity; people play videogames in different places, on different media, in different situations, alone or with partners, online or offline. Unsurprisingly, they thereby share space (physically or virtually) with other playing or non-playing people. The special issue investigates through different contexts and settings how non-players become participants of the gaming interaction and how players and non-players co-construct presence. The introduction provides a problem-related context for the individual contributions and then briefly presents them.
This paper investigates situations in French videogame interactions where non-players who share the same physical space as players, participate in the gaming activities as spectators. Through a detailed multimodal and sequential analysis, we show that being a spectator is a local achievement of all co-present participants - players and non-players.
Among the many peculiarities of the German tense system which make its description or reconstruction such a difficult task to perform, there is one outstanding stumbling-block, viz. the relation between the - morphologically simple - Preterite and the - compound - (Present-)Perfect. Disregarding problems of variety - in spoken German in the South, the Preterite either doesn't exist or is restricted to modal verbs, e.g.
In the first part of this contribution, we will present, as a starting point for the following discussions, a simple formal language P containing one stative predicate. We will then discuss, on an intuitive level, how a treatment of predicates of change could be conceived, and how the progressive could be rendered in a formal language.
We will then give a formal definition of a language, TP1, based on P, and we will construct a semantics for TP1, which incorporates the ideas discussed.
As the Web ought to be considered as a series of sources rather than as a source in itself, a problem facing corpus construction resides in meta-information and categorization. In addition, we need focused data to shed light on particular subfields of the digital public sphere. Blogs are relevant to that end, especially if the resulting web texts can be extracted along with metadata and made available in coherent and clearly describable collections.
While adjusting to the COVID-19 pandemic, people around the world started to talk about the “new normal” way of life, and they conveyed feelings and thoughts on the topic through social networks and traditional communication channels resorting to a set of specific linguistic strategies, such as metaphors and neologisms. The vocabulary in different domains and in everyday speech was expanded to accommodate a complex social, cultural, and professional phenomenon of changes. Therefore, this new life gave birth to a new language – the “coronaspeak”. According to Thorne (2020), the “coronaspeak” has three stages: first, it emerged in the way medical aspects were communicated in everyday language; secondly, it occurred when speakers verbalized the experiences they had undergone and “invented their own terms”; finally, this “new” way of speaking emerged in the government and authorities’ jargon, to ensure that the new rules and policies were understood, and that population adopted socially responsible behaviours.
In this paper, we will focus on the second stage, because we intend to take stock of how speakers communicate and verbalize this new way of living, particularly on social networks, for example. Alongside, we are interested in the context in which the neologism – be it a new word, a new meaning, or a new use – emerged, is used, and understood, through the observation of the occurrence of the new word(s) either on social networks or through dissemination texts (press) to confront it with the ones that Portuguese digital dictionaries have attested so far. Different criteria regarding the insertion of new units, the inclusion date, and the lexicographic description of the entries in the dictionaries will be debated.
Linguistics is facing the challenge of many other sciences as it continues to grow into increasingly complex subfields, each with its own separate or overarching branches. While linguists are certainly aware of the overall structure of the research field, they cannot follow all developments other than those of their subfields. It is thus important to help specialists but also newcomers alike to bushwhack through evolved or unknown territory of linguistic data. A considerable amount of research data in linguistics is described with metadata. While studies described and published in archived journals and conference proceedings receive a quite homogeneous set of metadata tags — e.g., author, title, publisher —, this does not hold for the empirical data and analyses that underlie such studies. Moreover, lexicons, grammars, experimental data, and other types of resources come in different forms; and to make things worse, their description in terms of metadata is also not uniform, if existing at all. These problems are well-known and there are now a number of international initiatives — e.g., CLARIN, FlareNet, MetaNet, DARIAH — to build infrastructures for managing linguistic resources. The NaLiDa project, funded by the German Research Foundation, aims at facilitating the management and access to linguistic resources originating from German research institutions. In cooperation with the German SFB 833 research center, we are developing a combination of faceted and full-text search to give integrated access through heterogeneous metadata sets. Our approach is supported by a central registry for metadata field descriptors, and a component repository for structured groups of data categories as larger building blocks.
The long road to a historical dictionary of Lower Sorbian. Towards a lexical information system
(2022)
The Sorbian Institute has been taking preparatory steps for a historical-documentary vocabulary information system for Lower Sorbian for about 10 years. To this end, the entire extant written material (16th–21st centuries) of this strongly endangered European minority language is to be systematically evaluated. An attempt made a few years ago to organise and finance the project as a long-term scientific project was not successful in the end. Therefore, it can only be advanced step by step and via some detours. The article informs about the interim status of the project, especially with respect to the creation of a reliable database.
This paper contributes to the growing body of knowledge on current listeners' responses in talk-in-interaction. In particular, it complements earlier findings on double sayings of German JA by describing some additional prosodic-phonetic parameters and a visual feature of its realization in institutional and semi-private interaction (doctor-patient interaction, Big Brother, TV talk shows). These include pitch contour, pitch range and phonetic ending, on the one hand, and nodding on the other. The paper shows that JAJA is a truly multimodal phenomenon, with the individual features accomplishing interactional functions across sequence-organizational habitats, including re)claiming epistemic priority in an aside, making continuation relevant, agreeing/ acknowledging with reservation and aligning with the continuation of a sequence. Lack of nodding is suggested to have situational as well as misalignment reasons. On the basis of its observations, the paper also raises the question whether it is the applicability of response token variants across action and sequence types which makes them memorizable despite their variability.
This paper presents the concept of the "participant perspective" as an approach to the study of spoken language. It discusses three aspects of this concept and shows that they can offer helpful tools in spoken language research. Employing the participant perspective provides us with an alternative to many of the approaches currently in use in the study of spoken language in that it favours small-scale, qualitative research that aims to uncover categories relevant for the participants. Its results can usefully complement large-scale studies of phenomena on all linguistic dimensions of talk.
In conversation, speakers need to plan and comprehend language in parallel in order to meet the tight timing constraints of turn taking. Given that language comprehension and speech production planning both require cognitive resources and engage overlapping neural circuits, these two tasks may interfere with one another in dialogue situations. Interference effects have been reported on a number of linguistic processing levels, including lexicosemantics. This paper reports a study on semantic processing efficiency during language comprehension in overlap with speech planning, where participants responded verbally to questions containing semantic illusions. Participants rejected a smaller proportion of the illusions when planning their response in overlap with the illusory word than when planning their response after the end of the question. The obtained results indicate that speech planning interferes with language comprehension in dialogue situations, leading to reduced semantic processing of the incoming turn. Potential explanatory processing accounts are discussed.
When humans have a conversation with one-another, they generally take turns speaking one after the other without overlapping each others talk or leaving silence between turns for long stretches of time. Previous research has shown that conversation is a structured practice following rules that help interlocutors to manage the flow of conversation interactively. While at the beginning of a conversation it remains open who will speak when about what and for how long, interlocutors regulate the flow of conversation as it unfolds. One basic set of rules that interlocutors operate with governs the allocation of speaking turns, with the central rule stating that whoever starts speaking first at a point in time when speaker change becomes relevant has the rights and obligations to produce the next turn. The organization of turn allocation, therefore, is one reason for conversational turn taking to be so remarkably fast, with the beginnings of turns most often being quite accurately aligned with the ends of the previous turns. Observations of this outstanding speed of turn taking gave rise to a number of questions concerning language processing in conversational situations. The studies presented in this thesis investigate some of these questions from the perspective of the current listener preparing to be the next speaker who will respond to the current turn.
The study presented in Chapter 2 investigates when next speakers begin to plan their own turn with respect to two points in time, (i) the moment when the incoming turn’s message becomes clear enough to make response planning possible and (ii) the moment when the incoming turn terminates. Results of previous studies were inconclusive about the timing of language planning in conversation, with evidence in favour of both late and early response planning. Furthermore, previous studies presented both evidence as well as counter evidence indicating that response planning depends or does not depend on an accurate prediction of the timing of the incoming turn’s end. The study presented here makes use of a novel experimental paradigm which includes a dialogic task that participants need to fulfil in response to critical utterances by a confederate. These critical utterances were structured, on the one hand, so that their message became clear either only at the end of the turn or before the end of the turn, and, on the other hand, so that it was either predictable or not predictable when exactly the turn would end. Participant’s eye-movements as well as their response latencies indicated that they always planned their next turn as early as possible, irrespective of the predictability of the incoming turn’s end. The presented results provide evidence in favour of models of turn taking that predict speech planning to happen in overlap with the incoming turn.
Having established that next speakers begin to plan their turn in overlap, the study presented in Chapter 3 goes more into detail investigating to which depth language planning progresses while the incoming turn is still unfolding. To this end, a number of psycholinguistic paradigms were combined. In the study’s main experiment, participants had to fulfil a switch-task in which they switched from picture naming in response to an auditorily presented question to making a lexical decision. By manipulating the relatedness of the word for lexical decision with the picture that was prepared to be named before the task-switch it was possible to draw inferences on which processing stages were entered during the speech production process in overlap with the incoming turn. Participants’ behavioural responses in the lexical decision task revealed that they entered the stage of phonological encoding while the incoming turn was still unfolding, showing that planning in overlap is not limited to conceptual preparation but includes all sub-processes of formulation.
Given that speech production regularly enters the stages of formulation in overlap with the incoming turn, as shown in Chapters 2 and 3, the question arises whether planning the next turn in overlap is cognitively more demanding than during the gap between turns. This question is approached in the study presented in Chapter 4 by measuring pupillometric responses of participants in a dialogic task. An increase in pupil diameter during a cognitive task is indicative of increased processing load, and pupillometric responses to planning in overlap with the incoming turn were found to be greater than responses to planning in the gap between turns. These results show that planning in overlap is more demanding than planning during the gap, even though it is highly practiced by speakers.
After Chapters 2 to 4 investigated the timing and mechanisms of speech planning in conversation, Chapter 5 turns towards the timing of articulation of a planned turn, asking the question what sources of information next speakers use to time the articulation of a planned utterance to start closely after the incoming turn comes to an end. In this Chapter’s study, participants taking turns with a confederate responded to utterances containing or not containing different cues to the location of the incoming turn’s end. Participants made use of lexical and turn-final intonational cues, but not of turn-initial intonational cues, responding faster when the relevant cues were present than when they were not present. These results show that the timing of turn initiation in next speakers depends on the recognition of the incoming turn’s point of completion and not merely on the progress in planning the next turn.
All evidence presented in Chapters 2 to 5 is summed up and bundled together in a cognitive model of turn taking, which is being presented in Chapter 6. This model assumes, centrally, that the planning of a turn and the timing of its articulation are separate cognitive processes that run in parallel in any next speaker during conversation. Planning generally starts as early as possible, often in overlap with the incoming turn, while the timing of articulation depends on the next speaker’s level of certainty that speaker change has become relevant at a particular moment, with a number of cues to the end of the incoming turn leading to an increase of certainty. Next turns are assumed to often be planned down to fully formulated utterance plans including their phonological form as early as possible on the basis of anticipations of the incoming turn’s message, which are created with the help of the general and situational knowledge about the world, the current speaker and her intentions, as well as the input that has been received so far. The level of certainty that speaker change becomes relevant rises or decreases as lexico-syntactic, prosodic, and pragmatic projections about the development of the current turn are fulfilled or not fulfilled. As the incoming turn progresses towards its end as was projected by the current listener, he becomes certain that speaker change becomes relevant and will initiate articulation of the prepared next turn. Viewing these two processes, planning a next turn and timing of its articulation, as separate makes it possible to explain the observable fast timing of turn taking while still modelling the allocation of turns as interactionally managed by interlocutors — a considerable advantage of the presented model compared to more traditional perspectives on turn taking and conversation.
To ensure short gaps between turns in conversation, next speakers regularly start planning their utterance in overlap with the incoming turn. Three experiments investigate which stages of utterance planning are executed in overlap. E1 establishes effects of associative and phonological relatedness of pictures and words in a switch-task from picture naming to lexical decision. E2 focuses on effects of phonological relatedness and investigates potential shifts in the time-course of production planning during background speech. E3 required participants to verbally answer questions as a base task. In critical trials, however, participants switched to visual lexical decision just after they began planning their answer. The task-switch was time-locked to participants' gaze for response planning. Results show that word form encoding is done as early as possible and not postponed until the end of the incoming turn. Hence, planning a response during the incoming turn is executed at least until word form activation.
In conversation, turn-taking is usually fluid, with next speakers taking their turn right after the end of the previous turn. Most, but not all, previous studies show that next speakers start to plan their turn early, if possible already during the incoming turn. The present study makes use of the list-completion paradigm (Barthel et al., 2016), analyzing speech onset latencies and eye-movements of participants in a task-oriented dialogue with a confederate. The measures are used to disentangle the contributions to the timing of turn-taking of early planning of content on the one hand and initiation of articulation as a reaction to the upcoming turn-end on the other hand. Participants named objects visible on their computer screen in response to utterances that did, or did not, contain lexical and prosodic cues to the end of the incoming turn. In the presence of an early lexical cue, participants showed earlier gaze shifts toward the target objects and responded faster than in its absence, whereas the presence of a late intonational cue only led to faster response times and did not affect the timing of participants' eye movements. The results show that with a combination of eye-movement and turn-transition time measures it is possible to tease apart the effects of early planning and response initiation on turn timing. They are consistent with models of turn-taking that assume that next speakers (a) start planning their response as soon as the incoming turn's message can be understood and (b) monitor the incoming turn for cues to turn-completion so as to initiate their response when turn-transition becomes relevant.
Speech planning is a sophisticated process. In dialog, it regularly starts in overlap with an incoming turn by a conversation partner. We show that planning spoken responses in overlap with incoming turns is associated with higher processing load than planning in silence. In a dialogic experiment, participants took turns with a confederate describing lists of objects. The confederate’s utterances (to which participants responded) were pre-recorded and varied in whether they ended in a verb or an object noun and whether this ending was predictable or not. We found that response planning in overlap with sentence-final verbs evokes larger task-evoked pupillary responses, while end predictability had no effect. This finding indicates that planning in overlap leads to higher processing load for next speakers in dialog and that next speakers do not proactively modulate the time course of their response planning based on their predictions of turn endings. The turn-taking system exerts pressure on the language processing system by pushing speakers to plan in overlap despite the ensuing increase in processing load.
In conversation, interlocutors rarely leave long gaps between turns, suggesting that next speakers begin to plan their turns while listening to the previous speaker. The present experiment used analyses of speech onset latencies and eye-movements in a task-oriented dialogue paradigm to investigate when speakers start planning their responses. German speakers heard a confederate describe sets of objects in utterances that either ended in a noun [e.g., Ich habe eine Tür und ein Fahrrad (“I have a door and a bicycle”)] or a verb form [e.g., Ich habe eine Tür und ein Fahrrad besorgt (“I have gotten a door and a bicycle”)], while the presence or absence of the final verb either was or was not predictable from the preceding sentence structure. In response, participants had to name any unnamed objects they could see in their own displays with utterances such as Ich habe ein Ei (“I have an egg”). The results show that speakers begin to plan their turns as soon as sufficient information is available to do so, irrespective of further incoming words.
Having found their way onto the computer screens, comics soon branched into webcomics. These kept a lot of the characteristics of print comic books, but gradually adapted new unexplored modes of representation. Three relatively new ‘enhancements’ to the medium of comics are presented in this article: webcomics enhanced through the use of the infinite canvas, as proposed by Scott McCloud, those enhanced with videos and/or sound, and lastly those enhanced with interactive and ludic elements. All of the mentioned push the medium of comics into new waters, and by doing so they add new layers of meaning and modify their structure based on the make-up of the implemented features. Infinite canvas manages to lift some limitations of print comics without changing the overall feel too drastically, while animated and voiced webcomics, as well as interactive or game comics, have a much higher inclination to transgress into domains of other media and transform themselves in order to accommodate and integrate these novel foreign features.
In this paper, we present first results of training a classifier for discriminating Russian texts into different levels of difficulty. For the classification we considered both surface-oriented features adopted from readability assessments and more linguistically informed, positional features to classify texts into two levels of difficulty. This text classification is the main focus of our Levelled Study Corpus of Russian (LeStCoR), in which we aim to build a corpus adapted for language learning purposes – selecting simpler texts for beginner second language learners and more complex texts for advanced learners. The most discriminative feature in our pilot study was a lexical feature that approximates accessibility of the vocabulary by the second language learner in terms of the proportion of familiar words in the texts. The best feature setting achieved an accuracy of 0.91 on a pilot corpus of 209 texts.
The goal of the MULI (MUltiLingual Information structure) project is to empirically analyse information structure in German and English newspaper texts. In contrast to other projects in which information structure is annotated and investigated (e.g. in the Prague Dependency Treebank, which mirrors the basic information about the topic-focus articulation of the sentence), we do not annotate theory-biased categories like topic-focus or theme-rheme. Trying to be as theory-independent as possible, we annotate those features which are relevant to information structure and on the basis of which typical patterns, co-occurrences or correlations can be determined. We distinguish between three annotation levels: syntax, discourse and prosody. The data is based on the TIGER Corpus for German and the Penn Treebank for English, since the existing information on part-of-speech and syntactic structure can be re-used for our purposes. The actual annotation of an English example sequence illustrates our choice of categories on each level. Their combination offers the possibility to investigate how information structure is realised and can be interpreted.
We present the annotation of information structure in the MULI project. To learn more about the information structuring means in prosody, syntax and discourse, theory- independent features were defined for each level. We describe the features and illustrate them on an example sentence. To investigate the interplay of features, the representation has to allow for inspecting all three layers at the same time. This is realised by a stand-off XML mark-up with the word as the basic unit. The theory-neutral XML stand-off annotation allows integrating this resource with other linguistic resources such as the Tiger Treebank for German or the Penn treebank for English.
We present an approach on how to investigate what kind of semantic information is regularly associated with the structural markup of scientific articles. This approach addresses the need for an explicit formal description of the semantics of text-oriented XML-documents. The domain of our investigation is a corpus of scientific articles from psychology and linguistics from both English and German online available journals. For our analyses, we provide XML-markup representing two kinds of semantic levels: the thematic level (i.e. topics in the text world that the article is about) and the functional or rhetorical level. Our hypothesis is that these semantic levels correlate with the articles’ document structure also represented in XML. Articles have been annotated with the appropriate information. Each of the three informational levels is modelled in a separate XML document, since in our domain, the different description levels might conflict so that it is impossible to model them within a single XML document. For comparing and mining the resulting multi-layered XML annotations of one article, a Prolog-based approach is used. It focusses on the comparison of XML markup that is distributed among different documents. Prolog predicates have been defined for inferring relations between levels of information that are modelled in separate XML documents. We demonstrate how the Prolog tool is applied in our corpus analyses.
The KorAP project (“Korpusanalyseplattform der nächste Generation”, “Corpus-analysis platform of the next generation”), carried out at the Institut fUr Deutsche Sprache (IDS) in Mannheim, Germany, has as its goal the development of a modem, state-of-the-art corpus-analysis platform, capable of handling very large corpora and opening the perspectives for innovative linguistic research. The platform will facilitate new linguistic findings by making it possible to manage and analyse extremely large amounts of primary data and annotations, while at the same time allowing an undistorted view of the primary un-annotated text, and thus fully satisfying expectations associated with a scientific tool. The project started in July 2011 and is funded till June 2014. The demo presentation in December will be the first version following a preliminary feature freeze, and will open the alpha testing phase of the project.
The present article describes the first stage of the KorAP project, launched recently at the Institut für Deutsche Sprache (IDS) in Mannheim, Germany. The aim of this project is to develop an innovative corpus analysis platform to tackle the increasing demands of modern linguistic research. The platform will facilitate new linguistic findings by making it possible to manage and analyse primary data and annotations in the petabyte range, while at the same time allowing an undistorted view of the primary linguistic data, and thus fully satisfying the demands of a scientific tool. An additional important aim of the project is to make corpus data as openly accessible as possible in light of unavoidable legal restrictions, for instance through support for distributed virtual corpora, user-defined annotations and adaptable user interfaces, as well as interfaces and sandboxes for user-supplied analysis applications. We discuss our motivation for undertaking this endeavour and the challenges that face it. Next, we outline our software implementation plan and describe development to-date.
The present contribution addresses an infrastructural issue of universal relevance, addressed in the specific context of the TEI. We describe a combination of open-source tools and an open-access approach to creating knowledge repositories that have been employed in building a bibliographic reference library for the “TEI for Linguists” special interest group (LingSIG). The authors argue that, for an initiative such as the TEI, it is important to choose open, freely available solutions. If these solutions have the advantage of attracting new users and promoting the initiative itself, so much the better, especially if it is done in a non-committal way: no one using the LingSIG bibliographic repository has to be a member of the LingSIG or a “TEI-er” in general.
The paper presents best practices and results from projects in four countries dedicated to the creation of corpora of computer-mediated communication and social media interactions (CMC). Even though there are still many open issues related to building and annotating corpora of that type, there already exists a range of accessible solutions which have been tested in projects and which may serve as a starting point for a more precise discussion of how future standards for CMC corpora may (and should) be shaped like.
The paper presents best practices and results from projects in four countries dedicated to the creation of corpora of computer-mediated communication and social media interactions (CMC). Even though there are still many open issues related to building and annotating corpora of that type, there already exists a range of accessible solutions which have been tested in projects and which may serve as a starting point for a more precise discussion of how future standards for CMC corpora may (and should) be shaped like.
Converting and Representing Social Media Corpora into TEI: Schema and best practices from CLARIN-D
(2016)
The paper presents results from a curation project within CLARIN-D, in which an existing lMWord corpus of German chat communication has been integrated into the DEREKO and DWDS corpus infrastructures of the CLARIN-D centres at the Institute for the German Language (IDS, Mannheim) and at the Berlin-Brandenburg Academy of Sciences (BBAW, Berlin). The focus is on the solutions developed for converting and representing the corpus in a TEI format.
The paper reports the results of the curation project ChatCorpus2CLARIN. The goal of the project was to develop a workflow and resources for the integration of an existing chat corpus into the CLARIN-D research infrastructure for language resources and tools in the Humanities and the Social Sciences (http://clarin-d.de). The paper presents an overview of the resources and practices developed in the project, describes the added value of the resource after its integration and discusses, as an outlook, to what extent these practices can be considered best practices which may be useful for the annotation and representation of other CMC and social media corpora.
The paper presents an XML schema for the representation of genres of computer-mediated communication (CMC) that is compliant with the encoding framework defined by the TEI. It was designed for the annotation of CMC documents in the project Deutsches Referenzkorpus zur internetbasierten Kommunikation (DeRiK), which aims at building a corpus on language use in the most popular CMC genres on the German-speaking Internet. The focus of the schema is on those CMC genres which are written and dialogic―such as forums, bulletin boards, chats, instant messaging, wiki and weblog discussions, microblogging on Twitter, and conversation on “social network” sites.
The schema provides a representation format for the main structural features of CMC discourse as well as elements for the annotation of those units regarded as “typical” for language use on the Internet. The schema introduces an element <posting>, which describes stretches of text that are sent to the server by a user at a certain point in time. Postings are the main constituting elements of threads and logfiles, which, in our schema, are the two main types of CMC macrostructures. For the microlevel of CMC documents (that is, the structure of the <posting> content), the schema introduces elements for selected features of Internet jargon such as emoticons, interaction words and addressing terms. It allows for easy anonymization of CMC data for purposes in which the annotated data are made publicly available and includes metadata which are necessary for referencing random excerpts from the data as references in dictionary entries or as results of corpus queries.
Documentation of the schema as well as encoding examples can be retrieved from the web at http://www.empirikom.net/bin/view/Themen/CmcTEI. The schema is meant to be a core model for representing CMC that can be modified and extended by others according to their own specific perspectives on CMC data. It could be a first step towards an integration of features for the representation of CMC genres into a future new version of the TEI Guidelines.
Taking a usage-based perspective, lexical-semantic relations and other aspects of lexical meaning are characterised as emerging from language use. At the same time, they shape language use and therefore become manifest in corpus data. This paper discusses how this mutual influence can be taken into account in the study of these relations. An empirically driven methodology is proposed that is, as an initial step, based on self-organising clustering of comprehensive collocation profiles. Several examples demonstrate how this methodology may guide linguists in explicating implicit knowledge of complex semantic structures. Although these example analyses are conducted for written German, the overall methodology is language-independent.
Empirical synchronic language studies generally seek to investigate language phenomena for one point in time, even though this point in time is often not stated explicitly. Until today, surprisingly little research has addressed the implications of this time-dependency of synchronic research on the composition and analysis of data that are suitable for conducting such studies. Existing solutions and practices tend to be too general to meet the needs of all kinds of research questions. In this theoretical paper that is targeted at both corpus creators and corpus users, we propose to take a decidedly synchronic perspective on the relevant language data. Such a perspective may be realised either in terms of sampling criteria or in terms of analytical methods applied to the data. As a general approach for both realisations, we introduce and explore the FReD strategy (Frequency Relevance Decay) which models the relevance of language events from a synchronic perspective. This general strategy represents a whole family of synchronic perspectives that may be customised to meet the requirements imposed by the specific research questions and language domain under investigation.
The paper discusses from various angles the morphosyntactic annotation of DeReKo, the Archive of General Reference Corpora of Contemporary Written German at the Institut für Deutsche Sprache (IDS), Mannheim. The paper is divided into two parts. The first part covers the practical and technical aspects of this endeavor. We present results from a recent evaluation of tools for the annotation of German text resources that have been applied to DeReKo. These tools include commercial products, especially Xerox' Finite State Tools and the Machinese products developed by the Finnish company Connexor Oy, as well as software for which academic licenses are available free of charge for academic institutions, e.g. Helmut Schmid's Tree Tagger. The second part focuses on the linguistic interpretability of the corpus annotations and more general methodological considerations concerning scientifically sound empirical linguistic research. The main challenge here is that unlike the texts themselves, the morphosyntactic annotations of DeReKo do not have the status of observed data; instead they constitute a theory and implementation-dependent interpretation. In addition, because of the enormous size of DeReKo, a systematic manual verification of the automatic annotations is not feasible. In consequence, the expected degree of inaccuracy is very high, particularly wherever linguistically challenging phenomena, such as lexical or grammatical variation, are concerned. Given these facts, a researcher using the annotations blindly will run the risk of not actually studying the language but rather the annotation tool or the theory behind it. The paper gives an overview of possible pitfalls and ways to circumvent them and discusses the opportunities offered by using annotations in corpus-based and corpus-driven grammatical research against the background of a scientifically sound methodology.
Our paper describes an experiment aimed to assessment of lexical coverage in web corpora in comparison with the traditional ones for two closely related Slavic languages from the lexicographers’ perspective. The preliminary results show that web corpora should not be considered ― inferior, but rather ― different.
A comparison between morphological complexity measures: typological data vs. language corpora
(2016)
Language complexity is an intriguing phenomenon argued to play an important role in both language learning and processing. The need to compare languages with regard to their complexity resulted in a multitude of approaches and methods, ranging from accounts targeting specific structural features to global quantification of variation more generally. In this paper, we investigate the degree to which morphological complexity measures are mutually correlated in a sample of more than 500 languages of 101 language families. We use human expert judgements from the World Atlas of Language Structures (WALS), and compare them to four quantitative measures automatically calculated from language corpora. These consist of three previously defined corpus-derived measures, which are all monolingual, and one new measure based on automatic word-alignment across pairs of languages. We find strong correlations between all the measures, illustrating that both expert judgements and automated approaches converge to similar complexity ratings, and can be used interchangeably.
Ein integriertes Datenbank-, Such- und Tagging-Tool (IDaSTo) wird vorgestellt, das sich besonders für Variablenanalysen, für Paralleltexte und für diachronische Untersuchungen eignet. Relevante Kategorien bzw. Variablen können individuell definiert, Tags frei im Text und auf verschiedenen Wegen gesetzt und ihre Häufigkeiten in den verlinkten Statistiken direkt abgerufen werden.
Conversation is usually considered to be grammatically simple, while academic writing is often claimed to be structurally complex, associated primarily with a greater use of dependent clauses. Our goal in the present paper is to challenge these stereotypes, based on the results of large-scale corpus investigations. We argue that both conversation and professional academic writing are grammatically complex but that their complexities are dramatically different. Surprisingly, the traditional view that complexity is realized through extensive clausal embedding leads to the conclusion that conversation is more complex than academic writing. In contrast, written academic discourse is actually much more ‘compressed’ than elaborated, and the complexities of academic writing are realized mostly as phrasal embedding rather than embedded clauses.
This is a study of how aspects of information structure can be captured within a formal grammar of Spanish, couched in the framework of Head-Driven Phrase Structure Grammar (HPSG, Pollard
and Sag 1994). While a large number of morphological, syntactic and semantic aspects in a variety of languages have been successfully analysed in this theory, information structure has not been paid the same attention in the HPSG literature. However, as a theory of signs, HPSG should include all
levels of description without which the structural descriptions offered by the grammar would ultimately remain incomplete. Languages often explicitly mark the information-structural partitioning of utterances. Depending on the particular language, linguistic resources used for this purpose include
prosody (stress/intonation), syntax (e. g. constituent order, special syntactic constructions) and morphology (e. g. special affixes). In HPSG, phonological, syntactic, semantic and pragmatic information is represented in parallel, which would seem to be a well-suited architecture for modelling
the sort of interfaces called for.
Language of Responsibility. The Influence of Linguistic Abstraction on Collective Moral Emotions
(2017)
Two experiments investigated the effects of linguistic abstractness on the experience of collective moral emotions. In Experiment 1 participants were presented with two scenarios about ingroup misbehavior, phrased using descriptive action verbs, interpretative action verbs, adjectives or nouns. The results show that participants experienced slightly more negative moral emotions with higher levels of linguistic abstractness. In Experiment 2 we also tested for the influence of national identification on the relationship between linguistic abstractness and emotional reactions. Additionally, we expanded the number of scenarios. Experiment 2 replicated the earlier pattern, but found larger differences between conditions. The strength of national identification did not moderate the observed effects. The results of this research are discussed within the context of the linguistic category model and psychology of collective moral emotions.
Uncertain about Uncertainty: Different ways of processing fuzziness in digital humanities data
(2014)
The GeoBib project is constructing a georeferenced online bibliography of early Holocaust and camp literature published between 1933 and 1949 (Entrup et al. 2013a). Our immediate objectives include identifying the texts of interest in the first place, composing abstracts for them, researching their history, and annotating relevant places and times. Relations between persons, texts, and places will be visualized using digital maps and GIS software as an integral part of the resulting GeoBib information portal. The combination of diverse data from varying sources not only enriches our knowledge of these otherwise mostly forgotten texts; it also confronts us with vague, uncertain or even conflicting information. This situation yields challenges for all researchers involved – historians, literary scholars, geographers and computer scientists alike. While the project operates at the intersection of historical and literary studies, the involved computer scientists are in charge of providing a working environment (Entrup et al. 2013b) and processing the collected information in a way that is formalized yet capable of dealing with inevitable vagueness, uncertainty and contradictions. In this paper we focus on the problems and opportunities of encoding and processing fuzzy data.
XML-based technologies offer powerful resources for open source applications in the field of e-learning. The paper describes a model of hypertext as interlinked structures that can be intertwined by cross-annotation linking. This infrastructure integrates multiple perspectives and allows creating a personal learning environment. We exemplify the approach in a case study: the Hamlet project. In the course of this project, several German translations of William Shakespeare’s Hamlet have been collected and annotated. Two different annotation layers are used to achieve a cross-linking reference between the various German translations. We will describe the theoretical background of cross-annotation linking and the actual technological implementation of the system. Additionally, we will use the personas method to gain insights into the potential benefit of the system as a personal learning environment.
We report results from an exploratory study of college students’ conceptions of poetry in which we asked them to name three things they expect from a poem. Frequency- and list-based analyses of their responses revealed that they primarily expect poems to rhyme, but they also identified a number of form-, content-, and reception-related genre expectations, which we discuss in relation to relevant previous research. We propose that rhyme’s predominance in college students’ genre expectations reflects its perceptual and cognitive salience during incremental poetry comprehension rather than its frequency in contemporary poetic practice. Our results characterize the genre conceptions of the population that empirical studies of poetry comprehension typically investigate, and thus provide relevant background information for the interpretation of empirical
findings in this field.
We examined genre-specific reading strategies for literary texts and hypothesized that text categorization (literary prose vs. poetry) modulates both how readers gather information from a text (eye movements) and how they realize its phonetic surface form (speech production). We recorded eye movements and speech while college students (N = 32) orally read identical texts that we categorized and formatted as either literary prose or poetry. We further varied the text position of critical regions (text-initial vs. text-medial) to compare how identical information is read and articulated with and without context; this allowed us to assess whether genre-specific reading strategies make differential use of identical context information. We observed genre-dependent differences in reading and speaking tempo that reflected several aspects of reading and articulation. Analyses of regions of interests revealed that word-skipping increased particularly while readers progressed through the texts in the prose condition; speech rhythm was more pronounced in the poetry condition irrespective of the text position. Our results characterize strategic poetry and prose reading, indicate that adjustments of reading behavior partly reflect differences in phonetic surface form, and shed light onto the dynamics of genre-specific literary reading. They generally support a theory of literary comprehension that assumes distinct literary processing modes and incorporates text categorization as an initial processing step.
This paper describes the lexical database tool LOLA (Linguistic-Oriented Lexical database Approach) which has been developed for the construction and maintenance of lexicons for the machine translation system LMT. First, the requirements such a tool should meet are discussed, then LMT and the lexical information it requires, and some issues concerning vocabulary acquisition are presented. Afterwards the architecture and the components of the LOLA system are described and it is shown how we tried to meet the requirements worked out earlier. Although LOLA originally has been designed and implemented for the German-English LMT prototype, it aimed from the beginning at a representation of lexical data that can be reused for other LMT or MT prototypes or even other NLP applications. A special point of discussion will therefore be the adaptability of the tool and its components as well as the reusability of the lexical data stored in the database for the lexicon development for LMT or for other applications.
Connectives are conjunctions, prepositions, adverbs and other particles which share the function of encoding semantic relations between sentences, or rather, between semantic objects some of which can be meanings of sentences. The relata linked by any such relation will fall into one of four distinct categories: they will be physical objects, states of affairs, propositions, or pragmatic options (the atoms of human interaction). Physical objects constitute the conceptual domain of space, states of affairs the domain of time, propositions the epistemic domain, and pragmatic options the deontic domain. The relations encodable in any of these domains can be divided into four basic types: similarity relations, situating relations, conditional relations, and causal relations. Conceptual domains and types of relations define the universe of possible connections between semantic objects.
Connectives differ as to the interpretations they permit in terms of conceptual domains and types of relations. Very few connectives are specialized on relata of one certain category and relations of one certain type. Possible examples in German are später (‘later on’) and zwischenzeitlich (‘in the meantime’), which encode situating relations between states of affairs. Other connectives are specialized on relata of one certain category, but are underspecified with respect to the type of relation. An example is German sobald (‘as soon as’), which can only connect states of affairs, but accepts situating, conditional and causal readings. Connectives of a third group are specialized on relations of a certain type, but are underspecified with respect to the category of the relata. Examples of this kind are German weil (‘because’) and trotzdem (‘nevertheless’), which encode causal relations, but accept states of affairs, propositions and pragmatic options as their relata. Connectives of a fourth group are underspecified both for the category of relata and the type of relation. An example is German da (‘there’), which accepts relata of any category and allows for situating, conditional and causal readings. Connectives like und (‘and’) and oder (‘or’) exhibit an even higher degree of under specification, in that they allow for all kinds of relations and relata.
This article discusses the question whether the distinction between subordination and coordination is parallel in syntax and discourse. Its main thesis is that subordination and coordination, as they are commonly understood in the linguistic literature, are genuinely syntactic concepts. The distinction between hierarchical and non-hierarchical connection in discourse structure, as far as it is defined clearly in the literature, is of a quite different nature. The syntax and semantics of connectives (as the most prominent morphosyntactic means by which subordination and coordination are encoded) offers little evidence to support the assumption of a structural parallelism between syntax and discourse. As a methodological consequence, sentence and discourse structure should not be mixed up in linguistic analysis.
A formal narrative representation is a procedure assigning a formal description to a natural language narrative. One of the goals of the computational models of narrative community is to understand this procedure better in order to automatize it. A formal framework fit for automatization should allow for objective and reproducible representations. In this paper, we present empirical work focussing on objectivity and reproducibility of the formal framework by Vladimir Propp (1928). The experiments consider Propp’s formalization of Russian fairy tales and formalizations done by test subjects in the same formal framework; the data show that some features of Propp’s system such as the assignment of the characters to the dramatis personae and some of the functions are not easy to reproduce.
This paper demonstrates systematic cross-linguistic differences in the electrophysiological correlates of conflicts between form and meaning (“semantic reversal anomalies”). These engender P600 effects in English and Dutch (e.g. Kolk et al., 2003, Kuperberg et al., 2003), but a biphasic N400 – late positivity pattern in German (Schlesewsky and Bornkessel-Schlesewsky, 2009), and monophasic N400 effects in Turkish (Experiment 1) and Mandarin Chinese (Experiment 2). Experiment 3 revealed that, in Icelandic, semantic reversal anomalies show the English pattern with verbs requiring a position-based identification of argument roles, but the German pattern with verbs requiring a case-based identification of argument roles. The overall pattern of results reveals two separate dimensions of cross-linguistic variation: (i) the presence vs. absence of an N400, which we attribute to cross-linguistic differences with regard to the sequence-dependence of the form-to-meaning mapping and (ii) the presence vs. absence of a late positivity, which we interpret as an instance of a categorisation-related late P300, and which is observable when the language under consideration allows for a binary well-formedness categorisation of reversal anomalies. We conclude that, rather than reflecting linguistic domains such as syntax and semantics, the late positivity vs. N400 distinction is better understood in terms of the strategies that serve to optimise the form-to-meaning mapping in a given language.
Hierarchical predictive coding has been identified as a possible unifying principle of brain function, and recent work in cognitive neuroscience has examined how it may be affected by age–related changes. Using language comprehension as a test case, the present study aimed to dissociate age-related changes in prediction generation versus internal model adaptation following a prediction error. Event-related brain potentials (ERPs) were measured in a group of older adults (60–81 years; n = 40) as they read sentences of the form “The opposite of black is white/yellow/nice.” Replicating previous work in young adults, results showed a target-related P300 for the expected antonym (“white”; an effect assumed to reflect a prediction match), and a graded N400 effect for the two incongruous conditions (i.e. a larger N400 amplitude for the incongruous continuation not related to the expected antonym, “nice,” versus the incongruous associated condition, “yellow”). These effects were followed by a late positivity, again with a larger amplitude in the incongruous non-associated versus incongruous associated condition. Analyses using linear mixed-effects models showed that the target-related P300 effect and the N400 effect for the incongruous non-associated condition were both modulated by age, thus suggesting that age-related changes affect both prediction generation and model adaptation. However, effects of age were outweighed by the interindividual variability of ERP responses, as reflected in the high proportion of variance captured by the inclusion of by-condition random slopes for participants and items. We thus argue that – at both a neurophysiological and a functional level – the notion of general differences between language processing in young and older adults may only be of limited use, and that future research should seek to better understand the causes of interindividual variability in the ERP responses of older adults and its relation to cognitive performance.
This paper presents the application of the <tiger2/> format to various linguistic scenarios with the aim of making it the standard serialisation for the ISO 24615 [1] (SynAF) standard. After outlining the main characteristics of both the SynAF metamodel and the <tiger2/> format, as extended from the initial Tiger XML format [2], we show through a range of different language families how <tiger2/> covers a variety of constituency and dependency based analyses.
TripleA is a workshop series founded by linguists from the University of Tübingen and the University of Potsdam. Its aim is to provide a forum for semanticists doing fieldwork on understudied languages, and its focus is on languages from Africa, Asia, Australia and Oceania. The second TripleA workshop was held at the University of Potsdam, June 3-5, 2015.
This paper first argues that the distinction between Propositions and States-of-Affairs is significant for understanding a number of linguistic contrasts, including contrasts between nominalizations, complement clauses, readings of modal infinitives, raising constructions, illocutions and moods, relative clauses, and nouns. Subsequently, the paper outlines a cognitive linguistic model of the distinction, according to which Propositions and States-of-Affairs differ in terms of construal. Both prompt Langackerian “processes”, but only Propositions prompt a construal of these processes as referential. The paper argues that this model has a number of advantages over a traditional, denotational understanding of the distinction.
In mechanical speech synthesis from the 18th up to the 20th century, reed pipes were mainly used for the generation of the voice and the organ stop vox humana was central in this process. This has been described in different historical documents which report that the vox humana in some organs sounded like human vowels. In this study, tones of four different voces humanae were recorded to investigate their similarity to human vowels. The acoustical and perceptual analysis revealed that some, though not all, tones show a high similarity to selected vowels.
The paper reports on experiments with acoustic recordings of a self-built replica of the historic speaking machine of Wolfgang von Kempelen. Several possibilities of the reed as the glottal excitation mechanism were tested. Perception tests with naïve listeners revealed that the machinegenerated words 'mama' and 'papa' were partially recognised as an authentic child voice – as it was also the case in von Kempelen's demonstrations in the late 18th century.
In this paper, we report on an effort to develop a gold standard for the intensity ordering of subjective adjectives. Rather than pursue a complete order as produced by paying attention to the mean scores of human ratings only, we take into account to what extent assessors consistently rate pairs of adjectives relative to each other. We show that different available automatic methods for producing polar intensity scores produce results that correlate well with our gold standard, and discuss some conceptual questions surrounding the notion of polar intensity.
Should events be conceived of as primitive or should they be decomposed into more basic elements with certain syntax? This talk presents new evidence for the latter view: If events are represented as contradictory propositional meanings representing their pre- and post states, a uniform analysis of certain eventive and certain too- comparative constructions is possible; this is wanted given striking parallels between the two types of structure. The analysis goes some way, among other, toward explaining ‘repetetive/restitutive’ asymmetries familiar from eventive constructions (von Stechow 1996) but similarly arising in too- comparative constructions.
The principal claim of this dissertation is that there is a unique structural core shared by Double Object, Dative Experiencer and Existential/Presentational constructions. This core is argued to take the form of a Cipient Predication structure, `cipient covering traditional notions like (affected) source/goal, recipient, indirect object or dative experiencer. Central questions arising in defining Cipient Predication are: How are cipients thematically licensed, and what is the role of there in argument-structural terms? What is the structural locus of cipients/there? What is the role and nature of dative case? How can the possessive interpretation, the blocking and definiteness effects associated with the above-mentioned constructions be explained? Cipients are presented as external arguments and logical subjects (location individuals) of predicates derived from a propositional meaning embedded in the VP, the predicate formed by a lower tense head `little t that is overtly realized as there. Little t is argued to encode a distinction at the reference time level, structural dative hinging on a tense property like structural nominative. The cipient relates as a whole to a part to a VP-internal location argument that together with the theme furnishes the propositional meaning (`possession ). As logical subjects, cipients anchor the predicate to the utterance context, forcing its interpretation in extralinguistic terms (`blocking effects ). It is proposed that lacking structurally encoded subjects, Existential/Presentational constructions are not saturated expressions in syntax, precluding the interpretation of certain quantifiers (most/every, vide `definiteness effects ). Cipient Predication, couched in terms of the Minimalist Program (in particular, Chomsky 1999) and a semantics relying on tense and the ontological distinction of locations as well as scalar and part-whole structure, should be of interest to scholars working on datives, argument structure, and the syntax/semantics/pragmatics interface more generally.
The paper gives an analysis of productively occurring dative constructions in German, attempting to unify what are known traditionally as Double Object and Experiencer Datives. The datives in question - cipients as we call them - are argued to be licensed under two conditions: One, predicates licensing cipients project a theme and a location argument internally; two, interpretation of the predication as a whole involves reference to two dissociated temporal intervals, or more generally, indexical truth intervals. It is argued that the location argument is needed because it provides the variable that is bound by the cipient argument - the variable in question ranges over superlocations of the location argument referent. Reference to two truth intervals is forced because interpreting the cipient structure involves evaluation of two propositional meanings that would contradict each other in a single context. The first propositional meaning is embedded in the predicate; it encodes that something is at a certain location (in quality space). The second propositional meaning is projected as a presupposition that corresponds just to the negation of the first one. The cipient, functioning as the logical subject of the construction, accommodates this second presuppositional meaning; this makes the construction as a whole interpretable. The analysis applies uniformly to what appear to be the two major contexts licensing cipients: ‘eventive’ and ‘foo-comparative’ predications, thereby accounting for some striking parallels between them.
Patterns pertaining to 'strong' DMPs and scope in presentational there-sentences (henceforth: PTSs) have received much attention, and many attempts have been made to derive them. Building on the account of Heim 1987, this paper proposes a novel account based on temporal reference encoding and general assumptions concerning the nature of the interface between the computational system of syntax (CS) and the systems of sound and meaning (Chomsky 1999).
The effects of different forms of predication have been insightfully (and almost exclusively) studied for 'simple' cases of predication, of which the 'presentational sentence' is maybe the paradigm instantiation. It is the aim of this paper to show that thc same kind of effects as well as in fact the same kind of structures are present at embedded levels in thematically and otherwise more complex structures. Beyond presentational sentences, 'unaccusative' experiencing constructions involving a dative subject, 'double object constructions' and - to a lesser extent - spraylload constructions are discussed. For all of these, it is argued that they comprise a predication encoding the ascription of a transient temporal property to a location. On this basis, a proposal is made as to how the scope asymmetry between the two arguments involved in the colistructions can be explained. Furthermore, a proposal is made as to how what has been called 'argument shift' is motivated.
Grammatische Strukturen verbinden Systeme des Denkens und Systeme des Sprechens und Zeigens, deren jeweilige Bedingungen kaum zueinander zu passen scheinen. Der Reparaturansatz betrachtet den regulären Umgang mit Übersetzungsproblemen innerhalb des grammatischen Systems und an seinen Schnittstellen als konstitutiv für Expressivität und Ökonomie der Sprache. Reparaturen sind produktive Wiedergutmachungs- und Anpassungsmechanismen, die linguistische Phänomene als Reflex der Kompensation für derivationelle oder interpretative Schäden erklären.
The present investigation targets the phenomenon commonly called control. Many languages including German and Polish employ non-finite clauses (besides finite clauses) as propositional complements. The subject of these complement clauses is left unexpressed and must generally be interpreted co-referentially with the subject or object of the matrix clause (subject or object control). However. there are also infinitive-selecting verbs that do not allow for a co- referential interpretation of the embedded subject - semantically, the embedded infinitives of these anti-control verbs are thus less dependent on or less unifiable with the matrix proposition. In Polish anti-control constructions, non-finite complements are overtly marked with the complementizer zeby, suggesting that they are structurally more complex (namely. containing a C-projection) than the non-finite complements in control constructions lacking zeby (modulo special contexts. viz. 'control switch'). In a comparative perspective, the paper brings corpuslinguistic and experimental evidence to bear on the question whether surface appearances notwithstanding, the infinitival complements of anti-control verbs in German should similarly be analyzed as truly sentential, i.e., C-headed structures.
While written corpora can be exploited without any linguistic annotations, speech corpora need at least a basic transcription to be of any use for linguistic research. The basic annotation of speech data usually consists of time-aligned orthographic transcriptions. To answer phonetic or phonological research questions, phonetic transcriptions are needed as well. However, manual annotation is very time-consuming and requires considerable skill and near-native competence. Therefore it can take years of speech corpus compilation and annotation before any analyses can be carried out. In this paper, approaches that address the transcription bottleneck of speech corpus exploitation are presented and discussed, including crowdsourcing the orthographic transcription, automatic phonetic alignment, and query-driven annotation. Currently, query-driven annotation and automatic phonetic alignment are being combined and applied in two speech research projects at the Institut für Deutsche Sprache (IDS), whereas crowdsourcing the orthographic transcription still awaits implementation.
The naturalness of synthetic speech depends strongly on the prediction of appropriate prosody. For the present study the original annotation of the German speech database “Kiel Corpus of Read Speech” was extended automatically with syntactic features, word frequency, and syllable boundaries. Several classification and regression trees for predicting symbolic prosody features, postlexical phonological processes, duration, and F0 were trained on this database. The perceptual evaluation showed that the overall perceptual quality of the German text-to-speech system MARY can be significantly improved by training all models that contribute to prosody prediction on the same database. Furthermore, it showed that the error introduced by symbolic prosody prediction perceptually equals the error produced by a direct method that does not exploit any symbolic prosody features.
In this study we investigate the intonational characteristics of the four utterance types statement, wh-question, yes/no-question and declarative question. Readings of two German scripted dialogues were examined to ascertain characteristic features of the F0 contour for each utterance type. Final boundary tone, nuclear pitch accent, F0 offset, F0 onset, F0 range, and the slopes of a topline and a bottomline were determined for each utterance and compared for the four utterance types. Results show that for an average speaker, the final boundary tone, the F0 range, and the slope of the topline can be used to distinguish between the four utterance types. However, speakers may deviate from this pattern and exploit other intonational means to distinguish certain utterance types or choose not to mark a syntactic difference at all.
We present an XML-based metadata standard for the documentation of speech and multimedia corpora that was developed at the Institute for German Language (IDS) in Mannheim, Germany. The IDS is one of the major institutions providing German speech and language corpora to researchers. These corpora stem from many different sources and were previously documented in a rather heterogeneous fashion using a variety of data models and formats. In order to unify the documentation for existing and future corpora, the IDS- internal Archive for Spoken German collaborated with several projects and developed a set of standardised XML metadata schemas. These XML schemas build on existing internal and external documentation schemas (such as IMDI) and take into account the workflow of speech corpus production. In order to minimise redundancy, separate schemas were designed for projects, speakers, recording sessions, and entire corpora. The resulting schemas are tested in ongoing speech and multi-media projects at the IDS and are regularly revised. They are accompanied by element definitions, guidelines, and examples. In addition, a mapping to IMDI will be provided.
The research project “German Today” aims to determine the amount of regional variation in (near-)standard German spoken by young and older educated adults and to identify and locate regional features. To this end, we compile an areally extensive corpus of read and spontaneous German speech. Secondary school students and 50-to-60-year-old locals are recorded in 160 cities throughout the German speaking area of Europe. All participants read a number of short texts and a word list, name pictures, translate words and sentences from English, answer questions in a sociobiographic interview, and take part in a map task experiment. The resulting corpus comprises over 1000 hours of speech, which is transcribed orthographically. Automatically derived broad phonetic transcriptions, selective manual narrow phonetic transcriptions, and variationalist annotations are added. Focussing on phonetic variation we aim to show to what extent national or regional standards exist in spoken German. Furthermore, the linguistic variation due to different contextual styles (read vs. spontaneous speech) shall be analysed. Finally, the corpus enables us to investigate whether linguistic change has occurred in spoken (near-)standard German.