English German Research Articles

A Hierarchy-to-Sequence Attentional Neural Machine Translation Model

Although sequence-to-sequence attentional neural machine translation NMT has achieved great progress recently, it is confronted with two challenges: learning optimal model parameters for long parallel sentences and well exploiting different scopes of contexts. In this paper, partially inspired by the idea of segmenting a long sentence into short clauses, each of which can be easily translated by NMT, we propose a hierarchy-to-sequence attentional NMT model to handle these two challenges. Our encoder takes the segmented clause sequence as input and explores a hierarchical neural network structure to model words, clauses, and sentences at different levels, particularly with two layers of recurrent neural networks modeling semantic compositionality at the word and clause level. Correspondingly, the decoder sequentially translates segmented clauses and simultaneously applies two types of attention models to capture contexts of interclause and intraclause for translation prediction. In this way, we can not only improve parameter learning, but also well explore different scopes of contexts for translation. Experimental results on Chinese–English and English–German translation demonstrate the superiorities of the proposed model over the conventional NMT model.

Indo-European and common Germanic roots of the preterito-presentia verbs

The article addresses the Indo-European and common Germanic roots of the preterito-presentia verbs. The words of the ancient Indian, Greek, Latin, Old Slavonic, Lithuanian, Gothic, Old English, Old High German, Old Saxon, and Old English languages are analysed within the framework of the comparative-historical approach. The analysis is carried out on the phonological, morphological and semantic levels. The main attention is paid to the discovery of common latent archaic roots that were obscured in the process of historical development and the separation of the Germanic language group from the rest of the Indo-European language family.

Shallow features as indicators of English–German contrasts in lexical cohesion

Abstract This paper contrasts lexical cohesion between English and German spoken and written registers, reporting findings from a quantitative lexical analysis. After an overview of research aims and motivations we formulate hypotheses on distributions of shallow features as indicators of lexical cohesion across languages and modes and with respect to register ranking and variation. The shallow features analysed are: highly frequent words in texts, lexical density, standardized type-token-ratio, top-frequent content words of the language within individual registers and texts, and several types of Latinate words. Descriptive analyses of the corpus are then presented and statistically validated with the help of univariate and multivariate analyses. The results are interpreted relative to our hypotheses and related to the following properties of texts in terms of lexical cohesion: semantic variability, cohesive strength, number and length of nominal chains, degree of specification of lexis, and degree of variation along all of these properties.

Neural versus phrase-based MT quality: An in-depth analysis on English–German and English–French

Within the field of statistical machine translation, the neural approach (NMT) is currently pushing ahead the state of the art performance traditionally achieved by phrase-based approaches (PBMT), and is rapidly becoming the dominant technology in machine translation. Indeed, in the last IWSLT and WMT evaluation campaigns on machine translation, NMT outperformed well established state-of-the-art PBMT systems on many different language pairs. To understand in what respects NMT provides better translation quality than PBMT, we perform a detailed analysis of neural versus phrase-based statistical machine translation outputs, leveraging high quality post-edits performed by professional translators on the IWSLT data. In this analysis, we focus on two language directions with different characteristics: English–German, known to be particularly hard because of morphology and syntactic differences, and English–French, where PBMT systems typically reach outstanding quality and thus represent a strong competitor for NMT. Our analysis provides useful insights on what linguistic phenomena are best modelled by neural models – such as the reordering of verbs and nouns – while pointing out other aspects that remain to be improved – like the correct translation of proper nouns.

Does translation involve structural priming?

When asked to translate utterances, people might merely make sure that their translations have the same meaning as the source, but they might also maintain aspects of sentence form across languages. We report two experiments in which English-German and German-English bilinguals (without specialist translator training) repeated German ditransitive sentences whose meaning was compatible with more than one grammatical form or translated them into English. Participants almost invariably repeated the sentences accurately, thereby retaining the grammatical structure. Importantly, Experiment 1 found that they tended to repeat grammatical form across languages. Experiment 2 included a condition with sentences that had no grammatical equivalent form in English; here participants tended to persist in the order of thematic roles. We argue that cross-linguistic structural priming plays a major role in the act of translation.

Asymmetrical Priming Effects: An Exploration of Trilingual German-English-French Lexico-Semantic Memory.

The growing number of multilingual speakers poses an interesting question as to the way in which three or more languages are represented in the memory of a language user. The Revised Hierarchical Model (Kroll and Stewart in J Mem Lang 33: 149-174, 1994) or the Sense Model (Finkbeiner et al. in J Mem Lang 51(1), 1-22, 2004) skillfully capture the prediction regarding two languages, with the lexical level being separate and the conceptual one being unified or distributed to a varying degree. In this set of experiments, we employed primed animacy decision tasks to address the lexico-semantic representation of trilingual German-English-French speakers. The comparison of reaction times has revealed priming effects from L1 to L2 and from L2 to L1, both with prime duration of 100 and 50ms; a priming asymmetry effect between the L2 and L3 language directions; and no interaction between L1 and L3. The aggregated findings point to a hybrid representation, with both compound and coordinate representations being possible.

Assessing Performative Competence in German ELF-Classrooms – The Task of Teachers and Learners

When it comes to categorizing what pupils in a German English language classroom do when they are taking over roles and enacting scenes, a variety of terms is usually applied, e.g. acting-out, role-play, scenic play, drama-based education, scenic improvisation, to name but a few. Drama in English Language Teaching (ELT) enables pupils to change perspectives and put themselves into the shoes of other personae thereby learning English in a holistic and creative manner that equally appeals to all senses. The question arises how teachers (and learners) can assess drama activities in an EFL-context. In order for assessment formats to yield conclusive insights into the achievements to be gained in performative EFL-classrooms, the former have to be based upon a sound understanding of the essential elements the construct of performative competence comprises. Therefore, firstly a definition of performative competence will be presented and practical examples of drama activities related to different school levels within the German school system will be provided. Secondly, the teacher’s tasks during drama activities in the EFL-classroom will be discussed, and thirdly, the agents and goals of assessing drama activities in ELT will be outlined before one generic assessment sheet will be introduced and analysed in detail.

Uuaregat selfiu uuaregot – A Personal Note on Anglo-German Philological Relations Past and Present

This essay explores the relation between the studies of early English and early German during the past two centuries. It outlines the gradual segregation between the modern philologies within the context of nineteenth-century nationalism and studies the cultural roots for prevailing prejudices on either side. Outlining common areas of research and integrative curricula it attempts to argue for a new perspective on Early Germanic Studies, one that overcomes the artificial boundaries between disciplines, in order to save the subject and its sub-disciplines from academic and political misappropriations.

Chinese–Spanish neural machine translation enhanced with character and word bitmap fonts

Recently, machine translation systems based on neural networks have reached state-of-the-art results for some pairs of languages (e.g., German–English). In this paper, we are investigating the performance of neural machine translation in Chinese–Spanish, which is a challenging language pair. Given that the meaning of a Chinese word can be related to its graphical representation, this work aims to enhance neural machine translation by using as input a combination of: words or characters and their corresponding bitmap fonts. The fact of performing the interpretation of every word or character as a bitmap font generates more informed vectorial representations. Best results are obtained when using words plus their bitmap fonts obtaining an improvement (over a competitive neural MT baseline system) of almost six BLEU, five METEOR points and ranked coherently better in the human evaluation.

The bi-personal bilingual: a study of the perceived feeling of a changed self

ABSTRACTIt has been a topic in psychiatry, it is referred to by translingual writers, and bilinguals report it frequently – the feeling of becoming a different person when using a different language. The present study was set out to investigate the reasons behind this feeling. We looked at the effects of biculturalism and personality traits, as well as introspective data from 88 German–English bilinguals. The analyses revealed no significant effects of biculturalism. However, an effect was found between high levels of Agreeableness and an affirmative answer to the feeling of a changed self. Furthermore, an integrated approach suggests that four main categories play a vital role in giving rise to this feeling, namely cultural differences, language proficiency, ‘breaking free’ from the L1 personality, and changes in personality due to reactions of interlocutors.

Inside phrasal verb constructions

This study investigates 2,909 phrasal verbs [PVs] in (in)transitive constructions across native English and French- and German-English interlanguages (ILs). Recent work by Gilquin (2014) shows that, as verb-particle combinations, PVs associate more or less strongly with particular syntactic constructions. Further, those association patterns vary across native and learner Englishes. Building on Gilquin (2014), the present study digs deeper into PV constructions by assessing degrees of mutual attraction between verbs and particles as well as between PVs and their semantic uses. Degrees of association are measured using a co-varying collexeme analysis approach (Stefanowitsch & Gries 2005). Generally, it emerges that within given syntactic constructions (e.g. verb-particle-object constructions), different verbs and particles (i.e. bring and up in the case of build up or bring and about for bring about) combine in different degrees, suggesting that, as cognitive routines, those combinations are not equally entrenched. In addition, French and German English learners seem to operate at different levels of semantic complexity when using PVs in their L2. Ultimately, the study bears important pedagogical implications, namely the need (i) to focus on phrasal verbs’ aspectual uses to help learners develop a more confident use of PVs in completive, inceptive and continuative semantic contexts and (ii) to develop resources that help learners improve their uses of PVs in verb-object-particle constructions.

The priming of word order in second language German

ABSTRACTThe present study investigates the priming and subsequent production of word order variation (adverb–verb–subject vs. subject–verb–adverb order) with temporal phrases (Experiment 1) and locative phrases (Experiment 2) among intermediate English–German second language learners. Participants exhibited comparable short-term priming for adverb-first word order in both experiments. In the initial baseline phase, participants produced adverb-first sentences with temporal phrases but not locative phrases, and only temporal phrases led to significant long-term priming, as measured in a postpriming phase. This suggests that at lower proficiency levels, long-term, but not short-term, priming may depend on the stability of specific semantically constrained constructions rather than more generalized syntactic representations and that such cumulative effects may be shaped by preferences for a particular construction in the native language.

Evaluating the predictions of three syntactic frameworks for mixed determiner–noun constructions

Abstract This paper presents a comparative evaluation of three linguistic frameworks, the Minimalist Programme (MP), Word Grammar (WG) and the Matrix Language Frame Model (MLF), regarding their predictions of possible combinations in a corpus of 187 German–English code-switched (CS) determiner–noun constructions. The comparison revealed a significant difference in the accuracy of the predictions between the MP and WG, but not between the other frameworks. We draw attention to the fact that while WG and MP deal with the processes of feature agreement between determiner and noun, the MLF is concerned with a broader notion of agreement in language membership. We suggest that advances in our understanding of grammaticality in code-switching will be achieved by combining the insights of all three frameworks instead of considering them in isolation.

Sentence alignment using local and global information

Parallel corpora are essential resources for statistical machine translation (SMT) and cross language information retrieval (CLIR) systems. Creating parallel corpora is highly expensive in terms of both time and cost. In this paper, we propose a novel approach to automatically extract parallel sentences from aligned documents. To do so, we first train a Maximum Entropy binary classifier to compute the local similarity between each two sentences in different languages. To consider global information (e.g., the position of sentence pairs in the aligned documents), we define an objective function to penalize the cross alignments and then propose an integer linear programming approach to optimize the objective function. In our experiments, we focus on English and Persian Wikipedia articles. The experimental results on manually aligned test data indicate that the proposed method outperforms the baselines, significantly. Furthermore, the extrinsic evaluations of the corpus extracted from Wikipedia on both SMT and CLIR systems demonstrate the quality of the extracted parallel sentences. In addition, Experiments on the English–German language pair demonstrate that the proposed ILP method is a language-independent sentence alignment approach. The extracted English–Persian parallel corpus is freely available for research purposes.

Rising to the bilingual challenge: self-reported experiences of managing life with two languages

ABSTRACTA recent surge of findings on bilingual cognitive advantage has attracted attention from both researchers and the media. An advantage has been demonstrated with regard to, inter alia, inhibiting, switching, monitoring and updating. However, Paap et al. argue that the advantage does not exist or is only limited to executive functioning. Both sides of the debate are well grounded in research. Nevertheless, what seems to be missing from the current discussion is the voice of bilingual speakers, that is , how they manage life with two languages and if they perceive any cognitive, social or economic advantages or indeed, disadvantages of knowing two languages more than one language. Here, I report on an extensive questionnaire in which German–English participants reflected upon their use of languages and the particular benefits as well as the challenges they face. Constructs, such as inhibition, retrieval of words language mixing were explored. It has been elicited that all participants considered knowledge of two languages to be highly advantageous despite encountering occasional problems in the form of: tip of the tongue states, code switching or relying on literal translation. In general, all participants agreed that the positive aspects of being bilingual greatly outweigh the negatives.

Sensitivity to parasitic gaps inside subject islands in native and non-native sentence processing

We report the results from an eye-movement monitoring study that investigated late German–English bilinguals’ sensitivity to parasitic gaps inside subject islands. The online reading experiment was complemented by an offline scalar judgement task. The results from the offline task confirmed that for both native and non-native speakers, subject island environments must normally be non-finite in order to host a parasitic gap. The analysis of the reading-time data showed that, while native speakers posited parasitic gaps in non-finite environments only, the non-native group initially overgenerated parasitic gaps, showing delayed sensitivity to island-inducing cues during online processing. Taken together, our findings show that non-native comprehenders are sensitive to exceptions to island constraints that are not attested in their native language and also rare in the L2 input. They need more time than native comprehenders to compute the linguistic representations over which the relevant restrictions are defined, however.

A systematic study of knowledge graph analysis for cross-language plagiarism detection

Cross-language plagiarism detection aims to detect plagiarised fragments of text among documents in different languages. In this paper, we perform a systematic examination of Cross-language Knowledge Graph Analysis; an approach that represents text fragments using knowledge graphs as a language independent content model. We analyse the contributions to cross-language plagiarism detection of the different aspects covered by knowledge graphs: word sense disambiguation, vocabulary expansion, and representation by similarities with a collection of concepts. In addition, we study both the relevance of concepts and their relations when detecting plagiarism. Finally, as a key component of the knowledge graph construction, we present a new weighting scheme of relations between concepts based on distributed representations of concepts. Experimental results in Spanish–English and German–English plagiarism detection show state-of-the-art performance and provide interesting insights on the use of knowledge graphs.

Structural effects of English–German language contact in translation on concessive constructions in business articles

AbstractStudies on a variety of languages have observed a shift away from hypotactic, hierarchical structures towards paratactic, incremental structures, and have attributed this to language contact with English in translation. This paper investigates such a shift towards parataxis as the preferred structure of concessive constructions in German business articles. To this effect, a diachronic corpus method that has been applied to popular science articles in existing studies is adopted and applied to business articles, in an attempt to reproduce existing findings for this genre. This method is complemented by a corpus of manuscripts which allow to control for the effect of editing on the translated texts. Based on the analysis of hypotactic and paratactic translations of English concessive conjunctions between 1982/83 and 2008, I argue that hypotactic structures are indeed used less frequently in translated texts, but that this development is restricted to translated language. In non-translated texts, the use of hypotactic conjunctions has increased. The use of sentence-initial conjunctions, however, does seem to spread in this genre (as was reported for popular science), which may be further evidence for it to be a case of language change through contact in translation.

The Origin of the Article in Indo-European Languages of Western Europe

This paper is concerned with the origin of the article in Indo-European languages of Western Europe. Several hypotheses concerning the origin of the article are studied, including the hypothesis of spontaneous and independent development, the Arabic-origin hypothesis, the Hebrew-origin hypothesis and the Biblical-origin hypothesis. We suggest that the main source of article borrowing into the ancient languages of Western Europe (Germanic and Romance) was the Bible. Supposedly, the grammatical category in question penetrated into the languages when the Bible was translated into national languages. We present a historical analysis of literary monuments in Old French, Old Spanish, Old German, and Old English. This shows that these languages had acquired the article before the Bible was translated into the mentioned national languages. It allows us to suppose that Ulfilas’ Gothic Bible, which appeared earlier, was the source of penetration of the article into Western European languages. This assumption is based on the analysis of literary monuments in ancient languages spoken in Europe, as well as on the comparison of the geographical spread of the article in European languages and the map of Gothic conquests in the 6th century AD. Some of the research draws upon the electronic linguistic resources WALS (http://wals.info), the “Languages of the World” database of Institute of Linguistics of Russian Academy of Sciences, and the data of ASJP project (http://asjp.clld. org/). DOI: 10.5901/mjss.2015.v6n5s4p61

Exploring the functional locus of language switching: Evidence from a PRP paradigm

To identify the stages of language processing at which language switching takes place, we examined the effect of partially temporally overlapping language processing on language switch costs in a dual-task study, in which we varied the stimulus onset asynchrony (SOA), the language (German–English), and the language sequence (language repetition trials and language switch trials). Using language comprehension tasks, the data revealed an effect of SOA (i.e., the “PRP” effect) and language switch costs. However, we did not find any influence of SOA on language switch costs, which suggests that language switches occur at the stages of lexical selection or post-selection processing. Thus, the findings of the present study provide evidence for models that assume language control to occur during or after lexical selection.