Korpuslinguistik
Refine
Year of publication
Document Type
- Part of a Book (59) (remove)
Language
- English (59) (remove)
Has Fulltext
- yes (59)
Keywords
- Korpus <Linguistik> (48)
- Deutsch (12)
- Annotation (11)
- Automatische Sprachanalyse (5)
- Computerunterstützte Kommunikation (5)
- Gesprochene Sprache (5)
- Grammatik (5)
- Syntax (5)
- Deutsches Referenzkorpus (DeReKo) (4)
- Forschungsdaten (4)
Publicationstate
- Veröffentlichungsversion (40)
- Zweitveröffentlichung (14)
- Postprint (3)
Reviewstate
- Peer-Review (35)
- (Verlags)-Lektorat (20)
- Peer-review (2)
Publisher
The project Referenzkorpus Altdeutsch (‘Old German Reference Corpus’) aims to es- tablish a deeply-annotated text corpus of all extant Old German texts. As the automated part-of-speech and morphological pre-annotation is amended by hand, a quality control system for the results seems a desirable objective. To this end, standardized inflectional forms, generated using the morphological information, are compared with the attested word forms. Their creation is described by way of example for the Old High German part of the corpus. As is shown, in a few cases, some features of the attested word forms are also required in order to determine as exactly as possible the shape of the inflected lemma form to be created.
The availability of electronic corpora of historical stages of languages has been wel- comed as possibly attenuating the inherent problem of diachronic linguistics, i.e. that we only have access to what has chanced to come down to us - the problem which was memorably named by Labov (1992) as one of “Bad Data”. However, such corpora can only give us access to an increased amount ot historical material and this can essentially still only be a partial and possibly distorted picture of the actual language at a particular period of history. Corpora can be improved by taking a more representative sample of extant texts if these are available (as they are in significant number for periods after the invention of printing). But, as examples from the recently compiled GerManC corpus of seventeenth and eighteenth century German show, the evidence from such corpora can still fail to yield definitive answers to our questions about earlier stages of a language. The data still require expert interpretation, and it is important to be realistic about what can legitimately be expected from an electronic historical corpus.
Multi-faceted alignment. Toward automatic detection of textual similarity in Gospel-derived texts
(2015)
Ancient Germanic Bible-derived texts stand in as test material for producing computational means for automatically determining where textual contamination and linguistic interference have influenced the translation process. This paper reports on the results of research efforts that produced a text corpus; a method for decomposing the texts involved into smaller, more directly comparable thematically-related chunks; a database of relationships between these chunks; and a user-interface allowing for searches based on various referential criteria. Finally, the state of the product at the end of the project is discussed, namely as it was handed over to another researcher who has extended it to automatically find semantic and syntactic similarities within comparable chunks.
In this paper we present some preliminary considerations concerning the possibility of automatic parsing an annotated corpus for N-N compounds. This should in prin- ciple be possible at least for relational and stereotype compounds, if the lemmatization of the corpus connects the lemmata with lexical entries as described in Höhle (1982). These lexical entries then supply the necessary information about the argument structure of a relational noun or about the stereotypical purpose associated with the noun’s referent which can be used to establish a relation between the first and the head constituent of the compound.
The relative order of dative and accusative objects in older German is less free than it is today. The reason for this could be that speakers of the direct predecessor of Old High German organized the referents according to the Thematic Hierarchy. If one applies a Case Hierarchy Nom>Acc>Dat to this, the order Nom - Dat - Acc falls out. It becomes apparent that the status of the Thematic Hierarchy is not a factor governing underlying word order, but a factor inducing scrambling. Arguments from binding theory, whose validity is discussed, indicate that the underlying order is ‘accusative before dative’
The International Comparable Corpus (ICC) (Kirk/Čermáková 2017; Čermáková et al. 2021) is an open initiative which aims to improve the empirical basis for contrastive linguistics by compiling comparable corpora for many languages and making them as freely available as possible as well as providing tools with which they can easily be queried and analysed. In this contribution we present the first release of written language parts of the ICC which includes corpora for Chinese, Czech, English, German, Irish (partly), and Norwegian. Each of the released corpora contains 400k words distributed over 14 different text categories according to the ICC specifications. Our poster covers the design basics of the ICC, its TEI encoding, a demonstration of using the ICC via different query tools, and an outlook on future plans.
Similar to the European Reference Corpus EuReCo (Kupietz et al. 2020), ICC follows the approach of reusing existing linguistic resources wherever possible in order to cover as many languages as possible with realistic effort in as short a time as possible. In contrast to EuReCo, however, comparable corpus pairs are not defined dynamically in the usage phase, but the compositions of the corpora are fixed in the ICC design. The approaches are thus complementary in this respect. The design principles and composition of the ICC are based on those of the International Corpus of English (ICE) (Greenbaum (ed.) 1996), with the deviation that the ICC includes the additional text category blog post and excludes spoken legal texts (see Čermáková et al. 2021 for details). ICC’s fixed-design approach has the advantage that all single-language corpora in the ICC have the same composition with respect to the selected text types and that this guarantees that the selected broad spectrum of potential influencing variables for linguistic variation is always represented. The disadvantage, however, is that this can only be achieved for quite small corpora and that the generalisability of comparative findings based on the ICC corpora will often need to be checked on larger monolingual corpora or translation corpora (Čermáková/Ebeling/Oksefjell Ebeling forthcoming). Arguing that such issues with comparability and representativeness are inevitable, in one way or the other, and need to be dealt with, our poster will discuss and exemplify the text selections in more detail.
In many European languages, propositional arguments (PAs) can be realized as different types of structures. Cross-linguistically, complex structures with PAs show a systematic correlation between the strength of the semantic bond and the syntactic union (cf. Givón 2001; Wurmbrand/Lohninger 2023). Also, different languages show similarities with respect to the (lexical) licensing of different PAs (cf. Noonan 1985; Givón 2001; Cristofaro 2003 on different predicate types). However, on a more fine-grained level, a variation across languages can be observed both with respect to the syntactic-semantic properties of PAs as well as to their licensing and usage. This presentation takes a multi-contrastive view of different types of PAs as syntactic subjects and objects by looking at five European languages: EN, DE, IT, PL and HU. Our goal is to identify the parameters of variation in the clausal domain with PAs and by this to contribute to a better understanding of the individual language systems on the one hand and the nature of the linguistic variation in the clausal domain on the other hand. Phenomena and Methodology: We investigate the following types of PAs: direct object (DO) clauses (1), prepositional object (PO) clauses (2), subject clauses (3), and nominalizations (4, 5). Additionally, we discuss clause union phenomena (6, 7). The analyzed parameters include among others finiteness, linear position of the PA, (non) presence of a correlative element, (non) presence of a complementizer, lexical-semantic class of the embedding verb. The phenomena are analyzed based on corpus data (using mono- and multilingual corpora), experimental data (acceptability judgement surveys) or introspective data.
This paper presents an extended annotation and analysis of interpretative reply relations focusing on a comparison of reply relation types and targets between conflictual pages and neutral pages of German Wikipedia (WP) talk pages. We briefly present the different categories identified for interpretative reply relations to analyze the relationship between WP postings as well as linguistic cues for each category. We investigate referencing strategies of WP authors in discussion page postings, illustrated by means of reply relation types and targets taking into account the degree of disagreement displayed on a WP talk page. We provide richly annotated data that can be used for further analyses such as the identification of interactional relations on higher levels, or for training tasks in machine learning algorithms.
This paper presents the IVK-Ler corpus, a longitudinal, annotated learner corpus of weekly writings produced by a group of 18 adolescents in a preparatory class. The corpus consists of 117 student texts collected between 2020 and 2021 and has a structure layered by student and text number. It includes metadata that enables researchers to analyze and track individual student progress in terms of syntactic competence and literacy. The annotation schema, manual and automatic annotation processes, and corpus representation are described in detail. The corpus currently includes target hypotheses and gold standard part-of-speech tags. Future work could include additional annotation layers for topological fields and dependency relations, as well as semantic and discourse annotations to make the corpus usable for tasks beyond syntactic evaluations.
This article details the process of creating the Nottinghamer Korpus deutscher YouTube-Sprache ('The Nottingham German YouTube Language Corpus' - or NottDeuYTSch corpus) and outlines potential research opportunities. The corpus was compiled to analyse the online language produced by young German-speakers and offers significant opportunity for in-depth research across several linguistic fields including lexis, morphology, syntax, orthography, and conversational and discursive analysis. The NottDeuYTSch corpus contains over 33 million words taken from approximately 3 million YouTube comments from videos published between 2008 to 2018 targeted at a young, German-speaking demographic and represent an authentic language snapshot of young German speakers. The corpus was proportionally sampled based on video category and year from a database of 112 popular German-speaking YouTube channels in the DACH region for optimal representativeness and balance and contains a considerable amount of associated metadata for each comment that enable further longitudinal cross-sectional analyses. The NottDeuYTSch corpus is available for analysis as part of the German Reference Corpus (DeReKo).
In this article, we examine the current situation of data dissemination and provision for CMC corpora. By that we aim to give a guiding grid for future projects that will improve the transparency and replicability of research results as well as the reusability of the created resources. Based on the FAIR guiding principles for research data management, we evaluate the 20 European CMC corpora listed in the CLARIN CMC Resource family, individuate successful strategies among the existing corpora and establish best practices for future projects. We give an overview of existing approaches to data referencing, dissemination and provision in European CMC corpora, and discuss the methods, formats and strategies used. Furthermore, we discuss the need for community standards and offer recommendations for best practices when creating a new CMC corpus.
This paper describes a method for extracting collocation data from text corpora based on a formal definition of syntactic structures, which takes into account not only the POS-tagging level of annotation but also syntactic parsing (syntactic treebank model) and introduces the possibility of controlling the canonical form of extracted collocations based on statistical data on forms with different properties in the corpus. Specifically, we describe the results of extraction from the syntactically tagged Gigafida 2.1 corpus. Using the new method, 4,002,918 collocation candidates in 81 syntactic structures were extracted. We evaluate the extracted data sample in more detail, mainly in relation to properties that affect the extraction of canonical forms: definiteness in adjectival collocations, grammatical number in noun collocations, comparison in adjectival and adverbial collocations, and letter case (uppercase and lowercase) in canonical forms. The conclusion highlights the potential of the methodology used for the grammatical description of collocation and phrasal syntax and the possibilities for improving the model in the process of compilation of a digital dictionary database for Slovene.
This paper describes a method for automatic identification of sentences in the Gigafida corpus containing multi-word expressions (MWEs) from the list of 5,242 phraseological units, which was developed on the basis of several existing open-access lexical resources for Slovene. The method is based on a definition of MWEs, which includes information on two levels of corpus annotation: syntax (dependency parsing) and morphology (POS tagging), together with some additional statistical parameters. The resulting lexicon contains 12,358 sentences containing MWEs extracted from the corpus. The extracted sentences were analysed from the lexicographic point of view with the aim of establishing canonical forms of MWEs and semantic relations between them in terms of variation, synonymy, and antonymy.
When comparing different tools in the field of natural language processing (NLP), the quality of their results usually has first priority. This is also true for tokenization. In the context of large and diverse corpora for linguistic research purposes, however, other criteria also play a role – not least sufficient speed to process the data in an acceptable amount of time. In this paper we evaluate several state of the art tokenization tools for German – including our own – with regard to theses criteria. We conclude that while not all tools are applicable in this setting, no compromises regarding quality need to be made.
Enabling appropriate access to linguistic research data, both for many researchers and for innovative research applications, is a challenging task. In this chapter, we describe how we address this challenge in the context of the German Reference Corpus DeReKo and the corpus analysis platform KorAP. The core of our approach, which is based on and tightly integrated into the CLARIN infrastructure, is to offer access at different levels. The graduated access levels make it possible to find a low-loss compromise between the possibilities opened up and the costs incurred by users and providers for each individual use case, so that, viewed over many applications, the ratio between effort and results achieved can be effectively optimized. We also report on experiences with the current state of this approach.
When comparing different tools in the field of natural language processing (NLP), the quality of their results usually has first priority. This is also true for tokenization. In the context of large and diverse corpora for linguistic research purposes, however, other criteria also play a role – not least sufficient speed to process the data in an acceptable amount of time. In this paper we evaluate several state-ofthe-art tokenization tools for German – including our own – with regard to theses criteria. We conclude that while not all tools are applicable in this setting, no compromises regarding quality need to be made.
Dictionaries are often a reflection of their time; their respective (socio-)historical context influences how the meaning of certain lexical units is described. This also applies to descriptions of personal terms such as man or woman. Lexicographers have a special responsibility to comprehensively investigate current language use before describing it in the dictionary. Accordingly, contemporary academic dictionaries are usually corpus-based. However, it is important to acknowledge that language is always embedded in cultural contexts. Our case study investigates differences in the linguistic contexts of the use of man and woman, drawing from a range of language collections (in our case fiction books, popular magazines and newspapers). We explain how potential differences in corpus construction would therefore influence the “reality”1 depicted in the dictionary. In doing so, we address the far-reaching consequences that the choice of corpus-linguistic basis for an empirical dictionary has on semantic descriptions in dictionary entries.
Furthermore, we situate the case study within the context of gender-linguistic issues and discuss how lexicographic teams can engage with how dictionaries might perpetuate traditional role concepts when describing language use.
In this paper, we address two problems in indexing and querying spoken language corpora with overlapping speaker contributions. First, we look into how token distance and token precedence can be measured when multiple primary data streams are available and when transcriptions happen to be tokenized, but are not synchronized with the sound at the level of individual tokens. We propose and experiment with a speaker based search mode that enables any speaker’s transcription tier to be the basic tokenization layer whereby the contributions of other speakers are mapped to this given tier. Secondly, we address two distinct methods of how speaker overlaps can be captured in the TEI based ISO Standard for Spoken Language Transcriptions (ISO 24624:2016) and how they can be queried by MTAS – an open source Lucene-based search engine for querying text with multilevel annotations. We illustrate the problems, introduce possible solutions and discuss their benefits and drawbacks.
Validating the Performativity Hypothesis to Neg-Raising using corpus data: Evidence from Polish
(2021)
Twenty-two historical encyclopedias encoded in TEI: a new resource for the Digital Humanities
(2020)
This paper accompanies the corpus publication of EncycNet, a novel XML/TEI annotated corpus of 22 historical German encyclopedias from the early 18th to early 20th century. We describe the creation and annotation of the corpus, including the rationale for its development, suggested methodology for TEI annotation, possible use cases and future work. While many well-developed annotation standards for lexical resources exist, none can adequately model the encyclopedias at hand, and we therefore suggest how the TEI Lex-0 standard may be modified with additional guidelines for the annotation of historical encyclopedias. As the digitization and annotation of historical encyclopedias are settling on TEI as the de facto standard, our methodology may inform similar projects.
This paper discusses a theoretical and empirical approach to language fixedness that we have developed at the Institut für Deutsche Sprache (IDS) (‘Institute for German Language’) in Mannheim in the project Usuelle Worterbindungen(UWV) over the last decade. The analysis described is based on the Deutsches Referenzkorpus (‘German Reference Corpus’; DeReKo) which is located at the IDS. The corpus analysis tool used for accessing the corpus data is COSMAS II (CII) and – for statistical analysis – the IDS collocation analysis tool (Belica, 1995; CA). For detecting lexical patterns and describing their semantic and pragmatic nature we use the tool lexpan (or ‘Lexical Pattern Analyzer’) that was developed in our project. We discuss a new corpus-driven pattern dictionary that is relevant not only to the field of phraseology, but also to usage-based linguistics and lexicography as a whole.
The present paper outlines the projected second part of the Corpus Query Lingua Franca (CQLF) family of standards: CQLF Ontology, which is currently in the process of standardization at the International Standards Organization (ISO), in its Technical Committee 37, Subcommittee 4 (TC37SC4) and its national mirrors. The first part of the family, ISO 24623-1 (henceforth CQLF Metamodel), was successfully adopted as an international standard at the beginning of 2018. The present paper reflects the state of the CQLF Ontology at the moment of submission for the Committee Draft ballot. We provide a brief overview of the CQLF Metamodel, present the assumptions and aims of the CQLF Ontology, its basic structure, and its potential extended applications. The full ontology is expected to emerge from a community process, starting from an initial version created by the authors of the present paper.
Providing online repositories for language resources is one of the main activities of CLARIN centres. The legal framework regarding liability of Service Providers for content uploaded by their users has recently been modified by the new Directive on Copyright in the Digital Single Market. A new category of Service Providers, Online Content-Sharing Service Providers (OCSSPs), was added. It is subject to a complex and strict framework, including the requirement to obtain licenses from rightholders for the hosted content. This paper provides the background and effect of these changes to law and aims to initiate a debate on how CLARIN repositories should navigate this new legal landscape.
We present web services which implement a workflow for transcripts of spoken language following the TEI guidelines, in particular ISO 24624:2016 “Language resource management – Transcription of spoken language”. The web services are available at our website and will be available via the CLARIN infrastructure, including the Virtual Language Observatory and WebLicht.
Corpus REDEWIEDERGABE
(2020)
This article presents the corpus REDEWIEDERGABE, a German-language historical corpus with detailed annotations for speech, thought and writing representation (ST&WR). With approximately 490,000 tokens, it is the largest resource of its kind. It can be used to answer literary and linguistic research questions and serve as training material for machine learning. This paper describes the composition of the corpus and the annotation structure, discusses some methodological decisions and gives basic statistics about the forms of ST&WR found in this corpus.
The newest generation of speech technology caused a huge increase of audio-visual data nowadays being enhanced with orthographic transcripts such as in automatic subtitling in online platforms. Research data centers and archives contain a range of new and historical data, which are currently only partially transcribed and therefore only partially accessible for systematic querying. Automatic Speech Recognition (ASR) is one option of making that data accessible. This paper tests the usability of a state-of-the-art ASR-System on a historical (from the 1960s), but regionally balanced corpus of spoken German, and a relatively new corpus (from 2012) recorded in a narrow area. We observed a regional bias of the ASR-System with higher recognition scores for the north of Germany vs. lower scores for the south. A detailed analysis of the narrow region data revealed – despite relatively high ASR-confidence – some specific word errors due to a lack of regional adaptation. These findings need to be considered in decisions on further data processing and the curation of corpora, e.g. correcting transcripts or transcribing from scratch. Such geography-dependent analyses can also have the potential for ASR-development to make targeted data selection for training/adaptation and to increase the sensitivity towards varieties of pluricentric languages.
Entity framing is the selection of aspects of an entity to promote a particular viewpoint towards that entity. We investigate entity framing of political figures through the use of names and titles in German online discourse, enhancing current research in entity framing through titling and naming that concentrates on English only. We collect tweets that mention prominent German politicians and annotate them for stance. We find that the formality of naming in these tweets correlates positively with their stance. This confirms sociolinguistic observations that naming and titling can have a status-indicating function and suggests that this function is dominant in German tweets mentioning political figures. We also find that this status-indicating function is much weaker in tweets from users that are politically left-leaning than in tweets by right leaning users. This is in line with observations from moral psychology that left-leaning and right-leaning users assign different importance to maintaining social hierarchies.
Privacy by Design (also referred to as Data Protection by Design) is an approach in which solutions and mechanisms addressing privacy and data protection are embedded through the entire project lifecycle, from the early design stage, rather than just added as an additional layer to the final product. Formulated in the 1990 by the Privacy Commissionner of Ontario, the principle of Privacy by Design has been discussed by institutions and policymakers on both sides of the Atlantic, and mentioned already in the 1995 EU Data Protection Directive (95/46/EC). More recently, Privacy by Design was introduced as one of the requirements of the General Data Protection Regulation (GDPR), obliging data controllers to define and adopt, already at the conception phase, appropriate measures and safeguards to implement data protection principles and protect the rights of the data subject. Failing to meet this obligation may result in a hefty fine, as it was the case in the Uniontrad decision by the French Data Protection Authority (CNIL). The ambition of the proposed paper is to analyse the practical meaning of Privacy by Design in the context of Language Resources, and propose measures and safeguards that can be implemented by the community to ensure respect of this principle.
Making corpora accessible and usable for linguistic research is a huge challenge in view of (too) big data, legal issues and a rapidly evolving methodology. This does not only affect the design of user-friendly graphical interfaces to corpus analysis tools, but also the availability of programming interfaces supporting access to the functionality of these tools from various analysis and development environments. RKorAPClient is a new research tool in the form of an R package that interacts with the Web API of the corpus analysis platform KorAP, which provides access to large annotated corpora, including the German reference corpus DeReKo with 45 billion tokens. In addition to optionally authenticated KorAP API access, RKorAPClient provides further processing and visualization features to simplify common corpus analysis tasks. This paper introduces the basic functionality of RKorAPClient and exemplifies various analysis tasks based on DeReKo, that are bundled within the R package and can serve as a basic framework for advanced analysis and visualization approaches.
This paper analyses reply relations in computer-mediated communication (CMC), which occur between post units in CMC interactions and which describe references between posts. We take a look at existing practices in the description and annotation of such relations in chat, wiki talk, and blog corpora. We distinguish technical reply structures, indentation structures, and interpretative reply relations, which include reply relations induced by linguistic markers. We sort out the different levels of description and annotation that are involved and propose a solution for their combined representation within the TEI annotation framework.
A syntax-based scheme for the annotation and segmentation of German spoken language interactions
(2018)
Unlike corpora of written language where segmentation can mainly be derived from orthographic punctuation marks, the basis for segmenting spoken language corpora is not predetermined by the primary data, but rather has to be established by the corpus compilers. This impedes consistent querying and visualization of such data. Several ways of segmenting have been proposed,
some of which are based on syntax. In this study, we developed and evaluated annotation and segmentation guidelines in reference to the topological field model for German. We can show that these guidelines are used consistently across annotators. We also investigated the influence of various interactional settings with a rather simple measure, the word-count per segment and unit-type. We observed that the word count and the distribution of each unit type differ in varying interactional settings and that our developed segmentation and annotation guidelines are used consistently across annotators. In conclusion, our syntax-based segmentations reflect interactional properties that are intrinsic to the social interactions that participants are involved in. This can be used for further analysis of social interaction and opens the possibility for automatic segmentation of transcripts.
The paper describes preliminary studies regarding the usage of Example-Based Querying for specialist corpora. We outline an infrastructure for its application within the linguistic domain. Example-Based Querying deals with retrieval situations where users would like to explore large collections of specialist texts semantically, but are unable to explicitly name the linguistic phenomenon they look for. As a way out, the proposed framework allows them to input prototypical everyday language examples or cases of doubt, which are automatically processed by CRF and linked to appropriate linguistic texts in the corpus.
This paper discusses current trends in DeReKo, the German Reference Corpus, concerning legal issues around the recent German copyright reform with positive implications for corpus building and corpus linguistics in general, recent corpus extensions in the genres of popular magazines, journals, historical texts, and web-based football reports. Besides, DeReKo is finally accessible via the new
corpus research platform KorAP, offering registered users several news features in comparison with its predecessor COSMAS II.
In recent years, the availability of large annotated and searchable corpora, together with a new interest in the empirical foundation and validation of linguistic theory and description, has sparked a surge of novel and interesting work using corpus-based methods to study the grammar of natural languages. However, a look at relevant current research on the grammar of the Germanic, Romance, and Slavic languages reveals a variety of different theoretical approaches and empirical foci, which can be traced back to different philological and linguistic traditions. Still, this current state of affairs should not be seen as an obstacle but as an ideal basis for a fruitful exchange of ideas between different research paradigms.
The actual or anticipated impact of research projects can be documented in scientific publications and project reports. While project reports are available at varying level of accessibility, they might be rarely used or shared outside of academia. Moreover, a connection between outcomes of actual research project and potential secondary use might not be explicated in a project report. This paper outlines two methods for classifying and extracting the impact of publicly funded research projects. The first method is concerned with identifying impact categories and assigning these categories to research projects and their reports by extension by using subject matter experts; not considering the content of research reports. This process resulted in a classification schema that we describe in this paper. With the second method which is still work in progress, impact categories are extracted from the actual text data.
Linguistic query systems are special purpose IR applications. We present a novel state-of-the-art approach for the efficient exploitation of very large linguistic corpora, combining the advantages of relational database management systems (RDBMS) with the functional MapReduce programming model. Our implementation uses the German DEREKO reference corpus with multi-layer
linguistic annotations and several types of text-specific metadata, but the proposed strategy is language-independent and adaptable to large-scale multilingual corpora.
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.
In my talk, I present an empirical approach to detecting and describing proverbs as frozen sentences with specific functions in current language use. We have developed this approach in the EU project ‘SprichWort’ (based on the German Reference Corpus). The first chapter illustrates selected aspects of our complex, iterative procedure to validate proverb candidates. Based on our corpus-driven lexpan methodology of slot analysis I then discuss semantic restrictions of proverb patterns. Furthermore, I show different degrees of proverb quality ranging from genuine proverbs to non-proverb realizations of the same abstract pattern. On the one hand, the corpus validation reveals that proverbs are definitely perceived and used as relatively fixed entities and often as sentences. On the other hand, proverbs are not only interpreted as an interesting unique phenomenon but also as part of the whole lexicon, embedded in networks of different lexical items.
The paper reports on the results of a scientific colloquium dedicated to the creation of standards and best practices which are needed to facilitate the integration of language resources for CMC stemming from different origins and the linguistic analysis of CMC phenomena in different languages and genres. The key issue to be solved is that of interoperability – with respect to the structural representation of CMC genres, linguistic annotations metadata, and anonymization/pseudonymization schemas. The objective of the paper is to convince more projects to partake in a discussion about standards for CMC corpora and for the creation of a CMC corpus infrastructure across languages and genres. In view of the broad range of corpus projects which are currently underway all over Europe, there is a great window of opportunity for the creation of standards in a bottom-up approach.
As a consequence of a recent curation project, the Dortmund Chat Corpus is available in CLARIN-D research infrastructures for download and querying. In a legal expertise it had been recommended that standard measures of anonymisation be applied to the corpus before its republication. This paper reports about the anonymisation campaign that was conducted for the corpus. Anonymisation has been realised as categorisation, and the taxonomy of anonymisation categories applied is introduced and the method of applying it to the TEI files is demonstrated. The results of the anonymisation campaign as well as issues of quality assessment are discussed. Finally, pseudonymisation as an alternative to categorisation as a method of the anonymisation of CMC data is discussed, as well as possibilities of an automatisation of the process.
In this paper we present the results of an automatic classification of Russian texts into three levels of difficulty. Our aim is to build a study corpus of Russian, in which a L2 student is able to select texts of a desired complexity. We are building on a pilot study, in which we classified Russian texts into two levels of difficulty. In the current paper, we apply the classification to an extended corpus of 577 labelled texts. The best-performing combination of features achieves an accuracy of 0,74 within at most one level difference.
How (and when) do speakers generalise from memorised exemplars of a construction to a productive schema? The present paper presents a novel take on this issue by offering a corpus-based approach to semantic extension processes. Focusing on clusters of German ADJ N expressions involving the heavily polysemous adjective tief ‚deep’, it is shown that type frequency (a commonly used measure of productivity) needs to be relativised to distinct semantic classes within the overall usage spectrum of a given construction in order to predict the occurrence of novel types within a particular region of this spectrum. Some methodological and theoretical implications for usage-based linguistic model building are considered.
CMC Corpora in DeReKo
(2017)
We introduce three types of corpora of computer-mediated communication that have recently been compiled at the Institute for the German Language or curated from an external project and included in DeReKo, the German Reference Corpus, namely Wikipedia (discussion) corpora, the Usenet news corpus, and the Dortmund Chat Corpus. The data and corpora have been converted to I5, the TEI customization to represent texts in DeReKo, and are researchable via the web-based IDS corpus research interfaces and in the case of Wikipedia and chat also downloadable from the IDS repository and download server, respectively.
Many applications in Natural Language Processing require a semantic analysis of sentences in terms of truth-conditional representations, often with specific desiderata in terms of which information needs to be included in the semantic analysis. However, there are only very few tools that allow such an analysis. We investigate the representations of an automatic analysis pipeline of the C&C parser and Boxer to determine whether Boxer’s analyses in form of Discourse Representation Structure can be successfully converted into a more surface oriented event semantic representation, which will serve as input for a fusion algorithm for fusing hard and soft information. We use a data set of synthetic counter intelligence messages for our investigation. We provide a basic pipeline for conversion and subsequently discuss areas in which ambiguities and differences between the semantic representations present challenges in the conversion process.
Brown clustering has been used to help increase parsing performance for morphologically rich languages. However, much of the work has focused on using clustering techniques to replace terminal nodes or as a feature for parsing. Instead, we choose to examine how effectively Brown clustering is for unlexicalized parsing by creating data-driven POS tagsets which are then used with the Berkeley parser. We investigate cluster sizes as well as on what information (e.g. words vs. lemmas) clustering will yield the best parser performance. Our results approach the current state of the art results for the German T¨uBa-D/Z treebank when using parser internal tagging.
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.
We introduce our pipeline to integrate CMC and SM corpora into the CLARIN-D corpus infrastructure. The pipeline was developed by transforming an existing CMC corpus, the Dortmund Chat Corpus, into a resource conforming to current technical and legal standards. We describe how the resource has been prepared and restructured in terms of TEI encoding, linguistic annotations, and anonymisation. The output is a CLARIN-conformant resource integrated in the CLARIN-D research infrastructure.
Contemporary studies on the characteristics of natural language benefit enormously from the increasing amount of linguistic corpora. Aside from text and speech corpora, corpora of computer-mediated communication (CMC) Position themselves between orality and literacy, and beyond that provide in- sight into the impact of "new", mainly intemet-based media on language beha- viour. In this paper, we present an empirical attempt to work with annotated CMC corpora for the explanation of linguistic phenomena. In concrete terms, we implement machine leaming algorithms to produce decision trees that reveal rules and tendencies about the use of genitive markers in German.
This paper presents ongoing research which is embedded in an empirical-linguistic research program, set out to devise viable research strategies for developing an explanatory theory of grammar as a psychological and social phenomenon. As this phenomenon cannot be studied directly, the program attempts to approach it indirectly through its correlates in language corpora, which is justified by referring to the core tenets of Emergent Grammar. The guiding principle for identifying such corpus correlates of grammatical regularities is to imitate the psychological processes underlying the emergent nature of these regularities. While previous work in this program focused on syntagmatic structures, the current paper goes one step further by investigating schematic structures that involve paradigmatic variation. It introduces and explores a general strategy by which corpus correlates of such structures may be uncovered, and it further outlines how these correlates may be used to study the nature of the psychologically real schematic structures.
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.
This article reports about the on-going work on a new version of the metadata framework Component Metadata Infrastructure (CMDI), central to the CLARIN infrastructure. Version 1.2 introduces a number of important changes based on the experience gathered in the last five years of intensive use of CMDI by the digital humanities community, addressing problems encountered, but also introducing new functionality. Next to the consolidation of the structure of the model and schema sanity, new means for lifecycle management have been introduced aimed at combatting the observed proliferation of components, new mechanism for use of external vocabularies will contribute to more consistent use of controlled values and cues for tools will allow improved presentation of the metadata records to the human users. The feature set has been frozen and approved, and the infrastructure is now entering a transition phase, in which all the tools and data need to be migrated to the new version.
Discourse segmentation is the division of a text into minimal discourse segments, which form the leaves in the trees that are used to represent discourse structures. A definition of elementary discourse segments in German is provided by adapting widely used segmentation principles for English minimal units, while considering punctuation, morphology, sytax, and aspects of the logical document structure of a complex text type, namely scientific articles. The algorithm and implementation of a discourse segmenter based on these principles is presented, as well an evaluation of test runs.
Maximizing the potential of very large corpora: 50 years of big language data at IDS Mannheim
(2014)
Very large corpora have been built and used at the IDS since its foundation in 1964. They have been made available on the Internet since the beginning of the 90’s to currently over 30,000 researchers worldwide. The Institute provides the largest archive of written German (Deutsches Referenzkorpus, DeReKe) which has recently been extended to 24 billion words. DeReKe has been managed and analysed by engines known as COSMAS and afterwards COSMAS II, which is currently being replaced by a new, scalable analysis platform called KorAP. KorAP makes it possible to manage and analyse texts that are accompanied by multiple, potentially conflicting, grammatical and structural annotation layers, and is able to handle resources that are distributed across different, and possibly geographically distant, storage systems. The majority of texts in DeReKe are not licensed for free redistribution, hence, the COSMAS and KorAP systems offer technical solutions to facilitate research on very large corpora that are not available (and not suitable) for download. For the new KorAP system, it is also planned to provide sandboxed environments to support non-remote-API access “near the data” through which users can run their own analysis programs.