Corpus Linguistics 2013 Abstract Book Edited by Andrew Hardie and Robbie Love Lancaster: UCREL Table of Contents Plenaries What can translations tell us about ongoing semantic changes? The case of must KARIN AIJMER Taking a Language to Pieces: art, science, technology GUY COOK The textual dimensions of Lexical Priming MICHAEL HOEY, MATTHEW BROOK O’DONNELL No corpus linguist is an island: Collaborative and cross-disciplinary work in researching phraseology UTE RÖMER Papers A corpus-based study for assessing the collocational competence in learner production across proficiency levels MAHA N ALHARTHI ‘Sure he has been talking about coming for the last year or two”: the Corpus of Irish English Correspondence and the use of discourse markers CAROLINA P AMADOR-MORENO, KEVIN MCCAFFERTY 13 Developing AntConc for a new generation of corpus linguists LAURENCE ANTHONY 14 Bridging lexical and constructional synonymy, and linguistic variants – the Passive and its auxiliary verbs in British and American English ANTTI ARPPE, DAGMARA DOWBOR 16 An open-access gold-standard multi-annotated corpus with huge user-base and impact: The Quran ERIC ATWELL, NORA ABBAS, BAYAN ABUSHAWAR, CLAIRE BRIERLEY, KAIS DUKES, MAJDI SAWALHA, ABDULBAQUEE MUHAMMAD SHARAF 19 Triangulating levels of focus and the analysis of personal adverts on Craigslist PAUL BAKER 21 Robust corpus architecture: a new look at virtual collections and data access PIOTR BAŃSKI, ELENA FRICK, MICHAEL HANL, MARC KUPIETZ, CARSTEN SCHNOBER, ANDREAS WITT 23 Exemplar theory and patterns of production MICHAEL BARLOW 26 i The construction of otherness in the public domain: a CDA approach to the study of minorities in Ireland LEANNE BARTLEY, ENCARNACION HIDALGO-TENORIO 28 Exploring the Firthian notion of collocation SABINE BARTSCH, STEFAN EVERT 31 A corpus-based study of the Non-Obligatory Suppression Hypothesis(of Concepts in the Scope of Negation) ISRAELA BECKER 34 Integrating visual analysis into corpus linguistic research MONIKA BEDNAREK 37 Individual and gender variation in spoken English: Exploring BNC 64 VACLAV BREZINA 39 Automatically identifying instances of change in diachronic corpus data ANDREAS BUERKI 42 Reader engagement in Turkish EFL students’ argumentative essays DUYGU ÇANDARLI, YASEMIN BAYYURT, LEYLA MARTI 44 It was X that type of cleft sentences and their Czech equivalents in InterCorp ANNA ČERMÁKOVÁ, FRANTIŠEK ČERMÁK 45 The power of personal corpora: Students’ discoveries using a do-it-yourself resource MAGGIE CHARLES 48 Basic vocabulary and absolute homonyms: a corpus-based evaluation ISABELLA CHIARI 51 Using lockwords to investigate similarities in Early Modern English drama by Shakespeare and other contemporaneous playwrights JONATHAN CULPEPER, JANE DEMMEN 53 Not all keywords are created equal: How can we measure keyness? VÁCLAV CVRČEK 55 Context-based approach to collocations: the case of Czech VÁCLAV CVRČEK, ANNA ČERMÁKOVÁ, LUCIE CHLUMSKÁ, RENATA NOVOTNÁ, OLGA RICHTEROVÁ 57 A corpus-based study on the relationship between word length and word frequency in Chinese DENG YAOCHEN, FENG ZHIWEI 59 “Anyway, the point I'm making is”: relevance marking in lectures KATRIEN DEROEY 61 Visualizing chunking and collocational networks: a graphical visualization of words’ networks MATTEO DI CRISTOFARO 63 Using learner corpus tools in second language acquisition research: the morpheme order studies revisited ANA DÍAZ-NEGRILLO, CRISTÓBAL LOZANO 64 ii Risk, chance, hope – the lexis of possible outcomes and infertility KAREN DONNELLY 67 Scots online: Linguistic practices of a distinctive message forum FIONA M DOUGLAS 70 Linking adverbials in the academic writing of Chinese learners: a corpus-based comparison DU PENG 71 Public apologies and press evaluations: a CADS approach ALISON DUGUID 73 Using reference corpora for discourse analysis research: the case of class ROSA ESCANES SIERRA 75 Statistical modelling of natural language for descriptive linguistics STEFAN EVERT, GEROLD SCHNEIDER, HANS MARTIN LEHMANN 77 Literature and statistics – a corpus-based study of endings in short stories JENNIFER FEST, STELLA NEUMANN 80 Corpus Linguistics and English for Specific Purposes: Which unit for linguistic analysis? LYNNE FLOWERDEW 82 Corpus frequency or the preference of dictionary editors and grammarians?: the negative and question forms of used to KAZUKO FUJIMOTO 84 Discourse characteristics of English in news articles written by Japanese journalists: ‘Positive’ or ‘negative’? FUJIWARA YASUHIRO 86 Negotiating trust during a corporate crisis: a corpus-assisted discourse analysis of BP’s public letters after the Gulf of Mexico oil spill MATTEO FUOLI 89 Using corpus analysis to compare the explanatory power of linguistic theories: A case study of the modal load in if-conditionals COSTAS GABRIELATOS 92 Digital corpora and other electronic resources for Maltese ALBERT GATT, SLAVOMÍR CÉPLƯ 96 The role of the speaker’s linguistic experience in the production of grammatical agreement: A corpus-based study of Russian speech errors SVETLANA GOROKHOVA 98 Keywords, lexical bundles and phrase frames across English pharmaceutical text types: A corpus-driven study of register variation ŁUKASZ GRABOWSKI 100 Lexical density in writing assignments by university first year students CARMEN GREGORI-SIGNES, BEGOÑA CLAVEL-ARROITIA 104 Geographical Text Analysis: Mapping and spatially analysing corpora IAN GREGORY, ALISTAIR BARON, PATRICIA MURRIETA-FLORES, ANDREW HARDIE, PAUL RAYSON 105 iii The role of phonological similarity and collocational attraction in lexically-specified patterns STEFAN TH GRIES 108 A triangulated approach to media representations of the British women's suffrage movement KAT GUPTA 110 “Obvious trolls will just get you banned”: Trolling versus corpus linguistics CLAIRE HARDAKER 112 Lexical bundles performed by Chinese EFL learners: From quantity to quality analysis DICK KAISHENG HUANG 114 A complementary approach to corpus study: a text-based exploration of the factors in the (non-) use of discourse markers LAN-FEN HUANG 116 Lexical bundles in private dialogues and public dialogues: A comparative study of English varieties DORA ZEPING HUANG 119 SAE11: a new member of the family SALLY HUNT, RICHARD BOWKER 121 Bridging genres in scientific dissemination: popularizing the ‘God particle’ ERSILIA INCELLI 123 The TenTen Corpus Family MILOŠ JAKUBÍCEK, ADAM KILGARRIFF, VOJTECH KOVÁR, PAVEL RYCHLÝ, VIT SUCHOMEL 125 Imagining the Other: corpus-based explorations into the constructions of otherness in the discourse of tourism SYLVIA JAWORSKA 127 “Hold on a minute; where does it say that?” – Calculating key section headings and other metadata for words and phrases STEPHEN JEACO 130 Rape, madness, and quoted speech in specialized 18th and 19th century Old Bailey trial corpora ALISON JOHNSON 132 Family in the UK – risks, threats and dangers: a modern diachronic corpus-assisted study across two genres JANE HELEN JOHNSON 135 Reader comments on online news articles: a corpus-based analysis ANDREW KEHOE, MATT GEE 137 Collocation analysis and marketized university recruitment discourse BARAMEE KHEOVICHAI 139 Genre in a frequency dictionary ADAM KILGARRIFF, CAROLE TIBERIUS 142 A macroanalytic view of Swedish literature using topic modeling DIMITRIOS KOKKINAKIS, MATS MALM 144 iv Czech nouns derived from verbs with an objective genitive: Their contribution to the theory of valency VERONIKA KOLÁŘOVÁ 147 MotionML: Motion Markup Language – a shallow approach for annotating motions in text OLEKSANDR KOLOMIYETS, MARIE-FRANCINE MOENS 151 Use of dedicated multimodal corpora for curriculum implications of EAP/ESP programs in ESL settings MENIKPURA DSS KUMARA 154 Early Modern English vocabulary growth IAN LANCASHIRE, ELISA TERSIGNI 156 Detecting cohesion: semi-automatic annotation procedures EKATERINA LAPSHINOVA-KOLTUNSKI, KERSTIN ANNA KUNZ 160 Procedures for automatic corpus enrichment with abstract linguistic categories EKATERINA LAPSHINOVA-KOLTUNSKI, STEFANIA DEGAETANO-ORTLIEB, HANNAH KERMES, ELKE TEICH 163 The correlation between lexical core index, age-of-acquisition, familiarity and imageability JOHN HANHONG LI 167 Phraseological discourse actors in English academic texts JINGJIE LI, WENJIE HU 171 China English Corpus construction on an open corpus platform LI WENZHONG 173 Sparing a free hand: context-based automatic categorisation of concordance lines MAOCHENG LIANG 175 ‘What is the environment doing in my report?’ Analysing the environment-as-stakeholder thesis through corpus linguistics ALON LISCHINSKY 177 Using quantitative measures to investigate the relative roles of languages participating in codeswitched utterances CATHY LONNGREN-SAMPAIO 179 “The results demonstrate that …” A corpus-based analysis of evaluative that-clauses in medical posters STEFANIA M MACI 181 Reading Dickens’s characters: investigating the cognitive reality of patterns in texts MICHAELA MAHLBERG, KATHY CONKLIN 183 Experimenting with objectivity in corpus and discourse studies: expectations about LGBT discourse and a game of mutual falsification and reflexivity ANNA MARCHI, CHARLOTTE TAYLOR 184 Have – causative, or experiential? A parallel corpus-based study MICHAELA MARTINKOVÁ 186 v Annotating translation errors in Brazilian Portuguese automatically translated sentences: first step to automatic post-edition DÉBORA BEATRIZ DE JESUS MARTINS, LUCAS VINICIUS AVANÇO, MARIA DAS GRAÇAS VOLPE NUNES, HELENA DE MEDEIROS CASELI 189 Corpus-driven terminology and cultural aspects: studies in the areas of football, cooking and hotels SABRINA MATUDA, ROZANE REBECHI, SANDRA NAVARRO 192 Is there a reputational benefit to hosting the Olympics and Paralympics? A corpus-based investigation TONY MCENERY, AMANDA POTTS, RICHARD XIAO 195 Take a mirror and take a look: Reassessing usage of polysemic verbs with concrete and light senses SETH MEHL 197 A corpus linguistic study of ellipsis as a cohesive device KATRIN MENZEL 202 Student perceptions of university instructors: A multi-dimensional analysis of free-text comments on RateMyProfessors.com NEIL MILLAR 205 Hierarchical cluster analysis of nonlinear linguistic data HERMANN MOISL 208 An affix-based method for automatic term recognition from a medical corpus of Spanish ANTONIO MORENO-SANDOVAL, LEONARDO CAMPILLOS LLANOS, ALICIA GONZÁLEZ MARTÍNEZ, JOSÉ M GUIRAO MIRAS 214 Longitudinal development of L2 English grammatical morphemes: A clustering approach AKIRA MURAKAMI 217 Exploring intra-author variation across different modes of electronic communication using the FITT corpus MILLICENT MURDOCH 220 Integrating corpus linguistics and spatial technologies for the analysis of literature PATRICIA MURRIETA-FLORES, IAN GREGORY, DAVID COOPER, CHRISTOPHER DONALDSON, ALISTAIR BARON, ANDREW HARDIE, PAUL RAYSON 222 Citation in student assignments: a corpus-driven investigation HILARY NESI 225 Reporting the 2011 London riots: a corpus-based discourse analysis of agency and participants MARIA CRISTINA NISCO 228 Semantically profiling and word sketching the Singapore ICNALE Corpus VINCENT B Y OOI 230 Intimations of Spring? Political and media coverage – and non-coverage – of the Arab uprisings, and how corpus linguistics can speak to “absences” ALAN PARTINGTON 233 vi Using corpus data to calculate a rote-learning threshold for personal pronouns: You as a target for They and He LAURA LOUISE PATERSON 236 The identification of metaphor using corpus methods: Can a re-classification of metaphoric language help our understanding of metaphor usage and comprehension? KATIE PATTERSON 237 Stance adverbials in research writing MATTHEW PEACOCK 239 A pragmatic analysis of imperatives in voice-overs from a corpus of British TV ads BARRY PENNOCK-SPECK, MIGUEL FUSTER-MÁRQUEZ 242 A defence of semantic preference GILL PHILIP 244 Automated semantic categorisation of collocates to identify salient domains: A corpus-based critical discourse analysis of naming strategies for people with HIV/AIDS AMANDA POTTS 246 Linking qualitative and quantitative analysis of metaphor in end-of-life care PAUL RAYSON, ANDREW HARDIE, VERONIKA KOLLER, SHEILA PAYNE, ELENA SEMINO, ZSÓFIA DEMJÉN, MATT GEE, ANDREW KEHOE 249 Investigating orality in speech, writing, and in between INES REHBEIN, JOSEF RUPPENHOFER 251 It is surprising: participial adjectives after copular verbs form a special evaluative construction? OLGA RICHTEROVÁ 254 The empirical trend: ten years on GEOFFREY SAMPSON 256 Identifying discourse(s) and constructing evaluative meaning in a gender-related corpus (GENTEXT-N) JOSÉ SANTAEMILIA, SERGIO MARUENDA 259 Comparing morphological tag-sets for Arabic and English MAJDI SAWALHA, ERIC ATWELL 261 Comparing collocations in the totalitarian language of the former Czechoslovakia with the language of the democratic period VĚRA SCHMIEDTOVÁ 265 Linguistic means of knowledge transfer through knowledge-rich contexts in Russian and German ANNE-KATHRIN SCHUMANN 267 The discursive representation of animals ALISON SEALEY 271 Building a corpus of evaluative sentences in multiple domains JANA SINDLEROVÁ, KATERINA VESELOVSKÁ 273 vii Lexical, corpus-methodological and lexicographic approaches to paronyms PETRA STORJOHANN 275 Verbs with a sentential subject: A corpus-based study of German and Polish verbs JANUSZ TABOREK 277 “Criterial feature” extraction from CEFR-based corpora: Methods and techniques YUKIO TONO 280 Reflexivity of high explicitness metatext in L1 and FL research articles from the Soft and Hard Sciences: A corpus-based study NAOUEL TOUMI 282 Instrumental and integrative approaches to language in Canada: A cross-linguistic corpusassisted discourse study of Canadian language ideologies RACHELLE VESSEY 284 V wh semantic sequences: the communicating function BENET VINCENT 286 The role of corpus linguistics in social constructionist discourse analysis FANG WANG 289 Using life-logging to re-imagine representativeness in corpus design STEPHEN WATTAM, PAUL RAYSON, DAMON BERRIDGE 290 Code-mixing: exploring indigenous words in ICE-HK MAY L-Y WONG 293 Using corpora in forensic authorship analysis: Investigating idiolect in Enron emails DAVID WRIGHT 296 A multidimensional contrastive move analysis of native and nonnative English abstracts RICHARD XIAO, YAN CAO 299 The metaphoricity of fish: implications for part-of-speech and metaphor XU HUANRONG, HOU FULI 302 The structural and semantic analysis of the English translation of Chinese light verb constructions: A parallel corpus-based study JIAJIN XU, LU LU 305 The search for units of meaning in terms of corpus linguistics: The case of collocational framework “the * of” SUXIANG YANG 307 Posters New methods of annotation: The ‘humour’ element of Engineering lectures SIÂN ALSOP 313 Oxford Children’s Corpus: a corpus of children’s writing, reading, and education NILANJANA BANERJI, VINEETA GUPTA, ADAM KILGARRIFF, DAVID TUGWELL 315 viii PMSE: text categorization – a case study Jiří Mácha ÚČNK Jiří Václavík ÚČNK jm@petamem.com jv@petamem.com Introduction This poster presents a new corpus tool called PMSE and its first major application, categorization of textual documents PetaMem Scripting Environment The PetaMem Scripting Environment (PMSE) is a generic tool for batch processing of large text corpora and can also serve as middleware for other corpus processing tools The software suite is written in Perl and consists of components which enable the user to acquire statistical information from the given texts (corpora) Great attention was paid to maintaining as universal applicability of the tool as possible Similar to UNIX philosophy, PMSE is a building kit of small units that can be combined together in a different order, not a tool focused on one specific task or functionality The crucial parts of PMSE are represented by scripts intended for computing of languagestatistical information, e.g.:  Frequencies of token occurrences  Token probabilities  Generating of n-grams  Computing numerous association measures for n-grams of various length (e.g MI-score, t-score)  Computing various distance measures among pairs of n-grams The whole tool chain of PMSE is designed to process texts from the very beginning (downloading texts, converting formats, removing formatting etc.) to the final task – computation of the statistical characteristics of the data and their visualization PMSE is designed to be language independent, it works with plain texts encoded in UTF-8 PMSE is an ongoing effort and still under development, the available functionality, however, already allows for interesting real-world applications Future plans for development involve e g.: graphical web interface, conversion between various corpora formats and annotation tags Text Categorization The authors will present an application case for PMSE – a Text Categorization project (abbreviated: TextCat) The general task for the TextCat app is to categorize various documents in any language The authors will present an example of already finished categorization of parallel texts in about 20 European languages Resulting dendrograms (one for each language from the parallel corpora) show signs of similar structures The evaluation of individual dendrograms is part of future linguistic interpretation The modularity of the source code allows the user to change the behaviour of all procedural steps, especially since TextCat is extensible by simple plugins Also, a great attention was paid to the performance of the software (efficiency as well as parallel processing) The linguistic criteria addressed and delivered by TextCat may be defined by the user The categorization process has several steps: Extract text from all documents Pre-process all the texts, extract n-grams of any size Filter n-grams according to specific criteria Filter files, exclude the inconvenient (some files could be damaged, too short, too long, non-relevant etc.) Create and precompute all existing groups (each text belongs to one group) and compute distances for all the possible pairs Find the closest groups and join them into a parent group Repeat the previous two steps until only one group remains Visualize a binary tree representing relations among the texts (dendrogram) TextCat is a modular framework which could perform categorization on any criteria – that is why it has a high coverage It could be used for language identification, corpus sorting, forensic linguistics etc References Huang, A 2008 Similarity Measures for Text Document Clustering Available on-line at http://favi.com.vn/wpcontent/uploads/2012/05/pg049_Similarity_Measure s_for_Text_Document_Clustering.pdf 349 CLEG and “die Deutschen” Ursula Maden-Weinberger Lancaster University u.weinberger@lancs.ac.uk Introduction This presentation accompanies the public release of a new resource for learner corpus research: CLEG – a Corpus of LEarner German, and exemplifies its use in an exploration of British students’ views of Germany and the German people CLEG is a significant expansion of learner corpus resources for the German language At present, most of the research effort in learner corpora is focused on English and while the amount of learner corpus resources for English has been growing substantially over the last decade, other languages receive less attention – one such language is German (The learner corpus database of the Centre for English Corpus Linguistics at Louvain-la-Neuve lists 73 English learner corpora – most of which publicly available – as opposed to German learner corpora) To date, the only notable publicly available German learner corpus is the Fehlerannotiertes Lernerkorpus (FALKO) (error-annotated learner corpus), collected and managed at the Humboldt Universität, Berlin (Reznicek et al 2012) While research into German learner language is, of course, a valuable and insightful enterprise in itself, there is another reason why the expansion of resources and research involving languages other than English is crucial Research already carried out on CLEG data has provided evidence that could be used to triangulate findings from other L2 corpus studies and thus explore general, universal tendencies in learner language These are trends that are not just L1-, but also L2independent One such undertaking with CLEG data has revealed that learners, regardless of L1 or L2, exhibit a general tendency to overuse personalised expressions in writing (MadenWeinberger 2012) The CLEG corpus CLEG is an approx 300,000 word corpus of advanced second language learner writing Contributors and texts are tightly controlled for a number of relevant criteria The learners are undergraduate students of German at Lancaster University who had achieved an A-level in German (this equates to between 5-7 years of 350 school tuition in German and a CEFR level of B1 to B2) They are all native speakers of British English and between 18 and 22 years old Texts were collected over a period of four years from all three years of the undergraduate programme (called Year A, Year B and Year C) Most students in the first two years (Year A and Year B) have spent a few weeks in Germany on vacation or as part of a school exchange The students in the final year (Year C) have all spent between six and twelve months in a Germanspeaking country as part of their “year abroad”, which is a compulsory part of the degree scheme for all language majors in the third year of study The texts chosen for the corpus can be classified as “expository-argumentative” This is defined in a operational way as texts where the task instructions imply “the presentation and weighing up of arguments, writer’s criticism or systematic outlines of abstract concepts” (Lorenz 1999:12) Incidentally, expository-argumentative texts are also the kind of texts that learners are asked to produce most frequently throughout their study course at Lancaster University This means that this collection criterion yielded the largest amount of reasonably homogeneous texts from all year groups Metadata on learner and text profiles are stored in a database, which enables researchers to link each text to the relevant text and learner information and create sub-corpora according to different specifications if desired Truly longitudinal data As the data was collected over the course of four years, all of the data is quasi-longitudinal, however there is a core of one entire cohort of students whose data is truly longitudinal from their first year through to the end of the four year degree programme This is a distinctive feature of CLEG, which provides unique opportunities for corpus research on developmental aspects of learner writing CLEG and “die Deutschen” In a “taster” example, the development of students’ views on Germany and the German people are explored through the collocations around the item “deutsch*” in the three yeargroups The following examples provide a first glance at the kind of statements to be found in the student texts: Die Deutschen [the Germans ] sind sehr freundlich (Year A) [ are very friendly] .berichten viel über Fußball (Year B) [ give football a lot of coverage] sind Experten in Bezug auf technischen Qualität (Year C) [ are experts in technical quality] Although in its infancy at the moment, this study should provide some clear evidence of a development of opinions about Germans from the British students and it will be particularly interesting to investigate the impact of the time spent in Germany before the final year of study References Lorenz, G (1999) Adjective intensification – Learners versus Native Speakers A corpus study of argumentative writing Amsterdam – Atlanta: Rodopi Maden-Weinberger, U 2012 “Personal expressions in Learner German – it’s all about the bigger picture” Presentation at the Teaching and Language Corpora Conference, Warsaw, 11-14 July 2012 Reznicek, M., Lüdeling, A and Schwantuschke, F 2012 Das Falko-Handbuch: Korpusaufbau und Annotationen: Version 2.01 Berlin Conditionals in 18th-century philosophy texts: A corpus-based study1 Leida Maria Monaco Universidade da Coruña Luis Puente Castelo leidamaria.monaco @udc.es luis.pcastelo @udc.es Universidade da Coruña If a conditional clause can be defined as “a rhetorical device for gaining acceptance for one’s claims” (Warchal 2010: 141), it is not surprising that conditionals play a very important role in argumentative strategies in virtually any register within both spoken and written discourse This is especially the case in scientific register, where conditionals appear to be particularly frequent (Athanasiadou & Dirven 1997; Ferguson 2001: 69), not only in order to express the relationship between a phenomenon and its consequence, but also to state a hypothesis and/or speculate on possible outcomes of events Moreover, it has been noted that conditionals are frequently used as downtoning devices by means of which a claim would presumably sound less assertive or categorical, functioning thus as metadiscoursive strategies, or hedges, and therefore playing a mediating role in the relationship between the authors and their discourse community (Hyland 1994, 1998a, 1998b; Declerck & Reed 2001; Carter-Thomas & Rowley-Jolivet 2008; Warchal 2010: 141-142) Within scientific discourse, the argumentative function of conditionals may be noted best in highly speculative fields such as that of Philosophy, as “conditionals feature prominently in deductive argument – for example in the classical rules of inference known as modus ponens and modus tollens” (Ferguson 2001: 62) This argumentative function is expected to be particularly prominent in earlier philosophical texts, as these stem from the decaying roots of scholasticism, a knowledge framework which paid special attention to the logical form of the argument, frequently demanding the use of “common logical terms to organize discourse and build up the arguments” (Taavitsainen 1999: 249) On the other hand, conditionals are one of the This research was funded by the Consellería de Educación e Ordenación Universitaria (I2C plan, reference number Pre/2011/096, co-funded 80% by the European Social Fund) and the Ministerio de Ciencia e Innovación (FPU grant, reference number AP2009-3206) These grants are hereby gratefully acknowledged 351 linguistic parameters used in Biber’s (1988, 1995) multi-dimensional analysis of register variation, namely in his Dimension labelled “Overt expression of persuasion” According to Biber (1988: 111, 148-151), the linguistic features cooccurring in Dimension have a common underlying suasive function Likewise, in further multi-dimensional studies, such as Biber’s multiregister overview of 18th century texts (2001) and the new multi-dimensional analysis of present-day English academic writing (Biber et al 2004), conditional subordination appears to be grouped with linguistic features conveying an ‘oral’, vs a ‘literate’, type of discourse This may appear contradictory at first sight, as the relatively high use of conditionals in scientific discourse conflicts with its grouping with oral features in Biber’s most recent work A reason for such a characterisation of conditionals is presumed to lie in their inherent speculative function, which could be described otherwise as self-persuasion from an introspective point of view or as a mechanism to direct the discourse and to put forward the author’s ideas to the readers from a pragmatic perspective (Biber 1988: 111) The aim of this study is to describe the wideranging field of conditional sentences in eighteenth-century philosophy texts, describing their diachronic evolution, the uses of the different varieties and their pragmatic functions in the text and in the construction of scientific discourse For this purpose we have adopted Warchal’s (2010) functional classification, which covers up to eight different functions in the use of conditionals in scientific writing, ranging from conditionals which frame the logical argumentation of the discourse (i.e content conditionals) to the different types of hedging conditionals used to speculate or to downtone the certainty of claims (i.e epistemic and speech act conditionals) as is customary in the discourse of modern science Our working hypotheses are that there should be a correlation between the use of specific conditional subordinators in the different texts and the scores of those texts for Biber’s (1988) Dimension and that there should be a diachronic evolution in the importance of the different functions of conditional sentences along the century as scientific discourse drifts apart from the old scholastic trends towards the modern scientific methods This research is carried out on a ca 200,000word corpus corresponding to the eighteenthcentury half of CEPhiT (Corpus of English Philosophy Texts), a subcorpus of the Coruña Corpus of Scientific Writing (Moskowich & Crespo 2007; Moskowich 2011) The whole of 352 CEPhiT contains forty 10,000-word text samples spanning from 1700 to 1900, at a rate of two per decade We consider that CEPhiT is a representative corpus, as its compilation has been completed by taking into account different parameters such as the sex of the authors, their place of education, as well as the different text types used for writing Philosophy in English during the time span it covers Our corpus comprises a total of twenty eighteenth-century texts, among which thirteen are treatises, six are essays and only one is a textbook, written by three female and seventeen male authors Both the genre and the sex categories are borne in mind for the analysis of our results In order to find the different conditional uses, the corpus has been searched for selected conditional particles by using the Coruña Corpus Tool (Moskowich & Parapar 2008), a search engine specifically designed to work with texts from the Coruña Corpus, which allows for multiword and metadata-based searches However, given that CEPhiT had not yet been annotated for grammatical categories while this research was carried out, the lists of occurrences obtained automatically with the tool have had to undergo manual disambiguation in order to eliminate all non-conditional uses of the particles Once our search has been completed, the different conditional uses are examined from a double perspective to comply with both our objectives First, the function of each conditional structure in the text is described in order to examine the diachronic evolution of functions After that, the different occurrences of each subordinator are counted in an attempt to find correlations with each text score in Biber’s Dimension 4, using data from previous research conducted on eighteenth-century CEPhiT (Crespo 2011) This research is a starting point for a further comparative study of conditionals across different scientific registers References Athanasiadou, Angeliki and René Dirven 1997 On conditionals again Amsterdam: John Benjamins Biber, Douglas 1988 Variation across speech and writing Cambridge: Cambridge University Press Biber, Douglas 1995 Dimensions of register variation: A cross-linguistic comparison Cambridge: Cambridge University Press Biber, Douglas 2001 Dimensions of variation among eighteenth-century speech-based and written registers In Biber, Douglas, and Susan Conrad (eds.), Variation in English: Multi-Dimensional Studies, 200-214 Essex: Pearson Education Biber, Douglas, Susan Conrad, Randi Reppen, Pat Byrd, Marie Helt, Victoria Clark, Viviana Cortes, Eniko Csomay, and Alfredo Urzua 2004 Representing Language Use in the University: Analysis of the TOEFL 2000 Spoken and Written Academic Language Corpus Princeton, NJ: Educational Testing Service Carter-Thomas, Shirley, and Elizabeth Rowley-Jolivet 2008 If-conditionals in medical discourse: From theory to disciplinary practice Journal of English for Academic Purposes 7: 191-205 Crespo, Begoña 2011 Persuasion markers and ideology in eighteenth century philosophy texts Revista de Lenguas para Fines Específicos 17: 199228 Declerck, Renaat, and Susan Reed 2001 Conditionals: A comprehensive empirical analysis Berlin: Mouton de Gruyter Ferguson, Gibson 2001 If you pop over there: A corpus-based study of conditionals in medical discourse English for Specific Purposes 20: 61-82 Hyland, Ken 1994 Hedging in academic writing and EAP textbooks English for Specific Purposes 13(3): 239-256 Hyland, Ken 1998a Hedging in scientific research articles Amsterdam: Benjamins Hyland, Ken 1998b Persuasion and context: the pragmatics of academic metadiscourse Journal of Pragmatics 30: 437-455 Moskowich, Isabel 2011 “The golden rule of divine philosophy” exemplified in the Coruña Corpus of English Scientific Writing Revista de Lenguas para Fines Específicos 17: 167-197 Moskowich, Isabel, and Bega Crespo 2007 Presenting the Coruña Corpus: A Collection of Samples for the Historical Study of English Scientific Writing In Pérez Guerra, Javier et al (eds.) Of Varying Language and Opposing Creed: New Insights into Late Modern English 341-357 Bern: Peter Lang Moskowich, Isabel, and Javier Parapar 2008 Writing science, compiling science: The Coruña Corpus of English Scientific Writing In Lorenzo Modia, María Jesús (ed.), Proceedings from the 31st AEDEAN Conference 531-544 A Coruña: Universidade da Coruña Taavitsainen, Irma 1999 Dialogues in English Medical Writing In Jucker, Andreas H.; Gerd Fritz and Franz Lebsanft (eds.), Historical Dialogue Analysis 243-268 Amsterdam/Philadelphia: John Benjamins Warchal, Krystyna 2010 Moulding interpersonal relations through conditional clauses: Consensusbuilding strategies in written academic discourse Journal of English for Academic Purposes 9: 140150 The Czech preposition v/ve and its English equivalents Renata Novotná Charles University renata.novotna@ff.cuni.cz The linguistic data used for this study was searched in the English-Czech parallel corpus built as a part of the InterCorp project The number of occurences of the preposition v in this corpus is 69114 It was impossible to study this material as a whole, therefore 15 random samples of 100 sentences each were extracted automatically, i e 2% of the material The emphasis of this study was laid mainly on the valency functions, i.e adverbal and adnominal ones According to Syntagmatics and Paradigmatics of the Czech Word (Čerm k and Holub 2005) the adverbal function was divided into two groups: a) valency – cases such as spočívat v (to be based on), změnit se v (to change into) etc., where the preposition v expresses an abstract relation between the verb and the noun b) addition/complement – such as ležet v /blátě/ (to lie in the mud), in which the place, but not the abstract relation is expressed by the preposition; in this case the preposition v/ve can be replaced by other prepositions, such as na, e.g ležet na /zemi/ (to lie on the ground) etc The adnominal function was concentrated on nouns and adjectives, i e their valency As noun valency is slightly different from verbal valency, abstract and concrete nouns were treated separately: a) valency of abstract nouns, such as život v cizině (the life abroad) b) valency of concrete nouns, e g klíč v zámku (the key in the lock) The valency of the adjectives was studied as a separate group, e g obratný v řeči (skilful in conversation) While the valency of verbs, nouns and adjectives expressed by the preposition is the inherent attribute of the lexical item with the categorial feature and therefore the verb/noun/adjective with the valency preposition can be taken as one lexical item or one lexeme; the adverbial function of the preposition is independent and forms a collocation, e.g v létě – in summer Besides the groups given above, fixed collocations and idioms were also studied, such as v souvislosti s (in connection with), držet v zajetí (to hold captive), převracet něco v mysli (to turn 353 sth over in one's mind), mít v něčem prsty (have a hand in sth), být v tahu (to be a goner) etc According to the English equivalents given in the 15 extracted samples the groups of lexical valency, collocations and idioms were arranged to the following scheme: I – lexemes: 1) verb a) valency – e.g pokračovat v předčítání – to go on with the talk b) complement – e.g ležet v posteli – to lie in the bed 2) noun a) valency of the abstract noun – e.g záliba v určitém druhu krutosti – taste for a special kind of cruelty b) valency of the concrete noun – e.g kamery ve výtahu – cameras in the elevator In some cases the English equivalent is expressed not by a preposition but by an adjective, where the essential meaning of the adjective-noun combination in English is the purpose: c) valency of the concrete noun in Czech expressed by adjective in the attribute position in English, e.g pohovka v obývacím pokoji – living room sofa 3) adjective valency – e.g ponořený v sadech – buried in orchards II – collocations: 1) adverbials – e.g v listopadu – in November 2) fixed collocations – e.g v podstatě – in general III – idioms: This is also a separate group as it can include both collocations and sentences, e.g obrátit oči v sloup – to roll someone's eyes heavenward, při tom mi tuhla krev v žilách – it made my blood run cold IV – mixed group: This group incorporates the various possible differences in combinations of one-word and multi-word units between the two languages: 1) collocation – lexeme, e.g ve skutečnosti – really, v naději – hoping 2) idiom – lexeme – e.g mít v úmyslu – intend, říkat si v duchu – to wonder The last group is composed of quasi-idioms, i.e combinations of verbs and abstract nouns having one-word equivalents in the other language: 3) quasi-idiom – lexeme – e.g být v pokušení – to be tempted, být v rozpacích – to be embarassed, být v šoku – to be shocked V – Particular differences: As a separate group, cases of particular differences were also studied: The most typical examples are the addition of such collocations as v první chvíli (in the first 354 moment) etc The poster shows the variability of English equivalents, esp of the verbal and noun valency As an example we can show the variability of adverbials, such as v Londýně – in London, ve vlaku – on the train, v Rosině příbytku – at Rosa's; v září – in September, v pondělí – on Monday, v pět hodin – at five o'clock etc References Čerm k F 1991 Podstata valence z hlediska lexikologického, In Walencja czasownika a problemy leksykografii dwujęzycznej, ed D RytelKuc, Wrocław-Warszawa-Kraków: Wydawnictwo polskiej akademii nauk, 15-40 Čerm k F 1996 Systém, funkce, forma a sémantika českých předložek Slovo a Slovesnost 57, 30-46 Čerm k F and Křen M (eds) 2004 Frekvenční slovník češtiny Praha: Nakladatelství Lidové noviny Čerm k F 2005 Abstract Noun Collocations: Their Nature in a Parallel English-Czech Corpus In Meaningful Texts The Extraction of Semantic Information from Monolingual and Multilingual Corpora, Eds G Barnbrook, P Danielsson and M Mahlberg London,-New York: Continuum, 143153 Čerm k F and Holub J 2005 Syntagmatika a paradigmatika českého slova I Valence a kolokabilita Praha:Nakladatelství Karolinum Čerm k F (ed) 2007 Frekvenční slovník mluvené češtiny Praha: Nakladatelství Karolinum Čerm k F (ed.) 2007 Slovník Karla Čapka, 2007 Ed F Čerm k Praha: Nakladatelství Lidové noviny Čermák F (ed) 2009 Slovník české frazeologie a idiomatiky I-IV Academia, Praha Čerm k F and Cvrček (eds) 2009 Slovník Bohumila Hrabala Praha: Nakladatelství Lidové noviny Čerm kov A 2009 Valence českých substantiv Praha: Nakladatelství Lidové noviny Klégr A., Mal M and Šaldov P 2012 Anglické ekvivalenty nejfrekventovanějších českých předložek Praha: Nakladatelství Karolinum Kopřivov M 2006 Valence českých adjektiv Praha: Nakladatelství Lidové noviny Novotná R 2009 The Czech preposition na and its English Equivalents In: Intercorp: Exploring a Multilingual Corpus, eds F Čerm k, P Corness and A Klégr Praha: Nakladatelství Lidové noviny, 138145 Business ethics documents of French companies from an intercultural point of view: Example of a contrastive study of the French and American versions of Lafarge’s Principles of Action Emmanuelle Pensec University of South Brittany emmanuelle.pensec@univ-ubs.fr In the context of global markets, English has become the language used for communication in large international organizations However, globalization does not mean homogenization of outlooks and of culture It even seems that we are assisting at the maintenance of discursive and cultural communities Globalization of the economy and finance has developed alongside financial scandals such as those of Enron or Wordcom In their attempt to bring morality into their practices, companies have had to consider their stakeholders when making decisions This corporate social responsibility is clearly discernible through the writing of ethical documents for the stakeholders and civil society in general The companies aim to answer to the expectations of society These documents for external communication reflect publicly the company’s values In this context, how international companies communicate a common ethics policy to all their subsidiaries? Striking a balance between cultural diversity and business strategy is a real challenge for these companies It would seem that the communication, dissemination and the content of the business ethics documents changes depending on the different cultural communities We will try to understand the reasons for the needs for such changes while the situation would have us believe that globalization makes everybody think in the same way and respond to the same thinking and same consumer standards Through a contrastive analysis of the French and American versions of Lafarge’s Principles of Action, we will see how the company deals with cultural diversity while transmitting their own values and their corporate strategy We have analysed each of the two versions of the ethical principles with corpus linguistics tools such as Lexico and Antconc so as to determine the cultural and discursive changes The results of our contrastive study imply that there is a different cultural approach of the company to its different locations The use of Corpus Linguistics tools highlights that these differences address three dimensions (Hofstede, 2001): The position of the company in relation with its economic partners Relationships of the employees with hierarchy The community spirit within the company Each time the company has dealt with one of these three dimensions, we can observe several repeated words derived from one of those as in Table Dimension The position of the company in relation with one of its economic partners Relationships of the employees with hierarchy The community spirit within the company Key term in English Deliver Key term in French Répondre Provide Offrir Employees Collaborateurs Being a customer driven organization Members of our communities Orienter notre organisation vers le client Citoyen Table So as to underline these cultural and discursive differences, we have compared these key terms of our corpus to French and American national corpora so as to try to understand the cultural variations The results of the analysis imply that there is a greater tendency to community spirit (3rd dimension) both in the company and in American society than in the French version The French version adopts a more detached stance, showing its superiority compared to its employees (2nd dimension) and other companies (1st dimension) contrary to the American version in which we observe the company as a part of a whole (3rd dimension) We hypothesize that business culture can reflect the founding values of a company Lafarge, which can be qualified as a paternalistic company, was founded in the nineteenth century, and promotes values of mutual assistance and sharing Lafarge has opted for a communicative translation (Communicative translation attempts to produce on its readers an effect as close as possible to that obtained on the readers of the original (Newmark, 1981: 39)) when choosing to culturally adjust its ethical principles This communication strategy (corporate communication) can be explained by the company’s presence in the United States and the 355 number of its employees (14.6% of its global workforce in 2010) We can consider this as a strategic option for Lafarge The choice of this micro-corpus can be justified by the cultural and discursive interests in a bilingual approach Corpus linguistics tools have helped to demonstrate the importance of the intercultural dimension in the communication strategies of companies’ Corpus mining tools in the PLEC project Piotr Pęzik University of Łódź pezik@uni.lodz.pl References d'Iribarne, P 2009 L'épreuve des différences, L'expérience d'une entreprise mondiale Paris: Seuil Hofstede, G 2001 Culture's Consequences: Comparing Values, Behaviors, Institutions and Organizations Accross Nations London: Thousand Oaks Moirand, S., & Tréguer-Felten, G 2010 “Des mots de la langue aux discours spécialisés, des acteurs sociaux la part culturelle du langage: raisons et conséquences de ces déplacements” Asp 51-52 Available online at http://asp.revues.org/465 Newmark, P (1981) Approaches to translation New York: Oxford Schlegelmilch, C C (1990, 4th Qtr.) “Do Corporate Codes of Ethics Reflect National Character? Evidence From Europe and the United States” Journal of International Business Studies, pp 519539 Tréguer-Felten, G (2009) Le leurre de l'anglais lingua Franca; Une étude comparative de documents professionnels produits en anglais par des locuteurs chinois, franỗais et nord-amộricains Thesis, University of Paris III 356 PLEC The PELCRA Learner Corpus (PLEC) is a research project aimed at investigating the lexicogrammatical, phraseological and phonetic competence of Polish learners of English (Pęzik 2012) The project was launched in 2010 and it will run until late 2013 The corpus compiled in the project contains samples of learner English, such as essays, in-class and in-exam assignments, letters, MA theses and many other types written compositions authored by Poles using English as a foreign language (2.8 million words in total) It also features a time-aligned, error-annotated spoken subcorpus of learner English containing 200 000 word segments (which roughly corresponds to 25 hours of continuous recordings) These recordings are mostly informal interviews conducted with learners of English representing a variety of proficiency levels and social backgrounds Annotation There are several tiers of corpus annotation used in the PLEC corpus Texts and transcriptions of utterances are linked with author and speaker metadata, such as age, gender, education, proficiency level and language learning background Linguistic annotation includes not only part-of-speech annotation, but also automatic syntactic dependency metadata There are also two types of error annotation in PLEC Firstly, a general, learner error taxonomy was adopted for the manual annotation of errors in a selection of the corpus Secondly, the entire spoken component of the corpus has been annotated for word mispronunciations and used to compile an index of words commonly mispronounced by Polish learners of English (Zając & Pęzik 2012) Interestingly, the PLEC corpus also contains an automatic phraseological annotation tier Using the HASK collocation dictionary (pelcra.pl/hask_en) (Pęzik 2013) a BNC-based collocation tagger was developed to identify and annotate instances of selected native-like phraseological units in the learner data This type of annotation made it possible to estimate the socalled phraseological index of learner English samples, which is a measure of native-like idiomaticity in non-native texts Tools The different types of annotation available in the PLEC corpus can be mined using the online corpus mining tools developed within the project The linguistic tier of the corpus annotation can be explored through a dedicated search engine supporting complex part-of-speech queries and a syntactic dependency browser The search engine was implemented using a customized version of the Apache Lucene library and it can be used in other corpus projects as it scales well with the size of the collection up to billions of segments The online interface for the corpus provides multimodal access to the written and spoken data components; users can stream audio snippets for utterances matching their queries Learner and text profiles such as proficiency levels, domains, genres and register can be used as search criteria Error and phraseological annotation tiers can be explored through dedicated online tools Applying corpus techniques to climate change blogs Availablity The online corpus tools described in this abstract are available at http://pelcra/PLEC The spoken time-aligned component of PLEC has been released under a Creative-Commons license and it can be used in further research on spoken English learning Acknowledgments This research was funded in 2010-2013 by a grant from the Polish Ministry of Science and Higher Education References Pęzik, Piotr 2012 “Towards the PELCRA Learner English Corpus.” In Corpus Data Across Languages and Disciplines, ed Piotr Pęzik, 28:33–42 Łódź Studies in Language Peter Lang Pęzik, Piotr 2013 Forthcoming Graph-based analysis of native and learner phraseology in HASK Collocation Dictionaries pelcra.pl/hask_en Zając, Magda, and Piotr Pęzik 2012 “Annotating pronunciation errors in the PLEC spoken learner corpus.” In Proceedings ot TALC 10 Conference Warsaw Andrew Salway, Knut Hofland Uni Research, Bergen andrew.salway,knut.hofland@uni.no Samia Touileb University of Bergen samia.touileb@gmail.com Introduction The emergence of social media has created new opportunities for social scientists to investigate how organisations, individuals and media contribute to shaping public understanding of, and opinions about, important issues The work described here is about the application of corpus techniques to support investigations of complex and large-scale discourses in online social networks, e.g blogs about climate change These techniques are at the interface of corpus linguistics, text mining and information extraction Background In the NTAP project1 we are developing tools and methods for analysing the distribution, flow and development of knowledge and opinions across online social networks One innovation is that we treat the content of texts (blog posts) as key statements, rather than keywords; this contrasts with current tools for social media analysis that use word clouds and graphs of keywords over time Our system to identify key statements, e.g “climate change causes rising sea levels”, depends on the identification of information structures, e.g “X causes Y” Corpus techniques are crucial for us to ensure that the structures characterise how information about core concepts is typically expressed in the given domain and text type As well as supporting statement extraction, the distribution of information structures within corpora will be analysed in investigations of discourse style In a blog corpus, each blog post can have metadata including ‘time-date’, ‘author’ and ‘in/out hyperlinks’; there are also blogroll links between blogs Our second innovation is to fuse the analysis of text features and network features, NTAP: Networks of Texts and People, funded by the Norwegian Research Council, 2012-15, http://www.ntap.no 357 through interactive visualizations Given instances of a key statement in multiple posts, along with instances of supporting and opposing statements, we can explore the occurrence of these statements over the network of blogs, and over time Thus we aim to give social scientists new affordances for understanding what influences the diffusion of key statements; for more on this, see (Salway et al 2012) Progress to date We are currently crawling the blogosphere and have so far downloaded the complete content of about 3,000 English-language blogs that include posts mentioning “climate change” or a related term The crawl started from about 20 handpicked blogs: the terms to determine topic relevance were extracted from some of these, using a keyword list and word clusters Later, once we have a set of key statements relating to climate change, we plan to use a blog search engine to locate further material Effort is now focussed on transforming raw html into a database of blog posts with associated metadata (author, date, hyperlinks, etc.) In parallel, work has begun on the identification of information structures that characterise how information about climate change is expressed in blogs Using, for now, the Yahoo BOSS blog search engine, we are exploring both “top-down” and “data-driven” methods Working in a top-down manner, we assumed that statements about the causes and effects of climate change would be important A linguistic account of how causality can be expressed in English was used to generate a set of 14,000 queries: “causes climate change”, “is the result of climate change”, etc We assumed that the number of hits per query is a crude indication of the information structures that are used commonly Further, taking the snippets returned by the search engine, we collated the text around the query phrases, i.e where we expect to find the stated causes and effects Keyword and n-gram analysis of this text seems promising for elucidating further domain concepts, like “rising sea levels”, “greenhouse gas emissions, that represent typically stated effects and causes of climate change With a top-down method there is a danger of never knowing what information structures are missed Thus, we wish to work also in a datadriven manner and induce information structures directly from texts As a step in this direction, we obtained 3352 non-overlapping n-grams (n