847 research outputs found
Lectometry and latent variables
No full textEver since its first formulation in Geeraerts, Grondelaers & Speelman (1999), lectometry has been widely used to map distances between language varieties or ‘lects’. Often, these distances are given a geometrical representation in a low-dimensional space. Examples are the use of Multidimensional Scaling in Speelman, Grondelaers & Geeraerts (2003) and Ruette, Geeraerts, Peirsman & Speelman (2014) and of Correspondence Analysis in Plevoets (2008), Delaere, De Sutter & Plevoets (2012), Prieels, Delaere, Plevoets & De Sutter (2015) and Ghyselen (2016). Usually, the number of dimensions of the geometrical space is chosen on the basis of representativeness, leading to an approximate picture of the linguistic variation. However, the spatial dimensions can also be interpreted as underlying factors governing the variability of the data. This methodological paper will explore this functional interpretation of the geometrical dimensions by establishing the link between lectometry and Latent Variable Models. It will be shown that the dimensions of the lectal space can be considered as hidden variables which lay bare specific causal mechanisms. In particular, analyses of translation and interpreting data will demonstrate that the lectometrical dimensions can be made to correspond to various socio-cultural determinants. That opens up the possibility for lectometrical studies of determining the ‘social meaning’ of linguistic varieties and variants
Discourse markers and turn-planning at the pragmatics-prosody interface: the case of "allora" in spoken Italian
No full textIn recent research on discourse markers (DMs), it has been shown that these elements collaborate with prosody to convey crucial information about the organization of the discourse, both at the level of a single turn and the level of the overall conversation. In this article, we analyze the DM allora ‘then’, occurring in turn-initial position and followed by a silent pause in spontaneous monolingual dialogues between Italian native speakers. Various prosodic measures are taken comparing allora with the following intonational phrase (IP), to investigate the discursive relationship between the DM and the rest of the turn. The results show that the DM is peripheric as opposed to the general turn's prosodic planning: it is produced at a higher articulation rate and with a narrower range of pitch values than the following IP, and a pitch reset is visible between the two. The data suggest that allora and the following IP constitute two distinct units in the turn's planning. We argue that as m..
Lexis or parsing? A corpus-based study of syntactic complexity and its effect on disfluencies in interpreting
No full textCognitive load is probably one of the most cited topics in research on simultaneous interpreting, but it is still poorly understood due to the lack of proper empirical tests. It is a central concept in Gile’s (2009) Efforts Model as well as Seeber’s (2011) Cognitive Load Model. Both models invariably conceptualize interpreting as a dynamic equilibrium between the cognitive resources/capacities and cognitive demands that are involved in listening and comprehension, production and memory storage. In cases when the momentary demands exceed the interpreter’s available capacities, there is an information overload which typically results in a disfluent or erroneous interpretation. While Gile (2008) denies his Efforts Model is a theory that can be tested, Seeber & Kerzel (2012) put Seeber’s Cognitive Load Model to the test using pupillometry in an experimental interpretation task.
In a series of recent corpus-based studies Plevoets & Defrancq (2016, 2018) and Defrancq & Plevoets (2018) used filled pauses to investigate cognitive load in simultaneous interpreters, based on the widely shared assumption in the psycholinguistic literature that silent and filled pauses are ‘windows’ on cognitive load in monolingual speech (Arnold et al. 2000; Bortfeld et al. 2001; Clark & Fox Tree 2002; Levelt 1983; Watanabe et al. 2008). The studies found empirical support for increased cognitive load in simultaneous interpreting in the form of higher frequencies of filled pauses. However, the studies also showed that filled pauses in interpreting are caused mainly by problems with lexical retrieval. Plevoets & Defrancq (2016) observed that interpreters produce more instances of the filled pause uh(m) when the lexical density of their own output is higher. Plevoets & Defrancq (2018) demonstrated that the frequency of uh(m) in interpreting increases when the lexical density of the source text is also higher but it decreases when there are more formulaic sequences. This effect of formulaicity was found in both the source texts and the target texts. Other known obstacles in interpreting, such as the presence of numbers and rate of delivery do not significantly affect the frequency of filled pauses (although source speech delivery rate reached significance in one of the analyses). These results point to the problematic retrieval or access of lexical items as the primary source of cognitive load for interpreters. Finally, in a study of filled pauses occurring between the members of morphological compounds, Defrancq & Plevoets (2018) showed that interpreters produced more uh(m)’s than non-interpreters when the average frequency of the compounds was high as well as when the average frequency of the component members was high. This also demonstrates that lexical retrieval, which is assumed to be easier for more frequent items, is hampered in interpreting.
This study critically examines the results of the previous studies by analyzing the effect of another non-lexical parameter on the production of filled pauses in interpreting, viz. syntactic complexity. Subordinating constructions are a well-known predictor of processing cost (cognitive load) in both L1 research (Gordon, Luper & Peterson 1986; Gordon & Luper 1989) and L2 research (Norris & Ortega 2009; Osborne 2011). In interpreting, however, Dillinger (1994) and Setton (1999: 270) did not find strong effects of the syntactic embedding of the source texts on the interpreters’ performance. As a consequence, this paper will take a closer look on syntactic complexity and it will do so by incorporating the number of hypotactic clauses into the analysis.
The study is corpus-based and makes use of both a corpus of interpreted language and a corpus of non-mediated speech. The corpus of interpreted language is the EPICG corpus, which is compiled at Ghent University between 2010 and 2013. It consists of French, Spanish and Dutch interpreted speeches in the European Parliament from 2006 until 2008, which are transcribed according to the VALIBEL guidelines (Bachy et al. 2007). For the purposes of this study a sub-corpus of French source speeches and their Dutch interpretations is used, amounting to a total of 140 000 words. This sub-corpus is annotated for lemmas, parts-of-speech and chunks (Van de Kauter et al. 2013), and it is sentence-aligned with WinAlign (SDL Trados WinAlign 2014). The corpus of non-mediated speech is the sub-corpus of political debates of the Spoken Dutch Corpus (Oostdijk 2000). The corpus was compiled between 1998 and 2003, and it is annotated for lemmas and parts-of-speech. The political sub-corpus contains 220 000 words of Netherlandic Dutch and 140 000 words of Belgian Dutch.
The data are analysed with a Generalized Additive Mixed-effects Model (Wood 2017) in which the frequency of the disfluency uh(m) is predicted in relation to delivery rate, lexical density, percentage of numbers, formulaicity and syntactic complexity. Delivery rate is measured as the number of words per minute, lexical density as the number of content words per utterance length, percentage of numbers as the numbers of numbers per utterance length and formulaicity as the number of n-grams per utterance length. The new predictor, syntactic complexity, is measured as the number of subordinate clauses per utterance length. Because all five predictors are numeric variables, their effects are modelled with smoothing splines which automatically detect potential nonlinear patterns in the data. The observations are at utterance-level and are nested within the speeches, so the possible between-speech variation is accounted for with a random factor.
The preliminary results confirm the hypothesis: while lexical density and formulaicity show similar (positive, resp. negative) effects to what is reported in previous research, the syntactic complexity of the source text is ‘border-significant’ and the syntactic complexity of the target is non-significant. There are some sporadic differences among certain types of subordinate clauses, but the general conclusion is indeed that syntactic complexity is not such a strong trigger of cognitive load in interpreting in comparison to lexically-related factors. That calls for a model of interpreting in which depth of processing plays only a marginal role
Accepting Optimally in Automated Negotiation with Incomplete Information (abstract)
No full textIntelligent SystemsElectrical Engineering, Mathematics and Computer Scienc
Lexical variation in interlingual and intralingual audiovisual translation: a corpus-based, multivariate study of Belgian Dutch subtitles
No full textWith this study, we investigate whether the lexical choices of translators, subtitlers and original authors differ significantly. In order to achieve that goal, we gathered a set of linguistic variables and used a 500-million word balanced reference corpus for contemporary (1954-present) written Dutch (the SoNaR corpus; Schuurman et al. 2010). By means of a profile-based correspondence analysis (Plevoets, 2008), lexical distances between the translation and original varieties are measured and visualized in a two-dimensional plot. The results indeed reveal significant differences between interlingual and intralingual subtitles, between subtitles and written translations, and between subtitles and original texts. Influential variables explaining the variation are speakers' status and translation type
Is translated language more standardized than non-translated language? Using profile-based correspondence analysis for measuring linguistic distances between language varieties
No full textWith this article, we seek to support the law of growing standardization by showing that texts translated into Belgian Dutch make more use of standard language than non-translated Belgian Dutch texts. Additionally, we want to examine whether the use of standard vs. non-standard language can be attributed to the variables text type and source language. In order to achieve that goal, we gathered a diverse set of linguistic variables and used a 10-million-word corpus that is parallel, comparable and bidirectional (the Dutch Parallel Corpus; Macken et al. 2011). The frequency counts for each of the variables are used to determine the differences in standard language use by means of profile-based correspondence analysis (Plevoets 2008). The results of our analysis show that (i) in general, there is indeed a standardizing trend among translations and (ii) text types with a lot of editorial control (fiction, non-fiction and journalistic texts) contain more standard language than the less edited text types (administrative texts and external communication) which adds support for the idea that the differences between translated and non-translated texts are text type dependent
Recommended from our members
Episode #2: Interview with President Koen Lenaerts
No full textEpisode #2 of the Borderlines CJEU Series features CJEU President Koen Lenaerts in conversation with Professor Katerina Linos (Berkeley) and Professor Mark Pollack (Temple University). President Lenaerts has been re-elected to the Court’s top office by his peers three times since 2015, having served tirelessly since his nomination to the CJEU in 2003 by home country of Belgium. His interview traces the historic path to today’s Court of Justice, and illuminates differences from other courts, including the U.S. federal judicial system.Listeners will come away with an overview of the Court’s functions and structural methodology in interpreting the EU legal order, including the role of the President, the Judge rapporteur, and the Advocate General. Presidential responsibilities include assigning cases to Judges and presiding over the Grand Chamber to deal with the most important cases. New developments in case law, evolving technological access, and finding balance between unity and diversity, privacy and security, are addressed by the head of the EU’s judicial institution.President Lenaerts in total has spent 35 years as a European Union Judge, initially serving on the Court of First Instance of the European Communities (now the General Court) when it was established in 1989. He earned his legal degrees including a doctorate in Belgium and also obtained a Masters of Laws and a Masters in Public Administration from Harvard University. President Lenaerts is Professor of European Union Law at Leuven University in Belgium and a member of many legal and academic associations including the Academia Europaea, London; the Advisory Council of the British Institute of International and Comparative Law; the Advisory Board of the Centre of Law and Governance in Europe, University College London; the Governing Board of the Foundation of the Academy of European Law (ERA), Trier; and the Board of Trustees of the Max-Planck-Institut für ausländisches öffentliches Recht und Völkerrecht, Heidelberg. He is the recipient of numerous awards and author of a vast list of publications.</p
Reviving Modernist Housing: Arena District of Renaat Braem
No full textThe modernist housing project ‘Arenawijk’, in Deurne, Antwerp (BE), is currently threatened with (partial) demolition in order to make place for a new, more contemporary housing development. The initial plans were designed by Renaat Braem (1910–2001), one of the most prominent Belgian modernist architects and urban planners. The Arena district can be considered as an interesting cohesion between the internationally acknowledged CIAM principles and the, sometimes contradictorily, personal theories formulated by Braem himself. Hence, the site is a very significant part of Belgian modernist heritage. Nevertheless, the project is not protected. The demolition plans were initiated due to various reasons. This contribution addresses the historical richness and uniqueness of the Arena district as starting point for research-by-design. It is assumed that precise architectural interventions, based on a process of research-by-design, may herald new possibilities for reuse. The first part of the paper is based on the emergence of the site and illustrates Braem, as polemic figure, in a broader context. A second part will point out the value assessment of the unique characteristics of the site, based on the first historical situation and the new layers added by Braem. In conclusion, some further research suggestions will be made.Fonds Wetenschappelijk Onderzoek - 1116421
Correction to: Decisional Balance Inventory (DBI) Adolescent Form for Smoking: Psychometric Properties of the Persian Version
No full textCorrection After publication of the article [1], it has been brought to our attention that the first and last names of the third author were transposed in the original article. The author was published as “Ponnet Koen” where in fact the correct name is “Koen Ponnet”. The original article has been revised to reflect this
RESEARCH BY DESIGN AS METHOD FOR THE REGENERATION OF POST-WAR MODERNIST HOUSING ENSEMBLES
No full textThis paper is written in the light of a larger PhD research entitled BELGIAN MOMO HERITAGE ON THE RADAR – ‘Re-reading modernist housing estates: an inquiry into the value of threatened heritage sites and the possibilities of adaptive reuse as a method for re-evaluation’. Today, a considerable number of the urban settlements or high-rise estates, constructed according to Congrès Internationaux d'Architecture Moderne (CIAM) principles, are threatened with demolition following years of neglect and lack of maintenance. Notwithstanding the fact that they are all comprising discursive segments of the post-war time period, these estates are in poor (technical and social) condition. Additionally, none of them is protected, which means that the path for demolition is fully open. This research opens up some future visions by considering these ensembles as an opportunity to re-state and re-invest, instead of regarding them as an inescapable problem. The concept of research by design is used as a methodology to develop new insights. Within the framework of the PhD research architecture master students have tested the hypotheses of different regeneration scenarios on a larger scale of 13 case studies. The aim is to investigate the adaptive reuse potential of this modernist typology. This paper focuses on one of the case studies, specifically the Jan De Voslei housing complex designed by architect Jos Smolderen (Antwerp, BE). In the first part of the paper the general problem statement is introduced. In the second part, the objective of the exercise is explored, and an overview of the regeneration scenarios is explained. Subsequently, the students’ results that focus on an activation of the landscape values and the so-called parasite-concept of the Jan De Voslei housing complex are presented through drawings and isometries. In conclusion, these first outcomes of this exercise are discussed combined with a reflection and suggestion for further research.Fonds Wetenschappelijk Onderzoek (1116421N
- …
