Volumen 14 · publicación anual · 2019 http://polipapers.upv.es/index.php/rdlyla Volumen 14 publicación anual · 2019 http://polipapers.upv.es/index.php/rdlyla EDITORA Mª Luisa Carrió Pastor, Universitat Politècnica de València, España

COMITÉ EDITORIAL María José Labrador Piquer, Universitat Politècnica de València, España Mercedes López Santiago, Universitat Politècnica de València, España Josefa Contreras, Universitat Politècnica de València, España Debra Westall, Universitat Politècnica de València, España Penny MacDonald, Universitat Politècnica de València, España Hanna Skorczynska, Universidad Politécnica de Valencia, España Ricardo Mairal, Universidad Nacional de Educación a Distancia, España Jose Antonio Diaz Rojo, Consejo Superior de Investigaciones Científicas, España Maria Josep Cuenca, Universitat de València, España Miguel Casas Gómez, Universidad de Cádiz, España Teresa Cabré, Universitat Pompeu Fabra, España Francisco Ruiz de Mendoza, Universidad de La Rioja, España Francisco Yus Ramos, Universidad de Alicante, España Carmen Caldas-Coulthard, University of Birmingham, Reino Unido

COMITÉ CIENTÍFICO Francisco Alonso Almeida, Universidad de las Palmas de Gran Canaria, España Guadalupe Aguado de Cea, Universidad Politécnica de Madrid, España Josep Ballester, Universitat de València, España David Banks, Université de Bretagne Occidentale, Francia Pascual Cantos Gómez, Universidad de Murcia, España Berta Raposo Fernández, Universitat de València, España María Isabel González Cruz, Universidad de Las Palmas de Gran Canaria, España Marianna Chodorowska Pilch, University of Southern California, Estados Unidos María Tadea Díaz Hormigo, Universidad de Cádiz, España María Luisa Donaire Fernández, Universidad de Oviedo, España Juan Manuel Hernández Campoy, Universidad de Murcia, España Brigitte Lépinette, Universitat de València, España Begoña Clavel, Universitat de València, España María Belen Díez Bedmar, Universidad de Jaén, España Mercè Lorente Casafont, Universitat Pompeu Fabra, España Francisco Vizcaíno, Universidad Las Palmas de Gran Canaria, España María José Gómez Calderón, Universidad de Sevilla, España Aneider Iza Erviti, Universidad de La Rioja, España Barry Pennock, Universitat de València, España Mercedes Sanz Gil, Universitat Jaume I, España Noelia Ibarra, Universitat de València, España Ana Isabel Rodríguez-Piñero Alcalá, Universidad de Cádiz, España Sergio Maruenda Bataller, Universitat de València, España Stephan Patin, Université Paris Diderot Paris 7, Francia María Querol, Universidad Católica de Valencia, España Elena Quintana-Toledo, Universidad Las Palmas de Gran Canaria, España Chelo Vargas, Universidad de Alicante, España Isabel Moskowich-Spiegel, Universidad de La Coruña, España Margarita Mele Marrero, Universidad de La Laguna, España Begoña Bellés Fortuño, Universitat Jaume I, España Herbert Holzinger, Universitat de València, España Pascuala Morote, Universitat de València, España Susana Pastor Cesteros, Universidad de Alicante, España Carmen Sancho Guinda, Universidad Politécnica de Madrid, España Raquel Criado, Universidad de Murcia, España Jan Goes, Université d’Artois, Francia, Francia Ricardo Morant, Universitat de València, España Isabel Verdaguer Clavera, Universitat de Barcelona, España Chris Taylor, University of Trieste, Italia Ana Llinares García, Universidad Autónoma de Madrid, España Javier Martin Arista, Universidad de La Rioja, España REVISTA INDEXADA EN: ERIH European Reference Index for the Humanities: https://www2.esf.org/asp/ERIH/foreword/search.asp DIALNET: http://dialnet.unirioja.es/appl/ppal.jsp CINDOC: http://bddoc.csic.es:8080/index.jsp Base de conocimientos de SFX (ExLibris) METALIB: http://metalib.upv.es:3210/sfx_local/a-z?perform=textSearch DOAJ: http://www.doaj.org MLA: http://www.mla.org PORTAL DEL HISPANISMO: Ministerio de Cultura. Instituto Cervantes: http://www.hispanismo.es/revista.asp?DOCN=2604 LINGUISTICS ABSTRACTS ONLINE: http://www.linguisticsabstracts.com ELEKTRONISCHE ZEITSCHERIEFTENBIBLIOTHEZ: http://www.zblx1.uni-regensburg.de/ezeit/ LATINDEX: http://www.latindex.unam.mx JOURNALSEEK: http://journalseek.net DULCINEA: http://www.accesoabierto.net/dulcinea/consulta.php?directorio=dulcinea&campo=ID&texto=1647 EBSCO: http://www.ebsco.com/ RESH: http://epuc.cchs.csic.es/resh/ e-REVISTAS: http://www.erevistas.csic.es/ficha_revista.php?oai_iden=oai_revista935&anyo=2013

Más información: http://polipapers.upv.es/index.php/rdlyla Departamento de Lingüística Aplicada Universidad Politécnica de Valencia Camino de Vera, s/n 46022 Valencia

DEPÓSITO LEGAL: V-1613-2006 ISSN: 1886-2438 © Departamento de Lingüística Aplicada

Diseño y maquetación Enrique Mateo, Triskelion Diseño Editorial

EDITA Editorial UPV ÍNDICE

1 LOGISTIC CALIBRATED ITEMS (LCI) METHOD: DOES IT SOLVE SUBJECTIVITY IN TRANSLATION EVALUATION AND ASSESSMENT? Alireza Akbari

19 ¿COMPLIANCE O CUMPLIMIENTO NORMATIVO? ESTUDIO DE CORPUS CON FINES PROFESIONALES Y DIDÁCTICOS EN LA PRENSA ESPAÑOLA Elena Alcalde Peñalver Alexandra Santamaría Urbieta

29 ACTITUDES Y COMPETENCIAS DOCENTES EN PROFESORES DE ESPAÑOL DE LOS INSTITUTOS FEDERAIS (BRASIL) EN RELACIÓN CON LAS VARIEDADES DE LA LENGUA Bruno Rafael Costa Venâncio da Silva María Antonieta Andión Herrero

45 THE IMPACT OF APPLYING THE FONF PRACTICE MODEL ON DEVELOPING L2 LISTENING AND SPEAKING WITH A FOCUS ON INTENTIONAL AND INCIDENTAL VOCABULARY ACQUISITION IN CALL CONTEXT Akbar Bahari

59 THE SYNTACTIC PARSING OF ASD-STE100 ADVERBIALS IN ARTEMIS Francisco J. Cortés-Rodríguez Carolina Rodríguez-Juárez

81 LAS LENGUAS OFICIALES DEL ESTADO ESPAÑOL EN LOS TEXTOS LEGALES, ¿FOMENTO O RECONOCIMIENTO DEL PLURILINGÜISMO? Luis Escoriza Morera

91 ZOMBIES LOST IN TRANSLATION. THE TRANSLATION FROM ENGLISH TO SPANISH OF (DE)HUMANIZING PRONOUNS Linda Flores Ohlson

105 CORPUS-BASED STUDY OF L3 ACQUISITION ON SPANISH PAST TENSE: EVIDENCE FROM LEARNERS’ ORAL PRODUCTION Hui-Chuan Lu An Chung Cheng Shen Yun Hung

117 LA INFLUENCIA DE LA MEMORIA OPERATIVA Y DE LA INTELIGENCIA EMOCIONAL EN LA EXPRESIÓN DE LA EMOCIÓN EN ESPAÑOL COMO SEGUNDA LENGUA Irini Mavrou Fernando Bustos-López

131 THE LEXICAL REPRESENTATION OF ENGLISH VERBS OF ACTION. COMPLEX PREDICATES AND STRUCTURES Ana Elvira Ojanguren López

141 METAPHORS OF VICTORY AND DEFEAT IN SPORTS HEADLINES IN ENGLISH AND SPANISH Sara Quintero Ramírez

153 REMEDYING FOREIGN LANGUAGE ANXIETY THROUGH CLIL? A MIXED-METHODS STUDY WITH PUPILS, TEACHERS AND PARENTS Mathea Simons Claudio Vanhees Tom Smits Karen Van De Putte

RLyLA Vol. 12 (2017) | vii Revista de Lingüística y Lenguas Aplicadas Vol. 14 año 2019, 1-18 EISSN 1886-6298 https://doi.org/10.4995/rlyla.2019.11068

LOGISTIC CALIBRATED ITEMS (LCI) METHOD: DOES IT SOLVE SUBJECTIVITY IN TRANSLATION EVALUATION AND ASSESSMENT?

Alireza Akbari University of Isfahan, Iran

Abstract: This research paper introduces a model of translation evaluation called Logistic Calibrated Items method. The aim of LCI method is to maximize a translators’ performance and to identify top competent translators through detecting all parsing items within a source text. Parsing items are extracted by the application of Brat software. The purpose of LCI was to identify parsing items having optimal item difficulty and item discrimination values. LCI method involves six stages: (1) holistic scoring; (2) the application of Brat software to extract all parsing items; (3) the calculation of item difficulty; (4) the calculation of item discrimination; (5) the identification of items with optimal item difficulty and item discrimination values; and (6) the recalculation of scores. 125 translation students and 4 professional translation evaluators took part in this research. The final results showed that LCI method was more consistent when compared to holistic method. Limitations and implications were also discussed.

Keywords: translation evaluation product, Logistic Calibrated Items method, item difficulty, item discrimination, recalculation of scores.

1. INTRODUCTION

Traditionally, translation evaluation has been based on detecting and analyzing errors, a methodology which requires a remarkable investment in human resources when assessing a huge volume of translation drafts both in academic and professional settings (Schmitt, 2005). Thus far, research in the field of translation evaluation and assessment has predominately associated with descriptive and theoretical aspects and has concentrated on the following themes namely; criteria for good and poor translations (Newmark, 1991), the nature of translation and language errors (Gouadec, 1989), pragmatic and linguistic translation quality assessment (Nord, 2005), analyzing translation quality assessment based on text linguistic analysis (Larose, 1989), determining different textual levels and associating the significance of mistakes with these levels (Dancette, 1989), and various levels of translation competence (Stansfieldet al., 1992).

In the context of the above, the field of Translation Studies (hereafter TS) has vindicated the need for more experimental evidence for the assessment/evaluation of different translation tests (multiple-choice translation tests, expository translation exams, etc.) (Akbari and Segers, 2017b). Although educational and professional organizations have executed ‘the certification of translation skills’ (Eyckmanset al., 2013) according to different test administrations, the validity (e.g. concurrent validity, statistical conclusion validity) and reliability of them remain questionable. Therefore, translation evaluation and assessment is by some means depended on the codes of practice rather than empirical explorations (ibid.). The field of translation evaluation covers the following themes such as translation process, translation product (the target text), translation service (e.g. client, complaints, and invoicing), and translation competence accordingly. Nonetheless, the mentioned themes cannot be evaluated/ assessed/measured in the same way and necessitates different means of evaluation approaches.

According to Anckaert et al. (2008), there is no evaluation method which links the ‘performance indicator’ to the fundamental competence in a psychometric fashion. Lack of test development to assess translation competence

To cite this article: Akbari, A. (2019). "Logistic Calibrated Items (LCI) Method: does it solve subjectivity in translation evaluation and assessment?" Revista de Lingüística y Lenguas Aplicadas, 14, 1-18. https://doi.org/10.4995/rlyla.2019.11068 Correspondence author: [email protected]

Received: 2018-11-29 Accepted: 2019-03-26 | 1 Alireza Akbari Logistic Calibrated Items (LCI) Method: does it solve subjectivity in translation evaluation and assessment?

can be explained by two factors: (1) translation tests are not sufficiently valid and authentic to measure language ability (competence) for which this resulted in a definite overlooking of popularity during the era of Communicative Approach (CA) (Widdowson, 1978). This may be caused by the fact that translation tests are not laid open to ‘the same psychometric scrutiny as other language testing formats’ (e.g. cloze-test) (Eyckmans et al., 2013); (2) There exists an epistemological aperture between the human science (e.g. translation and interpreting studies (T/I studies), language and literature, language and linguistics, etc.) and hard science (biology, chemistry, etc.). The surmise that it is impossible to objectify the quality of translation while covering its very essence may be very resolute among language instructors and translation scholars/trainers/instructors whose ‘corporate culture exhibits a marked reticence towards the use of statistics’ (Eyckmans et al., 2012, 2013). In this direction, testing translation and interpreting approaches/skills have been more or less carried out by practitioners rather than of translation scholars and researchers. Thanks to different psychometric approaches, a fair amount of research in the field of reliability and validity of language tests has been recognized. Nonetheless, the field of T/I studies has been fallen back and requires more scrutinies.

Generally, translation evaluation focuses on issues namely; (i) translation process, (ii) translation, (iii) translator, (iv) linguistic, textual, and paralinguistic competences (Kockaert and Segers, 2017), and situatedness (Muñoz Martín, 2010). According to educational and professional settings, translation evaluation can be performed as maintained by a criterion-referenced approach (Schmitt, 2005) and can be gauged in respect of some ‘assessment grids’ to make translation evaluation more valid and reliable. These contexts can never sufficiently lessen the degree of subjectivity in translation evaluation. What’s more, the scoring system which is susceptible to be influenced by contrast effect menaces the reliability and validity of translation tests. Contrast effect refers to ‘a magnification or diminishment of perception as a result of previous exposure to something of lesser or greater quality, but of the same base characteristics’ (González, 2018).

In the light of the above, the present research is an attempt to introduce an objective translation evaluation method called Logistic Calibration Items (LCI hereafter). LCI method is specified by the total number of parsing items within a source text in accordance with translation relevance and norm-criterion referenced assessment methods. The term ‘calibration’ was used to characterize docimologically justified parsing items. Docimology alludes to the theory of the art of testing and docimologically justified items refer to such items which have optimal corrected item-total correlations (rit value hereafter) and appropriate item difficulty (p-value) based on 1-Parameter Logistic Model (1-PL) of item response theory (IRT) (the main target of LCI method). As is the case with the Preselected Items Evaluation (PIE) (Kockaert and Segers, 2017), correct and incorrect solutions are categorized for each parsing item in the source text in LCI method. LCI method consists of the following stages: (i) holistic scoring system based on evaluators’ impression/intuition/anticipation; (ii) the use of Brat Visualization software Stanford CoreNLP parser to detect each parsing item within a source text; (iii) the calculation of item difficulty by 1-Parameter Logistic Model; (iv) the calculation of rit value; (v) the extraction of docimologically justifed and calibrated items which have optimal rit (rit ≥ 0.30) and item difficulty (positive reliability coefficient and p-value < 0.05) values; and consequently (vi) the re-calculation of scores.

2. STATE OF THE ART 2.1. A Review of Translation Evaluation Translation evaluation is predominately represented by a criterion-referenced assessment. In terms of educational and professional settings, assessment and evaluative grids are used in an effort to make translation evaluation more reliable and valid. Conventionally, evaluative grids encompass ‘a near exhaustive taxonomy of different kinds of mistakes or bonuses’ (e.g. word choice, stylistic conventions, text coherence and cohesion) (Eyckmans et al., 2013). Although the use of the evaluative grids is caused by an evaluator’s desire to take into account different dimensions of translation competence, one has to affirm that these dimensions are unable to lessen the degree of subjectivity in translation evaluation and assessment (Anckaert et al., 2008). Apart from the subjective features of translation evaluation sub-competences, other elements may menace the reliability and validity of translation administration tests. Let us start with an evaluator’s consistency throughout the task of translation scoring during a specific time interval. Not only will the system of scoring be susceptible to a contrast effect, it is also compulsory to put forward a “sound testing practice” which discriminates good testing items from the bad ones. Moreover, all obtained scores must be docimologically acceptable and the scoring system must differentiate the standard quality of translations. For that reason, scholars and researchers from the field of translation quality assessment and evaluation (Conde Ruano, 2005; Kockaert and Segers, 2017; Akbari and Segers, 2017a) are now accrediting issues such as the degree of inter-intra-rater reliability (rater variability), construct, concurrent, and statistical conclusion validity as well as ecological validity. The methodology of educational and professional measurements along with the standpoint of theories of language testing are being transferred to the field of translation evaluation product and translation quality assessment in order to reach reliable and valid methods/approaches to measure translation competence. With that in mind, the dominant purpose of the present research paper is to emancipate translation evaluation from extraneous and irrelevant variables which have an effect on the outcome of assessment (Eyckmans et al., 2013).

RLyLA Vol. 14 (2019), 1-18 | 2 Alireza Akbari Logistic Calibrated Items (LCI) Method: does it solve subjectivity in translation evaluation and assessment?

2.2. Different Translation Evaluation Models 2.2.1. Holistic Method of Translation Evaluation Holistic method of evaluation is supposed to be an objective and accurate method of translation evaluation (Kussmaul, 1995). This method of evaluation has a short range of objectivity and resiliency due to an evaluator/ corrector’s anticipation/impression and the type of translation errors that students make throughout their translations. Actually, this method as ‘attending to the text as a whole’ (Hamp-Lyons, 1991:246) has been employed very differently by translation instructors and evaluators. Holistic method of translation evaluates the overall quality of a translation in terms of a translator’s impression (Mariana et al., 2015). This method is considered fast yet fully subjective since it is depended on the taste of an evaluator while scoring a translation. As Kockaert and Segers (2017:149) contend, ‘the value judgments of different holistic evaluators on the same translation can vary greatly’. For example, one evaluator takes into account one translation as excellent and acceptable; while another grader considers the same translation as fair or unacceptable (Eyckmans et al., 2012). Garant (2001) has put forward that ‘point-based error focused grading’ as a type of paradigm shift has been superseded by holistic method of assessment at the University of Helsinki. According to Kockaert and Segers (2017), translation is suitably evaluated based on ‘discourse level holistic evaluation’ rather than ‘grammar-like and analytical’ evaluations. This method also concentrates mainly on a ‘context-sensitive evaluation’ (Akbari and Segers, 2017b) and is surmised to shift from ‘exclusive attention to grammatical errors in translation tests’ (Kockaert and Segers, 2017:149).

Although holistic method of evaluation supposes to determine higher inter-coder (inter-rater) reliability (Barkaoui, 2011) and lead to ‘produce reliable and consistent assessment’ (Cumming et al., 2002:67), this method is ‘not necessarily an indicator of the raters actually applying the scale in a consistent way’ (Harsch and Martin, 2013). Higher degree of inter-rater reliability will conceal differences among evaluators ‘for the criterion scores’ (ibid.) and menace the degree of validity. According to Weigle (2002:114), ‘holistic scoring has also come under criticism in recent years for its focus on achieving high inter-rater reliability at the expense of validity’. In line with Weigle, Barkaoui (2010:516) has pointed out that evaluators can move away from ‘the criteria originally designed to define what is being assessed’. Therefore, ‘this can reduce score consistency across and within raters and, ultimately, change the meaning of the scores’ (ibid.).

Even though this method is acceptable, evaluators/graders are not always in a position of agreement when scoring translations. This fully shows that this method does not have sufficient objectivity. According to Bahameed (2016:144), holistic method of evaluation depends to a certain degree on ‘the corrector’s personal anticipation and appreciation’. By the same token, there do not exist any specific criteria available while scoring a translation based on holistic method. Bahameed (2016) continues that this method can never simply specify the top students as their scores ‘may reach one-third out of the whole translation class’. This is due to the fact that students are not responsible for minor mistakes (e.g. lexical, grammatical, etc.). These minor mistakes can never be disregarded by a grader since they initiate a matter in the quality of holistic evaluation that is too arduous to measure. The leniency of holistic evaluation can negatively reverberate on the quality of a translation and eventually on the teaching process (Akbari and Gholamzadeh Bazarbash, 2017). In the context of the above, holistic method of translation evaluation is not supportable.

2.2.2. Analytic Method of Translation Evaluation Analytic method of translation evaluation (also it is called evaluation grids method) is associated with error analysis and is maintained to be more valid and reliable in comparison with holistic method (Waddington, 2001:136). In analytic method of evaluation, an evaluator maintains a grid which includes a number of error levels and types. Therefore, the number of error types and error levels can be increased; nonetheless, this must be done with prudence. The reason is that an augment in error levels and error types can minimize the feasibility of this method. Analytic method evaluates the overall quality of a translation by investigating the text segments (e.g. individual words, clauses, paragraphs, etc.) in accordance with a certain number of criteria such as omission, addition, misinterpretation, and so forth. Eyckmans et al. (2013) have pointed out that translation errors should be identified with regard to ‘the evaluation grid criteria’. Furthermore, an evaluator should firstly ascertain the types of errors such as errors associated with translation or language and s/he must maintain the pertinent information in the margin with regard to the nature of errors accordingly.

To put it in a nutshell, a translator has ‘a better understanding of what is right and what is wrong in translation’ (Kockaert and Segers, 2017:150) in analytic method. Analytic method has a disadvantage that an evaluator focusing on the small text segment of the source text does not definitely have an exhaustive prospect of the target text. This method is also subjective and necessitates more time compared to holistic method.

2.2.3. Preselected Items Evaluation (PIE) Method The Preselected Items Evaluation (PIE) method is a system of scoring which was devised by Hendrik. J. Kockaert and Winibert Segers. This method is suitable for summative assessment. The purpose of a summative assessment is ‘to evaluate student learning at the end of an instructional unit by comparing it against some standards or

RLyLA Vol. 14 (2019), 1-18 | 3 Alireza Akbari Logistic Calibrated Items (LCI) Method: does it solve subjectivity in translation evaluation and assessment?

benchmarks’ (Eberly Center, 2016). In terms of functionality and time management, PIE method only investigates the number of preselected items within a source text. PIE method is considered to be a calibrated method since it checks the accuracy of ‘the measuring instrument’. Additionally, this method is called a dichotomous method because it vets the discrimination between correct and incorrect solutions (Kockaert and Segers, 2017:150). This calibrated and dichotomous method which does not distinguish between levels of errors, is suitable for all language combinations and comprises the following stages: (i) the preselection of a limited number of items in a source text based on evaluators’ expertise; (ii) the identification of correct and incorrect solutions to the preselected items; (iii) the calculation of p-docimology [cf. p-value] or item difficulty (the proportion of examinees provides a correct solution for an item); (iv) the calculation of d-index or item discrimination (candidates’ differentiations on the basis of the items being measured) based on the method of extreme group; and (v) the recalculation of scores in terms of optimal p-docimology and d-index values (PIE run scores). According to Lei and Wu (2007), the calculation of p-value and d-index associates with ‘the minimum number of items needed for a desired level of score reliability or measurement accuracy’. The ideal range of p-value in PIE method ‘should be higher than 0.20 and lower than 0.90’ (Kockaert and Segers, 2017). With this in mind, the larger the sample size of the participants provides a correct solution for an item, the easier the selected item will be. To calculate item discrimination (d-index), PIE method employs the method of extreme group (‘analysis of continuous variables sometimes proceeds by selecting individuals on the basis of extreme scores of a sample distribution and submitting only those extreme scores for further analysis’) (Preacher et al., 2005) through the application of 27% rule. That is to say, this method identifies the top 27% of candidates and the bottom 27% of candidates of the entire score ranking. The application of 27% rule will maximize differences in a normal distribution (Wiersma and Jurs, 1990). Items are preselected in terms of translation brief relevance and test specific criteria. Having administered a test, an evaluator calculates the difficulty of the selected items according to p-value and d-index. Items which are not responding to docimological standards (poor p-value and d-index) will be removed from the test.

Several studies have been conducted about PIE method to fine-tune it and proves its capacity toobjectify the evaluation of translation products (Kockaert and Segers, 2014, Kockaert and Segers, 2017, Akbari and Segers, 2017a,c,d). These articles mainly discuss PIE method and its application for English, Dutch, French, and Persian languages and analyze its reliability (no sign of validity) (Akbari and Segers, 2017b) compared to holistic and analytic evaluations. Nonetheless, more research must be carried out to identify and solve specific scientific aspects. For instance, the validity of PIE method has not so far been recorded and has been under critical questions. No vindications are proposed of why items within a source text are preselected as the most difficult or simple items. On what basis does an evaluator preselect an item within a text? What are the appropriate number of preselected items in a source text? Once a translation is evaluated, what happens to other mistakes in a text? And consequently, what typology of items must be preselected within a source text (linguistic items, extralinguistic items, etc.) when applying PIE method?

3. METHODOLOGY 3.1. Description of the Participants and Study Conditions This research paper took place in 2018. 125 translation students from the Bachelor of Arts in Translation Studies at the University of Sheikhbahaei and the University of Isfahan took part in this research. The participants were all Persian native speakers (L1) with average age of 22 years. They all passed the obligatory courses related to the literary, political, journalistic, translation of legal documents, and medical translation through which they were subjected to different translational texts. The participants were requested to translate a short text from English (L2) to Persian (L1). Even though the participants varied in their level of English language proficiency, the standard preconception was that it was by and large of good standard, since the registration in their study programs necessitated proof of passing compulsory credits such as medical, legal, economic, and political translation courses. The participants were asked to translate a short political text into Persian (L1) (see Appendix 1). They were all acquainted with political styles, terminologies, and structures as they passed the necessary courses related to the political translation. The length, type, and the difficulty of the selected political source text were considered indicative for the materials taught in the translation courses at the University of Isfahan and Sheikhbahaei University. Eventually, four representative translations made by four official translation agencies were provided for the graders; therefore, they would have access to different yet correct equivalents once evaluating and scoring translations. These official translation agencies had long-established experience (approximately 8 years) in assessing and translating political texts and documents.

3.2. Procedures The provided translations were handed to 4 translation evaluators and were asked to commonly score the drafts. The evaluators were selected in terms of their longstanding experience (nearly 10 years) in translation quality assessment and evaluation. They were selected from the University of Isfahan and Sheikhbahei University. At first sight, there were asked to score the translations holistically based on the Waddington’s framework (2001) (see Appendix 2). The reason to ask the evaluators to score holistically was just to show the difference between holistic

RLyLA Vol. 14 (2019), 1-18 | 4 Alireza Akbari Logistic Calibrated Items (LCI) Method: does it solve subjectivity in translation evaluation and assessment?

scoring and LCI scoring at the end of the research. The scores made by holistic and LCI scoring systems were all calculated up to 20. Having scored translations holistically, the evaluators were asked to identify parsing items with optimal item difficulty by the application of 1-Pramater Logistic Model of IRT (positive reliability coefficient and p-value < 0.05) (Stata, 2016) and then to detect discriminating items based on optimal rit values (0.30 and above for good and very good items) (SPSS, 2018). Moreover, the evaluators were apprised about the quasi-experimental design of the present research.

4. THE ADMINISTRATION OF LCI METHOD: FROM HOLISTIC CALCULATION TO SCORE RECALCULATION

- Stages of LCI Method (I) Holistic Scoring of Translations As mentioned, 125 translation students from two well-known universities were asked to translate a short political text (L2) into simple Persian (L1). When the translation task was done, four professional translation evaluators were assigned to commonly score the translation drafts based on the Waddington’s (2001) framework. The participants’ scores were as follows:

Table 1. Holistic Scoring (Docimologically Unjustified Scores).

Par Score Par Score Par Score Par Score Par Score (Holistic)/20 (Holistic)/20 (Holistic)/20 (Holistic)/20 (Holistic)/20 1 13 26 10 51 13 76 15 101 16 2 15 27 12 52 15 77 15 102 15 3 15 28 13 53 14 78 17 103 17 4 16 29 13 54 10 79 15 104 14 5 14 30 17 55 12 80 16 105 18 6 10 31 10 56 15 81 13 106 19 7 11 32 20 57 16 82 19 107 14 8 13 33 20 58 17 83 12 108 15 9 12 34 18 59 17 84 15 109 15 10 17 35 18 60 17 85 16 110 15 11 15 36 9 61 12 86 17 111 18 12 20 37 11 62 12 87 15 112 15 13 18 38 13 63 10 88 14 113 12 14 14 39 14 64 15 89 13 114 10 15 12 40 14 65 15 90 15 115 16 16 11 41 15 66 15 91 14 116 13 17 12 42 12 67 18 92 15 117 14 18 12 43 13 68 18 93 14 118 19 19 11 44 12 69 20 94 17 119 13 20 15 45 11 70 20 95 12 120 15 21 16 46 15 71 18 96 13 121 20 22 16 47 15 72 17 97 14 122 20 23 17 48 15 73 18 98 15 123 12 24 15 49 15 74 18 99 14 124 16 25 10 50 13 75 18 100 14 125 15

According to Table 1, participants [6], [25], [26], [31], [54], [63], and [114] obtained the lowest scores compared to the rest of translation participants. According to the evaluators’ comments and remarks, those participants largely applied literal translation (word-for-word translation), which end in ambiguous target text meanings for some parts of the source text. In this direction, those participants did not adapt optimal approaches when translating the source text. Moreover, they in question commit major semantic errors which caused their translations to a great extent ambiguous, unclear, and inaccurate. Those participants on many occasions lost the contextual function of the source text and resorted to word-for-word translation.

RLyLA Vol. 14 (2019), 1-18 | 5 Alireza Akbari Logistic Calibrated Items (LCI) Method: does it solve subjectivity in translation evaluation and assessment?

(II) The Application of Brat Visualization Stanford CoreNLP Parser Brat CoreNLP is a web-based tool which annotates and makes texts into parses. This tool was designed for structured annotation based on different NLP (Natural Language Processing) tasks. The aim of the Brat tool is to ‘support manual curation efforts and increase annotator productivity using NLP technique’ (Stenetorpet al., 2012).

Modern annotation and parsing tools are technically-directed and many of them present ‘little support to users beyond the minimum required functionality’ (ibid.). In this respect, tools with user-friendly interfaces can support human decisions and help to provide the quality of annotations or parsers as well as making them more accessible to non-technical users. Furthermore, these tools ameliorate parsing productivity and functionality. As a parsing tool, Brat software is based on previously STAV text annotation visualizer. STAV text annotation visualizer was devised to assist users to acquire an exhaustive comprehension of convoluted annotations including a great number of various ‘sematic types, dense, partially overlapping text annotations, and non-projective sets of connections between annotations’ (Stenetorp et al., 2011). This tool is thoroughly configurable and can be established to authenticate most text parsing tasks. Furthermore, the Brat software has been used to generate well-over 50,000 parses in thousands of documents. The purpose of LCI method for using the Brat Visualization CoreNLP software was to determine every parse or annotation (norm-referenced evaluation) within the source text and consequently identify such parses which were docimologically justified (criterion-referenced evaluation) (parses with optimal item difficulty and item discrimination). This tool categorizes every parse into specific classifications such as JJ (adjectives), NNS (common noun plural form), CC (coordinating conjunctions), NNP (proper nouns), MD (modal verb), VBN (verb past participles), IN (prepositions), amod (adjectival modifier), nmod (noun modifier), nsubjpass (passive nominal subject), aux (auxiliary), auxpass (auxiliary passive), and so forth. Figure 1 shows an extraction of the source text annotated by the Brat CoreNLP. Having imported the whole source text, the Brat software automatically exported 257 annotations (basic dependencies) based on the neighboring parses within a source text. At this phase, the four professional evaluators tried to extract all the annotations in a source text and then compared them to the participants’ translations. This comparison was conducted as per the representative translations.

Figure 1. Brat Annotation Tool.

(III) The Calculation of Item Difficulty (1-Parameter Logistic Model) The statistical models which are applied in the Item Response Theory (IRT) suppose that a participant’s probability of answering an item correctly is associated with the participant’s ability [cf. competence] and the features of an item being measured. The participant’s competence is considered the main feature of the person which is called competence parameter. This competence parameter is believed to be considered a fundamental and unobservable latent trait which assists an individual to provide a correct answer for an item. The application of 1-Parameter Logistic Model (1-PL) (also known as item difficulty or threshold parameter) is to determine and measure the degree of difficulty of items. 1-PL measures and evaluates the location of an item within a continuum. As Umobong (2017:129) puts forward,

The item parameter is believed to be a continuum with the upper end indicating greater proficiency in whatever is measured than the lower end. This means that items located towards the right side of the continuum demands an individual to possess greater proficiency (ability) in order to answer correctly, than items located towards the left side of the continuum.

This research paper used Stata ver.16 to calculate the difficulty of each parsing item. Under 1-PL, the probability of correct items that they differed regarding their degree of difficulty were determined. Due to space limitation, this paper only brought acceptable items which had positive reliability coefficient and p-value < 0.05 (95% confidence interval) (Stata, 2016). Mind that, such items in this section are not considered an absolute license for score recalculation. Therefore, the next stage is to calculate item discrimination. In this light, items which have optimal range of p-values (item difficulty) and appropriate amount ofit r values are considered docimologically justifiable items and must be included for the recalculation of scores. The purpose of 1-PL is to find out which parsing item is considered a simple item and which one is considered a difficult item.

RLyLA Vol. 14 (2019), 1-18 | 6 Alireza Akbari Logistic Calibrated Items (LCI) Method: does it solve subjectivity in translation evaluation and assessment?

Table 2. Acceptable Degrees of Items’ Difficulty Based on 1-PL (α = 0.05). Reliability Coefficient Std. Err. Z P < 0.05 ITEM 9 0.6135796 0.20727 2.96 0.003 ITEM 11 0.4636934 0.2036581 2.28 0.023 ITEM 12 0.5757467 0.2062605 2.79 0.005 ITEM 15 1.053717 0.2238635 4.71 0.000 ITEM 17 0.4636934 0.2036581 2.28 0.023 ITEM 20 0.5381685 0.2053231 2.62 0.009 ITEM 23 0.5381685 0.2053231 2.62 0.009 ITEM 24 0.500824 0.2044561 2.45 0.014 ITEM 30 0.7288296 0.2107522 3.46 0.001 ITEM 32 0.500824 0.2044561 2.45 0.014 ITEM 38 0.4636934 0.2036581 2.28 0.023 ITEM 50 0.1001953 0.1992226 0.50 0.005 ITEM 52 0.9282814 0.2182241 4.25 0.000 ITEM 54 1.011341 0.2218769 4.56 0.000 ITEM 56 0.9695425 0.2199988 4.41 0.000 ITEM 57 0.9282814 0.2182241 4.25 0.000 ITEM 59 0.8472268 0.2149666 3.94 0.000 ITEM 60 0.6900976 0.2095137 3.29 0.001 ITEM 61 0.9282814 0.2182241 4.25 0.000 ITEM 74 0.6900976 0.2095137 3.29 0.001 ITEM 79 0.500824 0.2044561 2.45 0.014 ITEM 81 0.7679104 0.2120719 3.62 0.000 ITEM 83 0.6135796 0.20727 2.96 0.003 ITEM 84 0.8472268 0.2149666 3.94 0.000 ITEM 89 0.5757467 0.2062605 2.79 0.005 ITEM 90 0.6900976 0.2095137 3.29 0.001 ITEM 91 0.6135796 0.20727 2.96 0.003 ITEM 93 0.5381685 0.2053231 2.62 0.009 ITEM 94 0.6135796 0.20727 2.96 0.003 ITEM 95 0.500824 0.2044561 2.45 0.014 ITEM 97 0.8472268 0.2149666 3.94 0.000 ITEM 98 0.6516889 0.2083537 3.13 0.002 ITEM 99 0.4636934 0.2036581 2.28 0.023 ITEM 103 0.7288296 0.2107522 3.46 0.001 ITEM 105 0.4636934 0.2036581 2.28 0.023 ITEM 106 0.8073666 0.2134757 3.78 0.000 ITEM 108 0.4267574 0.2029275 2.10 0.035 ITEM 109 0.5757467 0.2062605 2.79 0.005 ITEM 110 0.4636934 0.2036581 2.28 0.023 ITEM 111 0.500824 0.2044561 2.45 0.014 ITEM 120 0.6135796 0.20727 2.96 0.003 ITEM 122 0.4636934 0.2036581 2.28 0.023 ITEM 123 0.5757467 0.2062605 2.79 0.005 ITEM 126 1.053717 0.2238635 4.71 0.000 ITEM 128 0.4636934 0.2036581 2.28 0.023 ITEM 131 0.5381685 0.2053231 2.62 0.009 ITEM 134 0.5381685 0.2053231 2.62 0.009 ITEM 135 0.500824 0.2044561 2.45 0.014 ITEM 141 0.7288296 0.2107522 3.46 0.001 ITEM 143 0.500824 0.2044561 2.45 0.014 ITEM 149 0.4636934 0.2036581 2.28 0.023 Table 2, continues on the next page

RLyLA Vol. 14 (2019), 1-18 | 7 Alireza Akbari Logistic Calibrated Items (LCI) Method: does it solve subjectivity in translation evaluation and assessment?

Table 2, continues from the previous page

Reliability Coefficient Std. Err. Z P < 0.05 ITEM 161 0.6900976 0.2095137 3.29 0.001 ITEM 163 0.9282814 0.2182241 4.25 0.000 ITEM 165 1.011341 0.2218769 4.56 0.000 ITEM 167 0.9695425 0.2199988 4.41 0.000 ITEM 168 0.9282814 0.2182241 4.25 0.000 ITEM 170 0.8472268 0.2149666 3.94 0.000 ITEM 171 0.6900976 0.2095137 3.29 0.001 ITEM 172 0.9282814 0.2182241 4.25 0.000 ITEM 185 0.6900976 0.2095137 3.29 0.001 ITEM 190 0.500824 0.2044561 2.45 0.014 ITEM 192 0.7679104 0.2120719 3.62 0.000 ITEM 194 0.6135796 0.20727 2.96 0.003 ITEM 195 0.8472268 0.2149666 3.94 0.000 ITEM 200 0.5757467 0.2062605 2.79 0.005 ITEM 201 0.6900976 0.2095137 3.29 0.001 ITEM 202 0.6135796 0.20727 2.96 0.003 ITEM 204 0.5381685 0.2053231 2.62 0.009 ITEM 205 0.6135796 0.20727 2.96 0.003 ITEM 206 0.500824 0.2044561 2.45 0.014 ITEM 208 0.8472268 0.2149666 3.94 0.000 ITEM 209 0.6516889 0.2083537 3.13 0.002 ITEM 210 0.4636934 0.2036581 2.28 0.023 ITEM 214 0.7288296 0.2107522 3.46 0.001 ITEM 216 0.4636934 0.2036581 2.28 0.023 ITEM 217 0.8073666 0.2134757 3.78 0.000 ITEM 219 0.4267574 0.2029275 2.10 0.035 ITEM 220 0.5757467 0.2062605 2.79 0.005 ITEM 221 0.4636934 0.2036581 2.28 0.023 ITEM 222 0.500824 0.2044561 2.45 0.014 ITEM 231 0.6135796 0.20727 2.96 0.003 ITEM 233 0.4636934 0.2036581 2.28 0.023 ITEM 234 0.5757467 0.2062605 2.79 0.005 ITEM 237 1.053717 0.2238635 4.71 0.000 ITEM 239 0.4636934 0.2036581 2.28 0.023 ITEM 242 0.5381685 0.2053231 2.62 0.009 ITEM 245 0.5381685 0.2053231 2.62 0.009 ITEM 246 0.500824 0.2044561 2.45 0.014 ITEM 252 0.7288296 0.2107522 3.46 0.001 ITEM 254 0.500824 0.2044561 2.45 0.014

In this light, item 176 was considered the least difficult item compared to the whole items with regard to the reliability coefficient (-1.129195) and the p-value (0.000 < 0.05). As noted earlier, an item must have a positive reliability coefficient and the value lower than 5% to be considered an optimal item difficulty (Stata, 2016). Although item 176 has a value lower than 0.05, it has a negative reliability coefficient and it cannot be included as an optimal indicator for item difficulty. On the other hand, item 15 was taken into account the most difficult item in terms of the reliability coefficient (1.053717) and the p-value (0.000 < 0.05).

(IV) The Calculation of Rit Value (Item Discrimination)

Rit value (also known as corrected item-total correlation as well as item discrimination) is used to reverberate ‘the performance of the item versus the test as a whole’ (van Antwerpen, 2016). This value informs a researcher/ scholar to what degree an item assists to single out good participants (higher scorers) and weak participants

(lower scorers) from the entire pool of test takers. Simply put, the application of rit value shows the discriminating properties of an item. Moreover, rit value tells a researcher that to what extent items are correctly answered by high-performing participants compared to low-performing participants (positive discrimination index [between

RLyLA Vol. 14 (2019), 1-18 | 8 Alireza Akbari Logistic Calibrated Items (LCI) Method: does it solve subjectivity in translation evaluation and assessment?

0 and 1]). On the other hand, if the majority of low-performing participants choose a correct answer for an item compared to high-performing participants, then the item being measured has a negative discrimination index (between -1 and 0). By the same token, Eyckmans and Anckaert (2017) have noted that:

The rit value calculates the correlation between the item and the rest of the scale, without that item being considered as part of the scale; that is, it reflects the amount the item contributes to the test’s global reliability.

In order to detect the discriminating items, a researcher has to identify items with good discriminating power

(rit ≥ 0.30). In this vein, items with a rit value of 0.40 and above are indicators of very good items; items with a rit value of 0.30 to 0.39 are considered good discriminators; items with a rit value of 0.20 to 0.29 have fairly good discriminatory power; and consequently items with a rit value of 0.19 or less are considered poor discriminators (Anckaert et al., 2008). As a matter of fact, only items with good and very good discriminatory power are included within a test (Kockaert and Segers, 2017; Akbari and Segers, 2017a; Eyckmans and Anckaert, 2017). This research identified all items with fair and poor discriminatory power and excluded them from the test since their range of itr values were inferior to 0.30 (Table 3).

Table 3: Excluded Items from the Test (Rit < 0.30).

Rit Value (Corrected Item-Total Correlation ) Items Scale Variance if Item Corrected Item-Total Items Scale Variance if Corrected Item-Total Deleted Correlation Item Deleted Correlation ITEM 6 3332.265 0.291 ITEM 137 3333.525 0.282 ITEM 7 3334.639 0.255 ITEM 151 3334.550 0.251 ITEM 8 3334.313 0.256 ITEM 153 3333.061 0.277 ITEM 40 3334.550 0.251 ITEM 154 3332.509 0.286 ITEM 42 3333.061 0.277 ITEM 155 3348.903 0.003 ITEM 43 3332.509 0.286 ITEM 156 3349.794 -0.012 ITEM 44 3348.903 0.003 ITEM 157 3338.844 0.177 ITEM 45 3349.794 -0.012 ITEM 159 3343.018 0.105 ITEM 46 3338.844 0.177 ITEM 160 3344.519 0.079 ITEM 48 3343.018 0.105 ITEM 161 3337.813 0.204 ITEM 49 3344.519 0.079 ITEM 164 3347.957 0.020 ITEM 50 3337.813 0.204 ITEM 169 3335.960 0.227 ITEM 53 3347.957 0.020 ITEM 189 3334.811 0.247 ITEM 58 3335.960 0.227 ITEM 196 3335.455 0.236 ITEM 78 3334.811 0.247 ITEM 199 3332.361 0.289 ITEM 85 3335.455 0.236 ITEM 213 3350.794 -0.029 ITEM 88 3332.361 0.289 ITEM 215 3333.271 0.275 ITEM 102 3350.794 -0.029 ITEM 218 3333.351 0.274 ITEM 104 3333.271 0.275 ITEM 228 3332.265 0.291 ITEM 107 3333.351 0.274 ITEM 229 3334.639 0.255 ITEM 117 3332.265 0.291 ITEM 230 3334.313 0.256 ITEM 118 3334.639 0.255 ITEM 248 3333.525 0.282 ITEM 119 3334.313 0.256

(V) The Identification of Docimologically Justifed and Calibrated Items

Based on the obtained results from 1-PL (item difficulty) and rit values (item discrimination) in stages III and IV, 88 parsing items were categorized as docimologically calibrated parsing items. From the ninety acceptable items in 1-PL, 2 parsing items (parsing items 50 and 161) were not docimologically calibrated due lower rit values (below 0.30); therefore, they were excluded from the test. The rest had acceptable item difficulty and item discrimination. The complete list of the docimologically accepted parsing items is as follows:

RLyLA Vol. 14 (2019), 1-18 | 9 Alireza Akbari Logistic Calibrated Items (LCI) Method: does it solve subjectivity in translation evaluation and assessment?

Table 4. Docimologically Calibrated Parsing Items (Criterion-Referenced Evaluation).

Item P-value Rit Item P-value Rit Item P-value Rit 9 0.003 0.543 38 0.023 0.391 81 0.000 0.541 11 0.023 0.718 52 0.000 0.316 83 0.003 0.610 12 0.005 0.419 54 0.000 0.375 84 0.000 0.664 15 0.000 0.647 56 0.000 0.344 89 0.005 0.703 17 0.023 0.607 57 0.000 0.342 90 0.001 0.614 20 0.009 0.584 59 0.000 0.470 91 0.003 0.644 23 0.009 0.612 60 0.001 0.474 93 0.009 0.691 24 0.014 0.612 61 0.000 0.477 94 0.003 0.433 30 0.001 0.461 74 0.001 0.468 95 0.014 0.554 32 0.014 0.495 79 0.014 0.394 97 0.000 0.616

Item P-value Rit Item P-value Rit Item P-value Rit 98 0.002 0.752 122 0.023 0.718 163 0.000 0.316 99 0.023 0.634 123 0.005 0.419 165 0.000 0.375 103 0.001 0.397 126 0.000 0.647 167 0.000 0.344 105 0.023 0.389 128 0.023 0.607 168 0.000 0.342 106 0.000 0.390 131 0.009 0.584 170 0.000 0.470 108 0.035 0.478 134 0.009 0.612 171 0.001 0.474 109 0.005 0.499 135 0.014 0.612 172 0.000 0.477 110 0.023 0.500 141 0.001 0.461 185 0.001 0.468 111 0.014 0.402 143 0.014 0.495 190 0.014 0.394 120 0.003 0.543 149 0.023 0.391 192 0.000 0.541

Item P-value Rit Item P-value Rit 194 0.003 0.610 219 0.035 0.478 195 0.000 0.664 220 0.005 0.499 200 0.005 0.703 221 0.023 0.500 201 0.001 0.614 222 0.014 0.402 202 0.003 0.644 231 0.003 0.543 204 0.009 0.691 233 0.023 0.718 205 0.003 0.433 234 0.005 0.419 206 0.014 0.554 237 0.000 0.647 208 0.000 0.616 239 0.023 0.607 209 0.002 0.752 242 0.009 0.584 210 0.023 0.634 245 0.009 0.612 214 0.001 0.397 246 0.014 0.612 216 0.023 0.389 252 0.001 0.461 217 0.000 0.390 254 0.014 0.495

(VI) The Recalculation of Scores As mentioned earlier, the purpose of the first stage was to show the difference between holistic scoring system which has been using in majority of universities across the globe and LCI system of scoring to check which one was more objective and consistent:

RLyLA Vol. 14 (2019), 1-18 | 10 Alireza Akbari Logistic Calibrated Items (LCI) Method: does it solve subjectivity in translation evaluation and assessment?

Table 5. LCI Recalculation of Scores (Par: Participant).

Par Score LCI Par Score LCI Par Score LCI Par Score LCI Par Score LCI 1 13 13.863 26 10 8.863 51 13 13.636 76 15 13.863 101 16 17.727 2 15 14.772 27 12 10.227 52 15 12.50 77 15 15.227 102 15 15.909 3 15 16.363 28 13 12.954 53 14 12.045 78 17 17.50 103 17 17.045 4 16 17.727 29 13 13.863 54 10 9.318 79 15 15.454 104 14 14.545 5 14 15.909 30 17 16.818 55 12 10.909 80 16 14.772 105 18 18.409 6 10 8.636 31 10 8.409 56 15 15 81 13 15.681 106 19 17.954 7 11 11.590 32 20 19.090 57 16 17.727 82 19 18.181 107 14 14.090 8 13 12.50 33 20 19.318 58 17 18.181 83 12 11.818 108 15 12.272 9 12 12.272 34 18 19.772 59 17 17.954 84 15 12.50 109 15 13.409 10 17 18.863 35 18 19.545 60 17 17.50 85 16 15.454 110 15 15.227 11 15 14.318 36 9 7.50 61 12 8.863 86 17 15.227 111 18 17.727 12 20 18.409 37 11 8.636 62 12 9.545 87 15 17.727 112 15 14.772 13 18 16.363 38 13 12.045 63 10 10.909 88 14 14.090 113 12 10.909 14 14 12.045 39 14 12.954 64 15 13.181 89 13 15.454 114 10 10.681 15 12 11.818 40 14 16.363 65 15 15.681 90 15 17.50 115 16 17.272 16 11 12.727 41 15 14.545 66 15 16.136 91 14 12.272 116 13 14.772 17 12 12.954 42 12 13.181 67 18 17.045 92 15 13.181 117 14 12.272 18 12 9.318 43 13 15.681 68 18 18.636 93 14 15 118 19 19.090 19 11 10 44 12 9.545 69 20 18.636 94 17 18.409 119 13 10.681 20 15 17.045 45 11 9.090 70 20 19.545 95 12 11.136 120 15 12.954 21 16 17.272 46 15 14.090 71 18 17.954 96 13 11.363 121 20 19.545 22 16 16.136 47 15 15.454 72 17 16.363 97 14 12.954 122 20 19.090 23 17 15.681 48 15 16.136 73 18 17.50 98 15 15.227 123 12 8.409 24 15 16.363 49 15 17.045 74 18 17.045 99 14 15.681 124 16 14.090 25 10 5.227 50 13 15.909 75 18 18.636 100 14 14.090 125 15 14.318

According to Table 5, for instance, the outcome of this recalculation is the most crucial for participants [12] [13], [18], [25], [36], [96], and [123] (just to name a few) going from 20 (holistic scoring) to 18.409, 18 to 16.363, 12 to 9.318, 10 to 5.227, 9 to 7.50, 13 to 11.363, and 12 to 8.409 (LCI scoring) respectively. This is due to the fact that in spite of the overall quality of their translations, they had not been able to translate most of the docimologically calibrated parsing items correctly (88 parsing items) after calculating the degree of item difficulty by 1-PL andit r value (item discrimination). However, for example, participants [34], [35], [50], [87], [89], and [90] (just to name a few) obtained higher scores compared to the first calculation (holistic calculation) (18 vs. 19.772), (18 vs. 19.545), (13 vs. 15.909), (15 vs. 17.727), (13 vs. 15.454), and (15 vs. 17.50) respectively. This was on the grounds that they translated the justified parsing items correctly besides translating the total parses in a text (both justified and unjustified items).

5. DISCUSSION

5.1. The Confrontation between Item Response Theory (IRT) and Classical Test Theory (CTT): The Question of Item Difficulty Generally, a translation test consists of items which are both easy and difficult to translate. Moreover, some items within a source text discriminate high-performing students from low-performing ones. As explained earlier, stage (III) of LCI method dealt with the measurement of each item’s difficulty based on 1-Parameter Logistic Model of IRT. In layman’s terms, items difficulty refers to the proportion of correct answers which are provided for one item by the participants. In this direction, higher degree of item difficulty shows the condition when a small percentage of participants get an item correct and conversely lower degree of item difficulty demonstrates the condition that high percentage of participants get an item correct. Both IRT and CTT have been using and testing item difficulty. The question is that why LCI method selects the former approach for measuring item difficulty.

RLyLA Vol. 14 (2019), 1-18 | 11 Alireza Akbari Logistic Calibrated Items (LCI) Method: does it solve subjectivity in translation evaluation and assessment?

Unlike CTT which ignores the role of participants’ competence, IRT is applied to inspect the latent trait (in our case translation competence) which is associated with ‘a set of items within a test’ (Baker and Kim, 2004). In updated educational contexts, assessment and evaluation are considered inherent parts of curriculum design to evaluate and measure students’ proficiency and skill development (Le, 2013). Apart from calculating the total score, a researcher tends to find out whether testing and evaluation tools are adequately formulated to measure a number of specific aspects of students’ knowledge. In this direction, the purpose of IRT is to measure/assess/ evaluate the relevance of questions coupled with assessing the degree of participants’ competence (item difficulty and item discrimination). Unlike CTT, we supposed that a translation participant who translated the source text into plain Persian possessed some amount of translation ‘competence’ which more or less impacted the end results (see Table 5) (Hambleton, 1989; Kempf, 1983; Farmer et al., 2001; Finch and Edwards, 2015). In order to find out the high-performing translation students from the low-performing ones regarding their translation competence, a number of issues such as the degree of the item difficulty and degree of discriminatory power of each item must be calculated. With that idea, IRT takes into account participants’ translation competence and items’ characteristics based on item analysis or difficulty (p-value) of 1-Parameter Logistic Model. Unlike CTT, IRT does consider the number of correct items to measure a participant’s performance, nor does it suppose ‘equal contribution of the items (questions) to the overall scores’ (Le, 2013:13). Although items are different in terms of their difficulty and participants differ in terms of their competence or ability, IRT may provide accurate results compared to CTT (Baker, 2001; Zięba, 2013). This is chiefly on the grounds that that participants who obtain the same total scores in a test may vary in their degree of competence. For instance, participants [1], [8], and [28] commonly obtained 13 once the evaluators scored the translations holistically (stage I); however, when the evaluators applied LCI method of translation evaluation, their scores were changed to 13.863, 12.50, and 12.954 respectively. This showed that the participants’ level of competence differed with one another in that LCI method could demonstrate such difference. Fox (2010) has pointed out that if a difference between a participant’s competence and the level of difficulty of one item is positive or positively skewed (positive reliability coefficient and p-value < 0.05) (e.g. items [99], [134], [185]), the participant has a high chance of answering that item correctly (e.g. participants [40], [81], [94], and [115]). Conversely, if the difference between a participant’s competence and the level of difficulty of one item is negative or negatively skewed (negative reliability coefficient and p-value > 0.05) (e.g. items [7], [26], [257]), the participant has the low chance of answering that item correctly (e.g. participants [18], [36], [45], and [69]). As noted earlier, identifying the difficulty of each item is not the absolute condition. Additionally, a translation evaluator must then identify those parsing items which have good and very good discriminatory powers.

5.2. The Confrontation between Rit Value and Extreme Group Method: The Question of Item Discrimination

Two ways are available to calculate the item discrimination: (i) the use of rit value (corrected item-total correlation) and (ii) the extreme group method. Translation evaluation models such as PIE method and LCI method depend on discrimination indices to identify higher group of scorers from lower group of scorers. In this direction, in LCI method of evaluation, this stage is computed by the application of rit value function through SAS, SPSS, and Stata statistical packages. A question may arise concerning that why LCI method of evaluation selects the application of rit value to identify high-performing students from the lower ones. As noted earlier, PIE method calculates item discrimination through the application of the extreme group method which dates back to the pre-computer era (Pidgeon and Yates, 1968). Unfortunately, the purpose of using the method of extreme group in PIE method has not so far been explained and substantiated. According to Eyckmans and Anckaert (2017:45),

The Rit value has the advantage over the extreme-group method that every test-taker’s score is used to compute the discrimination coefficient, whereas only 54 percent of the test-takers’ results are used in the case of the extreme- group method (i.e. the 27% upper and the 27% lower scores).

Kockaert and Segers (2017) as the founder of PIE method believe that the application of 27% rule will maximize differences in a normal distribution; however, it is not the case in point. Selecting the sample size in a normal distribution from a distribution of a test score is of paramount importance. Conventionally, the size of the selected tails is considered an independent sample; nevertheless, the size of the selected tails is dependent and must contain about 21% instead of 27% (D’Agostino and Cureton, 1975). D’Agostino and Cureton (1975:40) have put forward that

The optimal tail size is around 0.27 if the correlation between the concomitant variable and the test scores is small (i.e. around 0.10). Under normality, this is implying independence. As the correlation increases, the optimal tail size deceases. From above it appears to follow that if the concomitant variable and the test scores have correlation one, then the optimal tail size is around 0.215.

Based on the findings of D’Agostino and Cureton (1975), the data are uncorrelated at the 27% level and yield inaccurate results. The application of rit value has an advantage over the method of extreme group in that it can be used for larger samples. Therefore, an evaluator will have access to the great number of good and very good discriminating items when calculating by rit value. On the other hand, item discrimination which is used in both PIE

RLyLA Vol. 14 (2019), 1-18 | 12 Alireza Akbari Logistic Calibrated Items (LCI) Method: does it solve subjectivity in translation evaluation and assessment?

method and CTT is determined based on surprisingly small sample size (e.g. from 10 test-takers to 20 test-takers) (Kockaert and Segers, 2014). Consequently, the obtained results must be analyzed and interpreted with great care.

In the context above, after having perused every annotation of the source text and having identified optimal item difficulties based on 1-PL, LCI method selected the application of rit value to identify the top scorers from the bottom ones since this research was dealing with both the large sample size (125 translation students) and the large scale purification, i.e. the process of removing items from multi-item scales due to their negative values or values below 0.30 (see Table 4).

5.3. Why Brat Text Annotator? As explained earlier, Brat rapid annotation tool is considered an impressionistic web-based tool for making a source text into a number of parses (Bunt et al., 2010). This tool is fully supported by Natural Language Processing (NLP). Brat is taken into account a shared task for detecting how factual statements and parses can be interpreted in terms of their textual contexts involving a hypothesis and an experimental result. The extraction of information into a number of annotations or parses is the fundamental task of representing information contained in a text through the Brat tool. Brat tool is employed to detect metaphor annotation by means of bottom-up identification (Stenetorp et al., 2012) chiefly concentrated on the linguistic metaphors within a source text and extrapolating the conceptual metaphors underlying them. The Brat Stanford CoreNLP tool has a number of features namely; (i) ‘high-quality annotation visualization’ (every parser is intuitively visualized based on the concept of ‘what you see is what you get’); (ii) ‘intuitive annotation interface’ (this tool is used to detect any parses using Uniform Recourse Identifiers (URLs) that empowers connecting to individual annotations for simplifying easy communication); (iii) ‘versatile annotation support’ (Brat is set up to support most annotation tasks by means of binary relations such as part of speech (POS) tagged tokens or chunks); and (iv) ‘corpus search functionality’ (this tool executes an exhaustive set of search functions which permits users to search through the collection of different documents for text span relations and correspondents) (Stenetorp et al., 2012). For example, in Figure 2, every annotation within a source text is visually linked to different colors. Moreover, the Brat tool detects relations between tokens or chunks so that an evaluator can simply spot the corresponding parsing items in a reciprocal language to scrutinize whether or not the detected parsing item is correctly translated.

LCI moves from the norm-referenced assessment to the criterion-referenced assessment. Unlike available translation evaluation methods such as PIE method, analytic method, etc. which focus on either criterion-referenced assessment or norm-referenced assessment, LCI method benefits from the amalgam of norm and criterion- referenced assessment methods by means of a feedback loop, including a norm-referenced assessment method (the whole parsing items in a sauce text), criterion-referenced assessment (the docimologically justified parsing items), and the actual evaluating. This feedback loop is used to remove any score inflation by the concomitant use of both norm- and criterion-referenced assessment methods.

Figure 2. Illustration of Brat Software Analysis.

For example, the term ‘infrastructure’ (NN) within the source text relates to the corresponding terms such as ‘enrichment-related’ (JJ-amod-NN) and ‘centrifuges’ (NNS-conj-NN) on the left side and ‘Natanz’ (NN-nmod-NNP) on the right side. In this vein, evaluators must search for the corresponding translations of the terms ‘enrichment- related-infrastructure’, ‘centrifuges- infrastructure’, and ‘infrastructure-Natanz’. Based on the four representative translations, the corresponding Persian translations are ‘zir sāxt’hāje qæni sāzi’je mærbouteh’, ‘sāntrifiouzh’ha va zir sāxt’, and ‘zir sāxt dar Nætænz’ respectively. Then, the evaluators compare students’ translations with the given corresponding translations to check their acceptability.

To take another example, the term ‘stored’ (VBN) corresponds to ‘centrifuge-stored’ (NNS-nsubjpass-VBN), ‘will-stored’ (MD-aux-VBN), and ‘be-stored’ (VB-auxpass-VBN) on the left side and ‘stored-monitoring’ (VBN- nmod-NN) and ‘stored-specified’ VBN-nmod-VBN( ) on the right side. The corresponding translations based on representative translations are ‘sāntrifiouzh’haje ænbār šode’, ‘ænbār kærdæn’, ‘ænbār xāhæd šod’, ‘tæhte nezāræt ænbār šode’ and ‘ænbār šode be næhve mæš’ruh’ respectively. These translations were approved by the

RLyLA Vol. 14 (2019), 1-18 | 13 Alireza Akbari Logistic Calibrated Items (LCI) Method: does it solve subjectivity in translation evaluation and assessment?

evaluators as correct corresponding translations. Last but not least, based on the extracted chunks, evaluators measured the acceptability of the translations (optimal p and rit values) to check whether or not a parsing item is tagged as a docimologically justified item.

6. CONCLUSION

6.1. Limitations of the Research One of the limitations of the present research is the relatively small sample size. As far as 1-PL of IRT is used for a larger sample size (Hambleton and Jones, 1993:43), the maintained findings of the present research paper are not adequately accurate when compared with a large sample size (above 500 participants). Also, the translation assignment (English to Persian translation) was carried out with paper and pen. In a recreation of the research paper with a larger sample size, much prudence is needed to permit the involved participants to conduct the assignment by computer or online-platforms. As far as the administration of 1-PL is manually time- consuming; thus, a good knowledge of statistical packages (e.g. Stata, SAS, and Winsteps) are pivotally needed for a researcher to precisely and meaningfully interpret and analyze the obtained results. Another limitation is that the application of LCI method is a time-consuming activity. To solve such a problem, a computerized platform is necessarily required to check responses and likewise a list of correct and incorrect solutions of annotations must be arranged.

6.2. Implications of the Research It is widely known that revising and scoring translations are time-consuming, tedious, repetitive, and subjective. A reviser cannot guarantee that he/she can find every mistake in translations of the same source text and give consistent feedback to translators or grade translations consistently. These are the challenges that translationQ solves in the field of translation evaluation and assessment. TranslationQ is a web and cloud-based productivity software that a translator can use for translation training and revision. This online platform ensures objectivity and works independent of language or domain. No matter which language pair is being used and no matter if a translator is using different text types. When a translator prepares a translation, the end product can be imported in the formats of SDL XLIFF and word files. TranslationQ segments a source text automatically. Also, translationQ tags errors for bilingual text through using TAUS Dynamic Quality Framework (DQF) error categories ‘that was developed as part of the (EU-funded) QTLaunchPad project (large-scale action for quality translation technology) based on careful examination and extension of existing quality models’ (TAUS, 2018). TAUS DQF refers to ‘a comprehensive set of tools, best practices, metrics, reports and data to help the industry set benchmarking standards’ (ibid.).

TranslationQ will automatically detect identical errors in other translations and this will save a huge amount of time. Blind revision of translations is another possibility in case where researchers and instructors tend to use translationQ for high-stake exams. In this light, LCI method has the potentiality to be employed in the translationQ platform, because this method can be used in various text types such as medical, political, cultural, and so forth. As far as the application of LCI method is a laborious activity, the automation of this method by the translationQ will improve the scoring system, add options throughout the Brat process (Stage II), and update all available corrections regularly.

In summary, this paper introduced a translation evaluation method called Logistic Calibrated Items (LCI). The purpose of LCI method was to objectify translation evaluation products. LCI method was an attempt to identify high proficient translators through the application of six phases as fully explained in section 4. Last but not least, LCI employed Brat CoreNLP software to identify all parses (norm-referenced assessment method) and then distinguish appropriate and justified items (optimal p and itr values) (criterion-referenced assessment method) within a source text.

REFERENCES

Akbari, A. & Gholamzadeh Bazarbash, M. (2017). “Holistic Assessment: Effective or Lenient in Translation Evaluation?” Skopos. Revista Internacional de Traducción e Interpretación, 8/1: 51-67. Akbari, A. & Segers, W. (2017a). “Translation Difficulty: How to Measure and What to Measure”.Lebende Sprachen, 62/1: 3-29. https://doi.org/10.1515/les-2017-0002 Akbari, A. & Segers, W. (2017b). “Translation Evaluation Methods and the End-Product: Which One Paves the Way for a More Reliable and Objective Assessment?” Skase Journal of Translation and Interpretation, 11/1: 2-24. Akbari, A. & Segers, W. (2017c). “Evaluation of Translation through the Proposal of Error Typology: An Explanatory Attempt”. Lebende Sprachen, 62/2: 408-430. https://doi.org/10.1515/les-2017-0022

RLyLA Vol. 14 (2019), 1-18 | 14 Alireza Akbari Logistic Calibrated Items (LCI) Method: does it solve subjectivity in translation evaluation and assessment?

Akbari, A. & Segers, W. (2017d). “The Perks of Norm and Criterion Referenced Translation Evaluation”. LICTRA, Leipzig, Germany, 20 March. Anckaert, P., Eyckmans, J. & Segers, W. (2008). “Pour Une Évaluation Normative De La Compétence De Traduction.” ITL - International Journal of Applied Linguistics, 155/1: 53-76. https://doi.org/10.2143/ITL.155.0.2032361 Bahameed, A. S. (2016). “Applying assessment holistic method to the translation exam in Yemen.” Babel, 62/1: 135-149. https://doi.org/10.1075/babel.62.1.08bah Baker, F. B. (2001). The Basics of Item Response Theory. 2nd ed. New York: ERIC Clearinghouse on Assessment and Evaluation. Baker, F. B. & Seock-Ho K. (2004). Item Response Theory: Parameter Estimation Techniques. 2nd ed. New York: Marcel Dekker. https://doi.org/10.1201/9781482276725 Barkaoui, K. (2010). “Explaining ESL essay holistic scores: A multilevel modeling approach”. Language Testing, 27/4: 515-535. https://doi.org/10.1177/0265532210368717 Barkaoui, K. (2011). “Effects of marking method and rater experience on ESL essay scores and rater performance”. Assessment in Education: Principles, Policy & Practice, 18/3: 279-293. https://doi.org/10.1080/096959 4X.2010.526585 Bunt, H., Merlo, P. & Nivre, J. (2010). Trends in Parsing Technology: Dependency Parsing, Domain Adaptation, and Deep Parsing. Netherlands: Springer. https://doi.org/10.1007/978-90-481-9352-3 Conde Ruano, T. (2005). “No me parece mal. Comportamiento y resultados de estudiantes al evaluar traducciones”. Unpublished doctoral dissertation. University of Granada, Granada. Cumming, Alister, Kantor, R. & Powers, D. E. (2002). “Decision making while rating ESL/EFL writing tasks: A descriptive framework”. The Modern Language Journal, 86/1: 67-96. https://doi.org/10.1111/1540-4781.00137 D’Agostino, R. B. & Cureton, E. E. (1975). “The 27 Percent Rule Revisited”. Educational and Psychological Measurement, 35/1: 47-50. https://doi.org/10.1177/001316447503500105 Dancette, J. (1989). “La faute de sens en traduction”. TTR : traduction, terminologie, rédaction, 2/2: 83-102. https:// doi.org/10.7202/037048ar Eberly Center. (2016). “What is the difference between formative and summative assessment?” Accesible at https://www.cmu.edu/teaching/assessment/basics/formative-summative.html [Last access: July 2019]. Eyckmans, J. & Anckaert, P. (2017). “Item-based assessment of translation competence: Chimera of objectivity versus prospect of reliable measurement”. Linguistica Antverpiensia, New Series: Themes in Translation Studies, 16/1: 40–56. Eyckmans, J., Anckaert, P. & Segers, W. (2013). “Assessing Translation Competence”. Actualizaciones en Comunicación Social. Centro de Lingüística Aplicada, Santiago de Cuba, 2, 513-515. Eyckmans, J., Segers, W. & Anckaert, P. (2012). Translation Assessment Methodology and the Prospects of European Collaboration. In Collaboration in Language Testing and Assessment, edited by D. Tsagari and I. Csépes, 171-184. Bruxelles: Peter Lang. Farmer, W. L., Thompson, R. C., Heil, S. K. R. & Heil, M. C. (2001). Latent Trait Theory Analysis of Changes in Item Response Anchors. Accesible at https://www.faa.gov/data_research/research/med_humanfacs/ oamtechreports/2000s/media/0104.pdf. [Last access: July 2019]. Finch, H. & Edwards, J. M. (2015). “Rasch Model Parameter Estimation in the Presence of a Nonnormal Latent Trait Using a Nonparametric Bayesian Approach”. Educational and Psychological Measurement, 76/4: 662-684. https://doi.org/10.1177/0013164415608418 Fox, J. (2010). Bayesian Item Response Modeling: Theory and Applications. Amsterdam: Springer. https://doi. org/10.1007/978-1-4419-0742-4 Garant, M. (2009). “A case for holistic translation assessment”. AFinLA-e: Soveltavan kielitieteen tutkimuksia, 5/2: 5-17. Gonzalez, K. (2018). “Contrast Effect: Definition & Example”. https://study.com/academy/lesson/contrast-effect- definition-example.html. Gouadec, D. (1989). “Comprendre, évaluer, prévenir : Pratique, enseignement et recherche face à l’erreur et à la faute en traduction”. TTR, 2/2: 35-54. https://doi.org/10.7202/037045ar Hambleton, R. K. (1989). Principles and selected applications of item response theory. In The American Council on Education/Macmillan series on higher education, edited by R. L. Linn, 147-200. New York, NY, England: Macmillan Publishing Co. Hambleton, R. K. & Jones, R. W. (1993). “Comparison of classical test theory and item response theory and their applications to test development”. Educational Measurement: Issues and Practice, 12/3: 38-47. https://doi.org/10.1111/j.1745-3992.1993.tb00543.x Hamp-Lyons, L. (1991). Scoring procedures for ESL contexts. In Assessing second language writing in academic contexts, edited by L. Hamp-Lyons, 241–76. Norwood, NJ: Ablex.

RLyLA Vol. 14 (2019), 1-18 | 15 Alireza Akbari Logistic Calibrated Items (LCI) Method: does it solve subjectivity in translation evaluation and assessment?

Harsch, C. & Martin, G. (2013). “Comparing holistic and analytic scoring methods: issues of validity and reliability”. Assessment in Education: Principles, Policy & Practice, 20/3: 281-307. https://doi.org/10.1080/096959 4X.2012.742422 Kempf, W. (1983). “Some Theoretical Concerns ahout Applying Latent Trait Models in Educational Testing”. Accesible at https://pdfs.semanticscholar.org/5909/0351c0bc109f28836a75eaa67e7eecaffa41.pdf. [Last access: July 2019]. Kockaert, H. & Segers, W. (2014). “Evaluation de la Traduction : La Méthode PIE (Preselected Items Evaluation)”. Turjuman, 23/2: 232-250. Kockaert, H. & Segers, W. (2017). “Evaluation of legal translations: PIE method (Preselected Items Evaluation)”. Journal of Specialized Translation, 27: 148-163. Kussmaul, P. (1995). Training the Translator. Amsterdam: John Benjamins Publishing Company. https://doi. org/10.1075/btl.10 Larose, R. (1989). Théories contemporaines de la traduction. Montréal: Presses de l’Université de Québec. Le, D. (2013). Applying item response theory modeling in educational research. Instructional Technology, Graduate College at Iowa State University Digital Repository. Lei, P. & Wu, Q. (2007). “CTTITEM: SAS macro and SPSS syntax for classical item analysis”. Behavior Research Methods, 39/3: 527-530. https://doi.org/10.3758/BF03193021 Mariana, V., Cox, T. & Melby, A. (2015). “The Multidimensional Quality Metrics (MQM) Framework: A New Framework for Translation Quality Assessment”. Journal of Specialized Translation, 23:137-161. Muñoz Martín, R. (2010). On Paradigms and Cognitive Translatology. In Translation and Cognition, edited by G. Schreve and E. Angelone, 169-187. Amsterdam and Philadelphia: John Benhamins. https://doi.org/10.1075/ ata.xv.10mun Newmark, P. (1991). About Translation. Clevedon: Multilingual Matters. Nord, C. (2005). Text Analysis in Translation. aMSTERDAM: Rodopi. Pidgeon, D. A. & Yates, A. (1968). An introduction to educational measurement. London: Routledge. Preacher, K. J. , Rucker, D. D., MacCallum, R. C. & Nicewander, W. A. (2005). “Use of the extreme groups approach: a critical reexamination and new recommendations”. Psycholological Methods, 10/2: 178-792. https://doi.org/10.1037/1082-989X.10.2.178 Schmitt, P. A. (2005). Qualitätsbeurteilung von Fachübersetzungen in der Übersetzerausbildung, Probleme und Methoden. Vertaaldagen Hoger Instituut voor Vertalers en Tolken, 16-17 March. SPSS, IBM. (2017). Available at https://www.ibm.com/analytics/us/en/technology/spss/. [Last access: July 2019]. Stansfield, C. W., Scott, M. L. & Kenyon, D. M. (1992). “The Measurement of Translation Ability”. The Modern Language Journal, 76/4: 455-467. https://doi.org/10.2307/330046 Stata. (2016). “Stata: Software for Statistics and Data Science”. Available at https://www.stata.com/. [Last access: July 2019]. Stenetorp, P., Pyysalo, S., Topic, G., Ohta, T., Ananiadou, S. & Tsujii, J. (2012). BRAT: a web-based tool for NLP- assisted text annotation. Proceedings of the Demonstrations at the 13th Conference of the European Chapter of the Association for Computational Linguistics, Avignon, France. Stenetorp, P., Topic, G., Pyysalo, S., Ohta, T., Kim, J. & Tsujii, J. (2011). BioNLP Shared Task 2011: Supporting Resources. BioNLP Shared Task 2011 Workshop, Portland, Oregon, USA. TAUS. (2018). “Measuring and Benchmarking Translation Quality”. Available at https://www.taus.net/quality- dashboard-lp. [Last access: July 2019]. Umobong, M. E. (2017). “The One-Parameter Logistic Model (1PLM) And its Application in Test Development”. Advances in Social Sciences Research Journal, 4/24: 126-137. Van Antwerpen, J. (2016). “P-, D-, and Rit values: a new start”. Available at http://www.andriesseninternational. com/p-d-and-rit-values-a-new-start/. [Last access: July 2019]. Waddington, C. (2001). “Different Methods of Evaluating Student Translations: The Question of Validity”. Meta, 46/2: 311-325. https://doi.org/10.7202/004583ar Weigle, S. C. (2002). Assessing writing. Cambridge: Cambridge University Press. https://doi.org/10.1017/ CBO9780511732997 Widdowson, H. G. (1978). Teaching Language as Communication Oxford: Oxford University Press. Wiersma, W. & Jurs, S. G. (1990). Educational Measurement and Testing. London: Allyn and Bacon. Zięba, A. (2013). “The Item Information Function in One and Two-Parameter Logistic Models- A Comparison and Use in the Analysis of the Results of School Tests”. Didactics of Mathematics, 10/14: 87-96. https://doi. org/10.15611/dm.2013.10.08

RLyLA Vol. 14 (2019), 1-18 | 16 Alireza Akbari Logistic Calibrated Items (LCI) Method: does it solve subjectivity in translation evaluation and assessment?

APPENDIX 1

Selected Source Text Iran’s long term plan includes certain agreed limitations on all uranium enrichment and uranium enrichment- related activities including certain limitations on specific research and development (R&D) activities for the first 8 years, to be followed by gradual evolution, at a reasonable pace, to the next stage of its enrichment activities for exclusively peaceful purposes, as described in Annex I. Iran will abide by its voluntary commitments, as expressed in its own long-term enrichment and enrichment R&D plan to be submitted as part of the initial declaration for the Additional Protocol to Iran’s Safeguards Agreement. Iran will begin phasing out its IR-1 centrifuges in 10 years. During this period, Iran will keep its enrichment capacity at Natanz at up to a total installed uranium enrichment capacity of 5060 IR-1 centrifuges. Excess centrifuges and enrichment-related infrastructure at Natanz will be stored under IAEA continuous monitoring, as specified in Annex I. Iran will continue to conduct enrichment R&D in a manner that does not accumulate enriched uranium. Iran’s enrichment R&D with uranium for 10 years will only include IR-4, IR-5, IR-6 and IR-8 centrifuges as laid out in Annex I, and Iran will not engage in other isotope separation technologies for enrichment of uranium as specified in Annex I. Iran will continue testing IR-6 and IR-8 centrifuges, and will commence testing of up to 30 IR-6 and IR-8 centrifuges after eight and a half years, as detailed in Annex I.

APPENDIX 2: WADDINGTON’S FRAMEWORK OF HOLISTIC TRANSLATION

Level Accuracy of Transfer of ST Quality of Expressions in TL Degree of Task Mark Content Completion Level 5 Complete transfer of ST Almost all the translation reads Successful 9,10 information, only minor revision like a piece originally written in needed to reach professional English. There may be minor standards. lexical, grammatical, and spelling errors. Level 4 Almost complete transfer; there Large sections read like a piece Almost completely 7,8 may be one or two insignificant originally written in English. successful inaccuracies; requires certain There are a number of lexical, amount of revision to reach grammatical, or spelling errors. professional standards. Level 3 Transfer of general ideas but with Certain parts read like a piece Adequate 5,6 a number of lapses in accuracy; originally written in English, but needs considerable revision to others read like a translation. reach professional standards. There are a considerable number of lexical, grammatical, or spelling errors. Level 2 Transfer undermined by serious Almost the entire text reads like Inadequate 3,4 inaccuracies; thorough revision a translation; there are continual required to reach professional lexical, grammatical, or spelling standards. errors. Level 1 Totally inadequate transfer of ST The candidate reveals a total Totally inadequate 1,2 content, the translation is not lack of ability to express himself worth revising. adequately in English.

RLyLA Vol. 14 (2019), 1-18 | 17 Alireza Akbari Logistic Calibrated Items (LCI) Method: does it solve subjectivity in translation evaluation and assessment?

RLyLA Vol. 14 (2019), 1-18 | 18 Revista de Lingüística y Lenguas Aplicadas Vol. 14 año 2019, 19-27 EISSN 1886-6298 https://doi.org/10.4995/rlyla.2019.10948

¿COMPLIANCE O CUMPLIMIENTO NORMATIVO? ESTUDIO DE CORPUS CON FINES PROFESIONALES Y DIDÁCTICOS EN LA PRENSA ESPAÑOLA

Elena Alcalde Peñalver Alexandra Santamaría Urbieta Universidad de Alcalá, España

Resumen: El objetivo de este artículo es analizar el uso del término compliance, considerado uno de los temas de moda en el ámbito jurídico-económico en 2017 y 2018, en la prensa española para determinar si existe una preferencia por su uso en inglés, su traducción al español o ambas opciones. Se ha utilizado una metodología de corpus para estudiar la frecuencia de uso del término en dos periódicos generalistas (El País y La Vanguardia) y otros dos especializados (Expansión y El Economista), durante el periodo comprendido entre enero de 2017 y octubre de 2018. Los resultados muestran una preferencia de uso del término en inglés, aunque esta tendencia se acentúa en la prensa especializada en comparación con la generalista. La realización de este análisis desde el punto de vista terminológico contribuye a la formación y profesión del traductor, puesto que supone una aportación para la adquisición de la competencia léxica traductora.

Palabras clave: compliance, corpus, terminología, prensa.

COMPLIANCE OR CUMPLIMIENTO NORMATIVO? A CORPUS STUDY WITH PROFESSIONAL AND DIDACTIC PURPOSES IN THE SPANISH PRESS Abstract: The aim of this paper is to analyze the use of the term “compliance”, considered one of the most popular topics in the legal-economic field in 2017 and 2018 in the Spanish press to determine whether there is a preference for its use in English, for its translation into Spanish or for both options. A corpus methodology has been used to study the frequency of use of the term in two generalist newspapers (El País and La Vanguardia) and two specialized ones (Expansión and El Economista), during the period between January 2017 and October 2018. The results show a preference for the use of the term in English, although this trend is accentuated in the specialized press in comparison with the general press. From a terminological perspective this analysis helps develop training and the profession of the translator since it is a contribution for the acquisition of the lexical competence necessary for professional translation.

Keywords: compliance, corpus, terminology, press.

1. INTRODUCCIÓN

Según un artículo publicado por el diario Expansión en junio de 2017, compliance es el término de moda entre los secretarios del consejo y los directores de las asesorías jurídicas de las empresas (Expansión, 2017). En octubre del 2018, el periódico El País también lo definió como el término de moda y añadía que su utilización en el ámbito jurídico responde al cambio de paradigma que se abre camino en la regulación de la actividad empresarial. Además, de acuerdo con una encuesta realizada por un grupo de expertos en este mismo año y publicada por esta misma fuente, el compliance se ha convertido en el tema jurídico del año. Aunque ya han pasado tres años desde la entrada en vigor de la última reforma del Código Penal (julio de 2015) y siete desde que se introdujo la responsabilidad penal de las empresas en el ordenamiento español, el cumplimiento normativo se ha convertido ahora en un tema de máxima actualidad (Expansión, 2017).

To cite this article: Alcalde Peñalver E., and Santamaría Urbieta, A. (2019). "Compliance or cumplimiento normativo? A corpus study with professional and didactic purposes in the Spanish press". Revista de Lingüística y Lenguas Aplicadas, 14, 19-27. https://doi.org/10.4995/rlyla.2019.10948 Correspondence authors: [email protected], [email protected]

Received: 2018-11-10 Accepted: 2019-04-09 | 19 Elena Alcalde Peñalver y Alexandra Santamaría Urbieta ¿Compliance o cumplimiento normativo? Estudio de corpus con fines profesionales y didácticos en la prensa española

A pesar de la recomendación de la Fundéu de utilizar su traducción “cumplimiento normativo” o simplemente “cumplimiento” como término utilizado para referirse a “los sistemas de detección de infracciones en las empresas” (Fundéu, 2014), es habitual observar el uso del término en inglés en la prensa española. La asociación española encargada de esta materia también utiliza el término en inglés (Asociación Española de Compliance) y su Libro Blanco sobre la función del Compliance (2017) hace uso en exclusiva de este.

El objetivo de este estudio es analizar el uso del término compliance en El País, La Vanguardia (periódicos de corte generalista), El Economista y Expansión (periódicos especializados en economía). Todos estos diarios se encuentran entre los más leídos en España, según indica el último Estudio General de Medios (EGM), que analiza el periodo comprendido entre octubre de 2017 y mayo de 2018 y se ha publicado recientemente. Abordamos el estudio desde un enfoque cualitativo, ya que lo que nos interesa es la observación del uso del término en su contexto más que cuantificarlo. Aunque, tal y como señala Sánchez Ramos (2017), el interés traductológico parece centrarse en la actualidad sobre todo a nivel textual y contextual, son las palabras las que constituyen el material del traductor y las que se consideran artífices de los textos (Anderman y Rogers, 1996). Además, tal y como señala esta misma autora, el componente léxico, sobre todo en ámbitos de especialidad, es de gran importancia para las tareas de traducción. Por ello, en este estudio nos centramos en realizar un análisis de corpus para observar el uso del término compliance con distintas combinaciones y obtener así datos sobre los contextos en los que se usa, si existen connotaciones en función del nivel de especialidad en el que se enmarca y a partir de ahí realizar una serie de aplicaciones prácticas que ayuden al buen uso del término por parte del traductor profesional y traductores en formación. Consideramos que la realización de análisis desde el punto de vista terminológico contribuye a la formación y profesión del traductor. Esto es debido a que hay casos en los que un término es de reciente aparición y no tiene un uso tan extendido en el ámbito general al circunscribirse a un área de especialidad. Por ello, es importante centrarnos en su uso real para la adquisición de la denominada competencia léxica traductora (Sánchez Ramos, 2013).

2. MARCO TEÓRICO

El término compliance puede incluirse tanto en el campo de la economía como en el derecho, ya que, aunque su origen sea jurídico por estar relacionado con el cumplimiento de una normativa, su ámbito de aplicación se centra en el mundo empresarial. El término compliance cuenta con una larga tradición en empresas de cultura anglosajona, ya que nació en EE. UU. en los años 70 y 80, cuando se produjeron grandes escándalos financieros que afectaron a empresas de gran influencia. Esto provocó que se dictara la Foreign Corrupt Practices Act o FCPA (1977), que incluyó una serie de especificaciones y prohibiciones en materia de sobornos y registros (Deloitte, 2018).

Tal y como indica Fernández (2018), compliance se puede resumir como una cultura para evitar riesgos y generar seguridad jurídica con el objetivo de mejorar las empresas, con reglas más transparentes para todos, una optimización de los procesos de compra a proveedores y un perfeccionamiento de los procedimientos y procesos internos. Su implantación implica un análisis y una evaluación muy exacta del funcionamiento de los departamentos corporativos para detectar ineficiencias, duplicidades e incoherencias y para desarrollar mecanismos de mejora.

Según este mismo autor, estas son algunas de las ventajas del compliance:

• Su adaptación en las empresas disminuye el coste de los seguros.

• Aumenta el valor de la compañía en procesos de venta y compra, ya que su implantación facilita los análisis de cumplimiento.

• Supone una ventaja competitiva para la licitación pública, sobre todo para proyectos europeos y firmar contratos con empresas internacionales.

• Aumenta la transparencia.

Si analizamos el término desde un punto de vista traductológico, podemos señalar que encontrar términos en inglés no es tan común en la terminología jurídica como en la empresarial. En efecto, tal y como indicamos en un estudio anterior (Alcalde Peñalver, 2015), la rapidez con la que se necesita transmitir la información financiera hace que la prensa española no se detenga a examinar en detalle el empleo de un término en concreto, por lo que en muchos casos aparece directamente en inglés, como ocurre con quantitative easing. Esto demuestra un dominio del inglés como lingua franca en el mundo empresarial. Para estos términos procedentes del inglés, la Fundéu suele reaccionar con bastante rapidez. Por ejemplo, para quantitative easing la Fundéu recomendaba el uso de “expansión cuantitativa”, aunque de nuevo la expresión en español tiene un uso mucho menos extendido en los

RLyLA Vol. 14 (2019), 19-27 | 20 Elena Alcalde Peñalver y Alexandra Santamaría Urbieta ¿Compliance o cumplimiento normativo? Estudio de corpus con fines profesionales y didácticos en la prensa española

medios de comunicación. Lo mismo ocurre para due diligence, para lo que la Fundéu recomienda “comprobaciones debidas”, aunque en raras ocasiones se observa este uso.

En este sentido, tal y como afirma Prieto-Ramos (2013), consideramos que el lenguaje no sería útil como instrumento comunicativo si no supiera aceptar las nuevas realidades que van surgiendo en distintos ámbitos. Por ello, el traductor tiene un papel destacado, ya que cuando se enfrente a términos que aún no tienen una opción léxica asentada, tendrá que intentar buscar una solución en función de las normas lingüísticas, el uso del término, la política de la institución en la que se trabaje o las preferencias del cliente. Por lo general, con el tiempo el término se acaba lexicalizando en español con sus correspondientes cambios ortográficos, pero, aun así, tanto entre expertos como en los medios de comunicación siguen existiendo excepciones como las antes mencionadas.

En cuanto a la terminología jurídica, tanto el inglés como el español destacan por el uso extendido de latinismos (ab initio, a posteriori, a priori, bona fide, corpus delicti, etc.), arcaísmos (otrosí, proveído, cualesquiera, futuro imperfecto de subjuntivo), dobletes y tripletes (seguido y conocido; se cita, llama y emplaza), expresiones fijas (con la venia, señoría), etc. En el caso del inglés, conviene mencionar el uso de términos propios del lenguaje jurídico francés, así como el uso de los verbos frasales con carácter técnico (parties enter into contracts, serve documents upon other parties, write off debts, etc.) (Alcaráz Varó y Hughes, 2009; Borja Albí, 2000).

No obstante, aunque como hemos señalado anteriormente el término compliance está a caballo entre la temática jurídica y económica, no podemos decir que su conocimiento se limite a su uso entre especialistas, ya que, tal y como observaremos a continuación en los resultados obtenidos tras el análisis de corpus, su uso se ha extendido a ámbitos en los que los usuarios no son totalmente expertos y aparece igualmente en prensa de corte generalista. En este sentido, tal y como señalaba Mayoral Asensio (2007:48), “as specialised knowledge has increasingly spread over the whole community, parcels of knowledge […] which were exclusive to specialists have become common ground for larger segments of population […]”. En efecto, los temas económicos han ido cobrando cada vez más importancia, sobre todo desde la crisis económica, lo que ha hecho que muchos de los temas que antes quedaban relegados a expertos en la materia sean ahora también frecuentes entre el público lego.

3. METODOLOGÍA

Para la metodología de nuestro estudio hemos adaptado la utilizada por García Luque en su artículo de 2010, en el que analizaba el uso del término “crisis” en la prensa española y francesa. No obstante, tal y como indicamos en la introducción, nuestro objetivo se limita al estudio del término en la prensa española para observar su frecuencia de aparición en distintas combinaciones y obtener datos para contribuir al correcto uso del término. Para ello, en primer lugar, hemos limitado el periodo de búsqueda al año 2017 y los diez primeros meses (enero-octubre) de 2018 para obtener información lo más reciente posible. Para los cuatro periódicos (El País, La Vanguardia, El Economista y Expansión) realizamos una búsqueda a través de la web del periódico de los artículos que incluyeran la palabra compliance. Asimismo, y para demostrar la frecuencia de uso durante este último año del término en la prensa española en comparación con años anteriores, llevamos a cabo una búsqueda entre el 1 de enero de 2010 y el 1 de enero del 2018. A continuación, se muestran los resultados de su frecuencia de uso durante el periodo en el que indicamos que estaba de moda (Expansión, 2017 y El País, 2018), en comparación con el periodo comprendido entre 2010 y 2018 en el que no era un tema tan recurrente.

Tabla 1: Comparativa de frecuencia de uso del término compliance en diferentes rangos de fechas.

1/1/2017 – 31/10/2018 1/1/2010 – 1/1/2018 El País 304 370 La Vanguardia 84 63 Expansión 460 450 El Economista 304 365

Estos datos dan muestra de cómo en un rango de fechas inferior la frecuencia de uso del término es mayor, lo que revela, tal y como se indicaba en los artículos de los periódicos citados, que el compliance ha sido en 2017 y 2018 un tema cada vez más recurrente en la prensa española. Los periódicos La Vanguardia y Expansión han aumentado incluso su frecuencia de uso. Esto justifica la necesidad de analizar por qué durante un periodo de ocho años el uso del término se asemeja, o incluso es menor, a su frecuencia en un periodo de búsqueda de un año y diez meses.

Para el periodo de duración inferior, en el caso de los periódicos de corte generalista, para El País la búsqueda arrojó un total de 304 coincidencias, mientras que para La Vanguardia el resultado fue de 84. En lo que respecta

RLyLA Vol. 14 (2019), 19-27 | 21 Elena Alcalde Peñalver y Alexandra Santamaría Urbieta ¿Compliance o cumplimiento normativo? Estudio de corpus con fines profesionales y didácticos en la prensa española

a los periódicos de corte económico, para Expansión obtuvimos un resultado de 460 coincidencias y para El Economista 304. Una vez identificados todos los resultados, procedimos a su análisis mediante la herramienta de análisis de corpus AntConc, que forma parte de un programa de software libre y sirve para realizar concordancias, clasificaciones de palabras, análisis del contexto y frecuencia de aparición. Las listas de frecuencia, herramienta en la que se basa este estudio, son de gran utilidad para identificar, tal y como apunta Hunston (2002:5),“possible differences between the corpora that can then be studied in more detail”.

Es necesario señalar que, a la hora de realizar la búsqueda, optamos también por utilizar el término “cumplimiento normativo” para determinar si el número de resultados variaba. En estos casos, o bien se trataba de artículos donde aparecía el término compliance, o bien versaban sobre un tema que nada tenía que ver con este ámbito.

En cuanto al corpus creado, se trata de lo que se denomina un corpus comparable monolingüe o corpus especializado en lengua meta (Rodríguez Inés, 2008). En el año 2002 Hunston establece una clasificación de ocho tipos de corpus entre los que se encuentra el corpus en el que se basa este estudio. Este tipo de corpus sirve para dar información sobre “unidades de significado” típicas de la lengua meta o de una sección especializada de la misma (Bernardini 2003). Entre sus ventajas, Aston (1999) menciona su facilidad de uso, su función informativa, ya que proporciona cifras de densidad léxica y repetitividad en los tipos de textos incluidos en el corpus, y su utilidad como medio para aprender sobre un área en la que el traductor necesita trabajar. Del mismo modo, Neubert (1992), Baker (1993) y Toury (1995) afirman que el corpus comparable constituye material fiable para la investigación de las características de los textos originales y, además, representa una fuente de información relevante para los traductores porque, en palabras de Corpas Pastor (1995:216), estas permiten a los investigadores observar y describir en detalle determinados patrones lingüísticos.

4. ANÁLISIS DE RESULTADOS

4.1. Diarios de corte generalista En primer lugar, analizaremos los resultados obtenidos en los dos diarios de corte generalista incluidos en el presente estudio.

En el diario El País, el término en inglés aparece en 304 ocasiones. En la gran mayoría de ellas, el término se incluye directamente en inglés [1] y únicamente hemos detectado 18 casos en los que aparece seguido de su equivalente en español, aunque cabe destacar que no se muestra homogeneidad en la traducción de este. Esto se debe a que, en tres ocasiones, se emplea “cumplimiento normativo” [2] y en quince se recurre únicamente a “cumplimiento” [3]. Cuando este diario se refiere al compliance officer solo en una ocasión parece seguido de su equivalente en español “director de cumplimiento normativo”.

Tabla 2: Ejemplos del uso del término compliance en el diario El País.

DIARIO El País [1] Tener un programa de compliance ha dejado de ser una moda para convertirse en una necesidad. (Febrero, 2018)

[2] Una de las piezas básicas sobre las que se asienta el buen funcionamiento del compliance (cumplimiento normativo) como método de combatir ciertas formas de delincuencia económica es el canal de denuncias. (Agosto, 2017)

[3] El término compliance (cumplimiento) está de moda. Su utilización en el ámbito jurídico responde al cambio de paradigma que se abre camino en la regulación de la actividad empresarial. (Octubre, 2018)

En lo que respecta al periódico La Vanguardia, la aparición del término es inferior a la del resto de los diarios analizados. Sin embargo, son más frecuentes las ocasiones (96) en las que se prefiere recurrir al término en español sin su equivalente en inglés [3]. En 84 ocasiones se emplea el término únicamente en inglés [1] y son 11 los casos en los que se acompaña el término en inglés con su equivalente en español [2]. En diez de las once ocasiones se opta por traducirlo por “cumplimiento normativo” y en una ocasión se recurre a equivalente español “cumplimiento”.

RLyLA Vol. 14 (2019), 19-27 | 22 Elena Alcalde Peñalver y Alexandra Santamaría Urbieta ¿Compliance o cumplimiento normativo? Estudio de corpus con fines profesionales y didácticos en la prensa española

Tabla 3: Ejemplos del uso del término compliance en el diario La Vanguardia.

DIARIO La Vanguardia [1] Así, y hasta la fecha, todas las resoluciones del Alto Tribunal han sido relativas a delitos cometidos por directivos, en la mayoría de los casos de empresas en las que no existe un modelo de Compliance. (Mayo, 2018)

[2] El Tribunal Supremo ha destacado en una sentencia la importancia de que las sociedades mercantiles implanten programas de cumplimiento normativo --conocidos como ‘compliance program’ -- para prevenir que sus trabajadores cometan delitos como estafa, apropiación indebida o administración desleal en el seno de las empresas. (Julio, 2018)

[3] La Comunidad de Madrid va a mejorar el procedimiento de concesión del sello de calidad Madrid Excelente al incorporar aspectos relativos al “cumplimiento normativo”, es decir, garantías de que se cumple el marco regulatorio. (Octubre, 2018)

4.2. Diarios de corte económico Al analizar el corpus de textos del diario Expansión, obtuvimos los siguientes resultados: El término se utiliza en 460 ocasiones en comparación con las 136 veces en las que aparece el término “cumplimiento normativo” en español. En seis ejemplos el término aparece entre comillas, como en “siete cosas sobre el ‘compliance’”. Solo en tres ocasiones se utiliza junto con el término en español: “política de cumplimiento normativo (compliance)”, “compliance (cumplimiento normativo)”, “por razones de compliance o cumplimiento normativo”. En tres ejemplos observamos también un uso combinado del término en español y en inglés: “han nombrado directores de cumplimiento normativo e incluso creado departamentos específicos para cumplir con los requisitos de compliance”; “ofrece servicios globales en materia de cumplimiento normativo, […] la práctica de compliance es una de las de mayor crecimiento de la firma. La figura del “compliance officer” se menciona tres veces para referirse a la función de “director de cumplimiento normativo”, traducción al español que solo acompaña una vez y entre paréntesis al término inglés. A continuación, se incluyen tres ejemplos que dan muestra de las tres tendencias principales de uso de este término en el diario Expansión: [1] su uso directamente en inglés, [2] su uso acompañado del equivalente en inglés y [3] su aparición en el periódico únicamente en español y sin su equivalente en inglés.

Tabla 4: Ejemplos del uso del término compliance en el diario Expansión.

DIARIO Expansión [1] Siete cosas sobre el ‘compliance’ que la mayoría de empresarios y CEO desconocen. (Mayo, 2018) [2] A lo largo de este mes de mayo, está previsto que se publique la norma española UNE 19601 que establece los requisitos para implantar un sistema de gestión de compliance o cumplimiento normativo penal. (Mayo, 2017) [3] El cumplimiento normativo se afianza en la cultura empresarial. (Enero, 2018)

Los resultados que arroja el análisis del diario El Economista son bastante similares a los observados y analizados en el periódico Expansión. El término se emplea un total de 304 veces únicamente en inglés [1], y solo en dos ocasiones se opta por incluirlo en inglés y en español [2]. Una de ellas es para referirse al “director de cumplimiento normativo” y otra incluyendo el término inglés “compliance program”. Por último, son 100 las veces en las que el periodista ha preferido el uso del término directamente en español [3].

Tabla 5: Ejemplos del uso del término compliance en el diario El Economista.

DIARIO El Economista [1] ¿Cuándo se regulará el ‘compliance’? (Agosto, 2017) [2] La figura del director de cumplimiento normativo o compliance officer es clave en todas las empresas, sea cual sea su tamaño, por la complejidad cada vez mayor del entorno normativo. (Diciembre, 2017) [3] Con estos programas de compliance, apunta el Tribunal Supremo, podrían evitarse estos delitos ad intra en el seno de las empresas para evitar una delincuencia societaria con grave perjuicio interno. (Agosto, 2017)

RLyLA Vol. 14 (2019), 19-27 | 23 Elena Alcalde Peñalver y Alexandra Santamaría Urbieta ¿Compliance o cumplimiento normativo? Estudio de corpus con fines profesionales y didácticos en la prensa española

A continuación, se incluye una tabla que muestra los resultados totales del uso del término objeto de estudio en los diarios mencionados.

Tabla 6: Uso del término compliance en la prensa digital española.

Usos El País Expansión La Vanguardia El Economista Apariciones del término en inglés. 304 460 84 304 Apariciones del término seguido del equivalente en español. 18 4 11 2 Aparición del término en español “cumplimiento normativo” 138 136 96 100 sin el término en inglés.

Aunque no se observan diferencias demasiado significativas entre los cuatro periódicos, sí detectamos una mayor recurrencia del término en inglés en los periódicos especializados y un menor uso de este seguido de su traducción al español. Esto se puede deber al mayor nivel de conocimiento especializado por parte de los lectores de estos periódicos, ya que, como comentaremos en el siguiente apartado, los términos en inglés suelen tener preferencia entre los usuarios expertos. Del mismo modo, la aparición del término en español sin su equivalente en inglés es más común en el diario El País y, en comparación con el número de veces que se emplea en inglés, el término en español se prefiere igualmente en el periódico La Vanguardia. Por lo tanto, en los periódicos de corte generalista existe una cierta preferencia por su uso en español, lo cual es previsible para el público que lee este tipo de prensa, menos especializado que los lectores de Expansión y El Economista.

5. DISCUSIÓN

Existen estudios previos que han hablado del uso de anglicismos en el campo económico-financiero. Aunque no son de reciente publicación, los recogemos a continuación para analizar algunas de las causas de su uso extendido cuando en ocasiones tienen un equivalente acuñado en español. Tal y como mencionamos anteriormente, es un hecho que, desde mediados del siglo XX el español es mucho más permeable a la inclusión de términos procedentes del inglés debido a la creciente globalización de la economía y la influencia de Estados Unidos a nivel mundial (Pizarro Sánchez 2010:116). Estas circunstancias hacen que en el lenguaje económico inglés se generen nuevos conceptos con sus correspondientes términos, que se exportan a diferentes lenguas y cuya adaptación no siempre es automática en la lengua meta, a pesar de que existan recomendaciones sobre su uso.

En este sentido, Alejo González (2000) realizó un estudio novedoso desde una perspectiva sociolingüística, teniendo en cuenta una muestra aleatoria de la población de economistas españoles entre los que distribuyó un cuestionario sobre el fenómeno de los anglicismos económicos en general. De los resultados obtenidos pudo concluir que la proliferación de los anglicismos en estos campos de especialización tiene que ver en gran medida con el hecho de ser una profesión en crecimiento y cuyos expertos buscan en los préstamos un reconocimiento de su nuevo estatus. Del mismo modo, Orts y Almela (2009:1070) afirmaron que el prestigio y el grado de especialización de los expertos ha impregnado el discurso económico en español de préstamos y adaptaciones de la lingua franca. En su estudio, distinguen entre neologismos de lujo y neologismos necesarios. Los primeros se incorporan por razones de prestigio, como es el caso de cash flow, hedge fundo call money, todos con un equivalente aceptado en español (flujo de caja, fondo de inversión libre y dividendo pasivo, respectivamente). Sin embargo, los expertos consideran el término en inglés más técnico o influyente. En el caso de los neologismos necesarios, el correspondiente en español no se encuentra tan ampliamente aceptado, como es el caso de joint venture o dumping. En el caso de compliance, podríamos afirmar, siguiendo lo indicado por los autores, que su uso en inglés sería lo que denominan neologismo de lujo, ya que, como hemos indicado, cuenta con equivalente acuñado en español.

Por otro lado, Ribas Pujol (1992:189) opina que los profesionales de los negocios se resisten a usar los equivalentes de los anglicismos por motivos más pragmáticos. De igual modo, Tapia Granados (2000) señala que, si la jerga económica es una de las más contaminadas por los ext ranjerismos, dentro de ella, la jerga financiera destaca por la profusión de términos y siglas incorporados directamente del inglés. Otros ejemplos relacionados son los que incluye en su estudio Pérez Berenguel (2003:621). Según afirma, entre los términos procedentes del inglés y que se han integrado definitivamente en el español encontramos algunos que, debido a su etimología latina, se han naturalizado con facilidad. Entre estos señala options (opciones), futures (futuro), bróker (corredor de bolsa), dumping (competencia desleal), leasing (alquiler con opción de compra), call (opción de compra), put (opción de venta). Por otro lado, Alejo González (1998) analiza los ejemplos de code switches en nuestra lengua, que mantienen una parte del término en inglés y otra en español, como es el caso de clubs de convergencia, efecto crossing-in, ejercicios de growth-accounting, modelo twin peaks, procesos de catch-up o análisis de shift-share.

RLyLA Vol. 14 (2019), 19-27 | 24 Elena Alcalde Peñalver y Alexandra Santamaría Urbieta ¿Compliance o cumplimiento normativo? Estudio de corpus con fines profesionales y didácticos en la prensa española

Siguiendo esta misma línea, Russo (2002) resalta en mayor medida, por su propia experiencia como revisora y traductora del FMI, el elevado grado de dinamismo de la terminología financiera y su consiguiente dependencia del inglés, lo que ilustra con numerosos ejemplos. Sin embargo, señala que la adopción de anglicismos representa una solución legítima desde el punto de vista práctico y morfológico, aunque en algunos casos resulte forzado desde el punto de vista semántico.

Por lo tanto, habrá que seguir observando la evolución del término compliance para analizar si su traducción al español termina contando con una mayor acogida entre los usuarios expertos, lo que a su vez tendría una repercusión en su uso entre el público lego y todas las publicaciones que surjan dirigidas a estos lectores.

6. APLICACIONES DIDÁCTICAS

En cuanto a las aplicaciones didácticas, coincidimos con las señaladas por García Luque en su artículo de 2010. Esta autora cita a Delisle (1988:208-210) para indicar que un ejercicio de este tipo de búsqueda de textos que contengan un término determinado sirve como acercamiento a la traducción jurídico-económica, la asimilación de terminología y fraseología de este ámbito, así como a la documentación a través de textos paralelos y fuentes fiables. Además, esto permite al alumno familiarizarse con textos con contenido híbrido, es decir, textos que pueden ser jurídicos, pero con contenido financiero, como era el caso de la mayoría de los archivos que añadimos para nuestro análisis. Esto está en consonancia con los resultados de un estudio previo de Alcalde Peñalver publicado en 2016 en el que se demostraba que, aunque siempre se habla de especialidades de traducción como si fueran compartimentos totalmente diferenciados, los textos híbridos que combinan contenidos del campo jurídico y económico son los más demandados y que por lo tanto habrá que hacer hincapié en la combinación de ambos en la formación de futuros traductores.

Por otro lado, esta mezcla de conceptos en inglés y posibles recomendaciones en español que no tienen un uso tan extendido puede resultar confusa y generar una falta de transparencia en la comprensión terminológica del ámbito económico y jurídico (Fernández Alonso y Alcalde Peñalver, 2017). Por lo tanto, extender el uso de una metodología de corpus en el aula de traducción especializada ofrece una gran posibilidad para que el alumno aprenda a recopilar de forma rápida información fiable y elaborar sus propios recursos documentarles y que le permitan enfrentarse con éxito a encargos de traducción. Una vez que realizan el corpus y observan el uso real de un determinado término, los alumnos pueden utilizar herramientas para la creación de bases de datos terminológicas como Multiterm e incorporar ejemplos de contexto. De hecho, los traductores deben cerciorarse de que el término no es únicamente válido para una comunidad específica, sino que además esta la emplea con frecuencia (Pimentel, 2017).

Por último, consideramos que llevar a cabo análisis de corpus de este tipo para determinados términos permite desarrollar en nuestros estudiantes las cuatro competencias señaladas por Cabré (1993) para cualquier trabajo terminológico: (1) competencia cognitiva, (2) competencia lingüística, (3) competencia sociofuncional y (4) competencia metodológica. En primer lugar, en relación con la competencia cognitiva, cabe señalar que la búsqueda y lectura de los textos del corpus para analizar el uso del término permite alcanzar un mayor conocimiento sobre el tema. Se trata de un tema de actualidad y es importante que el traductor que trabaja en este ámbito se mantenga al tanto de noticias relacionadas con el mismo. En relación con la segunda competencia, la lingüística, consideramos de gran importancia que los estudiantes sean conscientes de la evolución que puede tener el uso de un término proveniente del inglés y analizar si el equivalente acuñado en español termina por implantarse o, si, por el contrario, perdura el uso del anglicismo. Como hemos podido observar, las recomendaciones de la Fundéu apenas se tienen en cuenta en los ejemplos de los periódicos analizados, por lo que el traductor tendrá que conocer y discernir entre las distintas opciones que pueda utilizar en sus textos según el uso del término. En relación con la competencia sociofuncional, con este análisis hemos cumplido con los objetivos de observación de uso del término compliance en un medio determinado, como ha sido el caso de los diarios de corte generalista El País y La Vanguardia, por un lado, y los de corte económico Expansión y El Economista, por otro. Además, el corpus que ha conformado este estudio ha contado con un realismo máximo desde una perspectiva comunicativa, ya que se trata de textos que se utilizan en contextos financieros. Finalmente, la competencia metodológica establecida en el apartado de explicación de la metodología seguida nos ha permitido llevar a cabo nuestro estudio de forma rigurosa y de acuerdo con los objetivos que se perseguían.

7 CONCLUSIONES

En este artículo hemos analizado el uso del término compliance durante un periodo determinado mediante una metodología de corpus que nos ha permitido observar y describir su comportamiento lingüístico en periódicos de corte generalista y especializado. Los resultados nos muestran una mayor preferencia de su uso en inglés a pesar de las recomendaciones de la Fundéu de utilizar su equivalente en español. Estos resultados vienen justificados

RLyLA Vol. 14 (2019), 19-27 | 25 Elena Alcalde Peñalver y Alexandra Santamaría Urbieta ¿Compliance o cumplimiento normativo? Estudio de corpus con fines profesionales y didácticos en la prensa española

por la influencia del inglés en el mundo empresarial y el uso de anglicismos entre expertos de este ámbito en situaciones comunicativas en español. Como hemos señalado, el componente léxico, sobre todo en ámbitos de especialidad, es de gran importancia en la traducción, por lo que obtener datos de contextos reales de textos compilados en un corpus sirve de base para que el traductor en ejercicio y el traductor en formación puedan saber cómo proceder ante su traducción, sobre todo en casos en los que el término es novedoso y no cuenta con una gran expansión entre el público general. Igualmente, este tipo de metodología tiene una importante utilidad a nivel didáctico, ya que permite que el alumno elabore sus propios recursos con información fiable para términos que no van a aparecer actualizados en diccionarios en papel o en bases de datos. De esta forma, se contribuye a la adquisición de la competencia léxica traductora, lo que redunda en beneficio de la calidad de los numerosos textos que se traducen en un ámbito que cuenta hoy con tanta repercusión como es el tratado en este artículo.

BIBLIOGRAFÍA

Fernández Alonso, A. y Alcalde Peñalver, E. (2017). “Terminología y traducción para el mundo empresarial”. Formulación de los nuevos espacios docentes”, En Santos Martínez, C. J. (Ed.) Formulación de los nuevos espacios docentes. Madrid: Tecnos, 125-132. Alcalde Peñalver, E. (2016). La doble especialización en traducción jurídica y financiera: vínculos y necesidades formativas en España. MONTi. Monografía de Traducción e Interpretación, 8, 279-300. https://doi.org/10.6035/MonTI.2016.8.9 Alcalde Peñalver, E. (2015). La utilización del término quantitative easing en la prensa financiera española: el caso del diario Expansión. Lengcom, 8/3. Alcaraz Varó, E. y Hughes, B. (2009). El español jurídico (2.ª ed.). Barcelona: Ariel. Alejo González, R. (1998). “Code-Switching in Spanish Economic Discourse: Rethorical Strategies in Research Articles”, En Fortanet, I. et al. (Eds.) Genre Studies in English for Academic Purposes. Castelló de la Plana: Universitat Jaume I, 227-252. Alejo González, R. (2000). «Las actitudes lingüísticas de los economistas y su influencia en el uso de anglicismos en la terminología económica castellana». VII Simposio Iberoamericano de Terminología Riterm. Disponible en http://riterm.net [Último acceso: 23/4/12]. Anderman, G. y Rogers, M. (Eds.). (1996). Words, words, words. The translator and the language learner. Clevendon: Multilingual Matters. Asociación Española de Compliance (2017). Libro blanco sobre la función de Compliance. Disponible en https:// www.asociacioncompliance.com/new2017/wp-content/uploads/2017/08/Libro-Blanco-Compliance- ASCOM.pdf [Último acceso: 10/11/2018]. Aston, G. (1999). Corpora and language teaching. En: L. Burnard; T. McEnery (eds). Rethinking language pedagogy from a corpus perspective. Bern: Peter Lang. 7-17. Baker, M. (1993). “Corpus Linguistics and Translation Studies: Implications and Applications”, en: Baker, M., Francis, G. & Tognini-Bonelli, E. (eds.) Text and Technology: in Honour of John Sinclair, Amsterdam/Philadelphia, John Benjamins, 233-252. https://doi.org/10.1075/z.64.15bak Baskerville, R. y Evans, l. (2011). The darkening glass: Issues for translation of IFRS. Edimburgo: The Institute of Chartered Accountants of Scotland. Bernardini, S. (2003). “Designing a corpus for translation and language teaching: The CEXI experience”, En TESOL Quarterly (Special issue on Corpus Linguistics in TESOL), 37(3), 528-537. https://doi.org/10.2307/3588403 Borja Albi, A. (2000). El texto jurídico inglés y su traducción al español. Barcelona: Ariel. Cabré, M. T. (1993). La terminología: teoría, metodología, aplicaciones. Barcelona: Antártida/Empúries. Corpas Pastor, G. (1995). “The Role of Text Analysis in Corpus-Based Translation”, En Valero Garces, C. (ed.) Cultura sin fronteras. Encuentros en torno a la traducción. Madrid: Universidad de Alcalá de Henares, 215-222. Delisle, J. (1988). “L’initiation à la traduction économique”. Meta, 33/2, 204-2015. https://doi.org/10.7202/002515ar Deloitte (2018). La función de compliance en la prensa española. Disponible en https://www2.deloitte.com/es/ es/pages/governance-risk-and-compliance/articles/la-funcion-compliance-en-la-empresa.html [Último acceso: 10/11/2018]. El Mundo. (2016). EL MUNDO se acerca a ‘El País’ y supera a ‘ABC’ y ‘La Razón’ juntos. El Mundo. Disponible en: http://www.elmundo.es/television/2016/07/01/5775614e46163fe2038b461f.html [Último acceso: 10/11/2018]. Estudio General de Medios. (2018). Ránking de medios. Disponible en https://www.aimc.es/a1mc-c0nt3nt/ uploads/2018/06/resumegm218.pdf [Último acceso: 10/11/2018].

RLyLA Vol. 14 (2019), 19-27 | 26 Elena Alcalde Peñalver y Alexandra Santamaría Urbieta ¿Compliance o cumplimiento normativo? Estudio de corpus con fines profesionales y didácticos en la prensa española

Expansión. (2017). El ‘compliance’, tema jurídico del año. Expansión, p. 32. Disponible en https://www. globalcorporationcenter.com/wp-content/uploads/sites/157/2015/10/EXP27JNMAD-Nacional-Jurídico- pag-32.pdf [Último acceso: 10/11/2018]. Fernández, F. J. (2018). Siete cosas sobre el ‘compliance’ que la mayoría de empresarios y CEO desconocen. Disponible en http://www.expansion.com/juridico/opinion/2018/05/18/5afd9f74e5fdea984a8b456d.html [Último acceso: 10/11/2018] Fundéu. (2014). Cumplimiento normativo, mejor que compliance. Disponible en http://www.fundeu.es/ recomendacion/cumplimiento-normativo-observancia-compliance-officer-responsable/ [Último acceso: 10.11.2018] García Luque, F. (2010). “La palabra crisis en la prensa: análisis terminológico de cara a la enseñanza y al ejercicio profesional de la traducción socio-económica (francés-español)”. Anales de Filología Francesa, 18, 203-2015. Hunston, S. (2002). Corpora in Applied Linguistics. Cambridge: Cambridge University Press. https://doi. org/10.1017/CBO9781139524773 Mayoral Asensio, R. (2007b). “Specialised translation: a concept in need of revision”, Babel, 53/1, 48-55. https://doi.org/10.1075/babel.53.1.05may Neubert, A. (1992). “Lingüística del texto y traducción”, Sendebar, 3, 13-24. Orts, M. A. y Almela, A. (2009). “Translating the Spanish economic discourse of the crisis: Dealing with the inevitability of English loanwords”. IJES (International Journal of English Studies): Recent and Applied Corpus-Based Studies, 133-157. Pérez Berenguel, J. F. (2003). “Glosario de errores comunes en la traducción económica y financiera”, En Muñoz Martín, R. (Ed.) I AIETI. Actas del I Congreso Internacional de la Asociación Ibérica de Estudios de Traducción e Interpretación, Granada 12-14 de Febrero de 2003. Granada: AIETI, 619-628. Pimentel, J. (2017). “Specialized verbs and specialized uses of verbs in a comparable corpus of judgments produced in Canada, Portugal and Brazil”. En Drouin, P.; Francoeur, A.; Humbley, J. y Picton.A (Ed.) Multiple Perspectives on Terminological Variation, John Benjamins, 109-120. https://doi.org/10.1075/tlrp.18.05pim Pizarro Sánchez, I. (2010). Análisis y traducción del texto económico. La Coruña: Netbiblio. Prieto Ramos, F. (2013). Traducción institucional y (co)gestión de neologismos: entre la armonización y la congestión terminológicas. En C. Sinner (coord.), Comunicación y transmisión del saber entre lenguas y culturas (pp. 387-40), España: Peniope. Ribas Pujol, A. (1992). “Las innovaciones financieras y su traducción al español”. En Edo Julià, M. (Ed.)Actes del I Congrès Internacional sobre Traducciò. Barcelona: Universitat Autònoma de Barcelona, 183-190. Rodríguez Inés, P. (2008). Uso de corpus electrónicos en la formación de traductores (inglés-español-inglés). Tesis doctoral. Barcelona: Universidad Autónoma de Barcelona. Russo, A. (2002). “Préstamos del inglés en la jerga económica: ¿es posible evitar el sobreendeudamiento lingüístico?”, En Hernúñez, P. y González, L. (Coords.) El español, lengua de traducción: Actas del I Congreso Internacional. Luxemburgo: Comisión Europea. Sánchez Ramos, M. M. (2013). “El léxico en el aula de traducción: diseño de un modelo de adquisición de la competencia léxica traductora (inglés-español)”, Tonos. Revista de Estudios Filológicos, 24. Sánchez Ramos, M. M. (2017). “Metodología de corpus y formación en la traducción especializada (inglés-español): una propuesta para la mejora de la adquisición de vocabulario especializado”. Revista de Lingüística y Lenguas Aplicadas, 12, 137-150. https://doi.org/10.4995/rlyla.2017.6969 Tapia Granados, J. A. (2000). Sobre capital-riesgo y otros aspectos de la jerga económica. Puntoycoma, 64. Disponible en http://ec.europa.eu/translation/bulletins/puntoycoma/64/pyc64.pdf [Último acceso: 14/5/2013]. Toury, G. (1995). Descriptive Translation Studies and Beyond. Amsterdam: John Benjamins. https://doi.org/10.1075/ btl.4

RLyLA Vol. 14 (2019), 19-27 | 27 Revista de Lingüística y Lenguas Aplicadas Vol. 14 año 2019, 29-43 EISSN 1886-6298 https://doi.org/10.4995/rlyla.2019.10680

ACTITUDES Y COMPETENCIAS DOCENTES EN PROFESORES DE ESPAÑOL DE LOS INSTITUTOS FEDERAIS (BRASIL) EN RELACIÓN CON LAS VARIEDADES DE LA LENGUA

Bruno Rafael Costa Venâncio da Silva Instituto Federal do Rio Grande do Norte, Brasil María Antonieta Andión Herrero Universidad Nacional de Educación a Distancia, España

Resumen: Con el interés de estudiar la realidad de la enseñanza de ELE en Brasil y las variedades del español, 25 profesores de los Institutos Federais de los estados de Rio Grande do Norte y Roraima responden un exhaustivo cuestionario. A través de sus respuestas, definimos sus perfiles sociolingüísticos en relación con su formación, prejuicios, creencias y actuaciones frente a la variación y variedad de la lengua española, tanto de manera general como en el aula. El análisis de la información recogida muestra un complejo panorama de carencias formativas, sensibilidades, actitudes, acciones docentes y necesidades que recomiendan una seria reflexión y consecuente implicación de las autoridades académicas, así como de la colaboración de actores diversos del ámbito de ELE.

Palabras clave: variedades del español, profesores de español de Institutos Federais, español como lengua extranjera, creencias y actitudes, prejuicio lingüístico, Brasil.

ATTITUDES AND TEACHING COMPETENCIES IN TEACHERS OF SPANISH OF THE INSTITUTOS FEDERAIS (BRAZIL) IN RELATION TO THE VARIETIES OF THE LANGUAGE Abstract: With the intention of studying the reality of the teaching of Spanish as a foreign language in Brazil and its different varieties, 25 teachers who work at the Institutos Federais from the states of Rio Grande do Norte and Roraima have answered a thorough questionnaire. Thanks to their answers we can, both generally speaking and as teachers, define their sociolinguistic profiles in relation to their training, prejudices, beliefs and attitudes towards variation and varieties within the Spanish language. After analyzing all the collected information, we have come to observed a complex landscape of training deficits, sensitivities, attitudes, teaching approaches and needs that make necessary a profound reflection and the consequent involvement of the academic authorities at issue, as well as the cooperation of other actors involved in the realm of Spanish-FL.

Keywords: Spanish varieties, Spanish teachers at Institutos Federais, Spanish as a foreign language, attitudes and beliefs, linguistic prejudice, Brazil.

1. INTRODUCCIÓN

La importancia del español como lengua de comunicación internacional está avalada por su impactante demografía y extensión territorial. Aproximadamente 572 millones de personas lo hablan, ya sea como lengua materna, segunda (L2) o extranjera (LE) (Instituto Cervantes, 2017:5). Los hispanohablantes nos repartimos por 11 999 000 km² de superficie —casi el 9%— de la tierra, que abarca los cinco continentes (en Europa: España, Principado de Andorra; América: Hispanoamérica y los Estados Unidos de Norteamérica; en África: Guinea Ecuatorial, Sahara Occidental y los territorios españoles de Islas Canarias, Ceuta y Melilla; en Asia: Filipinas y sefardíes en Israel y Turquía; y en Oceanía: isla de Pascua).

Estudian español más de 21 millones de personas en diversos países del mundo (Ibídem:15). Entre ellos, Brasil es el segundo con mayor número de alumnos (por detrás de los Estados Unidos de América); más de seis

To cite this article: Silva, B.R.C.V. da, and Andion, M.A. (2019). "Attitudes and teaching competencies in teachers of spanish of the Institutos Federais (Brazil) in relation to the varieties of the language". Revista de Lingüística y Lenguas Aplicadas, 14, 29-43. https://doi.org/10.4995/rlyla.2019.10680 Correspondence authors: [email protected], [email protected]

Received: 2018-09-04 Accepted: 2019-03-27 | 29 Bruno Rafael Costa Venâncio da Silva y María Antonieta Andión Herrero Actitudes y competencias docentes en profesores de español de los Institutos Federais (Brasil) en relación con las variedades de la lengua

millones de brasileños estudian español en el Instituto Cervantes o en colegios públicos y privados de primaria y secundaria (Ibídem:14). Y su demanda sigue creciendo; según estimaciones del Gobierno, unos 30 millones de brasileños hablarán español como L2 en tan solo una década (Ibídem:15). Por todo ello, Brasil es ahora mismo una zona de interés estratégico en el ámbito de ELE.

Pero el español, por su extensión y pasado colonial, tiene una complejidad añadida que para Brasil, situado en medio de Hispanoamérica, resulta evidente: en sus comunidades florece el bilingüismo y el español recoge una diversidad lingüística1 con grados diferentes de relevancia. Y precisamente en la atención a la variación dentro de la enseñanza del español como lengua extranjera (ELE) se sitúa el interés de nuestro estudio, en el que partimos de la identificación del perfil lingüístico y sociolingüístico de 25 docentes de dosInstitutos Federais (IF) del nordeste brasileño para relacionarlos con el análisis de su formación académica, sus prejuicios, creencias, actitudes y actuaciones frente a la variación y variedad del español.

Centramos nuestra atención en este escenario, donde los docentes adquieren responsabilidades de transmitir principios y criterios para educar en actitudes de reconocimiento y respeto hacia las variedades del español, entendidas como el “conjunto de rasgos lingüísticos propios de una comunidad con validez geográfica determinada y que marcan el acento de sus hablantes” (Andión, 2007:24).

A través de un cuestionario con 28 preguntas abiertas y cerradas, obtuvimos la información necesaria de los docentes de los IF, para descubrir su edad, sexo/género, formación (grado y posgrado), experiencia docente, identidad dialectal nativa o adoptiva (en español), prejuicios lingüísticos, estudios sobre variación y variedad del español, (in)seguridades y preocupaciones en relación con estos temas en su ejercicio docente, así como expectativas y necesidades.

2. EL INSTITUTO FEDERAL COMO CONTEXTO DE LA INVESTIGACIÓN

En Brasil, a partir de la Ley N.º 11.892, de 29 de diciembre de 2008, se crean la Rede Federal de Educação Profissional, Científica e Tecnológica y los Institutos Federais de Educação, Ciência e Tecnologia en todos los estados, definidos en el Artículo 2.º como:

[…] instituições de educação superior, básica e profissional, pluricurriculares e multicampi, especializados na oferta de educação profissional e tecnológica nas diferentes modalidades de ensino, com base na conjugação de conhecimentos técnicos e tecnológicos com as suas práticas pedagógicas, nos termos desta Lei.

Al contar con un número amplio de niveles educativos, el Instituto Federal ofrece diversas posibilidades de estudio. Entre sus objetivos (Artículo 7.º), dos son de especial interés en la caracterización de sus docentes. El primero, “ministrar educação profissional técnica de nível médio, prioritariamente na forma de cursos integrados, para os concluintes do ensino fundamental e para o público da educação de jovens e adultos” (Ley N.º 11.892, 2008), ofrece una enseñanza secundaria con un currículo y asignaturas más o menos comunes para todo Brasil y una formación técnica en un área profesional.

Por su parte, el segundo objetivo, “ministrar em nível de educação superior”, cuenta con algunas especificidades interesantes:

a. cursos superiores de tecnologia visando à formação de profissionais para os diferentes setores da economia; b. cursos de licenciatura, bem como programas especiais de formação pedagógica, com vistas na formação de professores para a educação básica, sobretudo nas áreas de ciências e matemática, e para a educação profissional; c. cursos de bacharelado e engenharia, visando à formação de profissionais para os diferentes setores da economia e áreas do conhecimento; d. cursos de pós-graduação lato sensu de aperfeiçoamento e especialização, visando à formação de especialistas nas diferentes áreas do conhecimento; e e. cursos de pós-graduação stricto sensu de mestrado e doutorado, que contribuam para promover o estabelecimento de bases sólidas em educação, ciência e tecnologia, com vistas no processo de geração e inovação tecnológica (Ley N.º 11.892, 2008).

Así, muchos docentes de los Institutos Federais (IF) tienen un doble perfil: imparten enseñanza secundaria (general y técnica) y superior (grado y posgrado). Los profesores informantes de nuestra investigación, del Instituto

1 La variación lingüística está determinada por: (a) factores sociales como los diferentes sistemas de organización política, económica, social y geográfica; (b) factores individuales (idiolecto) como la edad, la raza, el sexo o el nivel de educación; (c) aspectos históricos, étnicos y culturales; y (d) la situación inmediata que rodea la inte- racción (contexto externo) (Corvalán y Arias, 2001).

RLyLA Vol. 14 (2019), 29-43 | 30 Bruno Rafael Costa Venâncio da Silva y María Antonieta Andión Herrero Actitudes y competencias docentes en profesores de español de los Institutos Federais (Brasil) en relación con las variedades de la lengua

Federal de Educação, Ciência e Tecnologia do Rio Grande do Norte (IFRN) y del Instituto Federal de Educação, Ciência e Tecnologia de Roraima (IFRR), nos interesaron por este doble perfil docente en una carrera de grado dirigida a la formación de profesores de Lengua española (modalidades presencial/a distancia).

Es importante destacar que desde 2006, cuando eran Centros Federais de Educação Tecnológica (CEFET), los IF de estos dos estados imparten carreras de grado para la formación del profesorado en Lengua española. Así, el IFRN de la ciudad de Natal, capital de Rio Grande do Norte, ofreció el primer grado en Lengua española, que luego replicó la Universidade Federal do Rio Grande do Norte (UFRN).

A partir de 2010, el IFRN también cuenta con un grado en Lengua española (modalidad a distancia), un paso fundamental para la expansión de profesionales en esta materia por el interior del estado y para que los profesores de español de los campus del interior puedan actuar en esta carrera.

2. NATURALEZA DEL ESTUDIO Y PREGUNTAS DE INVESTIGACIÓN

Nuestra investigación es descriptiva al contemplar la “descrição do objeto por meio da observação e do levantamento de dados” (Barros y Lehfeld, 1990:34); también porque sus objetivos especifican propiedades, características y perfiles de personas y grupos, y pretenden recolectar información sobre conceptos o variables y cómo se relacionan (Sampieri et al., 2013:102). Por otra parte, podemos calificarla de cuantitativa y cualitativa ya que es difícil entender ambos aspectos como dicotómicos (Moreira y Caleffe, 2008). Aunque en las humanidades lo cualitativo suele ser lo más frecuente, para la descripción e interpretación del fenómeno estudiado, incluimos datos cuantitativos que dan soporte al análisis. Asimismo, la investigación es no experimental por describir y explicar hechos y situaciones, evaluarlos y proponer innovaciones (Ibídem:73).

Abordamos a nuestros sujetos de estudio, los profesores de IFRN e IFRR, con el objetivo de responder a cuatro preguntas de investigación, con diferentes grados de complejidad y aspectos relacionados:

1. ¿Cuál es el perfil sociolingüístico de los profesores de los IF en cuanto asexo/género, edad, formación y experiencia docente?

2. ¿Cuál es su perfil lingüístico?, ¿son hablantes nativos?, ¿qué variedades tienen?

3. ¿Cómo se han visto afectados por el prejuicio lingüístico?, ¿cuáles son sus creencias y actitudes hacia las variedades del español?

4. ¿Qué formación han recibido sobre este tema?, ¿se sienten preparados para impartir contenidos de variación?, ¿lo hacen?

3. OBJETIVO Y METODOLOGÍA

Nuestro principal objetivo era responder a las preguntas anteriores y centrarnos en conocer la realidad formativa y docente de los profesores en relación con la variación y variedad del español, para así poder analizarla y orientar posibles respuestas a sus necesidades.

Para recoger los datos necesarios, elaboramos un cuestionario con preguntas especialmente orientadas hacia nuestro interés investigativo, que permitían obtener información a través de respuestas sin ayuda del investigador (Gressler, 2003:153). Respondieron al cuestionario 25 profesores del IFRN y del IFRR entre julio y agosto de 20162. Estos participantes fueron suficientes para el análisis ya que hay 32 profesores de español entre ambos IF, por lo que contamos con casi el 80% de los docentes de esta materia.

Tomando como base el cuestionario de Andión y Gil (2013) sobre el componente “variedades” como parte de la competencia docente, aplicado a profesores de ELE en centros del Instituto Cervantes del mundo, elaboramos un modelo adaptado al perfil de los encuestados.

El cuestionario contiene 28 preguntas, de respuesta cerrada y abierta3. A pesar de la riqueza que permiten estas últimas, optamos por más preguntas cerradas (21) para garantizar respuestas orientadas a los objetivos

2 Primero aplicamos un piloto a cuatro profesores del IFRN. 3 Las “cerradas” ofrecen alternativas, generalmente “sí” o “no”, adverbios de frecuencia (siempre, a veces, nunca) u otras opciones. En las “abiertas” se puede responder de forma más personal, espontánea y libre.

RLyLA Vol. 14 (2019), 29-43 | 31 Bruno Rafael Costa Venâncio da Silva y María Antonieta Andión Herrero Actitudes y competencias docentes en profesores de español de los Institutos Federais (Brasil) en relación con las variedades de la lengua

de la investigación y ofrecer dinamismo y agilidad. También incluimos preguntas abiertas (7) no obligatorias que permitían a los docentes-informantes añadir comentarios pertinentes4.

Dicho instrumento tiene tres partes: 1. Perfil sociolingüístico, que persigue identificar género, edad, formación, experiencia como profesor de español (niveles impartidos en los IF); 2. Perfil lingüístico, que identifica su procedencia y variedad lingüística (en español, nativa o no) y sus actitudes hacia el prestigio dialectal; y 3. Creencias y actitudes hacia la variación lingüística dentro la enseñanza del español, que permite conocer su formación, experiencia y conocimiento del componente “variedades”.

La plataforma Google Forms nos permitió enviar el cuestionario y recibir las respuestas, recoger y tratar los datos numéricos, así como elaborar los gráficos.

4. MARCO TEÓRICO

Nuestro marco teórico parte de las investigaciones realizadas en dos áreas, la variación y la enseñanza de lenguas, especialmente aquellas que atienden el componente “variedades” en diferentes ámbitos de la enseñanza de ELE, como el discurso de los profesores, el aprendizaje intercultural, el prejuicio lingüístico, etc. La información aportada por estos antecedentes nos ha permitido orientar las preguntas del instrumento de recogida de datos (cuestionario) y fijar las referencias bibliográficas para el análisis de los resultados.

Al ser el español una lengua de grandes proporciones, cuya mayor presencia está en el continente americano y al no ser este español “una variedad de la lengua, sino un conjunto de variedades diferentes entre sí y con una personalidad bien marcada y forjada a golpes de historia y geografía” (Moreno, 2010:48), su diversidad es una realidad innegable también para su enseñanza-aprendizaje como LE/L2.

En el cuestionario, abordamos la diversidad desde la propia identificación que pedimos a los docentes de su variedad del español. Los intentos por zonificar el español en geolectos o variedades de acuerdo con sus características fonético-fonológicas, gramaticales y léxicas han sido varios, aunque ninguno totalmente satisfactorio pues siempre prevalecen ciertos factores sobre otros (Araújo, 2013). Conscientes de ello y de que nuestro interés se centra en la enseñanza del español en Brasil, ofrecimos a los encuestados un catálogo de variedades apoyado en la propuesta de zonificación de Moreno (2000, 2010) y de Andión y Casado (2014), que para el español americano son versiones actualizadas de la que hiciera Henríquez Ureña (1921, apud Moreno, 2000)5. Así, el español americano estaría dividido en:

A.1. un área caribeña (representada, por ejemplo, por los usos de San Juan de Puerto Rico, La Habana o Santo Domingo), A.2. un área mexicana y centroamericana (representada, por ejemplo, por los usos de la ciudad de México y otras ciudades y territorios significativos), A.3. un área andina (representada, por ejemplo, por los usos de Bogotá, La Paz o Lima, A.4. un área rioplatense y del Chaco (representada por los usos de Buenos Aires, de Montevideo o de Asunción) y A.5. un área chilena (representada por los usos de Santiago) (Moreno, 2000:38).

Para la cuestión de las actitudes, debemos partir de que Brasil también posee una diversidad lingüística hacia cuyos acentos los brasileños manifiestan diferentes valoraciones sociolingüísticas. Aunque Bagno (2015:22) reconoce que la sociedad brasileña milita contra los prejuicios en general, “não tem atingido um tipo de preconceito muito comum na sociedade brasileira: o preconceito lingüístico”. Es tradicional que la escuela “age como se a língua culta fosse estática, pronta, inabalavelmente infensa a seu uso nos processos interlocutivos” (Geraldi, 1984:59). Esta postura ha alimentado estigmas hacia variantes y variedades del portugués, que se trasladan a las lenguas extranjeras aunque se encuentren fuera de sus comunidades lingüísticas (López Morales, 1989).

Es esperable, entonces, que Andión (2013) haya documentado en un estudio sobre variedades lingüísticas, identidad y enseñanza de ELE, la predilección de los docentes brasileños hacia la variedad castellana. Resulta paradójico, pues:

[E]n el caso de Italia o Portugal, la cercanía geográfica con España puede justificar la selección de la variedad castellana como modelo; parece, sin embargo, menos esperable en países como Brasil, cuyo contexto geográfico es hispanoamericano (Andión, 2013:168).

Ya Irala (2004) había reconocido, en un trabajo con profesores activos y en formación de una ciudad de Rio Grande do Sul, que la “pureza” y “corrección” eran habitualmente asociadas al español europeo. Aun siendo

4 Puede consultarse el cuestionario en el siguiente enlace: https://academicoifrnedu-my.sharepoint.com/:b:/g/personal/bruno_venancio_academico_ifrn_edu_br/ EVDbC_vfb_5GrVPKPZufo10B1YLoATa1abxghBIS7lT7FA?e=Fj92Wj. 5 Sus principales áreas coinciden con las grandes lenguas indígenas: náhualt (México), maya (Centroamérica), quechua (Andes), mapuche (Chile) y tupí-guaraní (Chaco y Río de la Plata).

RLyLA Vol. 14 (2019), 29-43 | 32 Bruno Rafael Costa Venâncio da Silva y María Antonieta Andión Herrero Actitudes y competencias docentes en profesores de español de los Institutos Federais (Brasil) en relación con las variedades de la lengua

de una ciudad fronteriza con Uruguay, los entrevistados no consideraron que el español austral o rioplatense fuera, en principio, el más adecuado a las necesidades de sus alumnos. No es extraño, entonces, que en nuestra investigación dos de los tres hablantes nativos rioplatenses confesaran haber sido víctimas de prejuicios lingüísticos.

Son las esperables consecuencias del lectocentrismo6, que “suele manifestarse más comúnmente entre geolectos de una misma lengua”, por lo que “la variedad favorecida termina calificada como ‘pura’ o ‘correcta’, mientras que las otras se consideran desviaciones” (Andión, 2017:133). En el mundo hispánico, las más beneficiadas por estos juicios extralingüísticos son “en España, la variedad castellana (septentrional peninsular), y en Hispanoamérica, la anterior y la colombiana” (Ídem:138). La enseñanza de ELE muestra también el impacto de estas creencias: “tras repasar el análisis de más de cuarenta manuales de ELE/L2, españoles y de otra factura, vemos que no son pocos los que responden, hasta en contextos poco justificados como el de Brasil, a un modelo lectocentrista” (Ídem).

En el mismo sentido, Zolin-Vesz (2013) reflexiona sobre el español como “lengua de encuentro”7, pues en su investigación sobre creencias en la escuela, España es el único lugar que se asocia a la lengua española. Al observar el discurso en clase de una profesora de español, el investigador constata que:

O relato da professora, certamente, sintetiza muito das aulas de espanhol que parecem dar as costas para outros países do mundo hispânico que não a Espanha. Se, por um lado, o livro didático de espanhol tem contribuído para reforçar a crença de superioridade da variedade europeia da língua, por outro os professores de espanhol parecem ter ‘comprado a ideia’ da superioridade da Espanha como a primeira e mesmo única opção para exemplificar aspectos culturais e mesmo estruturais da língua espanhola (Zolin-Vesz, 2013:59-60).

Es innegable la responsabilidad de los docentes en la percepción que los alumnos tengan sobre las sociedades hispanohablantes, y que las asocien o no con relaciones de poder que ocultan la diversidad lingüística. Afirmamos que esta debe tener un lugar en el aprendizaje de una lengua extranjera, en el que también debe encontrar un espacio la diversidad cultural que la acompaña.

Para alcanzar esa meta, nos apoyamos en el aprendizaje intercultural pues creemos que “para que a língua seja um bom instrumento é preciso considerá-la muito mais do que mero instrumento. Ela é a matéria-prima da constituição identitária” (Serrani, 2005:29). Un profesor de lenguas debe ser un agente intercultural y plantearse los siguientes objetivos:

a. estimular nos alunos o estabelecimento de pontes culturais com outras sociedades e culturas;

b. propiciar a educação à diversidade sócio-cultural e ao questionamento de etnocentrismos e exotismos;

c. dar ao componente cultural um peso significativo no planejamento de cursos de línguas (Serrani, 2005:22).

En este acercamiento de los aprendices a las diferentes manifestaciones de la cultura hispánica —y por ende, a su diversidad lingüística—, los recursos a través de Internet son inagotables. Lógicamente, es importante saber elegirlos de acuerdo con los objetivos que se desean alcanzar; para ello, el profesor necesita cierta formación en las variedades lingüísticas del español y su componente cultural.

5. CARACTERIZACIÓN DE LA MUESTRA, RESULTADOS Y REFLEXIONES

5.1. Perfil sociolingüístico de los profesores En cuanto al perfil sociolingüístico de los sujetos, en lo referido al sexo/género, la mayoría aplastante es femenina (18 encuestados, 72%), situación común en los grados de formación de profesores de Brasil, tanto de lengua materna (portugués) como extranjera. Los hombres son un 28% de la muestra (7). Esta casuística es importante porque las mujeres pueden marcar el comportamiento de los resultados ya que suelen estar más atentas al prestigio lingüístico y tienden al uso de rasgos normativos y conservadores (García Mouton, 2000:52).

Además, en los estudios de disponibilidad léxica, por ejemplo, “la producción léxica de las mujeres es mayor que la de los hombres, sobre todo en áreas temáticas de la ropa, las comidas y bebidas, la cocina, la escuela, los animales y las profesiones y oficios (Gómez Molina y Gómez Devis,apud Moreno, 2009:42).

6 Andión (2017:133) define este fenómeno glotopolítico como “la creencia de superioridad de una lengua o variedad (geolecto) sobre otras”. 7 Para Zolin-Vesz (2013:55), es “um instrumento de comunicação que possibilita o diálogo e a convivência harmônica próprios de uma pátria comum”.

RLyLA Vol. 14 (2019), 29-43 | 33 Bruno Rafael Costa Venâncio da Silva y María Antonieta Andión Herrero Actitudes y competencias docentes en profesores de español de los Institutos Federais (Brasil) en relación con las variedades de la lengua

En cuanto a la edad, la mitad de los encuestados (12) está en la franja “entre 31 y 40 años” (48%); ocho docentes tienen “entre 41 y 60 años” (32%) y cinco, “entre 21 y 30 años” (20%)8. La ausencia de menores de 21 años hace previsible que los docentes tengan hábitos lingüísticos maduros y suficiente consciencia lingüística, aspecto facilitado por su propia profesión.

En cuanto a la formación (Gráfico 1), el grado enLetras com Habilitação dupla (Português/Espanhol) fue la carrera más frecuente (9 encuestados, 36%). Conviene aclarar que, en Brasil, las carreras de grado en Letras pueden ser de habilitação única, que permite impartir solamente una lengua —por lo general, el portugués—, o habilitação dupla, que añade una lengua extranjera9.

Gráfico 1:Carrera de grado de los docentes de Español de los IF.

La mayoría ha cursado grados solo en Lengua española, a saber, Letras com Habilitação em Língua Espanhola e suas respectivas literaturas (8 encuestados, 32%), Licenciatura em Espanhol/Língua Espanhola (5, 20%), ofrecida exclusivamente por los IF, y Letras Espanhol (1, 4%). Además, hay dos sujetos con otros grados: Letras com Habilitação em Língua Portuguesa e suas respectivas literaturas (1, 4%) y Letras Inglês y Pedagogia (1, 4%). Algunos han seguido formándose a través de posgrados y proyectos dentro de los IF. Respecto a la titulación más alta de los sujetos, el gráfico 2 nos ofrece una representación de los resultados.

Gráfico 2:Nivel de las titulaciones de los docentes de Español de los IF.

La mayoría (16 encuestados, 64%) tiene como titulación de más nivel un máster (o maestría). Este dato es relevante si tenemos en cuenta que hay pocos programas de posgrado en Lengua española en Brasil y los de las áreas de Lenguaje y Educación cuentan con pocos profesores habilitados para dirigir trabajos finales de máster. Un 28% (7) tiene la especialización, y tan solo el 8% (2) es doctor. Ya que ninguno de los docentes se ha quedado únicamente con la carrera de grado, es positivo para ambos IF que un número elevado de sus profesores tengan posgrado10. La tendencia en nuestra muestra es que aumenten los másteres y doctores, pues muchos encuestados afirman que están matriculados en programas de posgrado.

8 Las franjas extremas (“menos de 21 años”, “más de 61 años”) no presentaron sujetos. 9 Esta habilitação dupla era común hasta 2005. Muchas universidades la han extinguido porque dudan de que ofrezcan formación de calidad a los futuros profesores de lenguas extranjeras (UFRG, 2006). 10 Es importante aclarar que los concursos públicos para plazas de profesores de Español en los IF no suelen exigir posgrados, como ocurre en las universidades públicas de Brasil, pues se entiende que los docentes actuarán en la enseñanza secundaria (técnico integrado) y la formación para ejercer en ese nivel es el grado.

RLyLA Vol. 14 (2019), 29-43 | 34 Bruno Rafael Costa Venâncio da Silva y María Antonieta Andión Herrero Actitudes y competencias docentes en profesores de español de los Institutos Federais (Brasil) en relación con las variedades de la lengua

Respecto del tiempo desempeñado como profesores de español, en toda su carrera y como parte del plantel de los IF, como se aprecia en el gráfico 3, la mayoría de los encuestados (64%) acumula una experiencia que supera una década en la enseñanza del español: entre 10 y 20 años. Es bastante variable el tiempo de ejercicio docente en los IF; el más frecuente es entre 5-20 años (9 encuestados, 36%). Las franjas “entre 2 y 5 años” y “menos de 2 años” tienen un 28% (7 cada una); este equilibrio se debe a la creación de nuevos campus de los IF, especialmente en ciudades interiores de Brasil.

Gráfico 3:Tiempo de desempeño como profesor de Español y como profesor en el IF.

Como ya comentamos, el equipo docente de los IF imparte asignaturas en diferentes niveles de enseñanza, desde secundaria (técnico integrado) hasta cursos superiores de formación de profesores. El gráfico 4 aporta los resultados en este sentido.

Gráfico 4:Experiencia docente en niveles de enseñanza en los IF.

En sus respuestas, los sujetos podían elegir más de un nivel de enseñanza; situación posible por el perfil múltiple de un profesor de IF. Hemos comprobado que el 92% de los encuestados impartían Español en secundaria (técnico integrado11). La enseñanza superior en los cursos de Licenciatura em Espanhol o de Letras / Español, modalidades presencial y a distancia, es el segundo nivel más frecuente (18, 72%).

Consideramos positivo que un número tan alto de profesores actúen, tanto en la enseñanza secundaria como en la superior, en cursos de formación de profesores de Español. Este profesional con doble perfil puede facilitar información sobre dos universos distintos de clase.

11 En esta especialidad, las clases no están dirigidas necesariamente a un área profesional concreta, ya que Español forma parte del currículo de secundaria (Ensino Médio). Como algunos cursos técnicos —por ejemplo, Turismo— ofrecen más carga horaria de Español, el profesor puede trabajar la lengua de especialidad.

RLyLA Vol. 14 (2019), 29-43 | 35 Bruno Rafael Costa Venâncio da Silva y María Antonieta Andión Herrero Actitudes y competencias docentes en profesores de español de los Institutos Federais (Brasil) en relación con las variedades de la lengua

5.2. Perfil lingüístico de los profesores Preguntar por la variedad lingüística (en español) de los sujetos permite deducir la caracterización dialectal del input al que estarán expuestos sus aprendices pues “os professores constituem o principal (quando não o único) modelo de expressão, é natural que os estudantes tendam a adotar a variedade à qual são expostos durante mais tempo” (Ministério da Educação, 2006:137).

La respuesta más frecuente a la pregunta sobre si el profesor era o no hispanohablante nativo dio como resultado que un 80% (20 encuestados) no lo era, mientras que un 20% (5) sí12. En la identificación dialectal que a continuación pide el cuestionario, los nativos —salvo excepción— tienen la variedad de su región de origen o del lugar donde han vivido más años. Los no nativos, por su parte, pueden o no identificarse con una variedad concreta. El gráfico 5 muestra la identificación dialectal de los encuestados.

Gráfico 5:Variedad lingüística de los profesores de Español de los IF.

Junto a las zonas lingüísticas americanas13, ofrecemos las variedades españolas (peninsulares e isleñas), que aparecen juntas en el gráfico anterior, el español ecuatoguineano y el de un “brasileño sin una variedad definida” (o estandarizado). Esta última fue la opción más elegida entre los encuestados (9, 36%). La segunda, muy próxima a la anterior, fue la del español de España en sus diversas variedades (7, 28%). La tercera variedad seleccionada fue la del español austral o rioplatense (5, 20%), cuya frontera por el sur es geopolíticamente estratégica para Brasil.

Es relevante, en la identificación de nuestros docentes de los IF de Rio Grande do Norte y Roraima, la preferencia por las variedades de España. Siete de los profesores brasileños afirman tener una variedad española. Por otra parte, de los cinco docentes que eligieron la variedad rioplatense, dos son nativos y uno es hablante de herencia. La variedad castellana es asumida por la gran mayoría de los docentes brasileños, quizás sea consecuencia del impacto de la inversión que España ha realizado en los últimos años en cursos de actualización para profesores de ELE en Brasil y del éxito de la distribución de sus materiales didácticos (Moreno, 2010:185). Esto hace que España ocupe un lugar central en la enseñanza de español en Brasil frente a otros países hispánicos (Zolin-Vesz, 2013)14.

Dos profesores han elegido el español mexicano o centroamericano; uno, el español andino y otro (nativo), el español caribeño15. Es llamativo que un nativo haya seleccionado el “español hablado por un brasileño sin variedad definida”; su larga permanencia en Brasil le hace creer que ha perdido los rasgos lingüísticos de origen.

Precisamente la variedad del profesor es el punto de partida para que los docentes planifiquen la presencia de otras variedades del español en sus clases. El alto índice de comunicabilidad del español permite incorporar inputs de diferentes regiones hispánicas.

12 Debemos reconocer que el tema “profesor nativo versus no nativo” se considera, en general, superado en Brasil (Figueredo, 2011). En el área de ELE se refleja en que, por ejemplo, las oposiciones para profesor no dan preferencia a los profesores nativos, como sí ocurre en las contrataciones para escuelas de idiomas. 13 La zonificación del español americano sigue la propuesta por Henríquez Ureña (1921,apud Moreno 2000:38-50), la más autorizada por la bibliografía sobre el tema. 14 Para Zolin-Vesz (2013:55), “o espanhol é considerado a língua comum do mundo hispânico, em que a Espanha desempenha papel central, servindo como modelo de expressão de múltiplas culturas. Simboliza o espírito de concórdia e democracia”. 15 Este informante es profesor del IFRR, un instituto situado en una región fronteriza con Venezuela, país caribeño.

RLyLA Vol. 14 (2019), 29-43 | 36 Bruno Rafael Costa Venâncio da Silva y María Antonieta Andión Herrero Actitudes y competencias docentes en profesores de español de los Institutos Federais (Brasil) en relación con las variedades de la lengua

5.3. Creencias y actitudes hacia las variedades del español Sin negar la unidad del español, en el marco intradialectal hispánico afloran creencias sobre el prestigio de sus variedades que les asocian las propias comunidades de habla. Ello se refleja también en el ámbito de ELE (Andión, 2017), los profesores tienen su opinión sobre cuál variedad puede ser considerada “mejor” o “peor” para la enseñanza de la lengua extranjera, o cómo deben actuar en clase cuando su variedad no coincide con el modelo preferente de lengua16.

Aunque los docentes estén dispuestos a aceptar principios como que la variación es inherente a las lenguas y sigue sus propias reglas, persiste la paradójica idea de que ciertos fenómenos variables son “desviaciones” respecto de un modelo ponderado como estándar (Andión, 2017) y, consecuentemente, deben ser evitados o —peor— censurados en el aula.

En el marco de estos complejos precedentes, preguntamos a los profesores si habían sido víctimas de prejuicios o actitudes negativas hacia su variedad (nativa o adoptada). 15 encuestados (60%) han respondido que no, mientras que 10 (40%) sí lo han sido. Aunque las respuestas de los afectados no sean mayoría en la muestra, el número es elevado, más aún si tenemos en cuenta que los docentes suelen ser referentes del “buen hablar” en la sociedad.

Si asociamos las respuestas anteriores a las se han dado a la pregunta sobre la variedad del profesor, existe correspondencia entre el 60% de profesores que no han sufrido lectocentrismo17 y el 28% de quienes se han identificado con una variedad del español de España, que suele tener un reconocido prestigio.

El reconocimiento y la estigmatización de las variedades del español se evidencian en los resultados del gráfico 6, que muestra las respuestas de los encuestados aplicadas a la enseñanza. Las respuestasb , c y d a la pregunta 12 del cuestionario ofrecen matices en el desequilibrio de prestigio entre las variedades del español en el marco del aula.

Gráfico 6:Prejuicio lingüístico de los profesores de los IF hacia las variedades lingüísticas.

La alternativa que las variedades del español “tienen todas el mismo prestigio” ha tenido un 24% de aceptación (6 encuestados). La alternativa b, elegida por un 28% (7), matiza entre prestigio y reconocimiento o fama en el contexto de ELE, sin que se niegue el primero pero aceptando que hay grados entre ellos.

Los profesores ampliaron su elección explicando cuáles eran esas variedades más reconocidas o de más fama. Todos los que eligieron la alternativa b opinaron que el español de España18 es el que recibe ese plus de prestigio, que uno de los docentes justifica por su “política editorial”. Las variedades rioplatense y mexicana aparecieron una vez mencionadas por dos docentes, la última asociada a las telenovelas. Un 44% (11 encuestados) eligió la c, asumiendo un estigma diferencial de prestigio entre las variedades del español. Algunas de sus respuestas son relevantes19:

16 Atendemos al concepto de Andión (2007:24) de modelo o variedad preferente como “geolecto del estándar ponderado dentro del programa de un curso de EL2/LE al cual le estamos diseñando un modelo lingüístico; es la variedad que hemos decidido presentar como modelo principal para la producción de los aprendices”. 17 Andión (2017: 133) define el lectocentrismo como “la creencia de superioridad de una lengua o variedad (geolecto) sobre otras. Suele manifestarse más comúnmente entre geolectos de una misma lengua, y la variedad favorecida termina calificada como 'pura' o 'correcta', mientras que las otras se consideran desviaciones”. 18 En realidad, se estaban refiriendo al español centro-norte peninsular (variedad castellana), que es el que identifican con España. 19 En los ejemplos se ha respetado la ortografía y redacción originales.

RLyLA Vol. 14 (2019), 29-43 | 37 Bruno Rafael Costa Venâncio da Silva y María Antonieta Andión Herrero Actitudes y competencias docentes en profesores de español de los Institutos Federais (Brasil) en relación con las variedades de la lengua

No por mí, como profe, pero por los materiales que utilizo en ellos casi siempre presentan algunas variantes solo como información o curiosidad y siguen con la variante ibérica o presenta variante de hispanoamérica de forma neutralizada. (PROFE5)20

Entiendo que el prestigio no es una cualidad intrínseca de los códigos lingüísticos, sino una cualidad atribuida por las comunidades que los usan. En este sentido, se atribuye más o menos prestigio a ciertos códigos en razón de condiciones sociohistóricas, por ejemplo: aquellas variedades de los centros de poder reciben más prestigio. (PROFE8)

A la que me encuadro21. Creo que tras una formación académica bajo diferentes variedades lingüísticas y sin haber vivido en países de habla hispana (realidad de muchos alumnos de los cursos de graduación), es muy difícil no mezclar estas variedades. (PROFE10)

Pienso que las variedades españolas suelen ser más prestigiosas que las hispanoamericanas, lo que para mí es un gran prejuicio. (PROFE20)

Ocho de las diez respuestas marcan el español de España como la variedad de más prestigio, mientras que tres también citan el español rioplatense (de La Plata). Es admirable que uno de los encuestados (PROFE8) reconozca que el prestigio no tiene carácter lingüístico sino social y económico, y es producto de las creencias de los hablantes. PROFE10, que eligió la opción c, parece defenderse de las dudas sobre el prestigio de su variedad: neutra, sin filiación dialectal evidente. Su respuesta a la pregunta 11 del cuestionario en la que reconoce haber sufrido prejuicio lingüístico así lo indica.

Llama la atención que la variedad rioplatense sea la segunda entre las variedades de más prestigio a pesar de que los tres nativos rioplatenses hayan afirmado haber sufrido prejuicio lingüístico. Ya que ningún profesor eligió la opción d (“algunas no tienen prestigio”), entendemos que se trata de una gradación de reconocimiento que no llega a marcas muy negativas.

Son tímidas las menciones de variedades americanas prestigiadas, a saber, rioplatense y mexicana; también como modelos lingüísticos de algunos informantes, que las reconocen como propias, caso de cuatro docentes brasileños, dos de la variedad rioplatense y dos de la mexicana-centroamericana.

5.4. Formación en variedades de los profesores de los IF Las respuestas a la pregunta 16 del cuestionario permitían conocer si los profesores habían cursado asignaturas sobre las variedades del español en sus estudios. La mayoría, un 72% (18 encuestados), no había recibido información específica en ningún nivel (grado, posgrado, máster o doctorado); es evidente que este tema no parece haber sido atendido en su formación.

Al 28% (7) que respondió que sí las había cursado, se le preguntó además qué asignatura o tema había atendido el componente “variedades”. A pesar de algún “no recuerdo” y la ambigüedad de “diversas otras en las que se contemplaba la riqueza de la variación lingüística”, las respuestas positivas fueron varias: Dialectología (del español), Pluralidad funcional, El tratamiento y sus formas, Cortesía y discurso hispánico, Variantes del español (Universidad de Granada, España).

En el mismo sentido se mostraron las respuestas a la pregunta 18 del cuestionario sobre si habían impartido asignaturas o temas sobre diversidad lingüística del español en un IF: un 80% (20 encuestados) no lo había hecho y un 20% (5) sí. Estos últimos, la minoría, podían precisar en la siguiente pregunta (19) la asignatura o el tema tratado. Las respuestas, en portugués y español, incluyen también algunas anotaciones de interés: Dialectología (de la lengua española), Dialetologia do Espanhol, Linguística Contrastiva (“a pesar de no tratar específicamente de la variación trae algunas reflexiones sobre el tema”). Un sujeto reconoce: “no la impartí, pero hice la revisión lingüística y pedagógica de una llamada Dialetologia da Língua Espanhola”. Otro aclara que el contenido sobre “variedades del español” se encuentra presente en otros componentes curriculares como fonética, semántica, traducción y las asignaturas de Lengua española.

Si profundizamos en los sujetos que responden a las preguntas 16 a la 19, resulta curioso —en realidad, incongruente— que de los cuatro docentes que han impartido la asignatura de Dialectología, solo uno afirme antes haber recibido formación relativa a las variedades del español. Los otros tres no contaban con información específica anterior. Como veremos más adelante, algunos salvan ese inconveniente con una preparación autodidacta; existe información bibliográfica suficiente, a saber, monografías, artículos y recursos audiovisuales sobre el tema. No obstante, esta falta de correspondencia entre formación y competencia docente hace dudar de su suficiencia para impartir estos contenidos.

20 Entre paréntesis anotamos los códigos de identificación de los informantes. 21 Se refiere a su propia variedad lingüística.

RLyLA Vol. 14 (2019), 29-43 | 38 Bruno Rafael Costa Venâncio da Silva y María Antonieta Andión Herrero Actitudes y competencias docentes en profesores de español de los Institutos Federais (Brasil) en relación con las variedades de la lengua

Es probable que, en esa situación, los docentes hayan sido presas de inseguridades. A través de las respuestas a la pregunta 20 pudimos conocer en qué medida se sienten/sintieron preparados para impartir asignaturas o contenidos sobre las variedades del español. El Gráfico 7 muestra la información obtenida.

Gráfico 7:Preparación de los profesores de los IF para impartir asignaturas o contenidos sobre variedades del español.

Casi la mitad de los docentes manifiesta inseguridad en su preparación: un 48% (12 encuestados) se siente poco o nada preparado para impartir una asignatura o contenidos sobre variedades del español en los dos IF. Así, en las respuestas a la pregunta 21, los profesores (nativos y no nativos) justifican sus necesidades:

Necesito acercarme más del tema, actualizarme, profundizar los estudios para así, sentirme más preparada para impartir clases. (PROFE1)

Porque necesito saber, estudiar más de las variedades del español. (PROFE2)

En el caso de que necesitara impartir clases sobre las variedades del español, seguramente tendría que prepararme por no haber estudiado nada en este sentido durante mi formación académica. Todo lo que sé sobre el asunto fue por informarme por mi esfuerzo propio. (PROFE4)

Poco preparada, como nunca he impartido clases con esta temática directamente, entonces. Aún no he investigado tan a fondo, solo para uso y a “título” de curiosidad. (PROFE5)

Creo que estar preparado no significa saber reproducir las variedades de la lengua sino hacerles conocer a los alumnos la existencia de estas variedades y mostrárselas a través de diversos recursos: textos, audios, películas, canciones, etc. Suelo hacerlo en mis clases del IFRN. (PROFE6)

Faltan más estudios, conocimientos y prácticas del contenido. (PROFE7)

Este tema siempre me ha interesado bastante. (PROFE8)

Tengo pocos conocimientos acerca de las muchas variantes del español. (PROFE9)

Haber impartido en el magisterio superior (UFRN) disciplinas que incluían las variedades entre su contenido me permitió estudiarlas y discriminarlas particularmente. (PROFE13)

Aunque sepa algo de las variedades del español, nunca me he dedicado al estudio con profundidad y, con eso, no estoy tan preparado para impartir una asignatura en nivel superior. Hace falta que me dedique más a estudiar sobre ellas. (PROFE14)

Por no haber estudiado una asignatura que tratase el tema de variedad del español. (PROFE15)

Creo que en estos años de experiencias vengo leyendo y reflexionando sobre las variaciones de mi lengua materna. Soy de un pequeño país del Caribe (Cuba) (…) con (…) una significativa variación lingüística. (PROFE17)

Aunque ya conozca un poco más de las variedades del español y tenga información de dónde puedo encontrar recursos para trabajar este tema en clase, no puedo decir que estoy “muy preparada” porque es algo muy amplio y para mí es una gran responsabilidad presentar a los alumnos la importancia de reconocer las variedades de la lengua española. Pero, estoy segura de que, con esfuerzo y dedicación, conseguiría cumplir esta misión con tranquilidad. (PROFE18)

RLyLA Vol. 14 (2019), 29-43 | 39 Bruno Rafael Costa Venâncio da Silva y María Antonieta Andión Herrero Actitudes y competencias docentes en profesores de español de los Institutos Federais (Brasil) en relación con las variedades de la lengua

Apesar de não haver recebido uma preparação específica para essa questão, a possibilidade de pesquisar sobre o assunto, de preparar as aulas e estudar sobre essa temática habilitam o docente, que já tem um conhecimento geral sobre o ensino da língua espanhola, ministrar um curso específico sobre esse assunto. Não percebo uma barreira que impeça um bom trabalho nesse sentido. (PROFE19)

Aunque tengo algunas lecturas y algún conocimiento sobre el tema, estoy poco preparada para impartir una asignatura de ese tipo. Esa no es efectivamente mi área de interés y actuación. (PROFE20)

Nunca estudei, formalmente, o assunto. (PROFE21)

Como hablante nativo, por tener una larga experiencia de vida (social, cultural y académica) en países hispanos, creo tener un conocimiento significativo de diversas variantes del español. Aparte, mi particular interés por la semántica, la lexicología y la sociolingüística, de modo general, ha hecho que dedicase no pocas lecturas al asunto. (PROFE25)

Por lo general, se percibe en los docentes una actitud de compromiso hacia el conocimiento de variedades del español asumiendo sus carencias y necesidades para impartir de una forma rigurosa el tema. En las respuestas a la pregunta 22 gradúan sus expectativas de formación al respecto (Gráfico 8).

Gráfico 8:Necesidades de formación sobre variedades del español de los profesores de los IF.

La mayoría de los docentes (22, 88%) cree que es muy necesario o necesario tener formación en este tema para impartir clases de español en el Ensino Médio. Tan solo tres de los encuestados rebajan su interés de formación hasta lo “normal”. Es reconfortante comprobar que ningún docente afirme que sea poco necesaria o innecesaria, lo que evidencia la importancia que conceden a este componente en la enseñanza de ELE.

Si tenemos en cuenta lo respondido antes por los profesores respecto a su formación y necesidades, las respuestas a la pregunta 23 muestran una realidad interesante y de compromiso con la presencia de la diversidad lingüística en el aula de español (Gráfico 9).

Gráfico 9:Presencia de las variedades en el aula de español.

Es admirable que todos los profesores de los IF, pese a la inconsistencia de formación ya comentada, se sientan comprometidos con presentar estos contenidos en función de las zonas geográficas hispánicas. Así,

RLyLA Vol. 14 (2019), 29-43 | 40 Bruno Rafael Costa Venâncio da Silva y María Antonieta Andión Herrero Actitudes y competencias docentes en profesores de español de los Institutos Federais (Brasil) en relación con las variedades de la lengua

un 68% (17 encuestados) responde que siempre lo hace, mientras un 32% (8) aprovecha la conveniencia de un contexto propicio para ello.

No queremos restar importancia al empeño de estos docentes de los IF, pero no deja de preocuparnos cómo consiguen abordar la diversidad del español en el marco de ELE (selección de rasgos, extensión y secuenciación por niveles) sin contar con la información lingüística, dialectológica y aplicada necesarias. Si nos detenemos, por ejemplo, en un rasgo como el voseo22, cuya presencia resulta evidente frente al tuteo (español estándar), su complejidad y diversidad morfológica, pragmática y geográfica precisaría que contaran con información como la siguiente:

Voseo [Tratamiento de vos para la segunda personal del singular, propio de situaciones informales.] con formas verbales propias, te y vos con preposición (vos te peinás solita; no quiere ir con vos) con consideraciones sociales y pragmáticas diferentes según zonas. [Notas del Tutor: Frente al tuteo, el voseo implica un mayor acercamiento al interlocutor.] [Con más detalle: Voseo verbal más frecuente en el presente de indicativo (sos un bárbaro) y en el imperativo (hacé caso); el voseo pronominal aparece como sujeto (vos no me querés), término de preposiciones (a vos, con vos, de vos, para vos…), de comparaciones (sabe más que vos). [Notas del Tutor: Ten presente que aunque el hablante vosee, utiliza tú para los complementos átonos (acordate, che) y su posesivo (pero vos no te hablás con tu hermano).] Las formas verbales predominantes son amás, temés, partís [en Chile: amái(s), temí(s), partí(s)] (Andión y Casado 2014).

Hispanoamérica: Andes (costas de Colombia y Ecuador, Mérida-Venezuela), Caribe (Panamá, occidente de Venezuela), Chile, México (zonas del sur) y Centroamérica (zonas), Río de la Plata (Argentina, Paraguay).

Los ocho encuestados (32%) que afirman aprovechar el contexto pueden dar a entender que no es un contenido planificado en la preparación de sus clases, tan solo aparecería en caso de dudas o situaciones espontáneas que puedan surgir en el aula.

También hemos preguntado por la variedad adoptada en el aula como modelo de la enseñanza-aprendizaje. Nos interesaba si los alumnos eran informados de la afiliación dialectal del mismo, es decir, sobre qué variedad iba a ser la preferente del curso. Los resultados aparecen en el gráfico 10.

Gráfico 10:Información dada a los alumnos sobre la variedad preferente del curso por parte de los profesores de los IF.

Las actuaciones docentes son diversas: un 48% (12 encuestados) afirma que siempre informa sobre qué variedad se va a enseñar como modelo; un 28% (7) lo hace frecuentemente; un 12% (3), solamente a veces; un 4% (1), casi nunca; y un 8% (2), nunca. Preocupa que más de la mitad de los docentes crea que esta información no sea imprescindible. El compromiso de informar sobre este particular es insoslayable, mucho más cuando la variedad del profesor no coincide con la preferente del curso23.

6. CONCLUSIONES

El análisis de la información obtenida a través del cuestionario aplicado a 25 profesores de dos IF de Rio Grande do Norte y Roraima nos permite perfilar, en la caracterización sociolingüística de los profesores, una muestra

22 Ese fenómeno de variación lingüística fue citado dos veces en las respuestas a la pregunta 27 del cuestionario. 23 Andión y Casado (2014: cap. 3) reflexionan sobre este supuesto y hacen interesantes recomendaciones de actuación docente.

RLyLA Vol. 14 (2019), 29-43 | 41 Bruno Rafael Costa Venâncio da Silva y María Antonieta Andión Herrero Actitudes y competencias docentes en profesores de español de los Institutos Federais (Brasil) en relación con las variedades de la lengua

mayoritariamente femenina (72% de mujeres); joven aunque de cierta madurez (48% entre 31 y 40 años); con formación de grado especializada preferentemente en Lengua española (Letras com Habilitação dupla [Português/ Espanhol], Letras com Habilitação em Língua Espanhola e suas respectivas literaturas, Licenciatura em Espanhol/ Língua Espanhola, Letras Espanhol), en la que muchos habían alcanzado niveles superiores de posgrado (máster, 64%; especialización, 7%; doctorado, 32%); y con experiencia docente variable, aunque en la mayoría es superior a una década (64%).

Por su parte, la caracterización lingüística define a una mayoría no nativa hispanohablante (80%), que en un 36% no se identificó con ninguna variedad y se enmarca en un español estandarizado o neutro. Junto a los nativos, les siguen los de variedad española (28%), austral o rioplatense (20%), mexicana-centroamericana, andina y caribeña. La adscripción de una mayoría no nativa al español de España refleja el impacto de la cooperación internacional en el ámbito educativo del gobierno español y la política de difusión de la marca España asociada a la lengua, así como el éxito de sus productos editoriales de ELE.

En general, los profesores de los IF manifestaron no haber sido víctimas de prejuicios por su variedad (nativa o adoptada) (60%). Lamentablemente, el otro 40 % sí se ha sentido discriminado a pesar de ser referentes de buen uso de la lengua por su propia profesión. La situación afecta sus creencias ya que solo el 24% de ellos considera que todas las variedades del español tienen prestigio. Sin negar esta cualidad, más profesores (28%) reconocen diferencias entre ellas en el marco de ELE y eligen el español de España como su referente más reconocido. Tímidamente se mencionan las variedades rioplatense y mexicana como prestigiosas. Un 44% asume la existencia de estigmas dialectales.

También debemos reconocer que es esperanzador que ningún encuestado piense que “algunas [variedades] no tienen prestigio”. El respeto hacia la competencia de los hablantes nativos parece mantener cierto reconocimiento por las hablas naturales del español.

La formación en temas sobre variedades del español de los profesores de los IF evidenció carencias importantes: un 72% no había recibido información específica en ningún nivel (grado, posgrado, máster o doctorado). Ese hecho incide en algunas cuestiones, como las preferencias y actitudes de los alumnos hacia un modelo de lengua específico, la presencia de rasgos lingüísticos de otras variedades en el curso y, consecuentemente, la selección de recursos didácticos que las muestren.

Tampoco una mayoría (80%) había impartido asignaturas o temas sobre diversidad lingüística del español. Es lamentable en cuanto a la formación, pero lo que sí resulta alarmante es que solo un profesor de estos últimos se encontraba entre los primeros (sí formados). Es lógica, entonces, la inseguridad que mostraron en su preparación sobre el tema y que casi un 90% afirme que necesitan formación.

Es evidente en los docentes una preocupación responsable sobre el tema, un reconocimiento a su importancia dentro de la enseñanza de ELE y un compromiso con que la variación y variedad del español forme parte de la competencia de sus aprendices. No obstante, son menos exigentes en cuanto a informar a los alumnos sobre a qué variedad corresponde su modelo de enseñanza.

Preocupa que más de la mitad de los docentes crea que esta información no sea imprescindible, pues como señalan Andión y Casado (2014:cap. 3), “lo primero que debe saber un profesor es qué rasgos fonéticos o gramaticales del español pertenecen a la variedad preferente de su curso, es decir, qué español va a enseñar y qué hablantes se identifican con ese modelo” y esa información debe transmitirse a los alumnos. A partir de ahí, se pueden presentar —seleccionadas y secuenciadas— otras variantes del español, pero esto no podría hacerse si no se identifica oportunamente la variedad preferente.

REFERENCIAS

Andión, M. A. (2013). “Los profesores de español segunda/lengua extranjera y las variedades: Identidad dialectal, actitudes y prácticas docentes”, Revista Signos 46/82: 155-189. https://doi.org/10.4067/ S0718-09342013000200001 Andión, M. A. (2017). “Etnocentrismo lingüístico vs. plurinormativismo. Consideraciones sobre la variación y variedad del español-LE/L2”, en Actas del XXVII Congreso Internacional de ASELE: Panhispanismo y variedades en la enseñanza del español L2-LE. La Rioja: ASELE-Fundación San Millán de la Cogolla, 131-140.

Andión, M. A. y Casado, C. (2014). Variación y variedad del español aplicadas a E-LE/L2. Madrid: Editorial UNED.

RLyLA Vol. 14 (2019), 29-43 | 42 Bruno Rafael Costa Venâncio da Silva y María Antonieta Andión Herrero Actitudes y competencias docentes en profesores de español de los Institutos Federais (Brasil) en relación con las variedades de la lengua

Andión, M. A. y Gil, M. (2013). “Las variedades del español como parte la competencia docente: Qué debemos saber y enseñar en ELE/L2”, en Actas del I Congreso Internacional de Didáctica del Español como Lengua Extranjera. Instituto Cervantes de Budapest y Embajada de España en Hungría, pp. 47-59. Araújo, L. S. (2013). “A dialetologia hispano-americana e a heterogeneidade linguística: o caso da Argentina”, Web- Revista Sociodialeto, 2/3: 28-48. Bagno, M. (2015). Preconceito lingüístico. São Paulo: Parábola Editorial. Barros, A. de J. P. de y Lehfeld, N. A. de S. (1990). Projeto de pesquisa: propostas metodológicas. Petrópolis: Vozes. Figueredo, C. J. (2011). “O falante nativo de inglês versus o falante não-nativo: representações e percepções em uma sala de aula de inglês”, Revista Linguagem e Ensino, 14/1: 67-92. García Mouton, P. (2000). Cómo hablan las mujeres. Madrid: Arco Libros. Geraldi, J. W. (1984). “Concepções de linguagem e ensino de português”, en J. W. Geraldi, (org.), O texto na sala de aula. São Paulo. Cascavel: Assoeste, pp. 39-46. Gressler, L. A. (2003). Introdução à pesquisa_ projetos e relatórios. São Paulo: Loyola. Irala, V. B. (2004). “A opção da variedade de Espanhol por professores em serviço e préserviço”, Linguagem & ensino, 7/II: 99-120. Ley N.º 11.892, Presidencia da República. Casa Civil. Subchefia de Assuntos Jurídicos, República Federativa do Brasil, 29 de diciembre de 2008. López Morales, H. (1989). Sociolingüistica. Madrid: Editorial Gredos. Ministério da Educação, Brasil. (2006). Orientações Curriculares Nacionais para o Ensino Médio (OCEM). Conhecimentos de Espanhol. Brasília, Secretaria de Educação Básica. Disponible en: http://portal.mec. gov.br/seb/arquivos/pdf/book_volume_01_internet.pdf. Acceso en: 24 de septiembre de 2016. Moreira, H. y Caleffe, L. G. (2008).Metodologia para o professor pesquisador. Rio de Janeiro: Lamparina. Moreno, F. (2000). Qué español enseñar. Madrid: Arco Libros. Moreno, F. (2009). Principios de sociolingüística y sociología del lenguaje. Barcelona: Ariel Letras. Moreno, F. (2010). Las variedades de la lengua española y su enseñanza. Madrid: Arco Libros. Sampieri, R. H., Collado, C. F. y Lucio, M. P. B. (2013). Metodologia de pesquisa. Porto Alegre: Penso Editora. Serrani, S. (2011). Discurso e cultura na aula de língua/currículo- leitura- escrita. Campinas: Pontes. Silva-Corvalán, C. y Enrique-Arias, A. (2001). Sociolingüística y pragmática del español. Washington D. C.: Georgetown University Press. Universidade Federal do Rio Grande do Norte (2006). Projeto político pedagógico do curso de Letras. Universidade Federal do Rio Grande do Norte: Natal. Zolin-Vesz, F. (2013). “A Espanha como o único lugar que se fala a língua espanhola - a quem interessa essa crença?”, en A (in)visibilidade da América Latina no Ensino de Espanhol. Campinas: Pontes, pp. 51-62.

RLyLA Vol. 14 (2019), 29-43 | 43 Revista de Lingüística y Lenguas Aplicadas Vol. 14 año 2019, 45-57 EISSN 1886-6298 https://doi.org/10.4995/rlyla.2019.10785

THE IMPACT OF APPLYING THE FONF PRACTICE MODEL ON DEVELOPING L2 LISTENING AND SPEAKING WITH A FOCUS ON INTENTIONAL AND INCIDENTAL VOCABULARY ACQUISITION IN CALL CONTEXT

Akbar Bahari University of Qom, Iran

Abstract: Drawing on dynamic systems theory, the FonF (focus on form) practice model was contextualized with respect to its impact on L2 learners’ incidental and intentional vocabulary acquisition in a CALL context. To this end, a mixed methods approach was conducted on a sample of 93 intermediate EFL learners in a CALL setting in keeping with the FonF practice model. Comparing incidental and intentional acquisition was the target of the study. The obtained results confirmed the effectiveness of the FonF practice model via form-, meaning-, and communication-oriented strategies to develop listening and speaking proficiency as well as incidental and intentional vocabulary acquisition among the experimental group. The main educational implication of the study is the effectiveness of the FonF practice model at unlocking the potential behind CALL affordances towards developing listening- speaking proficiency and developing intentional and incidental vocabulary while catering for nonlinear dynamic motivational factors at individual L2 learner level.

Keywords: Dynamic Systems Theory (DST), nonlinear dynamic motivation (NDM), intentional vocabulary acquisition, incidental vocabulary acquisition, CALL.

INTRODUCTION

The FonF (focus on form) model of practice introduced by Bahari (2019a) was prepared and contextualized based on the effectiveness of FonF-based instruction for its potential for incidental and preplanned L2 learning (Bahari, 2018a; Fredricks, Blumenfeld, & Paris, 2004; Nassaji, 1999, 2016; Nassaji & Fotos, 2007, 2011; Williams, 2005). The FonF practice model has proved effective in catering for learners’ nonlinear and dynamic motivational factors at individual level and taking advantage of the CALL affordances towards developing listening and speaking skills (Bahari, 2019a). In keeping with Nassaji and Fotos (2011) the FonF is considered as an optimal approach for learning which aims at mixing the best features of classroom L2 learning with computer-assisted L2 learning by using CALL tools and applications to facilitate L2 learning process. In contrast to previous L2 learning practice models sharing the feature of generality at group level, the FonF practice model addresses the nonlinearity and dynamicity of individual differences during learning process (Bahari, 2018a). This is in response to the call for integrating complex dynamic systems perspective (Larsen-Freeman & Cameron, 2008) towards implementing interactive-collaborative CALL environment. Under the FonF practice model, on the one hand, form, meaning and communication are addressed as critical learning components and on the other hand, individual learner’s motivation is catered to by selecting learning materials with respect to the nonlinearity and dynamicity of individual learners (Dörnyei & Ottó, 1998) in a systematic but flexible framework. Reflecting Vygotskian notions of regulation of learning and learning process to move from object to other and to self, the FonF practice model reflects the dynamic order of interrelated learning components which are dynamically and nonlinearly influenced by internal and external stimuli present in nonlinear and dynamic environments ranging from here-and-now environments to abstract ones to form engagement with meaning; to attain naturalness of L2 forms; to ensure unobstructive instruction; to provide communicative use of form/meaning to develop automaticity, accuracy and fluency (Norris & Ortega, 2000; Sheen, 2002). Given the above theoretical basis of the FonF practice model, it seemed significant

To cite this article: Bahari, A. (2019). "The Impact of Applying the FonF Practice Model on Developing L2 Listening and Speaking with a Focus on Intentional and Incidental Vocabulary Acquisition in CALL Context". Revista de Lingüística y Lenguas Aplicadas, 14, 45-57. https://doi.org/10.4995/rlyla.2019.10785 Correspondence author: [email protected]

Received: 2018-10-06 Accepted: 2019-06-28 | 45 Akbar Bahari The Impact of Applying the FonF Practice Model on Developing L2 Listening and Speaking with a Focus on Intentional and Incidental Vocabulary Acquisition in CALL Context to test its effectiveness for incidental and intentional vocabulary acquisition as well. The FonF practice model, drawing on nonlinearity and dynamicity of language and language learner (Larsen-Freeman & Cameron, 2008) was proposed to facilitate language learning by benefiting from the affordances provided by CALL (Colpaert, 2018). The model facilitates automatization of explicit knowledge by activities such as drills that focus on behaviors instead of structures (DeKeyser, 1998) and the necessary uses of L2 (Pawlak, 2006). The model consists of nine stages to practice listening-speaking strategies (see Figure 1). The first three stages of the model focus on theform , the second three stages focus on the meaning and the third three stages focus on the communication. To address the nonlinearity and dynamicity of L2 motivation, these stages multidirectionally engage learner in both listening and speaking activities via FonF approach-oriented strategies (Bahari, 2018b). The goal is to avoid modular approaches and single-factor effects (de Bot, 2008) and unidirectional listening activities (Chang, 2005; Macaro, Graham & Vanderplank, 2007) without directing learning activities towards communicative goals. All stages require task repetition with minimum intervals which reinforces interactions between conceptualization, formulation and articulation (Bygate & Samuda, 2005). This is to minimize the cognitive load by FonF rehearsal (Bahari, 2018c) in the early stages and facilitate connecting form and meaning before producing communicative output.

Figure 1. Graph of the FonF Practice Model by Bahari (2019a). Reprinted from Bahari (2019a).

To test the effectiveness of the FonF practice model, a mixed methods approach was conducted on a sample of 54 female and 39 male intermediate EFL learners who kindly cooperated in a three-month study. The study observed the sample’s ethical standards (e.g. informed consent, respect for anonymity and confidentiality). With an average age of 17, the participants were randomly assigned into experimental group (41 female and 29 male) and control group (13 female and 10 male). While the experimental group received the FonF practice model-based educational intervention via CALL affordances (e.g. videos, podcasts, games, and quizzes), the control group did not receive any educational intervention and observed the ordinary schedule of the classroom. To involve the agents of change two expert teachers were invited to monitor the assessment process of the obtained results from qualitative data collection instruments (interview and questionnaire) as well as listening, speaking, and incidental and intentional vocabulary tests. To test the effectiveness of the FonF practice model, several strands of data collection were employed (see Figure 2) in response to the following research questions:

RQ1: Is there any significant relationship between applying the FonF practice model and developing listening- speaking proficiency in a CALL context?

RQ2: Is there any significant relationship between applying the FonF practice model and developing incidental and intentional vocabulary in a CALL context?

RQ3: What relationships can be observed between EFL learners’ responses and improving listening-speaking and incidental and intentional vocabulary acquisition under the FonF practice model in a CALL context?

RLyLA Vol. 14 (2019), 45-57 | 46 Akbar Bahari The Impact of Applying the FonF Practice Model on Developing L2 Listening and Speaking with a Focus on Intentional and Incidental Vocabulary Acquisition in CALL Context

TRIANGULATIONS OF RESULTS

Findings re: RQ1 RQ2 RQ3

Figure 2. Visual Representation of the Study.

LITERATURE REVIEW

Review of the CALL literature revealed that a variety of CALL affordances (e.g. podcasts, videos, games, and quizzes) have a positive influence on L2 learning and teaching. Reported effectiveness of CALL affordances such as games (Chiu, 2013), podcasts (Winke, Gass, & Sydorenko, 2013), and quizzes (Arora, Evans, Gardner, Gulbrandsen, & Riley, 2015; Dumova, 2012; Lu, 2009) to develop different aspects of L2 learning (e.g. vocabulary learning, developing listening and speaking etc.) have attracted many scholars. According to these studies, CALL affordances have the potential to promote engagement among L2 learners (Wilson, Hainey &, Connolly, 2013) and to improve learners’ attitudes towards CALL activities (Dashtestani, 2015). Despite the presence of significant studies elaborating on the effectiveness of CALL affordances, introducing and testing applicable practice models which can be adopted by teachers and learners to benefit the most from CALL affordances is the main gap in the literature. Such CALL-oriented practice models serve as a road map for CALL users and provide a tested, efficient and systematic practice model.

The FonF, as a language learning intervention can be described as a learner-oriented teaching approach (Long, 2000) which makes it possible to move away from merely meaning-focused instruction towards a form-, meaning-, and communication-focused instruction (Bahari. 2019a). The FonF involves learners in learning process via planned or incidental tasks which are interactive/non-interactive (Ellis, 2016). According to Bahari (2019a) FonF enables L2 teachers to address learners’ errors (via reactive FonF), to bring language into focus (via preemptive FonF), and to select linguistic devices to foster communicative accuracy (via collaborative FonF). Given the significant relevance between vocabulary learning and audiovisual materials (Vidal, 2003) and the significant role of input in developing L2 system (VanPatten, 2004), the present study adopted the FonF practice model to test its effect on developing incidental and intentional aspects of vocabulary acquisition. Incidental vocabulary acquisition in the present study refers to acquiring new vocabulary via exposure to audiovisual materials provided by CALL affordances, and intentional vocabulary acquisition refers to learning word lists via rote learning.

METHOD

Sample and procedure To test the effectiveness of the FonF practice model, a mixed methods approach was conducted on a sample of 54 female and 39 male intermediate EFL learners. The sample was drawn from boys and girls branches of a

RLyLA Vol. 14 (2019), 45-57 | 47 Akbar Bahari The Impact of Applying the FonF Practice Model on Developing L2 Listening and Speaking with a Focus on Intentional and Incidental Vocabulary Acquisition in CALL Context private language school in Tehran, Iran. To ensure the homogeneity of the participants in terms of the language proficiency, the outliers of the listening, speaking and vocabulary pretest (n=7) were excluded from the study and the study was conducted with 93 participants.

Learning Activities Administered during the educational intervention In the present study, four CALL affordances of podcasts, videos, quizzes and games presented by Merriam Webster Online Dictionary (MWOD), were used as a tool to observe the influence of CALL affordances under the FonF practice model on listening, speaking and intentional and incidental vocabulary acquisition. For example, the experimental group students selected videos from the MWOD (see Appendix E) and then practiced the video according to the stages outlined in Figure 1. To familiarize them with the procedure a pilot session was administered in advance of the main study. During the FonF-oriented educational intervention, the participants were asked to observe the suggested steps of the model while using CALL affordances of podcasts, videos, quizzes and games. They were asked to fill in the self-report form (see Table 1) and record their opinions about the effectiveness/ ineffectiveness of the educational intervention during the study (30 sessions).

Table 1. Self-report on the effectiveness/ineffectiveness of CALL affordances under the FonF practice model. *Title of Timetable showing the Learner’s self-report the video/ intervals between using about the effectiveness podcast/game/ CALL affordances under of educational quiz on MWOD the FonF practice Model Steps of using the FonF practice model intervention** Month Week Apply the following steps for every podcast, game, Effective

quiz and Ineffective

video effective Very Very ineffective Very

First First • Repeating while trying to mimic the form Second • Repeating while trying to mimic the Third pronunciation features Fourth • Repeating while trying to mimic the body Second First movements Third • Repeating with a focus on lexical aspect Third Fourth • Repeating with a focus on grammatical aspect • Repeating with a focus on meaning • Producing accuracy-oriented output • Producing fluency-oriented output • Producing output while negotiating for meaning *Please mention the title of the podcast, game, quiz or video (in terms of the vocabulary) **Dear fellow student, please explain in the box why you have selected a response.

Data Collection To elicit the required data for the first research question, the participants’ listening-speaking abilities were assessed before and after administering the educational intervention via pretest and posttest. General English proficiency test guidelines (e.g. difficulty level) were observed when developing the tests for the intermediate level participants under the study. The speaking part contained three tasks: listening-repeating; reading aloud; answering questions. In the listening part of the test the participants were asked to show their comprehension via multiple choice items. Both tests were scored on a 100-point score under the supervision of two experts to avoid unwanted biased assessment. Given the consistency of the obtained test results over the last terms for learners of the same language proficiency, the study confirms the reliability of the test items under similar conditions.

Given the multidimensionality and complexity of vocabulary knowledge (Nassaji, 2004), the present study elaborated on incidental and intentional vocabulary acquisition. The subjects were pretested prior to the study in terms of incidental vocabulary acquisition by means of an adapted scale (see Appendix A) in keeping with Vidal (2011). To this end, the subjects were asked to respond to prompts with respect to a list of 12 words (i.e. 4 nouns, 4 verbs, and 4 adjectives) randomly selected from the vocabulary section of the course book they were going to begin studying (i.e. Summit 1) and the obtained responses were scored on the basis of the scale (see Appendix A).

RLyLA Vol. 14 (2019), 45-57 | 48 Akbar Bahari The Impact of Applying the FonF Practice Model on Developing L2 Listening and Speaking with a Focus on Intentional and Incidental Vocabulary Acquisition in CALL Context

To collect the data with regard to the intentional vocabulary acquisition among the participants under study, the subjects were asked to memorize a list of 12 words (4 nouns, 4 verbs, and 4 adjectives randomly selected from the same course book) with the English definition, example sentence, and Persian translation in 15 minutes in keeping with Laufer (2006). To test their active knowledge, the participants were asked to provide the L2 translation of the presented L1 word and the responses were scored according to the scales adapted from Laufer (2006). Following the active knowledge which is an L1-L2 test, an L2-L1 test was administered with almost the same scoring scales (see Appendix B). Both tests were used as the pretest and posttest on the participants under the study in response to the second research question prior to the study and at the end of the study (with an interval of 80 days).

To elicit the required data for the third research question, two instruments were used:

First, FonF practice model’s questionnaire was prepared (see Appendix C) and administered among the members of the experimental group. It is a 16-item survey developed by the author to elicit the learners’ opinions on the effectiveness of the FonF practice model for developing listening-speaking with a focus on intentional and incidental vocabulary acquisition via four major affordances provided by MWOD (i.e. podcasts, videos, quizzes, and games) as part of the CALL syllabus. The items are rated along a 6-step Likert continuum (e.g., 1 = strongly agree to 6 = strongly disagree). The questionnaire took approximately 10–15 minutes to complete. To determine the internal consistency reliabilities of the subscales, the 12 subscales were subjected to a reliability test. The alphas are presented in keeping with (Wigfield & Guthrie, 1997) alphas in Table 2. The subscales (podcasts, videos, quizzes, and games) had reasonable reliabilities ranging from .78 to .81.

Table 2. Reliabilities for the questionnaire’s subscales.

Subscale No. of Items Reliability Podcasts 4 0.78 Videos 4 0.78 Quizzes 4 0.81 Games 4 0.79

Second, the interview (see Appendix D) is a 4-item survey developed to elicit the interviewees’ (only experimental group members) responses on the efficiency of using the FonF practice model to develop listening, speaking, and incidental and intentional vocabulary acquisition. Interview items are Why questions which can be considered as open-ended questions, however the responses are rated on the basis of the selected affordances (podcast, video, quiz, and game) to facilitate data analysis. To determine the internal consistency reliabilities of the subscales, the 4 subscales were subjected to a reliability test. Reliabilities are presented in keeping with Wigfield and Guthrie (1997) alphas and the subscales (podcasts, videos, quizzes, and games) had reasonable reliabilities ranging from 0.70 to 0.77 (see Table 3).

Table 3. Reliabilities for the interview subscales.

Subscale No. of Items Reliability Podcast 1 0.74 Video 1 0.70 Quiz 1 0.72 Game 1 0.77

RESULTS AND DISCUSSION

In response to the first research question, paired samples t-test was conducted to compare the listening scores of the experimental and control groups from pretest to posttest (see Table 4). The results of the study showed that the control group had a partially better listening record M=66 compared to the experimental group M=64.22. It can be argued that there was no significant difference in scores of the experimental (M=63.92, SD=3.35) and the control (M=65.14, SD=3.12) groups on the listening pre-test; t=0.618, p=0.518. This shows the partially equivalent listening abilities of the participants before the experiment. However, the experimental group (M=92.21, SD=4.97) displayed significant performance over the control group (M=65.15, SD=2.35) on the listening post-test; t=-23.20, p=0.000. Based on the obtained results it can be argued that students who received educational intervention (i.e. using CALL affordances under the FonF practice model) developed more prominently in listening abilities than those who received non-CALL instruction. It’s worth mentioning that during the posttest more confidence and less anxiety was observed among the experimental-group members. This can be interpreted as the effectiveness of

RLyLA Vol. 14 (2019), 45-57 | 49 Akbar Bahari The Impact of Applying the FonF Practice Model on Developing L2 Listening and Speaking with a Focus on Intentional and Incidental Vocabulary Acquisition in CALL Context practicing CALL affordances under the FonF practice model’s strategies to develop listening proficiency. The study confirms that the experimental group handled listening situations more efficiently and confidently compared to the control group following the educational intervention.

Table 4. Listening paired samples statistics Paired Samples Statistics Mean N Std. Deviation Std. Error Mean Pair 1 Ex-group Listening Pretest 64.22 70 3.358 0.430 Listening Posttest 92.21 70 4.975 0.765 Pair 2 Con-group Listening Pretest 66.14 23 3.121 0.665 Listening Posttest 65.15 23 2.350 0.565

Table 5 shows that following the educational intervention, the experimental group (M=92.21, SD=4.97) displayed significant performance in comparison to the control group (M=65.15, SD=2.35) on the listening post- test; t=-23.20, p=0.000. Therefore, based on the obtained results, the study confirms that students who received the special instruction based on the FonF practice model developed more prominently in listening abilities than those who received non-CALL instruction.

Table 5. Listening paired samples test. Paired Differences 95% Confidence Std. Interval of the Std. Error Difference Sig. Paired Samples Test Mean Deviation Mean Lower Upper t df (2-tailed) Pair 1 Ex-group Listening Pre - Listening Post -26.295 7.308 0.936 -28.167 -24.423 -23.201 60 0.000 Pair 2 Con-group Listening Pre -Listening Post 0.682 2.147 0.458 -0.270 1.634 1.490 21 0.151

Table 6 shows that there was no significant difference in scores of the experimental (M=61.64, SD=4.34) and the control group (M=62.14, SD=3.22) groups on the speaking pre-test; t=0.818, p=0.418. This shows the partially equivalent speaking abilities of the participants before the experiment. However, the experimental group (M=93.14, SD=6.33) displayed significant performance over the control group (M=62.12, SD=3.71) on the speaking post- test; t=-30.63, p=0.000. Based on the obtained results it can be argued that students who received educational intervention (i.e. using CALL affordances under the FonF practice model) developed more prominently in speaking proficiency than those who received non-CALL instruction.

Table 6. Speaking Paired Samples Statistics. Paired Samples Statistics Mean N Std. Deviation Std. Error Mean Pair 1 Ex-group Speaking Pretest 61.64 70 4.346 0.569 Ex-group Speaking Posttest 93.14 70 6.337 0.809 Pair 2 Con-group Speaking Pretest 62.14 23 3.221 0.665 Con-group Speaking Posttest 62.12 23 3.710 0.576

Table 7 shows that following the educational intervention, the experimental group (M=93.14, SD=6.33) displayed significant performance over the control group (M=62.12, SD=3.71) on the speaking post-test; t=-30.63, p=0.000. Therefore, based on the obtained results, the study confirms that students who received the special instruction based on the FonF practice model developed more prominently in speaking abilities than those who received non- CALL instruction.

Table 7. Speaking Paired Samples Test.

Paired Differences 95% Confidence Std. Interval of the Std. Error Difference Sig. Paired Samples Test Mean Deviation Mean Lower Upper t df (2-tailed) Pair 1 Ex-group Speaking Pretest –Speaking -30.295 7.480 0.958 -32.211 -28.379 -30.634 60 0.000 Posttest Pair 2 Con-group Speaking Pretest _ Speaking 1.773 2.742 0.585 0.557 2.988 3.033 21 0.016 Posttest

RLyLA Vol. 14 (2019), 45-57 | 50 Akbar Bahari The Impact of Applying the FonF Practice Model on Developing L2 Listening and Speaking with a Focus on Intentional and Incidental Vocabulary Acquisition in CALL Context

Exploring the relationship between applying the FonF practice model and developing incidental vocabulary acquisition in the CALL context was achieved by comparing the differences between the pretest and posttest state of incidental vocabulary acquisition between the experimental and control group. According to the obtained results at the pretest stage of the study (see Table 8) there was no significant difference in scores of the experimental (M=20.80, SD=2.61) and the control (M=21.45, SD=1.18) groups on the pretest incidental vocabulary acquisition test; t=0.718, p=0.818. This shows the equivalent state of incidental vocabulary acquisition among the participants before the experiment. However, the experimental group (M=26.69, SD=1.69) displayed significant performance over the control group (M=22.45, SD=1.81) on the posttest incidental vocabulary acquisition test; t=-20.26, p=0.000. Based on the obtained results it can be argued that students who received educational intervention based on the FonF practice model developed more prominently in incidental vocabulary acquisition than those who received other teaching schedules.

Table 8. Paired Samples Test Incidental Vocabulary Acquisition (VA).

Paired Differences 95% Confidence Std. Interval of the Std. Error Difference Sig. Paired Samples Test Mean Deviation Mean Lower Upper t df (2-tailed) Pair 1 Ex-Group Pretest-posttest -5.885 2.374 0.304 -6.493 -5.277 -20.261 60 0.000 Incidental VA Pair 2 Con-Group Pretest-Posttest 1.000 1.414 0.302 0.373 1.627 3.317 21 0.003 Incidental VA

Exploring the relationship between applying the FonF practice model and developing intentional vocabulary acquisition in a CALL context was done by comparing the differences between pretest and posttest state of intentional vocabulary acquisition between the experimental and control group. According to the obtained results at pretest stage, the experimental group with M=23.05 resembles the control group with M=23.5 in terms of the intentional vocabulary acquisition. However, there is a significant rise in the experimental group’sintentional vocabulary acquisition (M=37.15) following the administration of the FonF practice model-oriented educational intervention under a CALL context. The analysis of results revealed that there was no significant difference in scores of the experimental (M=23.05, SD=2.83) and the control (M=23.09, SD=2.22) groups on the pretest intentional vocabulary acquisition test; t=0.634, p=0.718. This shows the equivalent intentional vocabulary acquisition of the participants before the experiment. However, the experimental group (M=37.15, SD=2.63) displayed significant performance over the control group (M=24.09, SD=2.56) on the posttest intentional vocabulary acquisition; t=- 33.51, p=0.000. Based on the obtained results it can be argued that students who received educational intervention based on the FonF practice model developed more prominently in intentional vocabulary acquisition than those who received other L2 teaching instructions.

Table 9. Paired Samples Test Intentional Vocabulary Acquisition (VA).

Paired Differences 95% Confidence Std. Interval of the Std. Error Difference Sig. Paired Samples Test Mean Deviation Mean Lower Upper t df (2-tailed) Pair 1 Ex-group pretest-posttest -14.098 3.355 0.430 -14.958 -13.239 -33.519 60 0.000 Intentional VA Pair 2 Con-group pretest-posttest -1.000 2.047 0.436 -1.908 -0.092 -2.291 21 0.032 Intentional VA

The total M=92.17 of the elicited responses are strongly agree, agree, or slightly agree, which serves as evidence of the success of the suggested strategies to improve listening, speaking, and intentional and incidental Vocabulary Acquisition via CALL affordances under the FonF practice model. Such a high number of positive opinions on the efficiency of the suggested strategies not only reflects the perceived convenience (i.e. perceived usefulness and perceived ease of use) on the part of the learner, but also calls for more rigorous attention on the side of the scholars to delve more into the applicability of this model as part of general CALL instruction and its applicability as a practice model for other CALL affordances and apps.

RLyLA Vol. 14 (2019), 45-57 | 51 Akbar Bahari The Impact of Applying the FonF Practice Model on Developing L2 Listening and Speaking with a Focus on Intentional and Incidental Vocabulary Acquisition in CALL Context

Table 10. Questionnaire’s results on the effectiveness of the FonF practice model.

Affordances Level Videos (%) Podcasts (%) Games (%) Quizzes (%) Strongly agree 41 32 31 29 Agree 37.5 41 39 40 Slightly agree 14 21 23 21 Slightly disagree 3.5 2 3 6 Disagree 3.5 0.5 0.5 0.5 Strongly disagree 0.5 0.5 0.5 0.5

The results of the interviews administered among the experimental group at the end of the study were mixed and keyed into SPSS 22 in a 4-step Likert continuum (ranging from 1 very effective to 4 very ineffective). The results of the analysis revealed that the majority of the respondents had positive opinions on the efficiency of the administered educational intervention for developing listening, speaking and incidental and intentional vocabulary acquisition. It is worth mentioning that not a single very ineffective response was observed in the collected data which shows that, despite the deficiencies of the suggested model, there is a 90% plus consensus on the effectiveness of the FonF practice model to develop listening, speaking andintentional and incidental vocabulary acquisition. The triangulation of obtained data revealed more tendency on the part of the female participants of the study towards the FonF practice model compared to male participants which is consistent with the findings reported by previous studies for the significance of the relationship between gender and motivation (Ivey, 1999; Shapiro & Whitney, 1997). To capture different dimensions of the proposed model, methodological triangulation of the data was conducted with respect to the research questions. The triangulation of the elicited data from qualitative and quantitative methods supported the validity of the suggested strategies. This finding can serve as evidence of the conceptualization of the FonF practice model and the rationale to unlock the potential behind CALL affordances with a focus on incidental andintentional vocabulary acquisition.

PEDAGOGICAL IMPLICATIONS OF THE STUDY

The main pedagogical implication of the study is the effectiveness of using the FonF practice model to develop listening, speaking and intentional and incidental vocabulary acquisition via CALL affordances. Pedagogically, the proposed model with a focus on form facilitates learning in keeping with the prevalent trend of CALL as described by Clifford and Granoien (2008) where learning is considered as an informational construct. Accordingly, the findings of the study have important implications for English language teachers who avoid CALL affordances for a variety of reasons such as the lack of an applicable model with a focus on language skills. The use of the FonF practice model in CALL context, not only expands learners’ in-class and out-of-class exposure to authentic language which ensures sustainable learning (Tsou, Wang, & Tzeng, 2006), but also caters for a diverse range of motivational factors among the learners which creates a learner-friendly context (Bahari, 2019b).

FINAL THOUGHTS

Given the absence of an applicable practice model for the CALL context with a focus on nonlinearity and dynamicity, this nine-stage model was proposed to facilitate incorporating CALL affordances, particularly podcasts, videos, games, and quizzes to develop listening-speaking with a focus on intentional and incidental vocabulary acquisition. The results of the study confirmed the effectiveness of the model for developing listening-speaking proficiency as well as intentional and incidental vocabulary. This study has contributed to the field by offering an applicable practice model which can be adjusted with respect to any CALL affordances prepared for L2 teaching while catering for the nonlinear dynamic nature of motivation in L2 learners. The triangulation of the data revealed that CALL users express high levels of self-confidence and self-efficacy and lower anxiety when they are in charge of selecting the learning materials in line with their individual motivational factors which is strongly suggested under the FonF practice model. Besides that, practicing motivationally-selected contents via CALL affordances proved to be very effective for developing listening, speaking, and incidental and intentional vocabulary acquisition in the CALL context. Future studies are suggested to test the applicability of the FonF practice model for other aspects of L2 teaching-learning via CALL affordances and apps.

ACKNOWLEDGMENT

The author would like to thank the anonymous reviewers for their insightful comments and suggestions to improve the paper. He also wishes to thank his wife and daughter, Mona and Anahita for their support.

RLyLA Vol. 14 (2019), 45-57 | 52 Akbar Bahari The Impact of Applying the FonF Practice Model on Developing L2 Listening and Speaking with a Focus on Intentional and Incidental Vocabulary Acquisition in CALL Context

REFERENCES

Arora, A., Evans, S., Gardner, C., Gulbrandsen, K, & Riley, J. E. (2015). Strategies for success: Using formative assessment to build skills and community in the blended classroom. In S. Koç, X. Liu, & P. Wachira (Eds.), Assessment in online and blended learning environments (pp. 235-251). Charlotte, NC: Information Age Publishing. Bahari, A. (2018a). Nonlinear dynamic motivation-oriented telecollaborative model of language learning via formulaic sequences to foster learner autonomy. The Journal of Teaching English with Technology, 18/3: 65-81, http://www.tewtjournal.org Bahari, A. (2018b). Sacred text motivation for general l2 learners: a mixed methods study. Journal of Academic Ethics, 16/4: 1-31. https://doi.org/10.1007/s10805-018-9316-3 Bahari, A. (2018c). Developing listening and speaking via a psycho-socio-cultural learning model based on non- linear dynamic motivation. Jordan Journal of Modern Languages and Literature, 10/2: 115-135. Bahari, A. (2019a). "FonF practice model from theory to practice: CALL via focus on form approach and nonlinear dynamic motivation to develop listening and speaking proficiency".Computers & Education, 130/3: 40-58. doi.org/10.1016/j.compedu.2018.11.009 Bahari, A. (2019b). "Psychological reactance management via nonlinear dynamic motivation in classroom and telecollaborative second language learning contexts". Journal of educational, health and Community Psychology, 8/1: 40-58. https://doi.org/10.1016/j.compedu.2018.11.009 Bygate, M. & V. Samuda (2005). Integrative planning through the use of task-repetition. In R. Ellis (ed.), Planning and task performance in a second language. Amsterdam: John Benjamins, 37-74. https://doi.org/10.1075/lllt.11.05byg Chang, M. M. (2005). "Applying self-regulated learning strategies in a web-based instruction: An investigation of motivation perception". Computer Assisted Language Learning, 18/3: 217-230. https://doi.org/10.1080/09588220500178939 Chiu, Y.-H. (2013). "Computer-assisted second language vocabulary instruction: A metaanalysis". British Journal of Educational Technology, 44/2: E52-E56. https://doi.org/10.1111/j.1467-8535.2012.01342.x Clifford, R. & Granoien, N. (2008). Application of technology to language acquisition processes: What can work and why. In M. V. Holland & F. P. Fisher (Eds.), The path of speech technologies in computer assisted language learning: From research to practice (pp. 25-43). London, England: Routledge Language Learning. Colpaert, J. (2018). "Exploration of Affordances of Open Data for Language Learning and Teaching".Journal of Technology and Chinese Language Teaching, 9/1: 1-14. Dashtestani, R. (2015). "Examining the use of web-based tests for testing academic vocabulary in EAP instruction". Teaching English with Technology, 15/1: 48-61. de Bot, K. (2008). "Introduction: Second language development as a dynamic process". The Modern Language Journal, 92/2: 166-178. https://doi.org/10.1111/j.1540-4781.2008.00712.x DeKeyser, R. (1998). Beyond focus on form: Cognitive perspectives on learning and practicing second language grammar. In C. Doughty & J. Williams (Eds.), Focus on form in classroom second language acquisition (pp. 42-63). Cambridge: CUP. Dörnyei, Z. & Ottó, I. (1998). "Motivation in action: A process model of L2 motivation". Working Papers in Applied Linguistics (Thames Valley University, London), 4: 43-69. Dumova, T. (2012). The usability of online quizzes: Evaluating student perceptions. In S. Kelsey, & K. St. Amant (Eds.), Computer mediated communication: Issues and approaches in education (pp. 50-61). Hershey, PA: IGI Global. https://doi.org/10.4018/978-1-61350-077-4.ch004 Ellis, R. (2016). "Focus on Form: A Critical Review". Language Teaching Research, 20/3: 405-428. https://doi. org/10.1177/1362168816628627 Fredricks, J. A., Blumenfeld, P. C., & Paris, A. H. (2004). "School engagement: Potential of the concept, state of the evidence". Review of Educational Research, 74/1: 59-109. https://doi.org/10.3102/00346543074001059 Ivey, G. (1999). "A multicase study in the middle school: Complexities among young adolescent readers". Reading Research Quarterly, 34: 172-192. https://doi.org/10.1598/RRQ.34.2.3 Larsen-Freeman, D. & Cameron, L. (2008). "Research methodology on language development from a complex systems perspective". Modern Language Journal, 92: 200-213. https://doi.org/10.1111/j.1540-4781.2008.00714.x Laufer, B. (2006). "Comparing Focus on Form and Focus on FormS in Second-Language Vocabulary Learning". The Canadian Modern Language Review, 63/1: 149-166. https://doi.org/10.3138/cmlr.63.1.149 Long, M. (2000). Focus on Form in Task-Based Language Teaching. In Language Policy and Pedagogy, edited by R. H. Lambert and E. Shohamy, 179-192. Amsterdam: John Benjamins. https://doi.org/10.1075/z.96.11lon

RLyLA Vol. 14 (2019), 45-57 | 53 Akbar Bahari The Impact of Applying the FonF Practice Model on Developing L2 Listening and Speaking with a Focus on Intentional and Incidental Vocabulary Acquisition in CALL Context

Lu, H.-L. (2009). Pre-class online quizzing as a catalyst for pedagogical change. SoTL Commons Conference. Paper 56. Macaro, E., Graham, S., & Vanderplank, R. (2007). A review of listening strategies: Focus on sources of knowledge and on success. In A. D. Cohen & E. Macaro (eds.), Language learner strategies: 30 years of research and practice. Oxford: Oxford University Press, 165-185. Nassaji, H. (1999). "Towards integrating form-focused instruction and communicative interaction in the second language classroom: Some pedagogical possibilities". Canadian Modern Language Review, 55: 385-402. https://doi.org/10.3138/cmlr.55.3.386 Nassaji, H. (2004). "The relationship between depth of vocabulary knowledge and L2 learners’ lexical inferencing strategy use and success". The Canadian Modern Language Review, 61: 107-134. https://doi.org/10.3138/ cmlr.61.1.107 Nassaji, H. (2016). "Anniversary article: Interactional feedback in second language teaching and learning: A synthesis and analysis of current research". Language Teaching Research, 20/4: 535-562. https://doi.org/10.1177/1362168816644940 Nassaji, H., & Fotos, S. (2007). Current issues in form-focused instruction. In S. Fotos, & H. Nassaji (Eds.), Form focused instruction and teacher education: Studies in honor of Rod Ellis (pp. 7-15). Oxford: Oxford University Press. Nassaji, H. & Fotos, S. (2011). Teaching Grammar in Second Language Classrooms: integrating form focused instruction in communicative context. New York: Taylor & Francis. https://doi.org/10.4324/9780203850961 Norris, J., & Ortega, L. (2000). "Effectiveness of L2 instruction: A research synthesis and quantitative meta- analysis". Language Learning, 50: 417-528. https://doi.org/10.1111/0023-8333.00136 Pawlak, M. (2006). The place of form focused instruction in the foreign language classroom. Kalisz-Poznan´: Wydawnictwo WPA UAM. Read, J. (2000). Assessing Vocabulary. Cambridge University Press, Cambridge, UK. https://doi.org/10.1017/ CBO9780511732942 Shapiro, J., & Whitney, P. (1997). "Factors involved in the leisure reading of upper elementary school students". Reading Psychology, 18: 343-370. https://doi.org/10.1080/0270271970180402 Sheen, R. (2002). "Focus on form and focus on forms". ELT journal, 56/3: 303-305. https://doi.org/10.1093/ elt/56.3.303 Tsou, W., Wang, W., & Tzeng, Y. (2006). "Applying a multimedia storytelling website in foreign language learning". Computers & Education, 47/1: 17-28. https://doi.org/10.1016/j.compedu.2004.08.013 Vidal, K. (2003). "Academic listening: A source of vocabulary acquisition?" Applied Linguistics, 24/1: 56-86. https://doi.org/10.1093/applin/24.1.56 Vidal, K. (2011). "A Comparison of the Effects of Reading and Listening on Incidental Vocabulary Acquisition". Language Learning, 61/1: 219-258. https://doi.org/10.1111/j.1467-9922.2010.00593.x VanPatten, B. (2004). Input and output in establishing form-meaning connections. In B. VanPatten, J. Williams, & S. Rott (Eds.), Form-meaning connections in second language acquisition (pp. 29-48). Mahwah, NJ: Erlbaum. Vygotsky, L. S. (1997). The collected works of L. S. Vygotsky. Volume 4: The history of the development of higher mental functions (R. W. Rieber, Ed.). New York, NY: Plenum Press. Wigfield, A., & Guthrie, J.T. (1997). "Relations of children’s motivation for reading to the amount and breadth of their reading". Journal of Educational Psychology, 89/3: 420-432. https://doi.org/10.1037/0022-0663.89.3.420 Williams, J. (2005). Form-focused instruction. In E. Hinkel (Ed.), Handbook on research in second language teaching and learning (pp. 673-91). Mahwah, NJ: Lawrence Erlbaum Associates. Wilson, A., Hainey, T. & Connolly, T. M. (2013). "Using Scratch with Primary School Children: An Evaluation of Games Constructed to Gauge Understanding of Programming Concepts". International Journal of Games- Based Learning, 3/1: 25-41. https://doi.org/10.4018/ijgbl.2013010107 Winke, P., Gass, S., & Sydorenko, T. (2013). "Factors influencing the use of captions by foreign language learners: An eye-tracking study". The Modern Language Journal, 97/1: 254-75. https://doi. org/10.1111/j.1540-4781.2013.01432.x

RLyLA Vol. 14 (2019), 45-57 | 54 Akbar Bahari The Impact of Applying the FonF Practice Model on Developing L2 Listening and Speaking with a Focus on Intentional and Incidental Vocabulary Acquisition in CALL Context

APPENDIX A

INCIDENTAL VOCABULARY ACQUISITION SCALE ADAPTED FROM VIDAL (2011)

The subjects from both experimental and control groups were asked to respond to the following prompts: 1 Have you heard/seen this word before? If so, where/when? 2 Provide a full explanation (in Persian or in English) of all the meanings of the word you know 3 Provide a Persian translation of the word 4 Make a sentence in English using the word Scoring Scale Point Knowledge of the word -1 recognizes a nonword 0 does not recognize the word 1 recognizes having seen/heard the word 2 has a vague/partial idea of the meaning of the word 2.5 has a vague/partial idea of the meaning of the word but produces a clear example, similar to the one in the video/game/quiz/podcast 3 shows a full understanding of the meaning of the word 4 shows a full understanding of the meaning of the word and is able to provide a Persian translation or use the word in a sentence 5 shows a full understanding of the meaning of the word and is able to provide a translation and use the word in a sentence

APPENDIX B

INTENTIONAL VOCABULARY ACQUISITION SCALE ADAPTED FROM LAUFER (2006)

L1-L2 Test (Active knowledge test) Points Knowledge of the word 2 completely correct form 1 the correct word with a spelling error that did not interfere with the recognition of the word 0 a blank or an incorrect word L2-L1 Test Points Knowledge of the word 2 correct translation/explanation 1 semantically approximate explanation/translation 0 incorrect translation or a blank

RLyLA Vol. 14 (2019), 45-57 | 55 Akbar Bahari The Impact of Applying the FonF Practice Model on Developing L2 Listening and Speaking with a Focus on Intentional and Incidental Vocabulary Acquisition in CALL Context

APPENDIX C

Before administering the questionnaire, the abbreviations (e.g., the FonF practice model, MWOD etc.) and any technical term (e.g. incidental and intentional vocabulary acquisition, etc.) that could cause ambiguities were explained to the participants of the study.

Table 15. Scales for statement-type items.

Focus on Statement Strongly disagree Disagree Slightly disagree Slightly agree Agree Strongly agree Podcast I believe that using podcasts provided by MWOD under the Listening FonF practice model is helpful for developing listening Podcast I believe that using podcasts provided by MWOD under the Speaking FonF practice model is helpful for developing speaking Podcast & I believe that using podcasts provided by MWOD under the incidental FonF practice model is helpful for developing intentional vocabulary vocabulary acquisition Podcast & I believe that using podcasts provided by MWOD under intentional the FonF practice model is helpful for developing incidental vocabulary vocabulary acquisition acquisition Quizzes I believe that using quizzes provided by MWOD under the Listening FonF practice model is helpful for developing listening Quizzes I believe that using quizzes provided by MWOD under the Speaking FonF practice model is helpful for developing speaking Quizzes & I believe that using quizzes provided by MWOD under the incidental FonF practice model is helpful for developing incidental vocabulary vocabulary acquisition acquisition Quizzes & I believe that using quizzes provided by MWOD under the intentional FonF practice model is helpful for developing intentional vocabulary vocabulary acquisition acquisition Videos I believe that using videos provided by MWOD under the Listening FonF practice model is helpful for developing listening Videos I believe that using videos provided by MWOD under the Speaking FonF practice model is helpful for developing speaking Videos & I believe that using videos provided by MWOD under the incidental FonF practice model is helpful for developing intentional vocabulary vocabulary acquisition acquisition Videos & I believe that using videos provided by MWOD under the intentional FonF practice model is helpful for developing incidental vocabulary vocabulary acquisition acquisition Games I believe that using games provided by MWOD under the Listening FonF practice model is helpful for developing listening Games I believe that using games provided by MWOD under the Speaking FonF practice model is helpful for developing speaking

Table 15, continues on the next page

RLyLA Vol. 14 (2019), 45-57 | 56 Akbar Bahari The Impact of Applying the FonF Practice Model on Developing L2 Listening and Speaking with a Focus on Intentional and Incidental Vocabulary Acquisition in CALL Context

Table 15, continues from the previous page

Focus on Statement Strongly disagree Disagree Slightly disagree Slightly agree Agree Strongly agree Games & I believe that using games provided by MWOD under the incidental FonF practice model is helpful for developing incidental vocabulary vocabulary acquisition acquisition Games & I believe that using games provided by MWOD under the intentional FonF practice model is helpful for developing intentional vocabulary vocabulary acquisition acquisition

APPENDIX D

INTERVIEW

1. Which of the following computer-assisted affordances (i.e. games, quizzes, podcasts, and videos) practiced under the FonF practice model were more helpful for developing you listening proficiency? Why?

2. Which of the following computer-assisted affordances (i.e. games, quizzes, podcasts, and videos) practiced under the FonF practice model were more helpful for developing your speaking proficiency? Why?

3. Which of the following computer-assisted affordances (i.e. games, quizzes, podcasts, and videos) practiced under the FonF practice model were more helpful for developing your incidental vocabulary acquisition? Why?

4. Which of the following computer-assisted affordances (i.e. games, quizzes, podcasts, and videos) practiced under the FonF practice model were more helpful for developing your intentional vocabulary acquisition? Why?

APPENDIX E

Sample of Videos used as the CALL tools by the students from Merriam Webster to develop their incidental and intentional vocabulary

RLyLA Vol. 14 (2019), 45-57 | 57 Revista de Lingüística y Lenguas Aplicadas Vol. 14 año 2019, 59-79 EISSN 1886-6298 https://doi.org/10.4995/rlyla.2019.11007

THE SYNTACTIC PARSING OF ASD-STE100 ADVERBIALS IN ARTEMIS

Francisco J. Cortés-Rodríguez Universidad de La Laguna, Spain Carolina Rodríguez-Juárez Universidad de Las Palmas de Gran Canaria, Spain

Abstract: This paper seeks to offer an update of the syntactic representation of adverbials in the Lexical Constructional Model and the multipurpose lexico-conceptual knowledge base, FunGramKB, and to implement the conditions that are necessary for an effective parsing of such constituents within ARTEMIS, a NLP prototype within FunGramKB for the computational treatment of the syntax and semantics of sentences. As regards the status of adverbials in the Layered Structure of the Clause (LSC) proposed in Role and Reference Grammar, it is striking that they have been somewhat sidelined, despite the programmatic proposal in Van Valin (2005) for a distribution of adjuncts along the different layers in the LSC. Since then, almost no further contribution has been offered to fully expand this new scheme. Thus, our aim is to fully embody a layered approach for adverbials and to adapt our proposal to the conditions imposed by the Grammar Development Environment in ARTEMIS, by developing the rules that are necessary for the parsing of the catalogue of adverbials of a corpus written in the controlled natural language ASD-STE100.

Keywords: adjuncts, FunGramKB, ARTEMIS, parsing rules, simplified technical English.

1. INTRODUCTION

As it is already well established from a variety of studies (Periñán-Pascual, 2012, 2013; Fumero & Díaz, 2017; Mairal-Usón, 2017), for natural language processing (NLP) systems to be able to operatively and accurately develop NLP tasks, they have to be based on robust linguistic models which provide text processing applications with the necessary and adequate linguistic input for the parsing and production of utterances. This is the case of the knowledge base FunGramKB (Periñán-Pascual, 2013; www.fungramkb.com), which is “a user-friendly online environment for the semiautomatic construction of a multipurpose lexico-conceptual knowledge base for NLP systems” (Mairal-Usón, 2017:243). FunGramKB is grounded in two solid linguistic models, namely the functional theory of Role and Reference Grammar (RRG) (Van Valin & LaPolla, 1997; Van Valin, 2005) and the Lexical Constructional Model (LCM) (Ruiz de Mendoza, 2013; Ruiz de Mendoza & Galera, 2014), both of which contribute to the development of the lexical and grammatical modules of the knowledge base (see Figure 1). The lexical module (Lexicon) is based on the formalisms of RRG as regards the lexical representation of predicates (for example, Aktionsart ascription, macrorole assignment, status of variables, logical structures, etc.), whereas the grammatical module (Grammaticon) relies on the constructionist approach of the LCM for the classification of constructions into four different layers: argumental, implicational, illocutionary and discursive. FunGramKB also includes a language independent conceptual module which encodes deep conceptual semantic knowledge in the Onomasticon (encyclopaedic knowledge), the Cognicon (procedural knowledge) and the Ontology (hierarchical catalogue of concepts).

The scope of this paper is restricted to the study of adverbials in the English language and we seek, firstly, to offer an update of the syntactic representation of adverbials in the LCM and in FungramKB. Notwithstanding their centrality in any grammatical model, the status of adverbials in RRG has received, in our opinion, too little attention. Despite the radical reanalysis of such constituents proposed in Van Valin (2005), the impact of this proposal has been so scarce that most subsequent studies within the model still maintain the original single-layer analysis described in Van Valin and LaPolla (1997) (with some exceptions as are Toratani (2007) and Mora-Bustos

To cite this article: Cortés-Rodríguez, F.J., and Rodríguez-Juárez, C. (2019). "The syntactic parsing of ASD-STE100 adverbials in ARTEMIS". Revista de Lingüística y Lenguas Aplicadas, 14, 59-79. https://doi.org/10.4995/rlyla.2019.11007 Correspondence authors: [email protected], [email protected]

Received: 2018-11-18 Accepted: 2019-03-13 | 59 Francisco José Cortés-Rodríguez and Carolina Rodríguez-Juárez The syntactic parsing of ASD-STE100 adverbials in ARTEMIS

(2012)). Only recently has the work by Díaz-Jorge (2017) offered an extensive proposal for the analysis of adjuncts based on the programmatic description of Van Valin (2005).

Figure 1. The FunGramKB architecture (www.fungramkb.com).

Secondly, we aim to implement the conditions which are necessary for an effective parsing of adverbials within ARTEMIS. ARTEMIS stands for “Automatically Representing Text Meaning via an Interlingua-based System” and is a parsing device that incorporates RRG analytical tools and that FunGramKB exploits for the automatic generation of the syntactic and semantic representation of linguistic expressions in the form of a conceptual logical structure (CLS), which is an enhanced representation of RRG logical structures (Periñán-Pascual & Arcas-Túnez, 2014). We also aim to provide the rules that are needed for the computational parsing of adverbials and which are stored in the Grammar Development Environment (GDE) module within ARTEMIS. Since the computational workability of ARTEMIS is still to be tested, we have chosen to apply it to a Controlled Natural Language, namely ASD-STE100, with the assumption that it will help to validate the performance of our parser. Hence, the scope of our analysis will be confined to the catalogue of adverbials that can be found in the Airbus corpus written in this controlled language. This will necessarily have to be taken into account in the set of rules that we have devised, as the range of adverbials in such a controlled language is expected to be more restricted than that found in natural English.

We have organised the rest of this paper in 4 sections. Section 2 deals with the description of adverbials based on different innovative proposals in order to obtain an appropriate framework for the syntactic parsing of adverbials in ARTEMIS. The third section offers an overview of the methodological adaptations required for an effective formalization of adverbials in ARTEMIS. In doing so, special attention will be paid to the following issues: the integration of an intermediate constructional node between the core and the clause nodes in the layered structure of the clause (namely, an L1-CONSTR), and the replacement of the operator projection by feature-bearing matrixes and unification mechanisms. The description of the positional and peripheral preferences of adverbials in Airbus ASD-STE100, together with the set of rules that we have formulated, will be offered in the fourth section of this paper. Finally, some concluding remarks will be presented in Section 5.

2. ADVERBIALS IN RRG AND THE LCM: A VERY BRIEF HISTORY

2.1. RRG Syntactic Representations The Lexical Constructional Model (LCM) syntactic representation of sentences is primarily based on the Layered Structure of the Clause (LSC) as proposed in Role and Reference Grammar (RRG) (Van Valin & LaPolla, 1997; Van Valin, 2005; Pavey, 2010), but incorporates some variations motivated by the integration of constructional structures.

The LSC aims at capturing both universal and language-specific aspects of syntactic structures. With respect to universal features, two basic distinctions are considered: the first ones accounts for the difference between predicating elements and non-predicating elements, whereas the second concerns those elements that are arguments of the predicate and those which are not. This second opposition defines three syntactic units in the structure of the clause: the nucleus (which includes a verbal, an adjectival or a nominal predicate), the core (which

RLyLA Vol. 14 (2019), 59-79 | 60 Francisco José Cortés-Rodríguez and Carolina Rodríguez-Juárez The syntactic parsing of ASD-STE100 adverbials in ARTEMIS

contains the nucleus and its arguments), and the periphery (which includes constituents that are not predicate arguments). This is represented in the so-called Constituent Projection, as shown in Figure 2:

SENTENCE

CLAUSE

CORE PERIPHERY

(ARG) (ARG) NUCLEUS

PRED

XP XP X(P) XP/ADV

Figure 2. Formal representation of the LSC (Constituent Projection) (from Van Valin & LaPolla, 1997:31).

Concerning language-specific aspects, the Constituent Projection also incorporates two additional positions, the extra-core and the detached positions, which are both pragmatically motivated. Thus, in languages like English, fronted constituents and interrogative elements in questions occupy the PreCore Slot, whereas detached constituents are separated by a pause from the rest of the clause, a sign of their markedness for pragmatic purposes (Figure 3).

SENTENCE

(LDP) CLAUSE (RDP)

(PrCS) CORE (PoCS)

(ARG) (ARG) NUCLEUS

PRED

XP XP XP XP X(P) XP XP Figure 3. General format of the Constituent Projection (LSC) with extra-core and detached positions (from Van Valin & LaPolla, 1997:38)

Figure 4 is a representation of the Constituent Projection analysis of the sentence “Yesterday, what did Robin show to Pat in the library?” (Van Valin & LaPolla, 1997:36):

SENTENCE

LDP CLAUSE

PrCS CORE PERIPHERY

ARG NUC ARG

NP PRED PP

ADV NP V PP

Yesterday, what did Robin show to Pat in the library?

Figure 4. The LSC of a clause in English (Van Valin & LaPolla, 1997:36).

RLyLA Vol. 14 (2019), 59-79 | 61 Francisco José Cortés-Rodríguez and Carolina Rodríguez-Juárez The syntactic parsing of ASD-STE100 adverbials in ARTEMIS

With regard to the status of adverbials, they are described as peripheral elements, acting as modifiers of the CORE node, as shown in Figures 2 and 4. It is interesting to note that the periphery node is not analysed as a daughter of any other node, but is related to the core by an arrow, a sign of its status as modifier of this layer.

It should be highlighted that the Constituent Projection of the LSC only provides a syntactic analysis of content units (words and phrases); function words such as auxiliaries and grammatical morphemes are analysed as operators in a different projection within the LSC. Operators are grammatical categories like aspect, tense or illocutionary force and modify different layers of the clause. Since they are technically not part of the nucleus, core or periphery, but rather modify these layers, they are represented separately in a different projection within the LSC, as proposed by Johnson (1987) in the following projection grammar formalization (Figure 5):

V Aspect NUCLEUS Negation Directionals

Directionals CORE Event quant Modality Negation

Status Tense CLAUSE Evidentials Illocutionary SENTENCE Force

Figure 5: The Operator Projection (adapted from Van Valin, 2005:12).

A detailed syntactic description of a clause will then merge both the Constituent and Operator Projections, thus obtaining a fully detailed LSC analysis, as represented in Figure 6:

SENTENCE

(LDP) CLAUSE (RDP)

(PrCS) CORE (PoCS)

(ARG) (ARG) NUCLEUS

PRED

(XP) (XP) (XP) (XP) V (XP) (XP)

NUCLEUS Aspect NUCLEUS Negation NUCLEUS/CORE Directionals CORE Event quantification CORE Modality CORE Negation (Internal) CLAUSE Status CLAUSE Tense CLAUSE Evidentials CLAUSE Illocutionary Force SENTENCE

Figure 6. The Layered Structure of the Clause (Constituent and Operator Projections) (Van Valin, 2005:12).

RLyLA Vol. 14 (2019), 59-79 | 62 Francisco José Cortés-Rodríguez and Carolina Rodríguez-Juárez The syntactic parsing of ASD-STE100 adverbials in ARTEMIS

Figure 7 shows the LSC analysis of an English sentence:

SENTENCE CLAUSE CORE NP NUC PRED V Will they have to be leaving?

V

ASP NUC

MOD CORE

TNS CLAUSE

IF CLAUSE

SENTENCE Figure 7: English clause structure (Van Valin, 2005:14).

2.2. The syntax and semantics of adverbials in RRG As has been described in the preceding section, optional constituents are described in Van Valin & LaPolla (1997) as peripheral elements modifying the core layer in the Constituent Projection within the LSC. However, in Van Valin (2005:19-21) there is an important change as regards the description of such peripheral constituents since it is proposed that adverbials should be distributed as modifiers of the three different layers of the clause, namely, the nucleus, core and clause layers. Thus, not only must a core periphery be contemplated in the syntactic analysis, but also a nuclear periphery and a clause periphery. Van Valin (2005:19-20) claims that aspectual adverbs like completely and continuously modify the nucleus; pace and manner adverbials together with time and place adjuncts are core modifiers; andepistemic and evidential adverbs (like probably and obviously, respectively) modify the clause. The example (Van Valin, 2005:22) in Figure 8 illustrates the three-layered analysis of adverbs:

SENTENCE

PERIPHERY CLAUSE

PERIPHERY CORE

NP NUC PERIPHERY PP ADV NP ADV PRED ADV

V

Leslie has evidently been slowly immersing herself completely in the new language.

V

NUC ASP

ASP NUC

TNS CLAUSE

IF CLAUSE

SENTENCE Figure 8. The LSC of an English sentence including peripheral elements (Van Valin, 2005:22).

Despite the relevance of this new analysis, it is significant that its impact within the RRG model has been very scarce. A revision of the recent literature in this model reveals that most studies still maintain the original single- layer analysis of adverbs as core peripheral constituents, thus neglecting the new programmatic nature of the proposal in Van Valin (2005) for a distribution of adjuncts along the different layers in the LSC. To our knowledge, only the work by Mora-Bustos (2012) has applied the new proposal to the description of adverbials in Spanish. Even though Mora-Bustos (2012) adds some more types of adverbs to the brief classification proposed in Van Valin (2005) – mainly because of the type of semantic information that they convey – we believe that this work does not provide a fully-embodied account of the typology of optional constituents within the new RRG framework.

RLyLA Vol. 14 (2019), 59-79 | 63 Francisco José Cortés-Rodríguez and Carolina Rodríguez-Juárez The syntactic parsing of ASD-STE100 adverbials in ARTEMIS

Díaz-Jorge (2017) seems to be a more solid proposal for a complete implementation of the three-layered distribution of adverbials along the LSC, as it expands the typology and proposes a hierarchical arrangement of these optional elements by applying different criteria to support her classification. Díaz-Jorge (2017:13-14) explains that despite the fact that Van Valin’s (2005) new system of classification for adjuncts is a turning point as regards the status of these constituents in RRG, it still remains insufficient in order to give a complete description of the syntax of adjunct modifiers. Therefore, she proposes: (i) to establish a detailed typology of adverbials in English along the lines of this new analysis, and (ii) to establish some criteria to support the distribution of adverbials as peripheral modifiers along the three layers in the enhanced LSC.

Concerning the first issue, Díaz-Jorge (2017) offers a taxonomy of adverbials by combining three different sources: the types mentioned in Van Valin (2005), the typology of the so-called satellites from Dik’s (1997) Functional Grammar (FG), and the classification ofadjuncts and disjuncts from Quirk et al. (1985)1. Her classification is captured in Table 1:

Layers Adjunct Types Subtypes Examples CLAUSE Illocutionary briefly, frankly Evidential reportedly, presumably Epistemic cleverly, certainly CORE Concession despite X, in spite of X Contingency Purpose for X, in order to X Reason because of X, on account of X Company with X Process Beneficiary for X Manner silently, recklessly Means/Instrument (travel) by car, (shoot) with a rifle Pace quickly, slowly Direction towards X Space Path along X, across X Source from X Time position today, tonight Temporal Span since X, by X, recently Duration for X time NUCLEUS Focusing Restrictive/Limiters alone, just, only Additives further, also, even Degree still, completely Frequency Definite frequency hourly, once, twice Indefinite 1(High frequency) often, regularly Indefinite 1 (Usual/Mid frequency) on most days, generally, normally Indefinite 2 (Low frequency) hardly ever, seldom, scarcely Indefinite 2 (Zero frequency) never, no longer, nevermore Indefinite 3 (Continual/ Universal frequency) always, continually

As can be seen in Table 1, adjuncts are already distributed among the three layers of the LSC. The methodology developed by Díaz-Jorge (2017) in order to obtain this classification is based on two types of criteria, namely the semantic import of adverbials and their positional behaviour. With reference to the semantic aspects, this work establishes a correlation between the meaning of the different types of adjuncts and the operators of the LSC, in a similar fashion to the approach followed in the description of satellites in Functional Grammar (Dik, 1997). In FG, satellites are always depicted in lockstep with the operators so that both types of constituents are allocated to one layer by considering their semantic parallelism. In other words, both satellites and operators are considered as bearers of the same type of additional information pertaining to one of the layers in the semantic contour of clauses, their difference being morpho-lexical: satellites are lexical, and phrasal modifiers and operators are closed-class lexical units or morphemes. Van Valin (2005:19-20) also recognizes that “[a]dverbs are not operators, however,

1 Both Van Valin (2005) and Díaz-Jorge (2017) dispense with the distinction adjunct-disjunct and consider both as adjuncts; we will also use this as a cover term.

RLyLA Vol. 14 (2019), 59-79 | 64 Francisco José Cortés-Rodríguez and Carolina Rodríguez-Juárez The syntactic parsing of ASD-STE100 adverbials in ARTEMIS

but do interact with operators”. Díaz-Jorge (2017) applies this same rationale and establishes the equipollence between every semantic type of adverbial and a corresponding operator in the LSC. In fact, to ensure a proper equivalence between adverbials and operators, Díaz-Jorge (2017) not only considers the description of these units in RRG, but also pays much attention to the richer semantic descriptions offered by Quirket al. (1985) and Dik (1997). The latter brings about a further methodological difficulty that she manages to overcome: finding the correspondence between the three-layer analysis of clauses in RRG and the four-tier description of FG. Another consequence of this approach is the fact that reason, purpose and concessive adjuncts are classified as core adverbials, whereas Van Valin (2005:195-196) proposes taking them as clause peripheries. Van Valin’s analysis is based on the ordering co-occurrence restrictions that purpose and concession clauses have with regard to other core peripheral subordinate clauses: all these types appear consistently in clause final position but there is a preference to locate purpose and concession subordinate clauses after setting locative and temporal clauses in English. He further extends this classification to allpurpose and concessive adjuncts. We believe, however, that these restrictions are not enough to justify the analysis proposed for these types of adverbials, but merely show some ordering restrictions which occur within the group of core adjuncts, a phenomenon similar to that taking place within certain operators from the same layer.

Díaz-Jorge (2017) also uses the positional criterion in order to confirm the distribution of adjuncts along the LSC. By paying attention to the more frequent and, therefore, more natural positions in which adjuncts (especially adverbs) tend to occur, and their relative flexibility to appear in several positions, it was possible to further argue that they were modifiers of one layer or another. Van Valin (2005:21) refers to this criterion when he states that “unlike operators, which have fixed positions, adverbs may occur either before or after the verb, but in both cases the scope constraints require that the nuclear adverb be closer to the verb than the core adverb, and likewise for the core adverb with respect to the clausal adverb”. Following this line of reasoning, Díaz-Jorge (2017:18-19) draws some positional tendencies from her analysis based on an extensive sample of adverbials obtained from the British National Corpus (BNC), which can be summarized as follows:

• Nuclear adjuncts: all those adjuncts which tend to appear in mid position and are very rigid as it is not very common for them to appear in other spaces within the clause. For example, “You can regularly replace the fronts but not the straps” (BNC K6X), “She is reading a letter she has just written to her parents” (BNC FU4).

• Core adjuncts: all those adjuncts which tend to appear in end position and are very flexible as they are able to occur easily in other different positions as well. For example, “[…] a correspondingly green snake slithered silently across the wet patch.” (BNC AK6), “I’ve enjoyed my first year as a prodespite everything” (BNC C9E).

• Clausal adjuncts: all those adjuncts which tend to appear in initial positions and do not usually appear elsewhere. For example, “Frankly, until somebody finds a better way to build a GTi, we´re […]” (BNC CFS), “Presumably, it was felt that businessmen could take care of themselves” (BNC 6T).

Let us now consider how this classification can be integrated within ARTEMIS in order to automatically obtain an effective parsing of adjuncts in English.

3. ADVERBIALS IN ARTEMIS

Once the updated analysis of adverbials in RRG (and the LCM as well) has been attained (Section 2), this section describes how we have had to adapt this analysis to the conditions imposed by the Grammar Development Environment (GDE) in ARTEMIS, which necessarily implies the need to address two key issues: firstly, we will briefly comment on the adaptations that have had to be made to the Constituent Projection from RRG in order to integrate certain features from the LCM and FunGramKB, which are essential for the architecture of ARTEMIS and, specifically, of its Grammar Development Environment (GDE) module (Section 3.1). Secondly, we will have to devise in what way peripheries are to be integrated into the syntactic rules that form part of the GDE (Section 3.2).

3.1. A brief description of ARTEMIS and the Grammar Development Environment As already mentioned in the introductory section of this paper, ARTEMIS (“Automatically Representing Text Meaning via an Interlingua-based System”) is a NLP prototype that has been implemented within FunGramKB and is conceived as a syntactico-semantic parsing resource for the computational treatment of the semantics and syntax of linguistic expressions (Periñán-Pascual, 2013; Periñán-Pascual & Arcas-Túnez, 2014). ARTEMIS resorts to the shallow and deep semantic information that is stored in the different modules of FunGramKB (the lexical, grammatical and conceptual modules (see Section 1)) in order to understand fragments of natural language and

RLyLA Vol. 14 (2019), 59-79 | 65 Francisco José Cortés-Rodríguez and Carolina Rodríguez-Juárez The syntactic parsing of ASD-STE100 adverbials in ARTEMIS

bind them to their grammatical and semantic structure. In this process, three components are activated: the GDE, the Conceptual Logical Structure (CLS) constructor and the COREL-Scheme Builder (Figure 9):

Input Output > CLS representation > COREL Scheme à Re a s one r > Text Text

Grammar Development Conceptual Logical Structure (CLS) Constructor + Environment COREL Scheme Builder (GDE)

Figure 9. The architecture of ARTEMIS.

The GDE is responsible for decoding the morphosyntactic behaviour of predicates in the form of a parsed tree by applying the syntactic, constructional and lexical rules which are comprised within it. As explained in Cortés- Rodríguez (2016:81), the syntactic rules in the GDE “account for the generation/recognition of the underlying LSC”; constructional rules “guide the embedding of the structure of argument-predicate constructions (…) into the enhanced LSC”; and lexical rules “tokenize the abstract features encoded in the LSC by resorting to the information stored in the Lexicon and the Ontology of FunGramKB” (Cortés-Rodríguez, 2016:81).

In the next phase in the process, the shallow semantic representation of sentences is transformed into conceptually deeper structures in the CLS Constructor. As already mentioned, a CLS is an enhanced text meaning representation of RRG logical structures which includes Aktionsart ascription, the number of constructional variables of the predicate, thematic roles, macrorole functions (Actor/Undergoer), phrasal category, type of argument (argument, argument adjunct or a nucleus which contains the predicate), etc. The CLS is then turned into COREL (COnceptual REpresentation Language), which is the formal language that formalizes conceptual knowledge in FunGramKB (Cortés-Rodríguez, 2016:80; Fumero & Díaz, 2017). Example 1 by Fumero & Díaz (2017:38) illustrates this phase:

1. Louise had baked a cake for the kids.

CLS:

DECL PAST FBEN KER2 CACC >>>>> Extended COREL scheme:

+(e1: +BAKE_00 (x1: %LOUISE_00)THEME (x2: +CAKE_00)REFERENT (f1: (e2: +DO_00 (x1)AGENT (e1)REFERENT (f2: +CHILD_00)Beneficiary))Purpose)

As described in Periñán-Pascual & Arcas-Túnez (2014) and Cortés-Rodríguez & Mairal-Usón (2016), the RRG descriptive apparatus has undergone two necessary adaptations in the implementation process of ARTEMIS: (i) the integration of an intermediate constructional node, L1-CONSTR, in the layered structure of the clause between the CORE and the CLAUSE nodes; and (ii) the substitution of the operator projection by feature-bearing matrixes and unification mechanisms. The first modification is a direct consequence of the constructional orientation contributed by the Lexical Constructional Model (LCM) (Ruiz de Mendoza, 2013; Ruiz de Mendoza & Galera, 2014) in the design of both FunGramKB and ARTEMIS. According to the LCM, the syntactico-semantic contour of sentences is the joint result of a predicate-based approach (as the one from RRG) enhanced with the information from a repository of constructional structures (Constructicon) classified in terms of different levels of schematization. The four types of Constructicons are inspired in the four constructional layers of the LCM and are stored in the Grammaticon, which enriches FunGramKB in the sense that it can deal with both “the propositional and the non-propositional dimensions of meaning” (Mairal-Usón, 2017:246).

Therefore, the syntactic structure of sentences has to contemplate not only the projection of arguments from basic predicate frames in the Lexicon – described as kernel constructions in Periñán-Pascual (2013:214) – but also the possible modifications of such kernel frames (in terms of additions, reductions, shifts of constituents, or, even, as a result of differentAktionsart ascription) which involve the syntactic information pertaining to Level-1 (non-kernel) Constructions. Kernel constructions account for the configuration of the CORE in unmarked cases, as the core houses by default the arguments and the primary predicate of every clause. L1-Constructions, on the other hand, quite often introduce a secondary predicate (NUC-S), as, for instance, in resultatives; or an Argument- Adjunct (AAJ), as is the case of Beneficiary constituents like “for her sister” in “Ana caked a pumpkin pie for her

RLyLA Vol. 14 (2019), 59-79 | 66 Francisco José Cortés-Rodríguez and Carolina Rodríguez-Juárez The syntactic parsing of ASD-STE100 adverbials in ARTEMIS

sister”. There are even cases in which both types are simultaneously added as in “Bruno barks doors (AAJ) open (NUC-S)”, which is analysed in the following way (Cortés-Rodríguez & Mairal-Usón, 2016:95) (Figure 10):

SENTENCE

CLAUSE

CONSTR-L1

CONSTR-L1 AAJ NUC-S

CORE

ARG NUC

PRED

RP V RP MP

Bruno barks doors open

Figure 10: The enhanced LSC of an English sentence (Cortés-Rodríguez & Mairal-Usón, 2016:95).

The introduction of this new layer (CONSTR-L1) brings about some other changes, as is the case of the PreCore Slot Position (clause initial position), which has been re-described as a PreConstr-L1 position. In the case of adverbials, it seems feasible to reanalyze core adverbials as peripheral units modifying the CONSTR-L1 layer, according to the criteria proposed in Díaz-Jorge (2017): firstly, the semantic typology of L1-Constructional adjuncts would be the same as the one proposed for core adjuncts; secondly, their positional preferences are almost identical, as they tend to appear in clause final position2. Figure 11 illustrates the adaptations that have been incorporated as a result of the introduction of the CONSTR-L1 layer:

SENTENCE

CLAUSE

CONSTR-L1 PER_CONSTR_L1

PreConstr-L1 CONSTR-L1 AAJ NUC-S

CORE

ARG NUC ADJUNCT ADJUNCT

PRED

RP V RP MP RP PP

Who barks doors open every morning after its breakfast bone? Figure 11. Analysis of an English sentence with core adverbials as peripheral units modifying the CONSTR-L1 layer.

Arguably, there would be only one possible case to justify the distinction between core and L1-Construction adjuncts, and that is when an adverbial appears in core final position followed by another constituent introduced by a non-kernel L1-Construction, as in example (2):

2. [[I bake gluten-free bread] CORE (Kernel2) every Wednesday (Time Adjunct) for my celiac clientele (Beneficiary. AAJ)] - L1 CONSTRUCTION

2 This is especially true of prepositional phrase adjuncts; as explained in Van Valin (2005:21), they normally follow the core (or in our analysis, the L1-Construction), and, if they precede the core (or the L1-Construction), they must be in the PreC-L1 position or Left-detached position.

RLyLA Vol. 14 (2019), 59-79 | 67 Francisco José Cortés-Rodríguez and Carolina Rodríguez-Juárez The syntactic parsing of ASD-STE100 adverbials in ARTEMIS

However, we believe that this can be rather reinterpreted as a marked variant of the default final L1-Construction position (see Section 3.2 for a more detailed explanation).

The second methodological adaptation that is required for an effective formalization of adverbials in ARTEMIS consists of the replacement of the operator projection by feature-bearing structures called AVMs (Attribute-Value- Matrices), which encode the selectional and semantic information of different types of grammatical constituents in the format of attributes and values. This type of information cannot be retrieved from the Lexicon, the Grammaticon or the Ontology of FunGramKB (Cortés-Rodríguez, 2016:80-81). The AVM for the category Adjunct would be as follows:

The attributes for this category include: the type of Phrase, and in case it is a prepositional phrase, the type of preposition (Prep). An important attribute refers to Concept, which shall be retrieved from the Lexicon (in the case of adverbs) and will be crucial to ascertain the semantic type the adjunct belongs to. In the case of prepositional phrases, this attribute is obtained indirectly retrieved from the information of the corresponding preposition encoded in the Lexicon.

The AVMs for adverbs (ADV) and prepositions (PREP)3 are as follows:

The last attribute in the AVM for adjuncts (AdjunctRole) refers to the semantic type of adjunct; the list of available values for this attribute is encoded in its corresponding AVM, and is based on the classification proposed by Díaz-Jorge (2017) and presented in Section 2.2:

Additive Beneficiary Company Concession Definite frequency Degree Direction Duration Epistemic Evidential Illocutionary Indefinite frequency Instrument Limiter Manner Means Pace Path Purpose Reason Source Span Time position

3 Cf. Hernández-Pastor & Periñán-Pascual (2016) for an extensive description of the conceptual modelling of relations in FunGramKB’s Ontology, which accounts for the semantic intepretation of English prepositions.

RLyLA Vol. 14 (2019), 59-79 | 68 Francisco José Cortés-Rodríguez and Carolina Rodríguez-Juárez The syntactic parsing of ASD-STE100 adverbials in ARTEMIS

These values will also appear distributed in the syntactic rules that account for the realization possibilities of the different peripheries, thus restricting the array of possible semantic types admitted in each of them; the (partial) rule for Nuclear Peripheries would be as follows:

PER_NUC [Adjunct Role=Additive|Definite frequency|Degree|Indefinite frequency|Limiter]

àADJUNCT [Concept=[FIND: core > concept > concept | CHECK: %\w*], Phrase= [FIND: core > phraseset > tag], Prep= ?p, AdjunctRole= Additive|Definite frequency|Degree|Indefinite frequency|Limiter]

Figure 12 shows the refined tree of the LSC of an English sentence that includes the analysis of the additive adjunct also:

SENTENCE

CLAUSE

CONSTR-L1 PER_CONSTR-L1

CORE

ARG MODD PER NUC NUC ARG [AdRole:additive]

ADJUNCT PRED ADJUNCT [AdRole: additive]

RP MP V RP PP

The brake accumulator can also supply the Parking Brake System for 24 hours … Figure 12. Refined tree of the LSC.

3.2. Locating adjuncts in the GDE In order to develop the set of syntactic rules that form part of the Grammar Development Environment of ARTEMIS and that account for the ordering of constituents in those sentences that have optional adjuncts as constituents, it is necessary to further the explanation of the positional behaviour of adverbials offered in Díaz- Jorge (2017) by thoroughly identifying the different locations where they tend to appear within sentences, and, then, by showing the preferences for each type of periphery on a scale of positional markedness.

Following Greenbaum & Quirk (1990:161-162), and in addition to Diaz-Jorge’s (2017) basic clause positions (initial, medial and final positions), we have identified some subvariants of these three positions and distinguished the clause external detached positions that are also available for locating adverbials in English sentences. Figure 13 captures the set of possible positions available for the location of peripheral units:

SENTENCE

CLAUSE

CONSTR-L1

CORE

NUC

LDP, PrCONS-L1 PERIni ARG PERIniE AUX PERMed PRED PERMedE (ARG) (ARG) PERFinI (AJJ) (PRED-s) PERFin, RDP

PERIni: Initial-position PERiphery PERFinI: Final-position PERiphery PERIniE: Initial-position PERiphery (end variant) (initial variant) PERMed: Medial-position PERiphery PERFin: Final-position PERiphery PERMedE: Medial-position PERiphery (end variant) LDP (RDP): Detached positions4 Figure 13: Locating peripheries in the enhanced LSC in ARTEMIS.

4 On our scales of preferences we have only considered the LDP, as the RDP is subject to the same conditions of use as the LDP, though it seems to be less frequently used in English.

RLyLA Vol. 14 (2019), 59-79 | 69 Francisco José Cortés-Rodríguez and Carolina Rodríguez-Juárez The syntactic parsing of ASD-STE100 adverbials in ARTEMIS

We can now establish a scale of the positional preferences for the different types of adjuncts (especially adverbs; cf. note 2) described in Díaz-Jorge (2017) in terms of the positions established above (+ preferred > …>…>…> − preferred). On the scale, only the major positions (initial, medial, final and detached) are considered; therefore, subvariants are not taken as different positions. Hence, “by then” in the following sentences (Greenbaum & Quirk, 1990: 161) will be taken as occupying a medial position in all cases on the scale of preferences that we posit:

3. The book should by then (PERMed) have been returned to the library.

4. The book should have been returned by then (PERMedE) to the library.

This scale primarily refers to the position of adjuncts in declarative clauses, although it is also applicable to interrogatives and imperatives, with some minor considerations. For instance, in the case of imperatives, PERIni, PERIniE and PERMed are all neutralized since there is no intervening subject (ARG):

5. Only give these drinks with meals and dilute the juice by at least the same amount (BNC CH5).

Notice also that PERIniE and PERMed are neutralized in declaratives without an AUX constituent. On such occasions, the adverbial will be assigned to either of these two positions depending on its semantic type.

Clausal adjuncts will be analyzed as occupying PERIniE positions (example (6)), but CONSTR-L1 adjuncts will be parsed as PERMed adverbials (example (7)):

6. Although Mr Berghofer reportedly PERIniE said he was not competent […]. (BNC A4H)

7. Fergie and the Queen often rode together, but the troubled Duchess quickly PERMed found that her mother- in-law would shy away from anything too personal. (BNC CH1)

Manner adverbs in medial position can be sometimes interpreted as epistemic adjuncts; Van Valin (2005:20) explains that manner adverbs occurring before the tense operator can be construed as clausal modifiers. Thus, “cleverly” in sentence (8) has two interpretations (and possible positions): the first one shows that the manner in which she hid the cash was clever (hence the adverb occupies a PERMed position) whereas the second one shows that the fact that she hid the cash was clever.

8. Ruth cleverly hid the cash.

Notice, however, that, in our analysis, epistemic “cleverly” is occupying the PERIniE position. Only if there is an intervening tense auxiliary as in example (9), should the adjunct be analysed as occupying a PERMed position. This type of pattern of a clausal adjunct in mid position is quite infrequent.

9. The Security Council has wisely ruled out the death penalty. (BNC K5C)

Below we present the scales of positional preferences for the different types of adjuncts at each of the peripheral levels together with illustrating examples:

CLAUSAL PERIPHERY

ILLOCUTIONARY ADJUNCTS: PERIni > LDP

10. I frankly PERIniE don’t understand. (BNC J41)

11. Briefly, LDP there is nothing more I can do about it. (Quirk et al., 1985:616)

12. Totally unexpected because frankly PERIni no government, no civil service would have given us much money for (…). (BNC J9M)

EVIDENTIAL ADJUNCTS: PERIni > LDP

13. Although Mr Berghofer reportedly PERIniE said he was not competent […]. (BNC A4H)

14. In my experience, LDP such questions are seldom solved. (Dik, 1997:297)

15. What about the moralizing? Evidently PERIni there is something odd about the expression imperator noster divus Marcus. (BNC B2P)

RLyLA Vol. 14 (2019), 59-79 | 70 Francisco José Cortés-Rodríguez and Carolina Rodríguez-Juárez The syntactic parsing of ASD-STE100 adverbials in ARTEMIS

EPISTEMIC ADJUNCTS: PERIni > LDP > (PERMed)

16. Ashman wisely PERIniE reinserted the character into the film. (Web: thedissolve.com)

5 17. Ashman wisely PERMed reinserted the character into the film. (Web: thedissolve.com)

18. Fortunately, LDP we found him immediately. (Dik,1997:297)

19. Undoubtedly PERIni progress has been made, but it has been patchy. (BNC FT5)

CONSTR-L1 PERIPHERY

CONTINGENCY ADJUNCTS: PERFin > LDP > PERIni

20. They refused to leave on request and attempts to eject them failed on account of the sudden arrival of

reinforcements from the local station PERFin. (BNC ASB)

21. In order to redeem, LDP the mortgagor must give six months’ notice or pay six months’ interest. (BNC ABP)

22. Despite his personal battle against deafness PERIni he fought to help battered wives and rape victims. (BNC CH1)

PROCESS ADJUNCTS6: PERFin > PERMed /LDP > PERIni

23. Foley, of Finchley, north London, admitted driving recklessly PERFin while unfit through drink or drugs. (BNC CBF)

24. But London Ambulance Service categorically PERMed denied that it had plans to cut off phone lines. (BNC AAL)

25. Foolishly, LDP she had set them opposite each other. (BNC FBD)

26. I suddenly realised I had forgotten to bring my own umbrella to work that morning. Silently PERIni I cursed. (BNC HD7)

PACE ADJUNCTS: PERFin > PERMed /LDP > PERIni

27. Erm, it’s great having a neighbourhood area where all the traffic goesslowly PERFin. (BNC J44)

28. Fergie and the Queen often rode together, but the troubled Duchess quickly PERMed found that her mother- in-law would shy away from anything too personal. (BNC CH1)

29. Very hurriedly, LDP Charles tried to save a situation that threatened to end his whole career. (BNC ACE)

30. Slowly PERIni people are moving away from the marginal fringe. (BNC B1H)

SPACE ADJUNCTS: PERFin > PERMed /LDP > PERIni

31. Yes, I´ve got a show-stopper there PERFin. (BNC FU6)

32. Poor girl, what a sweet innocence is here PERMed destroyed. (BNC FU4)

33. Here, LDP to make amends for the fright I have put you in. (BNC FU4)

34. Outside PERIni it was a soft day -- and getting softer by the bucketful. (BNC CH1)

5 These examples ((16) and (17)) reveal that PERMed and PERIniE are neutralized in the absence of an AUX constituent. Compare it with the following variation, where the

neutralization dissolves: “Ashman has wisely PERMED reinserted the character into the film”. 6 Díaz-Jorge (2017:45-46) remarks the fact that, even though core (now CONSTR-L1) adverbials are usually in end positions, process adjuncts are very flexible and it is not strange for them to occupy middle positions. Pace, space and temporal adverbs can also occupy different places with relatively little stylistic or semantic effect (cf. Quirk et al., 1985:505, 519).

RLyLA Vol. 14 (2019), 59-79 | 71 Francisco José Cortés-Rodríguez and Carolina Rodríguez-Juárez The syntactic parsing of ASD-STE100 adverbials in ARTEMIS

TEMPORAL ADJUNCTS: PERFin > PERMed /LDP > PERIni

35. Today we are printing a special token to complete the voucher we gave you yesterday PERFin. (BNC CH1)

36. I think you should now PERMed begin to make another day in the week a happy one […]. (BNC FU4)

37. Nowadays, LDP there is no such thing as a bad job. (BNC BPE)

38. Yesterday PERIni we rang Frank at home only to be told by his mum: […]. (BNC CH1)

NUCLEAR PERIPHERY

FOCUSING ADJUNCTS: PERMed > PERFin > (PERIni)7

39. She is reading a letter she has justPERMed written to her parents. (BNC FUA)

40. Merely PERIni the experience of being governed was enough to provide a political education. (BNC HY9)

41. Who ever blamed Lucretia! All the shame lay on the ravisher only PERFin. (BNC FU4)

DEGREE ADJUNCTS: PERMed > PERFin

42. Costa Rica has partially PERMed recovered, though without solving its underlying economic problems. (BNC B12)

43. […] the pressure of the groundwater it contains forces apart the mineral grains and increases the volume

of the pores slightly PERFin. (BNC ANX)

DEFINITE FREQUENCY: PERFin > PERIni > (PERMed)

44. ELECTION fever is mounting hourly PERFin in the Stretchford Conurbation. (BNC AHN)

45. Once a week PERIni a ten-minute helicopter flight soars over the harbor. (BNC CH1)

46. A third would think twice PERMedE about moving home if it would upset their pets. (BNC CH5)

INDEFINITE FREQUENCY 1: PERFin / PERIni / LDP / PERMed

47. He used to help my brother out quite often PERFin. (BNC CH6)

48. Occasionally PERIni official syllabuses contain very detailed advice for teachers. (BNC BLY)

49. Occasionally, LDP the Princess would wince in horror as she heard how women often arrived. (BNC CH1)

50. […] that there will be dramatic changes in the lives of the favourite characters that have regularly

PERMed pulled in 16 million viewers over the last six years. (BNC CH1)

INDEFINITE FREQUENCY 2: PERMed > PERIni* > PERFin

51. Copeman of the Friends, told us that he seldom PERMed went out without seeing corpses lying about. (BNC A64)

52. Never before PERIni* has a news story triggered such sensational sales. (BNC CH1)

53. He was told he wasn’t wanted any longer PERFin. (BNC CH1)

7 It is very rare to findfocusing adjuncts in initial positions. Even more infrequent is it to finddegree adverbs in such a place in the sentence.

RLyLA Vol. 14 (2019), 59-79 | 72 Francisco José Cortés-Rodríguez and Carolina Rodríguez-Juárez The syntactic parsing of ASD-STE100 adverbials in ARTEMIS

INDEFINITE FREQUENCY 3: PERMed (>PERIni/LDP)

54. But Radio One will always PERMed be first and foremost a pop station. Just occasionally we will be more serious. (BNC CH1)

55. Continually PERIni adepts are warned that they must not strain themselves. (BNC CD4)

Frequency adjuncts are by far the most complex group in terms of positional preferences and restrictions. As a general rule, they tend to appear close to the predicate in rather rigid positions (Díaz-Jorge, 2017:23); furthermore, they are not expected to occupy clause external positions (i.e. left or right detached positions). However, different location tendencies can be identified which seem to be related to the specific semantic import of the adverbs. Because of this, we have established four subgroups, which correspond to a great extent to the distinctions posited in Díaz-Jorge (2017): firstly, we have differentiated betweendefinite and indefinite frequency adjuncts. Díaz-Jorge (2017:29) explains that definite frequency adjuncts (once, hourly, three times, weekly) behave differently from other frequency adverbials as they appear in initial and end positions; she refers to Ramaswamy’s (2007:349) words when he states that “they do not usually go in mid-positions”. We have marked this on the scale by placing PERMed at the end of the scale and between brackets; we proceeded similarly with initial positions in the case of focusing adjuncts.

Within the second group, three subtypes of indefinite frequency adverbials have been distinguished. In the first group, we have included the set offrequency adverbials that express usual (usually, normally), high (often, regularly) and mid frequency (sometimes, occasionally); the second group of indefinite frequency adjuncts will include those that convey a meaning of low (scarcely, seldom) to zero frequency (never, no longer). The two groups are distinguished because they have a different behaviour as regards their positional preferences. Group 1 adjuncts are characterized by being very flexible within the clause, whereas Group 2 are also flexible but they show some special features when in initial positions, as they trigger AUX inversion (which is marked with an asterisk on the scale). The third group includes adverbials that express continual or universal frequency (always, continually), which are unlikely to occur in positions other than mid positions.

Table 2 shows a summary of the adverbial positional preferences in each of the different peripheries:

Table 2. Scales of major positional preferences per periphery.

PERIPHERIES SCALES OF POSITIONAL PREFERENCES CLAUSAL PERIPHERY: PERIni > LDP > (PERMed) CONSTR-L1 PERIPHERY: PERFin > PERMed /LDP > PERIni NUCLEAR PERIPHERY (I): PERMed > PERFin /PERIni FOCUSING ADJUNCTS DEGREE ADJUNCTS INDEFINITE FREQUENCY 2 INDEFINITE FREQUENCY 3 NUCLEAR PERIPHERY (II): PERFin > PERIni / LDP / PERMed DEFINITE FREQUENCY INDEFINITE FREQUENCY 1

As can be observed, clausal peripheries quite often appear in initial positions and are frequent both in clause internal or clause external initial positions; they seem to be less likely to appear in final positions. Constr-L1 peripheries appear naturally in end positions, but they are not rare in clause external detached positions and medial positions; they are less frequently located in clause initial positions. This is especially true of contingency adjuncts, which are frequently realized by prepositional phrases and are normally placed at the end of the clause. As stated above, nuclear peripheral elements are more flexible, although two tendencies can be observed:focusing , degree and indefinite frequency group 2 (zero/low adverbs) and indefinite frequency group 3 (continual/universal) adverbials tend to be more rigid and quite often appear in medial positions; they can, however, also be located in final or initial clause internal positions, and some of them trigger inversion when placed initially. The rest of nuclear adjuncts, definite frequency and indefinite frequency group 1 (usual, high and mid frequency) adverbials are often in final positions, althoughindefinite group 1 adverbials are not infrequent in other positions.

We can now locate the different peripheries in the abstract LSC of an English sentence, arranging them in a hierarchy of preferences of occurrence in the positions described (we have excluded the detached positions as all types of periphery show an equal tendency to appear in these pragmatically marked positions). Table 3 shows

RLyLA Vol. 14 (2019), 59-79 | 73 Francisco José Cortés-Rodríguez and Carolina Rodríguez-Juárez The syntactic parsing of ASD-STE100 adverbials in ARTEMIS

the scale of peripheral preferences per position, whereas Table 4 presents the scale of positions in relation to their peripheral preferences:

Table 3. Scales of peripheral preferences per position

Position Peripheral preferences LDP (RDP) à PER-CL /PER-CONSTRL1/PER-NUC

PER Ini PER-CL > PER-NUC > PER-CONSTRL1 PERIniE

PERMed PER-NUC > PER-CONSTRL1 PERMedE

PERFinI PER-CONSTRL1> PER-NUC PERFin

Table 4. Scale of positions and peripheral preferences.

POSITIONS LDP PER Ini PERIniE PERMed PERMedE PERFinI PERFin RDP

+High PER_CL PER_NUC PER_CONSTRL1

± Mid all PER_NUC PER_CONSTRL1 PER_NUC all

−Low PER_CONSTRL1 (PER_CL) PERIPHERAL PREFERENCES

Table 4 can be especially helpful to decide in cases of ambiguity if used in combination with the ontological concepts which give form to the semantic information encoded in the lexical entries of adverbs and prepositions in the FunGramKB lexicons8. These concepts can be matched to the different semantic types identified in Section 2.2. Figure 14 presents the peripheries and positional preferences in the abstract LSC:

SENTENCE

CLAUSE

CONSTR-L1

CORE

NUC

LDP, PrC PER-CL ARG PER-CL (AUX) PER-NUC PRED PER-NUC (ARG) (ARG) PER-L1 (AJJ) (PRED-s) PER-L1 RDP

ALL PER-NUC PER-NUC PER-L1 PER-L1 PER-NUC PER-NUC ALL

PER-L1 PER-L1 PER-CL PER-CL Figure 14: Peripheries and positional preferences in the abstract LSC.

There is still one pending issue to be solved for the effective parsing of peripheral constituents in the GDE. The formalization of RRG in our computational prototype must meet one difficulty which arises from the syntactic status of adjuncts in this grammatical model. As can be seen in Figure 4, the peripheries are not subject to dominance relations; i.e. they are not analysed as daughters of any other syntactic node or layer, but are taken as modifiersof a given layer; such a relation is expressed by means of an arrow pointing to the layer modified, and not by a branch. There is no way to capture this type of modifying function in the format of the syntactic rules of the GDE. Following the format of context-free grammar rules, syntactic rules can only capture immediate dominance and linear ordering. To circumvent this problem, we follow a strategy similar to Kallmeyer & Osswald’s (2017) and

8 This is especially true for the occurrence of adverbs in clause internal positions. Thus, in pairs like the following: She sat there hopefully vs Hopefully it won’t rain tomorrow, the adverb hopefully will more naturally be understood as a clausal adjunct (PER_CL) if it appears in initial positions (as in the second sentence), but as a PER_CONSTRL1 adjunct if occurring in final positions (as happens in the first sentence). In the case of detached positions, it is much more difficult to decide upon ambiguous cases, since any kind of peripheral constituent can be assigned to these extraclausal positions for pragmatic purposes. It seems, however, that adverbs like ‘sadly’, ‘happily’, etc. which may be interpreted either as process (PER_CONSTRL1) or epistemic (PER-CL) adjuncts tend strongly to behave as clausal adverbials in such detached positions, as happens in Sadly, she sang (cf. She sang sadly, in which it behaves as a process adverbial), but there are also instances of process adjuncts in LDP, as in example 25 above. This is a pending issue which merits a more fine-grained analysis.

RLyLA Vol. 14 (2019), 59-79 | 74 Francisco José Cortés-Rodríguez and Carolina Rodríguez-Juárez The syntactic parsing of ASD-STE100 adverbials in ARTEMIS

Osswald & Kallmeyer’s (in press) and include peripheral nodes in our syntactic rules as if they were daughters of one of the layers in the LSC. For instance, the partial rule for the CLAUSE node proposed in Mairal-Usón & Cortés- Rodríguez (2017:16), which did not include any peripheral element9 (CL à (PreCONSTR-L1) CONSTR-L1) must be enlarged to account for the possibility of having optional constituents in PERIni position:

CL à (PreCONSTR-L1) (PER-CL) (PER-NUC) (PER-CONSTRL1) CONSTR-L1

This rule will generate a tree where some further computational operations must be done to obtain peripheral elements as modifier nodes with an arrow notation10.

Once assumed that the peripheries are to be described as daughter constituents of the central layers in the LSC, we still have to decide to which of those layers each of the peripheral positions in Figure 14 belong. In accordance with the tendencies observed on the scale of positional preferences and also taking into account that when there are several adverbs in a sentence they are subject to some iconic ordering constraints “in that adverbs related to more outer operators occur outside of adverbs related to more inner operators. In the simplest case ‘outside of’ means ‘farther from the verb’” (Van Valin, 2005:20), we propose the following general scheme (Figure 15):

SENTENC E

C L AUSE

C ONSTR -L1

C OR E

NUC

LDP, PrC PER-CL ARG PER-CL (AUX) PER-NUC PRED PER-NUC (ARG) (ARG) PER-L1 (AJJ) (PRED-s) PER-L1 RDP

ALL PER-NUC PER-NUC PER-L1 PER-L1 PER-NUC PER-NUC ALL

PER-L1 PER-L1 PER-CL PER-CL

Figure 15: Peripheries and positional preferences in the abstract LSC (with dominance relations).

Once the different positions are identified as daughters of a given layer, it is feasible to devise the syntactic rules that will spell out the ordering of constituents for each of the layers in the LSC. Since the computational workability of ARTEMIS is still to be tested, we have chosen to apply it to a Controlled Natural Language, namely ASD-STE100, with the assumption that it will help to validate the performance of our parser. Hence, the scope of our analysis will be confined to the catalogue of adverbials that can be found in a corpus written in this controlled language. The following section will deal with this issue.

4. ADVERBIALS IN AIRBUS ASD-STE100

Given the basic principle of Simplified Technical English (STE), “make texts easy to read and understand. […] write short sentences and use simple sentence structure” (2017: Specification Issue 7, page 1-4-1), and the nature of the texts written in the controlled natural language ASD-STE100, we should expect a reduced gamut of semantic types and of positional preferences and a tendency not to concatenate several adjuncts. As a matter of fact, with regard to the first issue, it is extremely rare to find clause adjuncts such as epistemic, evidential and illocutionary adjuncts, which tend to codify personal assessment, subjective opinions of the speaker or “comments on how the Speaker wishes the speech act to be taken or understood by the Addressee” (Dik, 1997:67). The semantic import of these types of constituents is logically absent in ASD-STE100 documents, since the texts produced by means of ASD-STE100 are procedural and descriptive texts and safety instructions, in which the voice of the speaker is irrelevant. This is confirmed in our sample of the Airbus Corpus11, with more than 17000 words, where only one instance of the epistemic adverb possibly has been found:

56. The aircraft is possibly not parallel to the ground. (DMC-AJ-A-32-11-76-00AAA-520A-A_027-00)

9 The actual rule is a bit more complex as it includes the AVMs for the constituent nodes; we have dispensed with this information to make the description easier to follow. 10 Within a Tree Adjoining Grammar formalism, Kallmeyer & Osswald (2017) and Osswald & Kallmeyer (in press) apply (sister) adjunction to their tree templates. This mode of composition is available if the template has a root label which coincides with the label of the target node. Some parallel tactic would probably be available in our model provided that the tags for our PER nodes include information on the layer they modify. 11 The Airbus corpus is a selection of texts from aircraft maintenance written following the ASD-STE100 specifications (2017).

RLyLA Vol. 14 (2019), 59-79 | 75 Francisco José Cortés-Rodríguez and Carolina Rodríguez-Juárez The syntactic parsing of ASD-STE100 adverbials in ARTEMIS

As far as positional preferences are concerned, it is also interesting to note that in our corpus we have not found cases of Level 1-Construction adjuncts in initial positions, which is also in consonance with their strong tendency to appear in final clause positions, followed by a more marked tendency to occur in medial positions. Here is a list of examples to illustrate the array of positions and types found in our sample selection from the Airbus corpus (Table 5):

Table 5. Examples from the Airbus corpus grouped by layers and peripheral positions.

SENTENCE LAYER: LDP PER_CONSTR-L1: Because of their different design, the structural assemblies are not interchangeable with each other and their position in the MLG is only one. PER_NUC: Also, the Nose Gear System supports the weight of the aircraft during ground maneuvers. RDP PER_NUC: This type of equipment can put water and moisture in parts, and can cause damage to equipment, specially to: Electrical equipment such as harnesses, proximity sensors and connectors. CLAUSE LAYER PER_Ini PER_NUC: Always the top bleed valve (2) of the related brake unit (1) must be used. CONSTR_L1 LAYER PER_IniE PER_NUC: They also must be at a minimum distance from the structure and adjacent elements. PER_FinI PER_CONSTR-L1: Let the lateral link (2) turn slowly by gravity until it is in its balanced position PER_NUC: If the test gives a fault message related to the CPIOM-G only: Replace the CPIOM-G related to the fault message. PER_Fin PER_CONSTR-L1: The gas rod must be lowered carefully. PER_NUC: The RTO label and the deceleration arrow come into view always. NUCLEUS LAYER PER_Med PER_NUC: When you operate the Ground Door Opening (GDO) system, you must always stay aft of the gear leg. PER_CONSTR-L1: The Normal Extension and Retraction System hydraulically operates the MLG leg actuators. PER_MedE PER_NUC: In emergency mode, each MLG door actuator moves only that mechanism to open the related door. PER_CONSTR-L1: This occurs because brake pressure […] goes directly to the shuttle valves.

Table 6 captures the reductions in both types and positional preferences:

Table 6. Reductions on the scale of positions and peripheral preferences in Airbus ASD-STE100.

POSITIONS LDP PER Ini PER IniE PER Med PER MedE PER FinI PER Fin RDP PER_ +High PER_CL PER_NUC PER_NUC CONSTRL1 PER_NUC PER_ PER_ PER_ ± Mid PER_NUC PER_NUC CONSTRL1 CONSTRL1 CONSTRL1 PER_CL PER_ PER_CL −Low (PER_CL)

PERIPHERAL PREFERENCES CONSTRL1

RLyLA Vol. 14 (2019), 59-79 | 76 Francisco José Cortés-Rodríguez and Carolina Rodríguez-Juárez The syntactic parsing of ASD-STE100 adverbials in ARTEMIS

Also in line with the spirit of ASD-STE100 to write short sentences, only in final positions have we found a tendency to amalgamate more than one adverbial, and when this takes place the maximum number of consecutive adjuncts seems to be three, as in the following sentences:

57. This operation can be possible during landing after a delay of twelve hours between a loss of blue hydraulic supply and the brake applications. (3 ADJUNCTS IN PER_CONSTRL1)

58. Let the lateral link (2) turn slowly by gravity until it is in its balanced position. (3 ADJUNCTS IN PER_CONSTRL1)

In our sample from the Airbus corpus there is one rare example of four adverbials in a final position:

59. The brake accumulator manifold supplies hydraulic blue power directly from the brake accumulator to the brake units through the PBSELV. (4 ADJUNCTS IN PER_CONSTRL1)

However, on most occasions, no more than two adverbials appear together in sequence, with a variegated array of possible combinations of formal types or realizations:

60. The MLG leg actuators also keep the MLG legs locked when the MLG legs are retracted because they have inner uplocks. (2 clauses; 2 adjuncts in PER_CONSTRL1)

61. It is necessary to do this step slowly to prevent air bubbles. (MP (adverb) + clause; 2 adjuncts in PER_CONSTRL1)

62. The HP filter measures the pressure dropacross the filter element to send the discrete signal. (PP (adverb) + clause; 2 adjuncts in PER_CONSTRL1)

63. The brake accumulator can also supply the Parking Brake System for 24 hours on a slope of 3 dega. (2 PP; 2 adjuncts in PER_CONSTRL1)

64. It is necessary to do the test one or two times to remove the air bubbles from the related hydraulic circuit. (RP + clause; 2 adjuncts in PER_CONSTRL1)

It is worth mentioning that catenation affects mainly Level 1-Construction adjuncts, whereas nuclear adjuncts tend to appear alone, a feature which will be reflected in their corresponding syntactic rules:

PER_NUC à ADJUNCT

PER_CONSTR-L1à ADJUNCT || ADJUNCT ADJUNCT || ADJUNCT ADJUNCT ADJUNCT|| ADJUNCT ADJUNCT ADJUNCT ADJUNCT

ADJUNCTà MP || PP || RP || CLAUSE

The last rule captures the realizational possibilities (Modifier Phrase, Prepositional Phrase, Referential Phrase or Clause) of any type of adjunct, irrespective of the peripheral constituent to which it will belong12.

5. CONCLUSIONS

The first part of this paper offered an updated description of the typology of adverbials within RRG, following Díaz-Jorge´s (2017) extensive overhaul of the schematic three-tier proposal in Van Valin (2005). In doing so, in the second part of our study, we set up the basic conditions for the treatment of such constituents within the so- called Grammar Development Environment, a syntactic parser which forms part of ARTEMIS. This involved, on the one hand, systematizing the locational tendencies exhibited by the different types of adverbials; this allowed us to establish a scale of probability of occurrence of the three types of peripheries in an array of positions within the clause. On the other hand, in our analysis we also dealt with the formalization problems concerning the status of the peripheries in RRG as modifiers, not subject to dominance relations within the Layered Structure of the Clause. We adapted the solution put forward in Kallmeyer & Osswald (2017) and Osswald & Kallmeyer (in press) in our

12 cf. Cortés-Rodríguez (2016) for a detailed description of these types of phrasal units both within RRG and ARTEMIS.

RLyLA Vol. 14 (2019), 59-79 | 77 Francisco José Cortés-Rodríguez and Carolina Rodríguez-Juárez The syntactic parsing of ASD-STE100 adverbials in ARTEMIS

proposal and assigned to each peripheral type a daughter status for the nodes that they modify, leaving it to the computational apparatus to relocate them as modifier units.

In order to carry this study, we restricted the scope of analysis to a sample selection of texts from the Airbus Corpus, written in ASD-STE100. Further work needs to be done to validate the positional behaviour of peripheral constituents in natural present-day English and to fully articulate the semantic classification of adverbials along the three-layered description within RRG and the LCM.

ACKNOWLEDGEMENTS

This research has been financially supported by the DGI, Spanish Ministry of Education and Science, project No. FFI2014-53788-C3-1-P.

REFERENCES

ASD-STE100. Simplified Technical English. (2017).Specification ASD-STE100. TM: International specification for the preparation of technical documentation in a controlled language. Issue 7. January 2017. Brussels: ASD. Cortés-Rodríguez, F. J. (2016). “Towards the computational implementation of Role and Reference Grammar: Rules for the syntactic parsing of RRG phrasal constituents”, Círculo de Lingüística Aplicada a la Comunicación, 65: 75-108. https://doi.org/10.5209/rev_CLAC.2016.v65.51982 Cortés-Rodríguez, F. J. & Mairal-Usón, R. (2016). “Building an RRG computational grammar”, Onomázein, 34: 86- 117. https://doi.org/10.7764/onomazein.34.22. Díaz-Jorge, V. (2017). Adjuncts in Role and Reference Grammar: The peripheries. (Final Degree Dissertation), Universidad de La Laguna, Spain. Dik, S. (1997). The Theory of Functional Grammar. Part 1: The Structure of the Clause. Berlin and New York: Mouton de Gruyter. https://doi.org/10.1515/9783110218367. Fumero Pérez, M. C. & Díaz Galán, A. (2017). “The interaction of parsing rules and argument–predicate constructions: Implications for the structure of the Grammaticon in FunGramKB”, Revista de Lingüísticas y Lenguas Aplicadas, 12: 33-44. https://doi.org/10.4995/rlyla.2017.5406 Greenbaum, S. & Quirk, R. (1990). A Student’s Grammar of the English Language. London: Longman. Hernández-Pastor, D. & Periñán-Pascual, C. (2016). “Developing a knowledge base for preposition sense disambiguation: A view from Role and Reference Grammar and FunGramKB”, Onomázein, 33: 251-288. https://doi.org/10.7764/onomazein.33.16 Johnson, M. (1987). “A new approach to clause structure in Role and Reference Grammar”, in Davis Working Papers in Linguistics, 2: 55-59. Davis: University of California. Kallmeyer, L. & Osswald, R. (2017). “Combining Predicate-Argument Structure and Operator Projection: Clause Structure in Role and Reference Grammar”, in Proceedings of the 13th International Workshop on Tree Adjoining Grammars and Related Formalisms, pp. 61–70. Mairal-Usón, R. (2017). “A computational implementation of idiomatic and non-idiomatic constructions”, Signos, 50/94, 241-264. https://doi.org/10.4067/S0718-09342017000200241 Mairal-Usón, R. & Cortés-Rodríguez, F. J. (2017). “Automatically Representing TExt Meaning via an Interlingua- based System (ARTEMIS). A further step towards the computational representation of RRG”, Journal of Computer-Assisted Linguistic Research, 1: 61-87. https://doi.org/10.4995/jclr.2017.7788 Mora-Bustos, A. (2012). “Los Adjuntos Periféricos”, in R. Mairal-Usón, L. Guerrero & C. González-Vergara (eds.) Gramática del Papel y la Referencia. Introducción, avances y aplicaciones. Madrid: Akal, pp. 203-223. Osswald, R. & Kallmeyer, L. (in press). “Towards a formalization of Role and Reference Grammar”, in R. Kailuweit, L. Künkel & E. Staudinger (eds.) Applying and Expanding Role and Reference Grammar. https://user.phil-fak. uni-duesseldorf.de/~osswald/publications/osswald_kallmeyer-rrg_formalization.pdf [retrieved: 15.3.2018]. Pavey, E. L. (2010). The Structure of Language: An Introduction to Grammatical Analysis. Cambridge: Cambridge University Press. https://doi.org/10.1017/CBO9780511777929 Periñán-Pascual, C. (2012). “En defensa del procesamiento del lenguaje natural fundamentado en la lingüística teórica”, Onomázein, 26: 13-48. Available at http://onomazein.letras.uc.cl/Articulos/26/1_Perinan.pdf [retrieved: 15.3.2018]. Periñán-Pascual, C. (2013). “A knowledge-engineering approach to the cognitive categorization of lexical meaning”, VIAL: Vigo International Journal of Applied Linguistics, 10: 85-104. Available at http://vialjournal.webs.uvigo. es/pdf/Vial-2013-Article4.pdf [retrieved: 15.3.2018].

RLyLA Vol. 14 (2019), 59-79 | 78 Francisco José Cortés-Rodríguez and Carolina Rodríguez-Juárez The syntactic parsing of ASD-STE100 adverbials in ARTEMIS

Periñán-Pascual, C. & Arcas-Túnez, F. (2014). “The implementation of the FunGramKB CLS Constructor”, in B. Nolan & C. Periñán-Pascual (eds.) Language Processing and Grammars: The Role of Functionally Oriented Computational Models. Amsterdam: John Benjamins, 165-196. https://doi.org/10.1075/slcs.150.07per Quirk, R., Greenbaum, S., Leech, G. & Scartvik, J. (1985). A Grammar of Contemporary English. London: Longman. Ramaswamy, E. S. (2007). Practical English Grammar. Chennai: Sura Books. Ruiz de Mendoza, F. (2013). “Meaning Construction, Meaning Interpretation and Formal Expression in the Lexical Constructional Model”, in B. Nolan & E. Diedrichsen (eds.) Linking Constructions into Functional Linguistics: The Role of Constructions in Grammar. Amsterdam: John Benjamins, 231-270. https://doi.org/10.1075/ slcs.145.09ib225 Ruiz de Mendoza, F. & Galera Masegosa, A. (2014). Cognitive modeling. A linguistic perspective. Amsterdam: John Benjamins. https://doi.org/10.1075/hcp.45 The British National Corpus, version 3 (BNC XML Edition). (2007). Distributed by Bodleian Libraries, University of Oxford, on behalf of the BNC Consortium. Available at http://www.natcorp.ox.ac.uk/ [retrieved: 15.3.2018]. Toratani, K. (2007). An RRG Analysis of Manner Adverbial Mimetics. York: York University. Van Valin, R. D. Jr (2005). Exploring the Syntax-Semantics Interface. Cambridge: Cambridge University Press. https://doi.org/10.1017/CBO9780511610578 Van Valin, R. D. Jr & LaPolla, R. J. (1997). Syntax, Structure, Meaning and Function. Cambridge: Cambridge University Press. https://doi.org/10.1017/CBO9781139166799

RLyLA Vol. 14 (2019), 59-79 | 79 Revista de Lingüística y Lenguas Aplicadas Vol. 14 año 2019, 81-89 EISSN 1886-6298 https://doi.org/10.4995/rlyla.2019.11033

LAS LENGUAS OFICIALES DEL ESTADO ESPAÑOL EN LOS TEXTOS LEGALES, ¿FOMENTO O RECONOCIMIENTO DEL PLURILINGÜISMO?

Luis Escoriza Morera Universidad de Cádiz, España

Resumen: La convivencia histórica de lenguas en un mismo espacio geográfico genera diferentes planteamientos en lo que respecta al control del uso de dichas lenguas en ámbitos que superan el espacio personal como la educación, los medios de comunicación o la Administración, desde posturas que priman el uso de la lengua mayoritaria hasta otras que inciden en la defensa de la lengua minoritaria, pasando por opiniones intermedias que abogan por la adopción de medidas que preserven la riqueza derivada del multilingüismo. El análisis del marco legal existente en España evidencia un desarrollo de los textos redactados en las comunidades bilingües para cubrir la ausencia de legislación concreta de ámbito general. Dichos textos nos permiten analizar sus ambigüedades y reflexionar acerca de cuál es su verdadero objetivo, el mero reconocimiento de la oficialidad de dos sistemas lingüísticos o el desarrollo de una planificación lingüística que garantice ambos en el uso público.

Palabras clave: bilingüismo, cooficialidad, planificación lingüística.

THE OFFICIAL LANGUAGES IN LEGAL TEXTS IN SPAIN, IS THIS PROMOTION OR RECOGNITION OF MULTILINGUALISM? Abstract: The use of different languages within the same geographical area usually generates different opinions about how these should be controlled within public ambits, such as education, media or public administration. Our main goal is to provide an analysis of the existing points of view with regard to the use of Spanish and the other co-official languages in Spain, stressing the legal framework that justifies them, as well as their underlying objectives and possible consequences. Our intention is to discuss if these texts are aimed to simply recognize the official character of two languages or if they pursue the development of a linguistic planning that guarantees the use of both within the public ambit.

Keywords: bilingualism, co-officiality, language planning.

1. ENFOQUES EN LA PLANIFICACIÓN DEL BILINGÜISMO SOCIAL

España puede ser considerada un Estado plurilingüe desde diferentes puntos de vista, pero es evidente que un hecho que destaca en su configuración lingüística es la coexistencia, en determinadas comunidades autónomas, de la lengua mayoritaria, el español o castellano1, y una segunda lengua, a la que podríamos denominar minoritaria en el sentido adoptado por la UE, quien define lengua minoritaria en la Carta Europea de las Lenguas Regionales o Minoritarias redactada en 2001 como una “lengua hablada tradicionalmente en un territorio de un Estado por nacionales de ese Estado que constituyen un grupo numéricamente inferior al resto de la población del Estado y diferente de la lengua oficial del Estado” (2001:34734). Nuestro objetivo principal es analizar diferentes textos legales que recogen la regulación del uso de esas dos lenguas. Nos centraremos en Cataluña, Galicia y País Vasco para comprobar en qué medida se fomenta el bilingüismo en esos territorios. En primer lugar describiremos las tres posibles posturas principales que existen, a nuestro juicio, a la hora de

1 Mucho se ha escrito sobre la denominación adecuada de la lengua común hablada en España y en torno a las connotaciones que las designaciones castellano y español pueden esconder, especialmente en algunas zonas del mundo hispánico, desde su coexistencia hace ya varios siglos. Probablemente el primer reflejo claro documentado de la polémica en torno a dichas denominaciones se esconde en los debates surgidos durante la redacción de la Constitución española de 1978. Para una información más detallada sobre dicho proceso y sobre las principales posturas acerca del uso de uno u otro vocablo cf. Montero Cartelle, 1979

To cite this article: Escoriza Morera, L. (2019). "The official languages in legal texts in Spain, is this promotion or recognition of multilingualism?"Revista de Lingüística y Lenguas Aplicadas, 14, 81-89. https://doi.org/10.4995/rlyla.2019.11033 Correspondence author: [email protected]

Received: 2018-11-20 Accepted: 2019-03-13 | 81 Luis Escoriza Morera Las lenguas oficiales del Estado español en los textos legales, ¿fomento o reconocimiento del plurilingüismo?

regular institucionalmente la convivencia entre dichas lenguas mayoritaria y minoritaria, ilustrándolas con tres manifiestos redactados en diferentes momentos en España; después haremos un repaso del escaso marco legal de carácter general o internacional existente en torno a dicho tema; recordaremos los ámbitos de uso sobre los que se puede hacer política y planificación lingüísticas, y finalmente analizaremos diferentes documentos de carácter legal desarrollados en el Estado español en las últimas décadas en el sentido expuesto para intentar concluir hasta qué punto dicho marco legal justifica la situación actual y si promueve realmente el fomento del bilingüismo en España.

Tres son, a nuestro juicio, los posibles enfoques en situaciones de bilingüismo del tipo que se plantea entre español o castellano, catalán, gallego y vasco en Cataluña, Galicia y País Vasco. En primer lugar, puede defenderse la necesidad de cuidar de manera preferente la lengua mayoritaria, basándose fundamentalmente en que esta es la única lengua común a todo el Estado y por tanto el mayor vehículo de comunicación nacional. En junio de 2008, un grupo de intelectuales del ámbito hispano redactaron en Madrid el denominado Manifiesto por la lengua común, que manifestaba esta idea y la necesidad de reivindicar su aplicación en España. Diferentes fragmentos de ese documento lo demuestran. Se describe al español como “la única lengua juntamente oficial y común de todos los ciudadanos españoles” y la “lengua principal de comunicación democrática en este país”. Se manifiesta que “hay una asimetría entre las lenguas españolas oficiales, lo cual no implica injusticia (?) de ningún tipo porque en España hay diversas realidades culturales pero sólo una de ellas es universalmente oficial en nuestro Estado democrático”, se afirma que el hecho de “que ciertas autoridades autonómicas anhelen como ideal lograr un máximo techo competencial bilingüe no justifica decretar la lengua autonómica como vehículo exclusivo ni primordial de educación o de relaciones con la administración pública". Y se concluye que “la lengua castellana es común y oficial a todo el territorio nacional, siendo la única cuya comprensión puede serle supuesta a cualquier efecto a todos los ciudadanos españoles”.

Frente a este planteamiento de la gestión de la convivencia de lengua mayoritaria y minoritaria en España, existe una tendencia diametralmente opuesta, que incide en la necesidad de fomentar la lengua minoritaria por ser esta la lengua propia de los territorios en los que se habla, símbolo inequívoco de identidad, subordinada, además, históricamente, a la lengua mayoritaria por decisiones de naturaleza política. Otro manifiesto, también redactado por un grupo de intelectuales, Per un veritable procés de normalització lingüística a la Catalunya independent, este en Barcelona en octubre de 2015, manifiesta este sentir en lo que respecta al catalán. Si bien es cierto que el texto se enmarca en un contexto independentista, la postura que recoge ha sido también defendida en diferentes comunidades españolas desde planteamientos no necesariamente independentistas, como veremos en el análisis de los textos legales, en la última parte de este trabajo. Este segundo manifiesto describe al catalán como “la llengua de Catalunya, en el sentit que és la llengua endògena del territori de Catalunya, on s’ha format i ha evolucionat històricament”, como “la llengua en què sempre ha parlat el poble català”. Por el contrario, el castellano es “a llengua de dominació, li disputa coercitivament aquest estatus de llengua territorial i ha intentat i continua intentant repressivament de desplaçar-la dels àmbits d’ús lingüístic general”. En este sentido, el “procés de bilingüització forçosa mitjançant la repressió politicojurídica de l’ús del català, l’ensenyament obligatori i l’extensió dels nous mitjans de comunicació, tots dos absolutament en castellà”, llegándose incluso a afirmar que “aquesta ideologia bilingüista no és res més que una forma d’encobrir i legitimar la subordinació d’una llengua a l’altra i el consegüent procés de substitució lingüística que pateix la societat catalana”. Por tanto, se concluye que “la degradació qualitativa, estructural, de la llengua no ha parat de créixer en el camí de convertir-se en una mena de dialecte del castellà” y que hay que “començar a solucionar-lo per les úniques vies possibles: a) la restitució al català de l’estatus de llengua territorial de Catalunya (i igualment per a l’occità a la Vall d’Aran), b) la reversió de la pràctica de la subordinació sistemàtica i generalitzada de l’ús del català (o de l’occità) a l’ús del castellà, i c) la recuperació progressiva de la genuïnitat de la llengua”.

Entre ambos planteamientos, existe una tercera postura conciliadora que reivindica la riqueza del plurilingüismo del Estado español y defiende, por tanto, la necesidad de desarrollar una planificación que conserve el bilingüismo al menos en las comunidades en las que este se da. También un manifiesto de similares características a los dos mencionados anteriormente, este redactado en Valencia en noviembre de 2015 y denominado España plurilingüe. Manifiesto por el reconocimiento y el desarrollo de la pluralidad lingüística de España, recoge este sentir. En dicho texto se recuerda que “España también es, y siempre ha sido desde su origen, un país multilingüe. Actualmente, el 47% de su población vive en una CA (comunidad autónoma) con más de una lengua”, y que “el plurilingüismo aspira a que todos y cada uno de los ciudadanos sean o se sientan plurilingües; es decir, puedan conocer más de una lengua o valoren su convivencia con más de una lengua. El fortalecimiento de una lengua no tiene por qué ir en detrimento de las otras”. Por consiguiente, “reconocer y desarrollar el plurilingüismo de España comporta determinar sobre su uso en las Cortes Generales y promulgar una Ley que regule específicamente esta cuestión”. Y concluye que “reconocer y desarrollar el plurilingüismo de España implica la corresponsabilidad presupuestaria de todas las administraciones y representa un coste perfectamente asumible que se puede y se debe abordar

RLyLA Vol. 14 (2019), 81-89 | 82 Luis Escoriza Morera Las lenguas oficiales del Estado español en los textos legales, ¿fomento o reconocimiento del plurilingüismo?

con recursos humanos y económicos disponibles, y muy especialmente con las tecnologías de la información y la comunicación”2.

Creemos que estos tres planteamientos ilustran de manera clara los distintos posicionamientos que pueden darse a la hora de gestionar la coexistencia de las lenguas mayoritaria y minoritaria en cualquier contexto en general, y en España, objeto de nuestro interés aquí, en particular. No tienen por qué ser posturas excluyentes entre sí y pueden fomentarse medidas que respondan a distintos intereses, pero a nuestro juicio ayudan a entender en gran medida el sentido de los textos que ocuparán la parte central de este trabajo.

2. DERECHOS LINGÜÍSTICOS Y ÁMBITOS DE APLICACIÓN

Centrándonos ya en el ámbito legal, es preciso mencionar en primer lugar, aspecto señalado por autores como R. E. Hamel (1995), la inexistencia histórica de referentes internacionales en lo que a los derechos lingüísticos respecta. En reconocidos textos legales de esta naturaleza la alusión a estos es meramente simbólica. La Declaración Universal de Derechos Humanos (1948:2), recoge en su artículo 2 la única mención a la no discriminación por razones lingüísticas cuando proclama “la fe en los derechos humanos fundamentales, en la dignidad y en el valor de la persona humana y en la igualdad de derechos de hombres y mujeres […] sin distinción de raza, color, sexo, lengua, religión, opinión política u otra”. El Pacto Internacional de Derechos Civiles y Políticos (1966: 9341), promulgado por Naciones Unidas solo recoge en su artículo 27 que

“en los Estados en que existan minorías, étnicas, religiosas o lingüísticas, no se negará a las personas que pertenezcan a dichas minorías el derecho que les corresponde, en común con los demás miembros de su grupo, a tener su propia vida cultural, a profesar y practicar su propia religión y a emplear su propio idioma”.

La Declaración sobre los Derechos de los Pueblos Indígenas (2007:6), aprobada también por Naciones Unidas después de más de 20 años de discusión, dice únicamente en su artículo 14 que

“los Estados adoptarán medidas eficaces, conjuntamente con los pueblos indígenas, para que las personas indígenas, en particular los niños, incluidos los que viven fuera de sus comunidades, tengan acceso, cuando sea posible, a la educación en su propia cultura y en su propio idioma”.

Este hecho explica la aparición en las últimas décadas de documentos que intentan precisar de modo general el alcance de los denominados derechos lingüísticos. Destacaremos aquí dos, la Carta Europea de las Lenguas Regionales o Minoritarias redactada en 1992 por la Unión Europea y la Declaración Universal de Derechos Lingüísticos aprobada en Barcelona en 1996 con el apoyo técnico de la Unesco. A pesar de la poca aplicación del segundo documento en la práctica y de la diferente interpretación que los Estados de la Unión Europea han realizado del primero, lo interesante es resaltar que ambos recogen de manera clara los ámbitos de regulación del uso de las lenguas desde una planificación lingüística institucional. Partiendo de la información aportada por los dos documentos podemos concretar los siguientes ámbitos comunicativos en los que puede concretarse la utilización de una o más lenguas por parte de la Administración:

Educación:

• Uso de la lengua como lengua vehicular o como objeto de estudio. • Existencia de recursos humanos y materiales.

Administración pública y organismos oficiales:

• Actuaciones judiciales y administrativas (tribunales de justicia). • Documentos públicos y privados (leyes, documentos notariales). • Atención en servicios públicos. • Formularios, impresos y modelos.

Onomástica:

• Uso de topónimos y antropónimos.

2 Uno de los firmantes de este manifiesto, Á. López García, ha advertido en otras ocasiones (cf. López García, 2009:126) de la necesidad de un consenso social para el desarrollo de la riqueza que el bilingüismo supone para cualquier comunidad, al tiempo que hace hincapié en que dicho consenso exigiría una actitud responsable por parte de los poderes políticos en el sentido manifestado en el documento que mostramos.

RLyLA Vol. 14 (2019), 81-89 | 83 Luis Escoriza Morera Las lenguas oficiales del Estado español en los textos legales, ¿fomento o reconocimiento del plurilingüismo?

Medios de comunicación y nuevas tecnologías:

• Medios de comunicación de su territorio. • Equipos y herramientas informáticos.

Cultura:

• Expresiones culturales. • Servicios (bibliotecas, videotecas, cines, teatros, museos, archivos, folklore) e industrias culturales. • Preservación del patrimonio cultural.

Ámbito socioeconómico:

• Actividades socioeconómicas. • Medios para el ejercicio de la actividad profesional (documentos, libros de consulta, formularios, instrucciones, impresos...). • Transacciones económicas (operaciones bancarias, seguros...). • Publicidad, rotulación, señalización exterior e imagen del país. • Productos y servicios de establecimientos comerciales (etiquetas, listados de ingredientes, publicidad, garantía, instrucciones de uso...). • Indicaciones públicas referentes a la seguridad. • Relaciones con las empresas, establecimientos comerciales y entidades privadas. • Ejercicio de las actividades laborales o profesionales.

Esa numeración de contextos explica tanto la complejidad de la gestión del uso de varias lenguas en un mismo contexto geográfico como la posible controversia a la hora de asumir determinaciones al respecto. Esos mismos ámbitos de uso aparecen reflejados en los textos legales redactados en el Estado español en el intento de regular el uso de las lenguas en las comunidades bilingües, de cuyo análisis nos ocuparemos a continuación.

3. EL MARCO LEGAL EN EL ESTADO ESPAÑOL

La Constitución española redactada en 1978 constituye, indudablemente, el máximo referente legal en España. La única mención a las lenguas habladas en el Estado español aparece en el artículo tercero del título preliminar que reproducimos a continuación:

“1. El castellano es la lengua española oficial del Estado. Todos los españoles tienen el deber de conocerla y el derecho a usarla. 2. Las demás lenguas españolas serán también oficiales en las respectivas Comunidades Autónomas de acuerdo con sus Estatutos. 3. La riqueza de las distintas modalidades lingüísticas de España es un patrimonio cultural que será objeto de especial respeto y protección.” (Constitución española, 1978:29315).

Hay dos hechos relevantes, a nuestro juicio, en esa formulación. En primer lugar, se define el alcance de la oficialidad del español (el ciudadano tiene el deber de conocer la lengua y el derecho a usarla) pero no se explicita dicha oficialidad cuando se convierte en cooficialidad, lo que permite formular diferentes interpretaciones (si hay dos lenguas oficiales, ¿el deber es conocer una al menos o ambas?). En segundo lugar, deriva hacia los Estatutos de las comunidades autónomas bilingües la regulación de dicha cooficialidad. No hay formulaciones legales relevantes posteriores en el ámbito nacional pero sí ha existido un desarrollo posterior al texto constitucional en aquellas comunidades bilingües mencionadas. Centraremos nuestro comentario, como indicábamos más arriba, en el País Vasco, Galicia y Cataluña, en ese orden, para ver algunos de los textos más significativos que se han redactado para completar y matizar lo expuesto en la Constitución. El orden de las comunidades no es aleatorio, ya que iremos de un menor a un mayor desarrollo legal en la explicitación de lo concerniente a la regulación de la convivencia de las lenguas mayoritaria y minoritaria3.

3 Hecho comprensible si admitimos que el catalán se encuentra en su comunidad en una mejor situación desde los puntos de vista lingüístico y político (cf. Subirats i Humet y Gomá, 1998:300-301).

RLyLA Vol. 14 (2019), 81-89 | 84 Luis Escoriza Morera Las lenguas oficiales del Estado español en los textos legales, ¿fomento o reconocimiento del plurilingüismo?

La Ley Orgánica 3/1979, de 18 de diciembre, de Estatuto de Autonomía para el País Vasco es el texto de referencia en esa comunidad. En él se observan aspectos comunes a los tres estatutos de autonomía que vamos a examinar. En primer lugar, obviamente, se reconoce la cooficialidad del español y de la lengua minoritaria, el euskera en este caso, si bien es necesario destacar que se especifica la denominación de lengua propia4 para la lengua minoritaria:

“El euskera, lengua propia del Pueblo Vasco, tendrá, como el castellano, carácter de lengua oficial en Euskadi, y todos sus habitantes tienen el derecho a conocer y usar ambas lenguas.” (Ley orgánica 3/1979:4).

El texto se limita a recoger la oficialidad del uso de ambas lenguas y anula de posibilidad de discriminación por la misma. Se señala, además, el conocimiento de la lengua minoritaria como mérito de acceso a la Administración pública:

“Las instituciones comunes de la Comunidad Autónoma, teniendo en cuenta la diversidad socio-lingüística del País Vasco, garantizarán el uso de ambas lenguas, regulando su carácter oficial, y arbitrarán y regularán las medidas y medios necesarios para asegurar su conocimiento. […]

Nadie podrá ser discriminado por razón de la lengua. […]

El nombramiento de los Magistrados, Jueces y Secretarios se efectuará en la forma prevista en las Leyes Orgánicas del Poder Judicial y del Consejo General del Poder Judicial, siendo mérito preferente el conocimiento del Derecho Foral Vasco y el del euskera, sin que pueda establecerse excepción alguna por razón de naturaleza o de vecindad”. (op. cit.: 4, 5 y 14).

Un planteamiento similar puede observarse en la Ley Orgánica 1/1981, de 6 de abril, de Estatuto de Autonomía para Galicia, aprobada dos años más tarde. Se reconoce al gallego el carácter de lengua propia, al tiempo que se garantiza el uso de ambas lenguas y se evita la discriminación por motivos de carácter lingüístico. Existe aquí una especificación significativa acerca del significado de la cooficialidad y se estima que los ciudadanos tienen derecho a conocer y usar ambos sistemas, pero no se plantea en término de deberes:

“La lengua propia de Galicia es el gallego. […]

Los idiomas gallego y castellano son oficiales en Galicia y todos tienen el derecho de conocerlos y usarlos. […]

Los poderes públicos de Galicia garantizaran el uso normal y oficial de los dos idiomas y potenciaran la utilización del gallego en todos los órdenes de la vida pública, cultural e informativa, y, dispondrán los medios necesarios para facilitar su conocimiento. […]

Nadie podrá ser discriminado por razón de la lengua”. (Ley orgánica 1/1981:6).

Esta redacción general, similar a la del texto estatutario del País Vasco, se ve matizada con posterioridad en otros documentos. Nos parece especialmente significativa laLey 3/1983, de 15 de junio, de normalización lingüística, aprobada por la Xunta de Galicia poco después del Estatuto. En ese texto vuelve a manifestarse el carácter de lengua propia de la lengua minoritaria pero se incluye, además del derecho a usarla, el deber de conocerla por parte de los ciudadanos. Se refleja una de las interpretaciones posibles en cuanto a las obligaciones de los ciudadanos respecto a las lenguas implicadas en contextos bilingües. Ese sintagma, precisamente, se mantiene en el texto con la aclaración de que ha sido declarado anticonstitucional por el Estado español, que no considera viable esa interpretación a pesar de la laxitud del texto constitucional del que partimos: “O galego é a lingua propia de Galicia. Tódo los galegos teñen o deber de coñecelo e o dereito de usalo”.

Por otra parte, se define con mayor detalle el concepto de lengua propia, lengua que se constituye en la lengua oficial de la comunidad, es decir, de sus instituciones, al tiempo que se promueve su uso como lengua normal de comunicación:

“O galego, como lingua propia de Galicia, é lingua oficial das institucións da Comunidade Autónoma, da súa Administración, da Administración Local e das Entidades Públicas dependentes da Comunidade Autónoma. […]

Os poderes públicos de Galicia promoverán o uso normal da lingua galega, oralmente e por escrito, nas súas relacións cos cidadáns. […]

O galego será a lingua usual nas emisoras de radio e televisión e nos demais medios de comunicación social sometidos a xestión ou competencia das institucións da Comunidade Autónoma”. (Ley 3/1983:2 y 5).

4 El significativo uso del adjetivopropia para identificar a una de las lenguas presentes en un ámbito multilingüe contiene implicaciones relacionadas con los derechos y deberes recogidos en los textos legales en los que aparece, como intentaremos demostrar en este texto. Sobre las distintas denominaciones utilizadas en política lingüística para designar a las lenguas y sus posibles implicaciones legales cf. Ruíz Vieytez (2004) o Süselbeck (2008).

RLyLA Vol. 14 (2019), 81-89 | 85 Luis Escoriza Morera Las lenguas oficiales del Estado español en los textos legales, ¿fomento o reconocimiento del plurilingüismo?

Queremos resaltar además dos ideas relacionadas con las lenguas en el contexto educativo:

“Os nenos teñen dereito a recibi-lo primeiro ensino na súa lingua materna. […]

Os alumnos non poderán ser separados en centros diferentes por razón da lingua. Tamén se evitará, a non ser que con carácter excepcional as necesidades pedagóxicas así o aconsellaren, a separación en aulas diferentes”. (op. cit.:4).

Se garantiza el derecho de los alumnos a recibir la educación en su lengua materna al tiempo que se impide la separación de los alumnos por lenguas en centros e incluso en aulas. A nuestro juicio, en contextos con hablantes de ambas lenguas como lengua materna ese planteamiento puede dar lugar a controversia, ya que si todos exigen el derecho a recibir la educación en su sistema materno pero no pueden ser separados, la única vía para no incumplir con la ley es garantizar una educación bilingüe completa en todos los niveles. Esa es la solución más enriquecedora desde el punto de vista lingüístico pero supone un mayor esfuerzo por parte de la Administración y los centros educativos. Este texto refleja un mayor desarrollo del marco legal en el ámbito gallego y una tendencia a reforzar la lengua minoritaria, más allá del reconocimiento de la cooficialidad.

El análisis de los textos redactados en la comunidad catalana permite observar una tendencia similar a la expuesta en la evolución de los dos textos reseñados en Galicia. Ya el Estatuto va mucho más allá de la mera consideración general de cooficialidad vista en los estatutos vasco y gallego, si bien es cierto que debemos basarnos en la Ley Orgánica 6/2006, de 19 de julio, de reforma del Estatuto de Autonomía de Cataluña, cuyo último texto consolidado, modificado y aprobado por el Parlamento catalán es de febrero de 2013, fecha muy posterior a la de los documentos analizados en País Vasco y Galicia. Hemos de partir de que el Estatuto catalán contempla la cooficialidad de ambas lenguas, castellano y catalán, recoge la no discriminación por razón de lengua y defiende el denominado derecho de opción lingüística en todos los usos institucionales. Además, aquí se interpreta que los ciudadanos tienen el deber de conocer y el derecho a usar ambas lenguas, interpretación más amplia y aceptada por el Tribunal constitucional:

“El catalán es la lengua oficial de Cataluña. También lo es el cas­tellano, que es la lengua oficial del Estado español. Todas las personas tienen derecho a utilizar las dos lenguas oficiales y los ciudadanos de Cataluña el derecho y el deber de conocerlas. Los poderes públicos de Cataluña deben establecer las medidas necesarias para facilitar el ejercicio de estos derechos y el cumplimiento de este deber. De acuerdo con lo dispuesto en el artículo 32, no puede haber discriminación por el uso de una u otra lengua. […]

Los ciudadanos tienen el derecho de opción lingüística. En las relaciones con las instituciones, las organizaciones y las Administraciones­ públicas en Cataluña, todas las personas tienen derecho a utilizar la lengua oficial que elijan”. (Ley orgánica 6/2006:3 y 8).

Pero ante fragmentos que reflejan estas ideas se define también aquí la lengua propia como lengua de uso normal y preferente en los usos regulados institucionalmente, planteamiento que puede resultar contradictorio con el anterior (hasta el punto de que el calificativo preferente también ha sido calificado de anticonstitucional por el Estado español):

“La lengua propia de Cataluña es el catalán. Como tal, el cata­lán es la lengua de uso normal y preferente de las Administraciones públicas y de los medios de comunicación públicos de Cataluña, y es también la lengua normalmente utilizada como vehicular y de aprendizaje en la enseñanza. (op. cit.:3).

En lo que respecta al ámbito educativo, la situación vista con anterioridad en la ley de normalización del gallego vuelve a repetirse, ya que se afirma la imposibilidad de separar al alumnado en grupos o clases distintos por razón de lengua, se recoge la consideración de la lengua minoritaria como lengua normal y vehicular de la educación y, aunque no se manifieste explícitamente en el artículo 35, dedicado a los derechos en el ámbito de la enseñanza, de la manifestación previa del derecho de opción lingüística cabe deducirse la posibilidad de recibir la educación en la lengua materna, sea esta la mayoritaria o la minoritaria:

“Todas las personas tienen derecho a recibir la enseñanza en ca­talán, de acuerdo con lo establecido por el presente Estatuto. El catalán debe utilizarse normalmente como lengua vehicular y de aprendizaje en la enseñanza universitaria y en la no universitaria. […]

Los alumnos tienen derecho a recibir la enseñanza en catalán en la enseñanza no universitaria. Tienen también el derecho y el deber de conocer con suficiencia oral y escrita el catalán y el castellano al finalizar la enseñanza obligatoria, sea cual sea su lengua habitual al incorporarse a la enseñanza. La enseñanza del catalán y el castellano debe tener una presencia adecuada en los planes de estudios. […]

Los alumnos tienen derecho a no ser separados en centros ni en grupos de clase distintos por razón de su lengua habitual”. (op. cit.:9).

RLyLA Vol. 14 (2019), 81-89 | 86 Luis Escoriza Morera Las lenguas oficiales del Estado español en los textos legales, ¿fomento o reconocimiento del plurilingüismo?

Esa aparente dualidad de planteamientos entre la afirmación de la cooficialidad y el fomento de la lengua minoritaria es aún mucho más notoria en otros textos, especialmente en la Ley 1/1998, de 7 de enero, de política lingüística aprobada por la Generalitat de Catalunya algunos años antes de la última versión del Estatuto que acabamos de mencionar. Se trata de un texto significativo y muy interesante, ya que representa uno de los documentos legales específicos sobre política lingüística en contextos bilingües más detallados en el contexto europeo y nos atreveríamos a afirmar que internacional.

Aquí también se afirma que la lengua minoritaria al ser la lengua propia debe ser la lengua de uso de los poderes públicos. Dicho uso se plantea en términos de obligación:

“El concepto de lengua propia aplicado a la catalana obliga a los poderes públicos y las instituciones de Cataluña a protegerla, a usarla de forma general y a promover su uso público en todos los niveles.

Las administraciones e instituciones catalanas deben utilizar de forma general el catalán, sin perjuicio del derecho de los ciudadanos y ciudadanas a dirigirse a las mismas en la lengua oficial que éstos escojan […]

El catalán [...] es la lengua de todas las instituciones de Cataluña, y en especial de la Administración de la Generalidad, de la Administración local, de las corporaciones públicas, de las empresas y los servicios públicos, de los medios de comunicación institucionales, de la enseñanza y de la toponimia”. (Ley 1/1998:13-15).

En el contexto educativo se mantiene el planteamiento ya mencionado, incluso de forma más explícita, ya que, junto a la ya recogida en otros textos posibilidad de elección de lengua por parte del ciudadano e imposibilidad de separación del alumnado por razón de lengua, se acentúa la obligación de los centros educativos de usar el catalán como lengua de uso normal. Ello induce a pensar en la inclusión de un programa bilingüe completo y obligatorio como la vía más fácil y segura de cumplimiento, aunque no sea esta siempre la realidad:

“El catalán, como lengua propia de Cataluña, lo es también de la enseñanza, en todos los niveles y modalidades educativos […]

Los centros de enseñanza de cualquier nivel deben hacer del catalán el vehículo de expresión normal en sus actividades docentes y administrativas.

-Los niños tienen derecho a recibir la primera enseñanza en su lengua habitual, ya sea ésta el catalán o el castellano […]

La enseñanza del catalán y del castellano debe tener garantizada una presencia adecuada en los planes de estudio, de forma que todos los niños, cualquiera que sea su lengua habitual al iniciar la enseñanza, han de poder utilizar normal y correctamente las dos lenguas oficiales al final de la educación obligatoria […]

En lo que se refiere a la enseñanza, la presente Ley garantiza a toda la población el pleno conocimiento de las dos lenguas y, a su vez, garantiza que el alumnado no sea discriminado ni separado en grupos diferentes por razón de la lengua, manteniendo el sistema de conjunción lingüística”. (op. cit.:14 y 20).

En determinados apartados de la ley, lo que constituye a nuestro juicio uno de los aspectos más interesantes del texto, se intenta definir de manera muy precisa la coexistencia de las lenguas mayoritaria y minoritaria. Este hecho es especialmente claro en lo que respecta a los medios de comunicación, en los que se llega a establecer cuotas concretas de presencia de ambas lenguas. Sin embargo, se vuelve a insistir en el fomento del uso exclusivo de la lengua minoritaria a través de subvenciones:

“Los medios de radiodifusión y televisión de concesión deben garantizar que como mínimo el cincuenta por ciento del tiempo de emisión de programas de producción propia de cualquier tipo y de los demás teleservicios que ofrecen sea en lengua catalana […]

Las emisoras de radiodifusión y televisión han de garantizar que en la programación de música cantada haya una presencia adecuada de canciones producidas por artistas catalanes y que como mínimo el veinticinco por ciento sean canciones interpretadas en lengua catalana o en aranés […]

El Gobierno de la Generalidad ha de fomentar y puede subvencionar las publicaciones periódicas de difusión general redactadas total o mayoritariamente en catalán”. (op. cit.:23-24).

Esta aparente contradicción se convierte en defensa clara del uso de la lengua minoritaria en otros ámbitos como las comunicaciones entre empresas públicas y clientes, la rotulación pública, la cartelería, etc.:

“Las empresas públicas han de utilizar normalmente el catalán en sus actuaciones y documentación internas y en la rotulación, las instrucciones de uso, el etiquetaje y embalaje de los productos o servicios que producen o (sic) ofrecen […]

RLyLA Vol. 14 (2019), 81-89 | 87 Luis Escoriza Morera Las lenguas oficiales del Estado español en los textos legales, ¿fomento o reconocimiento del plurilingüismo?

La señalización y los carteles de información general de carácter fijo y los documentos de oferta de servicio para las personas usuarias y consumidoras de los establecimientos abiertos al público deben estar redactados, al menos, en catalán. Esta norma no se aplica a las marcas, los nombres comerciales y los rótulos amparados por la legislación de la propiedad industrial”. (op. cit.:26-27).

4. CONCLUSIONES

De manera general, podemos afirmar que la regulación de la coexistencia del bilingüismo en el País Vasco, Galicia y Cataluña ha ido definiéndose progresivamente en las últimas décadas en los textos legales desarrollados en dichas comunidades, habida cuenta de la inexistencia de un marco legal en el ámbito nacional más allá del texto constitucional, cuya laxitud en materia de política lingüística ha quedado plasmada. La propia Constitución española delega dicha cuestión en los estatutos autonómicos. En los primeros textos, cercanos a la redacción del texto constitucional, simplemente se hace hincapié en el carácter cooficial de las lenguas mayoritaria y minoritaria en esas comunidades, impidiéndose la discriminación por razón de lengua. Son textos genéricos, sin especificación de los ámbitos de regulación del bilingüismo y cercanos a nuestro juicio a una postura de fomento de dicho bilingüismo. Es cierto que se plantea en ellos el carácter de lengua propia para la lengua minoritaria pero no se asocia ninguna medida concreta a dicha denominación.

La evolución de los textos legales regionales evidencia desde nuestro punto de vista un cierto cambio de postura o al menos un posicionamiento más claro inexistente en los primeros estatutos, derivado en gran medida del mayor desarrollo de esos textos, cuyo máximo exponente es la ley de política lingüística de enero de 1998 aprobada en Cataluña. Los principales cambios tienen que ver con la concreción de las implicaciones del carácter de lengua propia, ya que dicha lengua se convierte en lengua normal de uso, lengua oficial de la Administración pública y lengua de uso preferente u obligatorio en contextos oficiales (aspecto declarado anticonstitucional por el Estado español), como hemos visto. Pese a que el reconocimiento de la cooficialidad permanece, este convive con un planteamiento cercano al fomento de la lengua minoritaria, lo que implica cierta contradicción en los últimos textos analizados en Galicia y Cataluña siempre que no se apueste por un modelo bilingüe completo en todas las esferas del uso público (única manera de fomentar la lengua minoritaria sin eliminar el derecho de opción lingüística de los ciudadanos). Un ejemplo claro de esta situación se da en el ámbito educativo, tal y como hemos visto, donde se anima a los centros a usar la lengua minoritaria como lengua normal de uso al tiempo que se reconoce el derecho a elegir la lengua de la enseñanza a los ciudadanos pero se impide la separación en aulas por razón de lengua. Se echa en falta la especificación de la gestión de las lenguas mayoritaria y minoritaria en contextos oficiales de uso del mismo modo en que se especifica y se fomenta el uso de la lengua minoritaria en los mismos. Dicho fomento de la lengua minoritaria se fundamenta en el grado de identificación histórica de la misma con la comunidad y se refuerza con la idea de que dicha lengua haya sufrido un periodo de discriminación en el uso a favor de la lengua mayoritaria. No obstante, dicho planteamiento, que parece rayar en lo que podríamos denominar una discriminación positiva a favor de la lengua minoritaria, deja dudas, a nuestro parecer, a la hora de defender un modelo de sociedad bilingüe desde los textos legales, en los que el bilingüismo más que fomentarse como rasgo identitario parece simplemente reconocerse como derecho.

REFERENCIAS

Carta Europea de las Lenguas Regionales o Minoritarias (Boletín Oficial del Estado, número 222, de 15 de septiembre de 2001), pp. 34733-34749. Constitución española (Boletín Oficial del Estado, número 311, de 29 de diciembre de 1978). Declaración sobre los Derechos de los Pueblos Indígenas (2007). United Nations. Disponible en https://undocs.org/A/RES/61/295 [Último acceso: julio de 2019]. Declaración Universal de Derechos Humanos (1948). United Nations Disponible en http://www.un.org/es/universal- declaration-human-rights [Último acceso: julio de 2019]. Declaración Universal de Derechos Lingüísticos (1996). United Nations. Disponible en http://unesdoc.unesco.org/ images/0010/001042/104267S.pdf [Último acceso: julio de 2019]. España plurilingüe. Manifiesto por el reconocimiento y el desarrollo de la pluralidad lingüística de España. (2015) Disponible en https://socialismevedesocietat.files.wordpress.com/2015/12/manifest-cast.pdf [Último acceso: julio de 2019]. Hamel, R. E. (1995). “Derechos lingüísticos como derechos humanos: debates y perspectivas”, Alteridades, 5/10: 11-23. Ley 1/1998, de 7 de enero, de política lingüística (Diari Oficial de la Generalitat de Catalunya, número 2553, pág. 291, de 9.1.1998).

RLyLA Vol. 14 (2019), 81-89 | 88 Luis Escoriza Morera Las lenguas oficiales del Estado español en los textos legales, ¿fomento o reconocimiento del plurilingüismo?

Ley 3/1983, de 15 de junio, de normalización lingüística, aprobada por la Xunta de Galicia (Boletín Oficial do Parlamento de Galicia, número 128, de 30 de junio de 1983, pp. 1-7). Ley Orgánica 1/1981, de 6 de abril, de Estatuto de Autonomía para Galicia (Boletín Oficial del Estado, número 101, de 28 de abril de 1981, pp. 1-22). Ley Orgánica 3/1979, de 18 de diciembre, de Estatuto de Autonomía para el País Vasco (Boletín Oficial del Estado, número 306, de 22 de diciembre de 1979, pp. 1-19). Ley Orgánica 6/2006, de 19 de julio, de reforma del Estatuto de Autonomía de Cataluña (Boletín Oficial del Estado, número 172, de 20 de julio de 2006, pp. 27.269-27.310). López García, Á. (2009). La lengua común en la España plurilingüe. Lengua y sociedad en el mundo hispánico, 24. Madrid/Frankfurt: Iberoamericana/Vervuert. https://doi.org/10.31819/9783865279064 Manifiesto por la lengua común. Disponible enhttps://elpais.com/elpais/2008/06/23/ actualidad/1214209045_850215.html [Último acceso: julio de 2019]. Montero Cartelle, E. (1979). “Castellano o español a la luz de la Constitución. (La historia se repite)”, Senara, 1, 231-251. Pacto Internacional de Derechos Civiles y Políticos (Boletín Oficial del Estado, número 103, de 30 de abril de 1977, pp. 9337-9343). Parlament de Catalunya (2013). Estatuto de Autonomía de Cataluña. Texto consolidado. Barcelona: Parlament de Catalunya. Per un veritable procés de normalització lingüística a la Catalunya independent. Disponible en http://www.tribunacatalana.cat/data/documents/manifestkoine.pdf [Último acceso: julio de 2019]. Ruíz Vieytez, E. J. (2004). “Lenguas oficiales y lenguas minoritarias: cuestiones sobre su estatuto jurídico a través del derecho comparado”, en II Simposi Internacional Mercator: Europa 2004: Un nou marc per a totes les llengües?, Tarragona. Subirats i Humet, J. y Gomá, R. (coords.) (1998). Políticas públicas en España: contenidos, redes de actores y niveles de gobierno. Barcelona: Ariel. Süselbeck, K. (2008). “Lengua, nación e identidad en el discurso de la política lingüística de Cataluña”, en K. Süselbeck, U. Mühlschlegel y P. Masson (eds.) Lengua, nación e identidad: la regulación del plurilingüismo en España y América latina. Madrid/Frankfurt: Iberoamericana/Vervuert, 165-185.

RLyLA Vol. 14 (2019), 81-89 | 89 Revista de Lingüística y Lenguas Aplicadas Vol. 14 año 2019, 91-103 EISSN 1886-6298 https://doi.org/10.4995/rlyla.2019.10749

ZOMBIES LOST IN TRANSLATION. THE TRANSLATION FROM ENGLISH TO SPANISH OF (DE)HUMANIZING PRONOUNS

Linda Flores Ohlson University of Gothenburg, Sweden

Abstract: The present paper analyses which strategies are used in order to express the personal/inanimate pronoun contrast that serves the function of (de)humanizing zombies, when passages containing this linguistic feature in English are translated into Spanish. English has two sets of pronouns/adjectives, the ones that express personhood (he/his/him, she/her), and the inanimate ones (it/its). The explicit use of these pronouns is obligatory. Spanish on the other hand, has one set of pronouns (él, ella, su, lo, la) that are used both to express personhood as well as with inanimate references. The Spanish subject pronouns are normally used only when there is a need to highlight the subject or contrast it with another subject. Consequently, translators from English to Spanish face a challenge with regard to the translation of the (de)humanizing effect the pronoun contrast adds to the texts in English. The corpus contains examples of the English pronouns being translated with noun phrases, verb phrases, noun clauses, and pronouns, while in some cases the pronoun contrast is omitted, and therefore lost in the translation.

Keywords: English to Spanish translation, pronominalization, pronouns, dehumanization, fictional creatures, zombies.

1. INTRODUCTION

In Standard English, there are two sets of third person pronouns and possessive adjectives: the ones that express personhood (he/his/him, she/her), and the inanimate ones we use for inanimate objects and to a certain point, in reference to animals (it/its).The explicit use of the subject pronouns along with the verb is obligatory, and the two categories of pronouns and possessive adjectives and are normally distinct, i.e., in most contexts they cannot be used interchangeably (Biber et al., 1999:328). With regard to the use of the inanimate pronoun it, Joly (1975:260) states that it “clearly means that the referent is excluded from the sphere of humanity. […] Consequently, the referent is refused the full status of human personality.” This means that the inanimate pronoun can be used in order to dehumanize creatures, especially those creatures whose status as human/animal/monster is not clear cut. In fictional narratives, the prime example of this kind of creature is the zombie. Zombies, in most texts used to be human, occasionally someone close to us, like a friend or a family member, but the nature of their human/animal/monster status has then become ambiguous. This ambiguity is noticeably reflected in the use of pronouns referring to zombies, where frequent shifts between personal and inanimate forms are the norm (Flores Ohlson, 2018, Flores Ohlson, in press). This vacillation is, on the one hand, a reflection of the ambiguous nature of the zombie as a (non-)human creature. On the other hand, the change from one pronoun to another plays an important role in the creation of this creature. Translators from English to Spanish, of texts containing passages in which the contrast between personal and inanimate pronouns referring to zombies plays an important role, face a challenge with regard to the translation of the (de)humanizing effect the pronoun contrast adds to the texts.

This is due to the fact that, in Spanish, in contrast to English, the verb normally does not need to be accompanied by a subject pronoun. The pronouns are only used when there is a need to highlight the subject in order to distinguish it from another subject, or to emphasize it. When the pronoun does not serve the function of distinguishing or emphasizing, its use is redundant, and should be avoided (García Yebra, 1997:524-525). Moreover, the sets of third person singular personal and possessive pronouns/adjectives also differ substantially

To cite this article: Flores Ohlson, L. (2019). "Zombies Lost in Translation. The Translation from English to Spanish of (De)humanizing Pronouns". Revista de Lingüística y Lenguas Aplicadas, 14, 91-103. https://doi.org/10.4995/rlyla.2019.10749 Correspondence author: [email protected]

Received: 2018-09-26 Accepted: 2019-05-17 | 91 Linda Flores Ohlson Zombies Lost in Translation. The Translation from English to Spanish of (De)humanizing Pronouns

from the pattern found in English. Basically, there are only two subject pronouns (él/ella), as well as two direct object pronouns (lo/la) and these are used in reference to human beings as well as animals and inanimate objects. The third person singular possessive adjective (suyo/suya/su)1 is used for masculine and feminine humans and animals, as well as for inanimate objects.

Several studies have been carried out on the untranslatability of pronouns of address, for example of English you and Spanish tú and usted (Anderman, 1993), Russian ty and vy (Lewis, 2004:290), and French tu and vous in relation to Swedish du and ni (Künzli, 2009). In contrast, with regard to the English third person pronouns used to express personhood, and the inanimate pronoun, in relation to the Spanish counterparties, the process of translation has only been briefly mentioned in earlier studies. García Yebra (1997:528, 540), for example, states that it as a subject pronoun referring to inanimate objects and frequently animals is normally omitted in the translation. However, since he/she/it are generally not interchangeable, the fact that these pronouns are typically not translated at all in translation of English texts into Spanish does not generally lead to any loss in semantic content.

Clark (1992) discusses the use of pronouns in reference to animals in English, and notes that there is great inconsistency in their use. She argues that “[i]f the basic categories are indeed “personal” and “inanimate”, then such inconsistencies may not be only explicable but inevitable, in so far as subjects which are animate yet non- human fit uneasily into the pattern” (Clark, 1992:636). Clark further states that “[a]lthough choice of generic pronoun in reference to a particular species is not always consistent even within a single work, variation is not necessarily either random or unmotivated” (Clark, 1992:639), and she concludes by noting that “[t]he grammatical vacillation in fact mirrors the profound ambivalence characterizing all human attitudes towards animals” (Clark, 1992:644). Hence, in the case of animal pronominalization, one could argue that when the pronoun is not translated from English into Spanish, it could imply a loss in translation since the pronoun itself can add a certain meaning to the text. Nevertheless, in Clark’s examples, even though one and the same animal is referred to both with personal and inanimate pronouns within the same text, and sometimes even within the same sentence, the fact that the animal in question is a non-human creature is always obvious. That is, the choice of personal or inanimate pronouns says something about the ambivalence in the relation or attitude the speaker expresses towards the animal, but nothing about the nature of the animal in question.

By contrast, this is not the case when it comes to the zombie. An omission of the contrast between personal and inanimate pronouns in reference to this fictional creature in a translation into Spanish of an English text would most certainly involve a loss of semantic and/or pragmatic content in the translation. However, as stated by Toury (2012:116), “[i]t simply cannot be taken for granted that whenever a feature occurs in an SL text, be its position ever so high in the latter’s internal hierarchy, this feature will be picked by translators and retained in the translation.” Hence, in the present paper I ask the following question:

• Which strategies (if any) are used in order to express the personal/inanimate pronoun/adjective contrast that serves the function of (de)humanizing the zombie, when passages containing this linguistic phenomenon in English are translated into Spanish?

One could argue that the modern zombie was born in George A. Romero’s motion pictures from the 1970’s and 1980’s2 (ZRS, 2010). Hence, this creature is originated on the silver screen and in English. It should be noted, however, that the number of written zombie texts has increased dramatically in the last decades and currently the written genre is possibly even more productive than the audiovisual narratives. Although the number of zombie texts written in other languages is also increasing, the English language is still the primary source of zombie literature and motion pictures. Consequently, within the Spanish-language literature, the translated narratives, written or produced originally in English, can be considered to occupy the primary position within the polysystem hierarchy of zombie narratives.3 Even-Zohar (1990:46) argues that translated literature is not merely “an integral system within any literary polysystem, but [an] active system within, [and] it is by and large an integral part of the innovatory forces”. In the present analysis, I will comment on an example that could serve as an indication of English originals and their translations occupying the primary system of zombie narratives in Spanish.

The extracts analyzed in the present paper come from a corpus of 21 written and audiovisual zombie narratives. The titles, retrieved from sites on the internet that list the best and most popular zombie narratives, were selected due to their accessibility in the English original and the Spanish translation. However, due to the enormous number of zombie narratives published in the last decades, the selection was also to some extent random in nature.

The zombie represents one out of many different fictional creatures that could have been analyzed for the purposes of this study. In Flores Ohlson (2018), I study the pronominalization of the creatures in Guillermo del

1 The forms suyo/suya are used after the noun while the short form su is used before the noun. 2 Night of the Living Dead (1968), Dawn of the Dead (1978), Day of the Dead (1985). 3 The polysystem analyses the socio-semiotic phenomena, such as culture, language, and literature, as “networks of relations that can be hypothesized for a certain set of assumed observables” (Even-Zohar, 2005:1).

RLyLA Vol. 14 (2019), 91-103 | 92 Linda Flores Ohlson Zombies Lost in Translation. The Translation from English to Spanish of (De)humanizing Pronouns

Toro’s fictional work, which are trolls, fairies, and vampires. However, although interesting, the relation between the human characters and these creatures is not as complicated and ambiguous as in the case of the zombie. Furthermore, the popularity of the zombie in the last decades has been enormous. Cohen (2012:398) claims that zombies “thoroughly saturate pop culture, and assures that the “future belongs to the rotting, groaning, lumbering, hungry, herdlike walking dead.” Spooner (2015:183) states that the modern zombie is used to explore a variety of themes directly related to humankind such as capitalist consumption, viral pandemic, environmental disaster, military irresponsibility and celebrity culture. Additionally, even though fairies, trolls, vampires, and other fictional creatures exhibit human characteristics, distinguishing them from humans is generally uncontroversial. With regard to the zombie, on the one hand, Boon (2011:50) argues that the “proliferation of zombie mythology into mainstream culture during the past three decades has established the zombie as the predominant symbol of the monstrous other”. On the other hand, the zombies can at the same time be interpreted as a symbol of ourselves:

if you think about a monster like the vampire or the werewolf, you can see them as aspects of human behavior magnified and embodied; i.e. the vampire’s connection to various kinds of (taboo) eroticism has been explored ad infinitum, while the werewolf’s link to animal violence has also been recognized. With the zombie, what you get is us, pretty much as we are, maybe with a little damage, and we consume one another. No eroticism, no animal violence, just a single, overwhelming appetite. (Adams, 2008:77-81)

In analyses of zombie movies and literature, several writers observe that humans sometimes exhibit the same or similar characteristics as the zombies. Weinstock (1999:8), for example, in his discussion of The Night of the Living Dead (Romero 1968) notes that “there is little to distinguish the living from the dead” since some of the living characters are “as vapid and unemotional as the zombies they mercilessly pick off, one by one.” It should be noted that some of the survivors in zombie narratives even become cannibals, which is one of the most striking characteristics of the 21th century zombie. In fact, in the words of the character named Cassandra in Z Nation, humans can be “worse than Zs” (Engler and Schaefer, 2014-2018:season 1, episode 3). The nature of the zombie as a non-human or human creature, as the monstrous other or as a symbol of ourselves is thus highly ambiguous. This is clearly reflected in the pronominalization in reference to this creature. My previous studies of pronominalization of zombies in literary texts in English show that there is great inconsistency, both between different texts, and commonly also within one and the same text (Flores Ohlson, 2018, Flores Ohlson, in press). Hence, there is clearly no obvious choice of pronouns referring to the zombie. The use of the inanimate pronoun it excludes the zombie from the sphere of humanity and dehumanizes it. Dehumanizing a creature has great moral implications since, when a creature is not categorized as human, killing it generally does not imply committing murder, and there are fewer moral issues to be concerned about. Pifer (2011) reflects on one of numerous examples of this, found in the motion picture Shaun of the Dead (Wright, 2004):

when Ed hits a pedestrian with Pete’s car, Shaun is racked with guilt, cautiously calling out to the body in the street. When the corpse reveals its snarling zombie visage, Shaun says, “Oh, thank heavens for that,” and we can all breathe a sigh of relief. We thought for a moment that our protagonists had killed someone. (Pifer, 2011:169)

Hence, the zombie that was hit by the car was clearly not human (a someone), but rather a thing (a something).4 Greene and Mohammad identify three general justifications for killing zombies:

a) they are (or are very close to being) braindead; b) because of their radically altered life goals, what personality is left lacks the same personal identity as the original individual, releasing us from prior obligation; and c) after zombification they are usually homicidal cannibalistic killers. (Greene and Mohammad, 2010:28)

These justifications can possibly mean that the zombie could be considered a non-human, or a no-longer- human being. As we will see in some of the examples analyzed below, killing a he or a she is certainly not the same as killing an it.

Jakobson (2012:129) observes that “[l]anguages differ essentially in what theymust convey and not in what they may convey”, whereas Livia (2001:12), in regard to literary uses of linguistic gender, notes that, “because different languages have different structures that create different sets of oppositions, members of one linguistic community will not experience the world in the same way as members of another.” In the case of the discrepancies between English and Spanish in the structure and use of personal pronouns, a simple change of pronoun in an English text, such as I killed him/it, would normally not be captured in the Spanish version (Lo maté).5 The use of him or her in reference to the animal in order to express what Clark (1992:640) would call a “fellow-feeling” with the animal, would not change the content. I.e., it is still obvious that it was an animal killing and not a human murder.

4 Another clear example of this is found in Stephen King’s Cell (King, 2006) where one of the main characters insists that burning a large group of zombies “wouldn’t be murder, not really. It would be extermination.” 5 The contrast human/non-human could be reflected with third person singular pronouns in Spanish in texts that exhibit the pronoun phenomenon called leísmo. Leísmo happens when indirect object pronoun le is used instead of direct pronoun lo in regard to human masculine singular referents. A speaker of a variety of Spanish that exhibits this phenomenon would say Le vi (I saw him) and Lo vi (I saw it), while a non-leísta speaker would use the direct pronoun lo for both references. Leísmo is common among well-educated and prestigious speakers and writers of Spanish, therefore a phenomenon accepted by the Spanish Royal Academy (Real Academia Española 2005).

RLyLA Vol. 14 (2019), 91-103 | 93 Linda Flores Ohlson Zombies Lost in Translation. The Translation from English to Spanish of (De)humanizing Pronouns

Accordingly, in the Spanish version, there would be no doubt whether lo meant the killing of an animal or the murder of a human. However, it will be argued here that in the case of zombies, this type of pronominal shift has direct implications for the perception of its human/animal/monster status.

In my previous work of fictional creature pronominalization in written and audiovisual texts in English, two main types of pronominalization were identified, by closely analyzing pronoun use in reference to zombies, vampires, trolls and fairies in a corpus of around 20 narratives. The most common type is when there is a vacillation between personal and inanimate pronouns that cannot be interpreted as conveying an explicit meaning of (de)humanization of the creature in question, as the following example shows:

He brought the light up, right into its face, and the vampire flailed […]. Fet ducked him and got behind his back, stabbing the body guard in the back of its thick neck before shoving him hard down the stairs. (del Toro and Hogan, 2009b:102)

The pronoun use in the extract above can be interpreted as showing that the human character has an ambivalent attitude towards the vampire he is fighting against, hence, it is not clear to him whether this creature is human or not. However, in comparison with the following example, the contrast between the personal and inanimate pronouns in the example above is not used as a clear linguistic tool.

I tried to put her, it out of my mind. (Brooks, 2006:221)

In this example, the human character corrects herself, changing from the personal to the inanimate pronoun as she refers to a zombie girl. This correction is in line with the argument she has expressed about the importance of not thinking about zombies as humans. Hence, the contrast between the two pronouns explicitly conveys a certain meaning to the text since the pronoun contrast itself is used as a linguistic tool in order to dehumanize the zombie. For the purpose of analyzing the strategies used in the translation of pronouns in English zombie texts into Spanish, I consider the last type the most interesting. When a specific linguistic feature, such as pronominalization, conveys a certain meaning to the text, it should not be overlooked in the translation since an important message would be lost. Examples from the last category will therefore be analyzed in the following. A few comparisons with examples from the first type will also be briefly discussed.

As we will see in the analysis, one of the instances of pronominalization in my corpus can be interpreted as an example of formal equivalence translation, a type of translation which “attempts to reproduce several formal elements, including: (1) grammatical units […]. The reproduction of grammatical units may consist in: (a) translating nouns by nouns, verbs by verbs, etc.” (Nida, 2012:149). Other examples in my corpus are more associated with dynamic equivalence translation, which is “the closest natural equivalent to the source-language message” (Nida, 2012:151). This type of translation:

involves two principal areas of adaptation, namely, grammar and lexicon. In general the grammatical modifications can be made the more readily, since many grammatical changes are dictated by the obligatory structures of the receptor language. That is to say, one is obliged to make such adjustments as shifting word order, using verbs in place of nouns, and substituting nouns for pronouns. (Nida, 2012:151)

Four out of the seven cases analyzed in the present paper could be argued to be examples of dynamic equivalence translation.

2. ANALYSIS

In the corpus of the present paper, consisting of written as well as audiovisual texts originally produced in English and then translated into Spanish, seven cases of pronoun contrast as a clear linguistic tool have been identified. In these cases, zombie pronominalization is used as a tool in order to express a specific meaning in the text, i.e., to humanize or dehumanize the zombie, or to reveal an ambivalent attitude towards the nature or status of the zombie as a human/animal/monster being. In these seven cases, five different solutions with regard to the translation of the pronoun contrast have been found:

• Noun phrase (Z Nation, World War Z) • Verb phrase (Dead Heads) • Noun phrase + Noun clause (The Girl with All the Gifts) • Pronouns (Zombie Fallout) • Omission of pronoun contrast (Black Mirror, Cell, Dead Heads)

In the following, each solution will be discussed together with its respective examples.

RLyLA Vol. 14 (2019), 91-103 | 94 Linda Flores Ohlson Zombies Lost in Translation. The Translation from English to Spanish of (De)humanizing Pronouns

2.1. Noun phrase In the TV series Z Nation (Engler and Schaefer, 2014-2018), there is a scene in which the young character, who calls himself 10 K, is talking to the older Doc about the time when he had to kill his father who had been infected and turned into a zombie:

Doc: - What did you do? Doc: - ¿Qué hiciste? 10 K: - I killed it. 10 K: - Lo maté. Doc: - Damn kid, you had to put down your own dad? Doc: - Maldición, ¿cómo pudiste hacerlo? 10 K: - I didn’t kill him, I killed it. 10 K: - No asesiné a papá, fue a esa cosa. (Engler and Schaefer, 2014-2018:season 1, episode 4)

In this dialogue, Doc does not seem to react to 10 K’s first use of “it”, since he directly responds with “your own dad”. However, this is an example of non-explicit pronoun contrast. In the dubbed Spanish version, the object pronoun “lo” does not convey the same meaning, in as much as it could refer to a human as well as a non-human creature. As I have found in numerous cases, the meaning that the first “it” adds to the expression in this extract constitutes a use of pronoun contrast that would typically be ignored in the translation, as in the following example:

- You’ve seen him before. This thing. The Master. - ¿Lo has visto antes? ¿A esa cosa; al Amo? - Yes. - Sí. - You tried to kill it. -¿Intentase matarlo? (del Toro and Hogan, 2009b:444) (del Toro and Hogan, 2009a:363-364)

The example above does not present a use of the pronoun contrast as a clear linguistic tool. Although it does show a vacillation in the pronominalization that reflects the characters’ ambivalent attitude towards the creature they call The Master, the pronoun contrast is not explicit. In this extract, both “him” and “it” are translated with the expected masculine direct object pronoun “lo”, and consequently, the contrast is lost in the translation.

However, in the last sentence of the Z Nation dialogue, 10 K stresses the two pronouns in order to emphasize the difference between killing his father and the zombie his father had turned into. That is, 10 K uses the inanimate pronoun in line with Joly’s (1975:260) statement that reads that it “clearly means that the referent is excluded from the sphere of humanity.” Consequently, 10 K did not murder the human being his father used to be, but the non- human creature he had turned into. As we can see, the pronouns “him” and “it” are translated with the noun “papá” (dad) and the noun phrase “esa cosa” (that thing).

The noun phrase esa cosa in reference to zombies is frequently used in zombie apocalyptic texts originally written in Spanish, as in the following extracts:

salieron dos brazos y la cabeza de esa Out came that thing’s arms and head. cosa. Oh, eso de ahí no era Miguel, pero lo It wasn’t Miguel, but it had been. […] había sido hasta hacía muy poco. […] había Then one of those things had bitten him. conseguido que esas cosas lo mordieran. (Loureiro, 2012:81) (Loureiro, 2011:108)

Susana los observó con incrédula fascinación. Susana watched them with incredulous Eran esas cosas. Eran ésos de la televisión. fascination. They were those things. Eran gente muerta, o eso pensaba. Cosas The ones on television. They were dead muertas. Muertos vivientes. people. Dead things. The living dead. (Sisí, 2009a:19) (Sisí, 2009b:16)

One could consider whether the expression esa cosa is commonly used to refer to non-human creatures in discourse other than zombie apocalyptic texts. A search for this noun phrase in the Spanish Academy’s corpus of contemporary language (Real Academia Española, 2017), reveals 459 cases in 377 different texts. Of these cases, the noun phrase is used in reference to two fetus, one baby, three dead persons, two men, one woman, one langouste, one dog, two unspecified creatures, and one mermaid. In the remaining 445 cases, it is used in reference to non-living objects or abstract concepts. It is possible, therefore, that esa cosa in reference to “living” creatures is more frequent in zombie texts. Even-Zohar argues that when the translated literature

RLyLA Vol. 14 (2019), 91-103 | 95 Linda Flores Ohlson Zombies Lost in Translation. The Translation from English to Spanish of (De)humanizing Pronouns

assumes a central position, in the process of creating new, primary models, the translator’s main concern here is not just to look for ready-made models in his home repertoire into which the source texts would be transferable. Instead, he is prepared in such cases to violate the home conventions. Under such conditions the chances that the translation will be close to the original in terms of adequacy (in other words, a reproduction of the dominant textual relations of the original) are greater than otherwise. (Even-Zohar, 1990:50)

This is in line with my argument, based on the polysystem theory (Even-Zohar, 1990:46-47), that narratives translated from English occupy the primary position in the zombie narrative polysystem and therefore constitute the leading influence in the formation of new linguistics models for the target language, in this case Spanish.

In apocalyptic zombie narrations, zombies typically pose an important threat to the survival of the human race, i.e., they are “usually homicidal cannibalistic killers” (Greene and Mohammad, 2010:28). Therefore, differentiating them from humans is imperative to survival. This fact is emphasized in the following scene from the bestselling novel World War Z (Brooks, 2006):

The first G I saw was small, probably akid , I El primer monstruo que vi era pequeño, couldn’t tell. Its face was eaten off, the skin, probablemente una niña, no estoy segura. Tenía nose, eyes, lips, even the hair and ears […]. la cara comida: la piel, la nariz, los ojos, los labios, e incluso el pelo y las orejas […]. estaba It was stuck inside one of those long civilian metida dentro de uno de esos macutos de hiker’s packs, stuffed in there tight with the excursionista, allí atrapada, con el cordón que drawstring pulled right up around its neck. cerraba la bolsa apretado en torno al cuello.

[…]. it was splashing around […]. Its brain […] la cosa estaba chapoteando […]. Debía must have been intact […] It couldn’t moan, de tener el cerebro intacto […] No podía gemir, its throat had been too badly mangled, but tenía la garganta demasiado destrozada, pero the splashing might have attracted attention, el chapoteo podía llamar la atención, así que so I put it out of its misery, if it really was la liberé de su desdicha, si es que la sentía, e miserable, and tried not to think about it. […] intenté no pensar más en ello. […] don’t try to imagine who they used to be, how no intentes imaginarte cómo eran antes, ni they came to be here, how they came to be cómo llegaron hasta donde están, ni cómo se this. convirtieron en lo que son.

I know, who doesn’t do that, right? Who Lo sé, ¿quién no se lo pregunta, verdad? doesn’t look at one of those things and just ¿Quién es capaz de mirar a una de esas cosas naturally start to wonder? It’s like reading the y no empezar a preguntarse, sin quererlo? Es last page of a book…your imagination just como leer la última página de un libro…, tu naturally spinning. And that’s when you get imaginación empieza a dar vueltas, sin más. distracted, get sloppy, let your guard down Y es entonces cuando te distraes, cuando and end up leaving someone else to wonder te vuelves torpe, cuando bajas la guardia y what happened to you. acabas dejando que otro se pregunte qué te I tried to put her, it out of my mind. Instead, I pasó a ti. Intenté quitarme a la niña, al zombi, found myself wondering why it had been the de la cabeza, y entonces empecé a pensar en only one I’d seen. por qué era el único que había visto.

(Brooks, 2006:220-221) (Brooks, 2013:244)

To the character in the extract above, it is clear that it is important to see the zombies as non-human creatures. She refers to them by the letter G, which is probably an abbreviation of ghouls (ZRS, 2010) and Zack, an anthroponomy used in military fashion, similar to Charlie as a reference to the Vietnamese in the war in Vietnam.6 Yamamoto (1999:4), states that “plurality sometimes weakens the sense of animacy [i.e., the level of humanness we perceive] because the identity of the referent can be blurred”. The level of dehumanization is also clear when we look at the pronouns used, which are, with a sole exception, the inanimate ones. It can be observed how the character reminds herself of the zombies’ lack of feelings (“if it really was miserable”), and directly afterwards stresses the importance of not thinking about who they used to be when they were still human.

6 The Vietnamese army was named Viet Cong, abbreviated V.C., which in the American military alphabet was Victor Charlie (Moser and Drejer 1955).

RLyLA Vol. 14 (2019), 91-103 | 96 Linda Flores Ohlson Zombies Lost in Translation. The Translation from English to Spanish of (De)humanizing Pronouns

A comparison of the English original to the Spanish translation reveals clear differences in the use of the pronouns as well as the possessive adjectives between the two languages. The obvious dehumanization of the creature in question by the use of it/its, is largely lost. On the one hand, this loss is due to the total absence of subject pronouns in the translation. On the other hand, with regard to the possessive adjectives, the striking feature is not so much that his, her and its are all translated into su, but rather that the pronoun is avoided altogether, in line with the general principle in Spanish when referring to body parts or personal belongings (Real Academia Española, 2010:352). It is hence, the natural way of using, or rather not using the subjective and possessive pronouns/adjectives in Spanish that leads to the loss in translation. In a text where the human/animal/inanimate nature of the referent is clear, presumably the most frequent case, this difference in pronoun/adjective use would not imply a loss in translation. It should be noted that according to García Yebra (1997:531), the subject pronoun should be expressed whenever its omission could produce ambiguity or complicate the comprehension of the text. However, ambiguity or difficulty in comprehension might not be the case here, and expressing the subject pronoun in the Spanish text would often render the text unnatural.

There is in fact one attempt to compensate for this loss in the translation, where “it was splashing around” is not merely translated into estaba chapoteando, but into “la cosa estaba chapoteando”. Nevertheless, there are three inanimate subject pronouns, and five inanimate possessive adjectives that are ignored in the same translation. Still, up until the last sentence in the extract, the pronoun contrast is not explicit.

If we now consider the second last sentence of the scene, it can be noted that the character corrects herself, and contrasts the object pronoun her with the inanimate it. This pronoun use reinforces the character’s argument with regard to zombies’ non-human status and the importance of remembering that zombies are not human any longer. This pronoun contrast is thus a clear linguistic tool. This is the second of two examples in the corpus of the use of nouns and/or noun phrases in the translation. That is, her is translated with the noun phrase la niña (the girl), and it with al zombi (the zombie).

In the last sentence, the inanimate pronoun “it” is repeated, as to reinforce the character’s argument and conviction of the zombie’s non-human status. In the translation, since the masculine form of the only (“el único”), instead of the feminine form (la única) is used, it clearly refers to “el zombie”, and not to “la niña”. Hence, the same effect is achieved.

2.2. Verb phrase In the motion picture Dead Heads (Pierce and Pierce, 2011) there are two kinds of zombies: the traditional zombies that lack human conscience and do not speak, and the more humane zombies who resemble normal humans, except that their bodies are dead and decomposing. In the following dialogue, the humane zombie Mike wakes up to find his friend Bran playing with a more traditional zombie.

Mike: - The fuck is that? Mike: - ¿Qué coño es eso? Bran: - That? That’s Cheese. Bran: - ¿Este? Es Cheese. Mike: - That’s Cheese, ok. What’s he Mike: - Eh, es Cheese, vale. ¿Qué está doing here? haciendo aquí? Bran: - I’m teaching him to play fetch. Bran: - Estoy enseñándole a jugar. […] […] Mike: - What are you doing? That’s a Mike: - ¿Qué estás haciendo?, es un zombi. zombie. Bran: - We’re zombies! Bran: - ¡Somos zombis! Mike: - Yeah, but we’re good zombies. Mike: - Pero somos zombis buenos. Bran: - He’s a good zombie. Bran: - Es un zombi bueno. Mike: - Does he know that? What are we Mike: - ¿Lo sabe? ¿Qué pretendes hacer supposed to do with it? con él? Bran: - First off, that’s not anit , that’s Bran: - Para empezar, tiene nombre, es Cheese. Cheese. (Pierce and Pierce, 2011)

Although not a case of third person pronouns, the first line of this extract is nevertheless interesting. Mike says “The fuck is that”, presumably a short version of What the fuck is that. Thus, it could be interpreted as an immediate dehumanization of the zombie on part of this character, whereas Who the fuck is that would have humanized the zombie in question. This linguistic dehumanization is even more clearly expressed in the translation since “qué” (what) is used instead of quién (who), and “eso” (that) instead of ese (he/that one).

RLyLA Vol. 14 (2019), 91-103 | 97 Linda Flores Ohlson Zombies Lost in Translation. The Translation from English to Spanish of (De)humanizing Pronouns

In the second line, it is interesting to see how the humanization of the zombie on part of Bran is actually more reinforced in the Spanish version than in the English original. This is due to the fact that, while “that” in English can be used in reference to both human and inanimate referents, the demonstrative pronoun “este” is used in this context in order to present a person, and esto would be used to explain what something is (Este es Juan = This is Juan. Esto es un libro sobre Cervantes = This is a book about Cervantes).

In the lines that follow, Mike seems to accept the more or less human status of the zombie, and both characters use the personal pronouns in their conversation about Cheese. I use the expression more or less because it could be argued that Bran treats the zombie more like a dog than a human, in as much as he gives the zombie a name more fitting for a pet than a human, he plays fetch with Cheese, and the expressions he uses are typical of communication with dogs (“come here boy, go get it boy, come on buddy”). According to Chen (2012:35), “dehumanization insults hinge on the salient invocation of the nonhuman animal”, which means that the treatment of Cheese as a dog could in its own be interpreted as a form of dehumanization.

In the part that has been omitted in the transcription, Bran tries to show Mike how he has been teaching Cheese to dance like Michael Jackson in the music video of “Thriller”. This is something Mike strongly disapproves of, and in the following lines he suddenly attempts to dehumanize Cheese again, by switching from “he” to “it” (“What are we supposed to do with it?”). This line, together with Bran’s answer, are the examples I interpret as the use of pronoun contrast as a clear linguistic tool. The expression “that’s not an it” clearly shows that Bran disagrees with Mike’s attempt to dehumanize the zombie.

Surprisingly, the attempt of dehumanization in “What are we supposed to do with it?” is not translated in the Spanish version. The subject pronoun “él” could refer to a human or animal referent as well as to an inanimate one. Hence, Dead Heads also appears in the list of cases where the pronoun contrast has not been translated. However, even though I interpret this example as a case of a clear linguistic tool, it does not show such an explicit contrast between the pronouns as the last line does. Instead of translating the inanimate pronoun with a noun phrase, such as no es una cosa (it’s not a thing), a verb phrase is used (“tiene nombre” = he has a name). This is in line with the fact that Bran treats the zombie more like a pet than as a fellow human/zombie being. In other words, it would be more expected to talk about an animal not being an it, and having a name, than to talk about a human in such a way.

2.3. Noun clause In a similar manner as in the motion picture Dead Heads, in the novel The Girl with All the Gifts (Carey, 2014), there are two different kinds of zombies. When the surviving humans discover the existence of the second kind (the more humane one who can learn how to speak and does not merely act on cannibalistic instinct), they decide to capture a few of them for examination. The humane zombies are children and the scientists order the soldiers to start by bringing one of them to the camp.

Bring us one of those kids. Let’s take a good Traednos a uno de esos niños. Queremos long look at him/her/it. examinar a uno de esos críos, crías o lo que sea. (Carey, 2014:78) (Carey, 2015:76)

It is evident that the existence of the humane zombies challenges the previous dichotomy of human/zombie that the surviving humans had established. This fact is mirrored in the cited extract above. The humane zombies are obviously children, but are they to be considered human or not? In the Spanish version, the personal pronouns “him” and “her” are translated with the same strategy that we have seen in several of the previous examples analyzed: the noun. Since the Spanish noun crío (child) has a feminine version (cría), it works in the translation of “him” and “her”. What the expression in the English original conveys is exactly the idea of a creature whose nature as a human/non-human being is uncertain, and the use of the inanimate pronoun plays an important part in this message. Without it, the message would only come down to something like let’s examine one of them, boy or girl, it doesn’t matter which. The Spanish version of the expression uses the solution of a noun or relative clause: “o lo que sea” (or whatever /it is/). At this point, we find the neuter pronoun “lo” which, among other things, is used in reference to unspecified objects. This noun, together with the subjunctive of the verbser (to be), reinforces the meaning of reference to an unknown, unspecified object. Yet, it could also be interpreted in line with the English version without the inanimate pronoun: let’s examine one of them, boy or girl, it doesn’t matter which. To unmistakably convey the meaning of dehumanization, it would perhaps have been better to use the plural form of the verb: o lo que sean (or whatever they are).

RLyLA Vol. 14 (2019), 91-103 | 98 Linda Flores Ohlson Zombies Lost in Translation. The Translation from English to Spanish of (De)humanizing Pronouns

2.4. Pronouns In the novel Zombie Fallout (Tufo, 2010b), one could argue that the human character’s metalinguistic comment sums up the basic functions of the most frequent zombie pronominalization, namely dehumanization in order to justify their extinction as a necessary and morally accepted procedure rather than a cruel murder.

“I know, we see it too”. It amazed me that I “Lo sé, nosotros también lo vemos”, le susurré. was already able to call them “its”, instead Me asombró que ya estuviera llamándolos of “thems” to describe what was once “ello” en vez de “ellos”. “ello” parecía una human. But it was much easier this way. palabra tan impersonal para describir lo que una (Tufo, 2010b:audio edition 01:06:50-01:07-04) vez fuera humano, pero era mucho más fácil así. (Tufo, 2010a:552-559)

Although the extract presents a pronoun contrast, and a metalinguistic comment about the dehumanizing function of the pronoun it (“to describe what was once human”), it is not clear how the use of its and thems contributes to the dehumanization of the zombies, since them obviously can be used in reference both to animate and inanimate objects. In other words, them is not the personal counterpart of inanimate it. In consequence, to use them in reference to zombies does not necessarily mean that they are humanized. Furthermore, the inanimate pronoun it can be a subject and an object pronoun, and it is used in the singular, while them can only be an object pronoun and it is used in the plural.

As for the translation, it should be noted that the use of “lo” to translate “it” in the first sentence does not express the dehumanization found in the original.7 However, this example of pronominalization belongs to the category of non-explicit pronoun contrast, while the second sentence exhibits a case of pronoun contrast as a clear linguistic tool.

Similar to the fact that the contrast between “its” and “thems” in the original is not entirely successful in terms of its dehumanizing effect, the contrast between “ello” and “ellos” in the translation is equally ineffective. As stated by García Yebra (1997:537), it should be noted that ello is not a personal pronoun since it never refers to a person, but represents an action, a process or a state mentioned earlier in the text. This pronoun can therefore not be compared either in function or in frequency of use to the English inanimate pronoun it. Furthermore, “ellos” is used not merely in reference to living creatures, but to inanimate objects as well, especially after prepositions: Vimos unos edificios grandes y entramos en uno de ellos. (We saw some big buildings and we entered one of them) (Butt and Benjamin, 2013:132).

More importantly, however, the sentence “Me asombró que ya estuviera llamándolos ‘ello’ en vez de ‘ellos’” (It amazed me that I was already able to call them “its” instead of “thems”), has no logical connection to what was previously stated, since the pronoun “lo” does not convey the meaning of dehumanization, in line with the discussion above. That is, the character is amazed or surprised about something that has no relation to what has previously been said.

However, what is especially conspicuous is the metalinguistic comment about “ello” added to the translation: “‘Ello’ parecía una palabra tan impersonal para decribir lo que una vez fuera humano” (“Ello” seemed to be such an impersonal word to describe what was once human). This additional comment could be interpreted as serving the function of emphasizing the intended dehumanization of the zombies through the use of ello instead of ellos.

2.5. Omission of contrast between pronouns In the Netflix series Black Mirror (Brooker, 2011-), one of the episodes, Men Against Fire, is about pale, snarling, humanoid monsters called “roaches”. These creatures could be interpreted as a kind of zombies. In the scene below, a soldier has been in a close fight with one of these creatures and is now talking to a psychologist about the incident. It should be noted that the Spanish translation has two different versions, the first one being the dubbed, and the second the subtitled version.

7 It would only be in a text that exhibits leísmo, that lo could be considered dehumanizing. That is, in the variety of Spanish where third person indirect object pronoun le is used for masculine human direct objects in singular, the use of lo would be interpreted as referring to a non-human direct object. However, the following extract proves that this is not the case with this text: “Pero aún Travis me miró con ojos suplicantes, sin poder creer que su propio padre lo pondría en peligro” (Tufo, 2010a:155). A leista text would have written le podría en peligro.

RLyLA Vol. 14 (2019), 91-103 | 99 Linda Flores Ohlson Zombies Lost in Translation. The Translation from English to Spanish of (De)humanizing Pronouns

Soldier: - He was on the ground with me. Soldier: - El tío/Él estaba en el suelo Psychologist: - He? Psychologist: -¿Tío?/¿Él? Soldier: - Yeah, it was a he. […] He was on top of Soldier: - Sí, era un tío. /Era un hombre […] Sí, me, struggling, couldn’t reach my rifle so I had to encima, forcejeando conmigo, no llegaba al arma stick him, ah stick stick it, with my knife. así que tuve que pincharlo, pincharlo con mi (Brooker, 2011-:season 3, episode 5) cuchillo.

In this story the creatures are not really zombies but a community of people who live as outcasts and are rejected by the authorities. The soldiers who chase them have a neural implant that makes them see these individuals as terrifying monsters, which gives them the motivation and moral right to attack and kill them. No one in the story knows that the soldier in the extract above has a malfunctioning implant which allows him to see his victims for what they really are: ordinary humans. Therefore, the fact that the soldier first uses personal pronouns when referring to the individual he killed indicates some kind of problem, and the psychologist expresses his reaction both verbally and through his facial expression. The soldier seems to notice the psychologist’s disapproval, and in the following he dehumanizes the creature by switching to the inanimate pronoun.

The translated versions reveal a semantic extension of the noun “tío” that exists in some varieties of Spanish, where the word is not only used with the original meaning of uncle, but also with the meaning of guy (Real Academia Española, 2012).

If we only look at the first three lines, the use of the pronouns could suggest that the point the psychologist reacts to is the gender of the zombie, i.e., male rather than female zombie. Thus, the soldier’s dehumanization of the zombie by switching from the personal “he” to the inanimate “it” represents a crucial part of the dialogue. It is therefore conspicuous that the pronominalization in this important part of the dialogue has not been translated. In order to adapt the dubbed version to the original speech of the soldier, the same phrase is merely repeated with the direct object pronoun “lo” used twice. A strategy similar to the extract from Z Nation (Engler and Schaefer, 2014- 2018) could have been used in so much as the first “pincharlo” could have been translated intotuve que pinchar al tío/hombre (I had to stick the guy/man), and the second into a la cucharacha,8 for example. Consequently, as will be shown, this is one of only two examples in my corpus among the cases of pronoun contrast as a linguistic tool that fail in the sense that the human/inanimate contrast between the pronouns is lost at the point where it plays its most crucial role.

The creatures in King’s novel Cell are a clear example of how fuzzy and ambiguous the line between human and zombie can be. At the beginning of the novel, the zombies are the typical braindead creatures acting on cannibalistic instinct. As the narrative progresses, however, the zombies start to communicate, to collaborate, and their violent behavior partly wanes.

The Raggedy Man stopped about ten feet from El Hombre Andrajoso se detuvo a unos tres him. He - it - was standing on the Head’s grave. metros de él, sobre la tumba del director. (King, 2006:audio edition: 07:47:41-07:47:48) (King, 2012:3903)

In the extract above, we find the zombie that is referred to as The Raggedy Man by the human characters. He plays an important role in the story since he communicates with the human characters with some kind of mind reading/talking. As has been noted before (Footnote 4), one of the characters in the novel claims that killing zombies “wouldn’t be murder, not really. It would be extermination” (King, 2006:audio edition: 05:21_42-05:21:46), and the change of pronoun that dehumanizes The Raggedy Man is an important linguistic reminder of the moral standards the human characters act upon. However, this contrast between pronouns is lost in the translation.

3. CONCLUSIONS

The answer to my research question concerning which strategies are used in order to express the contrast between personal and inanimate pronouns and possessive adjectives that serves the function of (de)humanizing the zombie in English could now be answered as follows. When passages containing this linguistic feature are translated into Spanish, several different strategies are used, such as addition of a noun phrase, verb phrase, noun clause, and pronouns. Furthermore, in some cases no strategy at all is used, and the pronoun contrast is omitted.

8 The English version’s “roaches” is translated to “cucharachas” in the dubbed and in the subtitled versions in Spanish.

RLyLA Vol. 14 (2019), 91-103 | 100 Linda Flores Ohlson Zombies Lost in Translation. The Translation from English to Spanish of (De)humanizing Pronouns

The cases where the pronouns have been translated with other forms, such as noun or verb phrases, are in line with Nida’s (2012:142) view that “[o]nly rarely can one reproduce both content and form in a translation, and hence in general the form is usually sacrificed for the sake of the content.” Even if Nida (2012:144) does not refer directly to linguistic forms, but rather to cultural patterns, one could consider the noun phrase, the verb phrase, and the noun clause solutions as examples of Nida’s dynamic equivalence translations, which are translations that aim “at complete naturalness of expression.”

With regard to the extract where the English pronouns its and thems are translated with ello and ellos, it could be interpreted as an example of formal equivalence translation (Nida 2012) since pronouns are translated with pronouns. It could also be seen as a case of what Toury (2012:122) calls literal translation, since the strategy of translating English pronouns with Spanish ones seems to be a solution “sought on a level which is lower than the one on which [it] would have been selected in the case of a pragmatically equivalence, or ‘adequate’ translation”. The decision to use pronouns in the translation seems to have been determined mainly on linguistic grounds, while the last stage of Toury’s (2012:121) translation process, i.e. “the resulting entity was checked against the TL lexicon […] for its appropriateness and meaningfulness” has not been taken into consideration.

Although something is clearly lost in the translation in the three cases in my corpus where the pronoun contrast is omitted, it should be noted that:

[a] truly natural translation can in some respects be described more easily in terms of what it avoids than in what it actually states; for it is the presence of serious anomalies, avoided in a successful translation, which immediately strike the reader as being out of place in the context. (Nida, 2012:152)

Consequently, instead of creating a forced and unnatural translation of a pronoun contrast the Spanish language lacks, the dehumanization of the zombies is not expressed exactly in the same way as in the English original. However, an analysis of the translated texts as a whole, although outside the scope of this paper, would most likely show that other strategies are used in order to dehumanize the zombies.

Finally, as the translated zombie texts occupy the primary position in the polysystem, the translators have not felt constrained to follow Spanish literature models and in some cases they have broken conventions with regard to pronoun and noun use. The use of “cosa” as a translation of “thing” for a non-human monstrous being, not only in the translated texts but also in the texts originally written in Spanish, suggests that the influence of the English language on the Spanish translations has led to new models in original Spanish texts as well.

REFERENCES

Adams, J. J. (2008). The Living Dead. London: Orbit Books. Anderman, G. M. (1993). “Untranslatability: The Case of Pronouns of Address in Literature”, Perspectives, 1/1: 57- 67. https://doi.org/10.1080/0907676X.1993.9961200 Flores Ohlson, L. (2018). “Fictional Creature Pronominalization – The Use of He/She/It in Reference to Zombies, Vampires, Fairies, and Trolls in Guillermo del Toro’s Literary Work”, International Journal of Literary Linguistics, 7: 1-22. https://doi.org/10.15462/ijll.v7i2.95 Flores Ohlson, L (In press). “The Zombie as a Pronoun – What pronouns are used and why?” Biber, D., Johansson, S., Leech, G., Conrad, S., & Finegan, E. (1999). Longman Grammar of Spoken and Written English. Essex: Longman. Boon, K. (2011). The Zombie as Other: Mortality and the Monstruos in the Post-Nuclear Age, in D. Christie & S. J. Lauro (eds.) Better off Dead. The Evolution of the Zombie as Post-Human. New York: Fordham University Press, 51-60. https://doi.org/10.5422/fordham/9780823234462.003.0005 Brooker, C. (2011-). Black Mirror (Netflix series). http://www.imdb.com/title/tt2085059/ [retrieved: 14.5.2019] Brooks, M. (2006). World War Z: An Oral History of the Zombie War. New York: Broadway Paperbacks. Brooks, M. (2013). Guerra mundial Z: Una historia oral de la guerra zombi. Córdova: Almuzara. Butt, J. & Benjamin, C. (2013). A New Reference Grammar of Modern Spanish. London: Routledge. https://doi.org/10.4324/9780203783474 Carey, M. R. (2014). The Girl with All the Gifts. London: Orbit. Carey, M. R. (2015). Melanie. Barcelona: Minotauro. Chen, M. Y. (2012). Animacies. Biopolitics, Racial Mattering, and Queer Affect. Durham: Duke University Press. https://doi.org/10.1215/9780822395447

RLyLA Vol. 14 (2019), 91-103 | 101 Linda Flores Ohlson Zombies Lost in Translation. The Translation from English to Spanish of (De)humanizing Pronouns

Clark, C. (1992). Pets, Pests, and Pronouns. Aspects of Current English Pronominal Usage in Reference to Non-Human Animates, in T. Kirschner, D. Gutch, J. Gilbert & C. Blank (eds.) Language and Civilization: A Concerted Profusion of Essays and Studies in Honor of Otto Hietsch, I & II. Peter Lang, 634-644. Cohen, J. J. (2012). “Undead (A Zombie Oriented Ontology)”, Journal of the Fantastic in the Arts, 23/3: 397-412. del Toro, G. & Hogan, C. (2009a). Nocturna New York: Harper Collins. del Toro, G. & Hogan, C. (2009b). The Strain. London: Harper Collins Engler, C. & Schaefer, K. (2014-2018). Z Nation (Syfy series). https://www.syfy.com/znation [retrieved 14.5.2019]. Even-Zohar, I. (1990). “System and Repertoire in Culture: Polysystem Theory”, Poetics Today, 11/1: 9-175. Even-Zohar, I. (2005). “Polysystem Theory (Revised)”, Papers in Culture Research. Tel Aviv: Porter Chair of Semiotics. García Yebra, V. (1997). Teoría y práctica de la traducción. Madrid: Gredos. Greene, R. & Mohammad, S. K. (2010). Zombies, Vampires, and Philosophy. New Life for the Undead. Chicago: Open Court. Jakobson, R. (2012). On Linguistic Aspects of Translation, in L. Venuti (ed.) The Translation Studies Reader. London: Routledge, 126-131. Joly, A. (1975). Toward a Theory of Gender in Modern English, in A. Joly & T. Fraser (eds.) Studies in English Grammar. Paris: Université de Lille III: 227-284. King, S. (2006). Cell (Audio edition). USA: Scribner. King, S. (2012). Cell. Spanish Edition. Barcelona: Penguin Random House Grupo Editorial. Künzli, A. (2009). “Address Pronouns as a Problem in French-Swedish Translation and Translation Revision”, Babel, 55/4: 364-380. https://doi.org/10.1075/babel.55.4.04kun Lewis, T. J. (2004). “Untranslatable ‘You’ in Chekhov’s Lady with Lapdog”, Babel, 50/4: 289-297. https://doi.org/10.1075/babel.50.4.01lew Livia, A. (2001). Pronoun Envy. Literary Uses of Linguistic Gender. New York: Oxford University Press. Loureiro, M. (2011). Apocalipsis Z. El principio del fin. Barcelona: Penguin. Loureiro, M. (2012). Apocalypse Z. The Beginning of the End. Las Vegas: Amazon Crossing. Moser, H. M. & Drejer, J. J. (1955). “The Evaluation of the Military Alphabets”, Speech Monographs, 22/5, 256-265. https://doi.org/10.1080/03637755509375152 Nida, E. (2012). Principles of Correspondence, in L. Venuti (ed.) The Translation Studies Reader. London: Routledge, 141-155. Pierce, B. & Pierce, D. T. (2011). Deadheads (movie). http://www.imdb.com/title/tt1273207/ [retrieved 14.5.2019] Pifer, L. (2011). Slacker Bites Back: Shaun of the Dead Finds New Life for Deadbeats, in D. Christie & S. J. Lauro (eds.) Better off Dead. The Evolution of the Zombie as Post-Human. New York: Fordham University Press: 163-174. https://doi.org/10.5422/fordham/9780823234462.003.0011 Real Academia Española. (2017). CORPES XXI (Corpus del Español del Siglo XXI). http://www.rae.es. [retrieved 14.5.2019] Real Academia Española (2005). Diccionario panhispánico de dudas. http://www.rae.es/recursos/diccionarios/dpd. [retrieved 14.5.2019] Real Academia Española (2010). Nueva gramática de la lengua española. Madrid: Espasa Libros. Real Academia Española (2012). Diccionario de la Real Academia Española. http://www.rae.es. Romero, G. A. (1968). Night of the Living Dead (movie). https://www.imdb.com/title/tt0063350/ [retrieved 14.5.2019] Romero, G. A. (1978). Dawn of the Dead (movie). https://www.imdb.com/title/tt0077402/ [retrieved 14.5.2019] Romero, G. A. (1985). Day of the Dead (movie). https://www.imdb.com/title/tt0088993/ [retrieved 14.5.2019] Sisí, C. (2009a). Los caminantes. Palma de Mallorca: Dolmen. Sisí, C. (2009b). The Wanderers, Permuted Press. Spooner, C. (2015). Twenty-First-Century Gothic, in D. Townshend (ed.) Terror and Wonder. The Gothic Imagination. London: The British Library, 180-207. Toury, G. (2012). Descriptive Translation Studies - and Beyond. Amsterdam: John Benjamins. https://doi.org/10.1075/btl.100 Tufo, M. (2010a). Lluvia de zombis #1 (Kindle edition), DevilDog Press. Tufo, M. (2010b). Zombie Fallout #1 (Audio edition), Createspace Independent Publishing Platform.

RLyLA Vol. 14 (2019), 91-103 | 102 Linda Flores Ohlson Zombies Lost in Translation. The Translation from English to Spanish of (De)humanizing Pronouns

Weinstock, J. A. (1999). “Zombie TV”, Post Identity, 2/2. Wright, E. (2004). Shaun of the Dead (movie), http://www.imdb.com/title/tt0365748/. [retrieved 14.5.2019] Yamamoto, M. (1999). Animacy and Reference. Amsterdam: John Benjamins. https://doi.org/10.1075/slcs.46 ZRS (2010). (Zombie Research Society) “Romero Invented Flesh Eaters”. http://zombieresearchsociety.com/ archives/6863. [retrieved 14.5.2019]

RLyLA Vol. 14 (2019), 91-103 | 103 Revista de Lingüística y Lenguas Aplicadas Vol. 14 año 2019, 105-115 EISSN 1886-6298 https://doi.org/10.4995/rlyla.2019.10744

CORPUS-BASED STUDY OF L3 ACQUISITION ON SPANISH PAST TENSE: EVIDENCE FROM LEARNERS’ ORAL PRODUCTION

Hui-Chuan Lu National Cheng Kung University, Taiwan, Province of China An Chung Cheng University of Toledo, United States Shen Yun Hung National Cheng Kung University, Taiwan, Province of China

Abstract: This paper focuses on a corpus-based study on the acquisition of L3 Spanish past tense in oral production through a learners’ corpus. The main findings were: (1) Chinese-speaking learners in Taiwan demonstrated more accurate uses of the Spanish preterit than the imperfect in oral productions, same as the findings in their written productions and those in English- native speakers of Spanish; (2) These learners used telic verbs more correctly than those of activity and state in preterit form in both oral and written productions; (3) The developmental pattern of the Spanish past tense in the oral production of Chinese- speaking learners was similar to that of English-speaking learners of Spanish.

Keywords: tense, aspect, oral production, corpus, Spanish acquisition.

INTRODUCTION

It is known that second language learners of Spanish at the early stages of development have difficulty mastering the preterit and imperfect morphology. The acquisition of past tense and aspect has been studied extensively in research on second language acquisition (cf., Salaberry, 2000; Bardovi-Harlig, 2002; Ayoun & Salaberry, 2005; Bonilla, 2013). This topic is also especially interesting in cross-linguistic studies with learners of various language backgrounds when the learners’ native language and target language share different systems in terms of marking tense and aspect. The situation is evident among Chinese-speaking learners of Indo-European languages. Chinese does not mark tense or aspect through morphological inflection as is the case in English and Spanish (e.g., Li, 2012; Lin, 2003). In English, the past tense is normally indicated by the use of a particular verb form – an inflected form of the main verb. Inflection may involve the use of affixes, such as the-ed ending that marks past tense of English regular verbs, e.g., ‘walk’ and ‘walked.’ On the other hand, Chinese expresses time references mainly by lexical means (adverbials, time phrases or context). Furthermore, Chinese aspect markers such as le and guo also place an action in past time. Time information can be conveyed through lexis of time or as a secondary feature by aspect markers. Thus, typological differences across languages impose substantial challenges for learners with different linguistic backgrounds as they acquire the tense and aspect of a target language. It is reasonable to assume that Chinese-speaking learners of Spanish have to learn new sets of mapping between the morphological forms and the functions/meanings of tense and aspect and that they also have to change their previous strategies for expressing tense and aspect with context, temporal adverbs, and aspectual markers (see also Cadierno, 1995; Chin, 2008).

Previous research on the acquisition of the Spanish past tense and aspect has examined different task conditions, including L2 learners’ oral and written texts, personal or impersonal narratives, and open- or closed- ended tasks (see also Bonilla, 2013). However, little research has been conducted to investigate the acquisition

To cite this article: Lu, H. C., Cheng, A. C., and Hung, S. Y. (2019). "Corpus-based Study of L3 Acquisition on Spanish Past Tense: Evidence from Learners’ Oral Production". Revista de Lingüística y Lenguas Aplicadas, 14, 105-115. https://doi.org/10.4995/rlyla.2019.10744 Correspondence authors: [email protected], [email protected], [email protected]

Received: 2018-09-25 Accepted: 2019-04-15 | 105 Hui-Chuan Lu, An Chung Cheng, and Shen Yun Hung Corpus-based Study of L3 Acquisition on Spanish Past Tense: Evidence from Learners’ Oral Production

of target forms from a corpus-based approach, which is an emerging area of inquiry in Spanish second language acquisition research (Mendikoetxea, 2014).

This corpus-based study, then, is an attempt to examine cross-linguistic influences that might affect the acquisition of the Spanish past tense and analyze learners’ accuracy rate of use and the lexical aspect classifications, using a created learners’ spoken and oral corpora and the assistance of an annotation tool. This paper, on the one hand, presents the construction of a learners’ oral corpus, in which the speech data provides better insights into spontaneous target language use for the study of multilingual acquisition in relation to written data. On the other hand, the present study focuses on a corpus-based study of oral productions containing occurrences of Spanish past tense and aspect by Chinese-speaking learners of Spanish in Taiwan. The participants learned Chinese as their first language (L1) and English at middle or high schools as their second language, and learned Spanish at senior high schools or colleges as their third language. Under the framework of the Lexical Aspect Hypothesis (Andersen 1986, 1989; Vendler, 1967), this study set out to analyze cross-linguistic influence that affects learners’ oral production in marking Spanish past tense and aspect and the relationship between the lexical aspect classification and the learner selections of tense and aspect. It is also intended to reveal the developmental pattern of the acquisition of Spanish tense and aspect by Taiwanese learners at different proficiency levels, in comparison with their written production. The results of this study are also compared with those of previous studies on this topic focusing on English-speaking learners of Spanish in order to explore the universality of the acquisition of tense and aspect.

This paper is organized as follows: In Section 2, previous research on learners’ corpus, past tense, grammatical and lexical aspects, and acquisition of L2 Spanish are discussed. Section 3 presents the research questions and methodology. Results are presented and discussed in Section 4. Section 5 provides the conclusions of this study.

PREVIOUS RESEARCH

Learner corpus The development of corpus linguistics has facilitated research in both theoretical and applied studies of language. As an essential and important source of research, different types of corpora have been created for various needs and purposes. Among these different types of corpora, the construction of learners’ corpus benefits research in language acquisition (Granger, 2003, 2009; Myles, 2005, among others). However, according to Weisser (2016), only 32 of the 360 exiting corpora are learners’ corpora, in which 81% (26/32) are related to English learning, 66% (21/32) are written corpora, 28% (9/32) are oral corpora, and 6% (2/32) are both written and oral corpora, whereas only four learner corpora are related to Spanish.

Among the constructed learners’ Spanish corpora, the Corpus Escrito del Español L2 (CEDEL21) and the Spanish Learner Language Oral Corpus (SPLLOC2) are renowned internationally. The data of these two corpora was all from English-speaking learners. They differ in that the former is a written corpus, and the latter is an oral one. Moreover, in terms of search functions, there is no public search interface for the CEDEL2, whereas the SPLLOC provides public access to search for words and phrases from compiled data. Furthermore, although several corpora related to Spanish acquisition in Taiwan have been constructed for different purposes (such as the Corpus Oral del Español en Taiwán COET3, Corpus de textos escritos por universitarios taiwaneses estudiantes de español4, Corpus Escrito de Aprendices Taiwaneses de Español de la Universidad Providence5), they are unfortunately not sharable resources, nor are they accessible for the general public.

In Taiwan, the learners’ written corpus, CEATE (Corpus Escrito de Aprendices Taiwaneses de Español / Taiwanese Learners’ Written Corpus of Spanish), which features on-line free access6 and multi-search functions7 was constructed by the corpus team of National Cheng Kung University. The purpose of building CEATE was to inform teaching and advance research on third language acquisition. In 2005-2011, the research team compiled 2,425 texts, with 446,694 words from written texts of L1 Chinese-speaking adult learners of Spanish as a third language (L3) after learning English as their second language (L2)8. The texts were collected from learners of Spanish at 15 universities in Taiwan. Then, the construction of a learners’ oral corpus of Spanish, COATE (the Corpus Oral de Aprendices Taiwaneses de Español / Taiwanese Learners’ Oral Corpus of Spanish), was started

1 The CEDEL2 was created by the Universidad Autónoma de Madrid and Universidad de Granada. 2 The SPLLOC was constructed by Southampton, Newcastle and York University in the UK. 3 Corpus Oral del Español en Taiwán (COET) was created by J. Pérez Ruiz and M. Rubio Lastra in 2004. 4 Lin, T.-J. (2005). Corpus de textos escritos por universitarios taiwaneses estudiantes de español. Lingüística en la Red, 3, 1-58. 5 Lu, L.-H. (2016). Corpus-based study of Spanish writing. Foreign Language Studies, 24, 145-169. 6 http://corpora.flld.ncku.edu.tw 7 The four major search functions include: (1) specific word, (2) word and its POS, (3) word, its POS, and the POS of the following word, and (4) verbal lemma. 8 It should be noted that these learners had learned L2 English previously as a school subject, so their knowledge of past tense morphology might affect their learning of L3 Spanish.

RLyLA Vol. 14 (2019), 105-115 | 106 Hui-Chuan Lu, An Chung Cheng, and Shen Yun Hung Corpus-based Study of L3 Acquisition on Spanish Past Tense: Evidence from Learners’ Oral Production

in 2013 with the same purpose as that of CEATE. Thus, the corpus of Taiwanese learners of Spanish, the CATE (Corpus de Aprendices Taiwaneses de Español) consists of two different types of learner production data from both the CEATE and COATE.

Past tense and aspects Tense indicates the position of a relationship between an event and a time on a timeline, whereas aspect defines the internal temporal feature of an event expressing the speaker’s point of view without it being associated with a timeline (Comrie, 1976). Spanish expresses tense by means of verbal inflections. The indicative past in Spanish has two forms, the preterit and imperfect, which are differentiated according to aspect. However, English does not mark past tense and aspect by means of inflectional morphology but rather by means of periphrastic constructions. The English sentences “I called my sister” and “I was calling my sister” convey perfective and progressive aspect, respectively. The perfective vs. imperfective distinction is often realized through grammaticalized affixes or auxiliaries in English and Spanish. Chinese tense and aspect, on the other hand, are expressed through temporal adverbs, context, and aspectual markers including guo, zai, zhe and le (See Li, 2012; Lin, 2003). For example, the verbal suffixle is often been characterized as a perfective marker indicating completion or termination of an action. A past event for “Ta chi yi-tiao yu” can be expressed as below.

Ta chi-le yi-tiao yu

he eat-Asp one-Cl fish

‘He ate a fish.’

Telic predicates such as “chi yi-tiao yu / to eat a fish” are interpreted perfectively, but if they are combined with zai, such as in zai chi yi-tiao yu ‘be eating a fish,’ they are interpreted imperfectively.

Vendler’s (1967) lexical aspect theory accounts for verbal structure in relationship to time. State, activity, accomplishment and achievement are four basic types that distinguish the internal lexical meanings of verbs. For example, “ver/to see,” “amar/to love,” “querer/to want,” and “esperar/to wait” express a stative state. Verbs of “activity” indicate durative actions without an endpoint, such as “correr/to run,” “cantar/to sing,” and “jugar/to play.” Verbs of “accomplishment” are associated with those durative actions with a clear endpoint, for instance, “leer una novela/read a novel,” and “cruzar la calle/to cross the street.” Finally, verbs of achievement are those that are associated with instantaneous actions with an endpoint, for example, “reconocer/to recognize,” “morir/ to die,” “encontrar/to find out.” Furthermore, Comrie (1976) and Andersen (1989, 1991) used three semantic features (dynamic, telic and punctual) to distinguish verbal aspects. Dynamic denotes that energy is required for the situation to exist or continue. Telic denotes having an inherent endpoint. Punctual denotes having no duration. Vendler’s categories are characterized by different combinations of the features punctual/durative, tetic/atelic, and dynamic/stative (Shirai and Andersen, 1995). These combinations are shown in the table below.

Feature analysis of the four verb classes.

STATE ACTIVITY ACCOMPLISHMENT ACHIEVEMENT PUNCTUAL ------+ TELIC -- -- + + DYNAMIC -- + + + (Source: Shirai and Andersen, 1995:744).

Research in the aspect hypothesis in the acquisition of past morphology In second language acquisition research, previous studies on the Spanish past tense and aspect have found that Spanish L2 learners initially use the preterit as a default tense marker, instead of relying on inherent aspectual distinctions (Salaberry, 1999, 2003; Salaberry and Ayoun, 2005; Salaberry and Shirai, 2002). For native English learners of Spanish, the preterit is considered the predetermined form used to express the past tense. Salaberry (1999) first proposed the Default Past Tense Hypothesis (DPTH) based on a study with L2 English-speaking learners of Spanish. DPTH predicts that beginning learners initially rely on the use of the preterit to mark tense distinctions and preterit marking will emerge across all aspectual classes. Ayoun and Salaberry (2005) hypothesize that learners initially are only able to mark tense, then gradually they become more sensitive to lexical aspect, then they begin to mark foreground/background distinctions.

Another leading hypothesis, Lexical Aspect Hypothesis (LAH) (Andersen, 1986, 1991; Andersen & Shirai, 1994; Bardovi-Harlig, 2000) argues that past marking emerges based on inherent aspectual categories. Certain form– meaning mappings (i.e. telic–Preterit and atelic–imperfect) guide the emergence of past tense forms in second

RLyLA Vol. 14 (2019), 105-115 | 107 Hui-Chuan Lu, An Chung Cheng, and Shen Yun Hung Corpus-based Study of L3 Acquisition on Spanish Past Tense: Evidence from Learners’ Oral Production

language acquisition. The acquisition of tense-aspect morphology can be explained by the interaction of two grammatical aspects (perfective & imperfective) and four inherent lexical aspects (stative, activity, accomplishment and achievement). The inherent lexical aspects of verbs can be characterized in terms of semantic features (punctual, telic, and dynamic) (Shirai & Andersen, 1995). The LAH suggests that the use of imperfective markers spreads from stative verbs to non-stative verbs, and the use of perfective markers spreads from punctual verbs (achievements) to non-punctual verbs (Andersen, 1986, 1991; Andersen & Shirai, 1994; Bardovi-Harlig, 2000). Research supporting the LAH has found a common pattern in which learners initially use present morphology for past context (Camps, 2000; Salaberry, 1999). At a later stage, preterit morphology emerges in telic predicates (accomplishments and achievements); imperfect morphology is almost nonexistent. However, it will be eventually extended to activity and stative verbs. When imperfect morphology emerges, it does so in state and then activity predicates, extending then to accomplishment verbs, and finally to achievement verb. That is, the preterit is associated more frequently with telic actions whereas the imperfect is used more with stative verbs. (Camps, 2002, 2005; Hasbún, 1995; Lopez-Ortega, 2000; Ramsay, 1990). In addition, Salaberry (2003) suggested that the LAH applied better to learners at more advanced levels. Advanced learners distinguished the use of preterit and imperfect contrast according to lexical aspectual classes. New evidence on the validity of the LAH in L2 Spanish was offered in a study by Domínguez et al. (2013). Temporal marking is used differently in terms of verbs and the proficiency level of learners. They found that beginner and intermediate beginner and intermediate speakers used preterit with event (dynamic) verbs but imperfect mainly with state (non-dynamic) verbs. The advanced learners used typical Preterit–telic associations in the least controlled oral tasks, as predicted by the LAH. In order to extend the application of LAH and to test the theory, Salaberry (2003) and Dalila and Salaberry (2005) suggested including Asian learners in research on this topic for further understanding of the processes and mechanisms involved in acquiring the past tense and aspect.

Moving the focus on English-speaking learners to learners of other languages, Lu, Cheng, & Hung (2015) investigated the development of the acquisition of L3 Spanish past tense in the written production (33,655 words) of 143 Chinese-speaking Taiwanese learners from six different universities. The result showed that Taiwanese learners demonstrated more accurate use of the Spanish preterit than that of the imperfect in written texts in the early stages of their development. From the perspective of lexical aspect, the results demonstrated the following acquisition order: telic (achievement and accomplishment) verbs preceded activity verbs, and finally stative verbs. However, the acquisition of the imperfect showed the opposite order regarding the lexical aspect of verbs. That is, stative verbs preceded activity verbs and then telic verbs (achievement and accomplishment). Furthermore, in the majority of cases, Taiwanese learners of Spanish demonstrated a similar pattern of acquisition of the Spanish past tense in terms of lexical aspect to that of English-speaking learners of Spanish.

To extend the scope of inquiry into the relationship between verbal morphemes and the types of lexical verbs in third language acquisition, the present study analyzes oral data compiled in an L3 Taiwanese learners’ corpus of Spanish under the framework of the Lexical Aspect Hypothesis (LAH) (Andersen, 1991).

Comparison between written and oral modalities Variability in learners’ production has long been discussed as important consideration in second language acquisition research. Previous studies have shown that there are differences in the oral and written production by second language learners (Dickerson & Dickerson, 1977; Tarone, 1979, 1985; Hsieh, 2005; Larsen-Freeman, 2006; Ellis, 2008). However, Cortés (2002) and Blanco Pena (2013) indicated that similar error patterns in written production can be observed in oral development. Furthermore, a case study by Hubert (2013) showed a strong correlation between speaking and writing performance.

Skehan (2009) argued that relationship between task design and complexity, accuracy and fluency in task performance had a trade-off effect due to learners’ limited attention and working memory. There was a tension between form (complexity and accuracy), on one hand, and fluency, on the other hand. The task conditions relevant to the oral and written narrative tasks used in this study of tense and aspect marking have found the effects on complexity, accuracy and fluency as reviewed by Skehan (2009). Pre-task planning overall seems to aid complexity and fluency. Also, narrative tasks have shown greater complexity but less accuracy and fluency.

Martelle’s (2011) study of learning L2 Russian by L2 English speakers showed that the DPTH was more supported in tasks with more planning time, such as written narratives, whereas the LAH was supported more by tasks that elicited oral narratives or conversations, which took less time.

Camps (2002, 2005) and Lubbers-Quesada (2006, 2007) are research on the acquisition of L2 Spanish tense and aspect examined the oral productions supported the LAH. However, Husbún (1995) focused on written production, and the results challenged the LAH developmental sequence proposed by Andersen (1991) because the acquisition of preterit did not move from telic verbs (accomplishment and achievement) to durative verbs (activity) and further to stative ones. On the other hand, Potowski (2005) analyzed both written and oral texts from bilingual Spanish L1, Spanish L2, and newly arrived native speakers of Spanish in two-way immersion classrooms.

RLyLA Vol. 14 (2019), 105-115 | 108 Hui-Chuan Lu, An Chung Cheng, and Shen Yun Hung Corpus-based Study of L3 Acquisition on Spanish Past Tense: Evidence from Learners’ Oral Production

The results showed similar distributions of preterit and imperfect by aspectual category in both written and oral productions. Ruiz-Debbe (2005) analyzed both types of production, and the results indicated that English learners of Spanish exhibit different acquisition sequences in oral and written production and across different proficiency levels. Based on the literature review, research related to learners’ written and oral data demonstrate different but conflicting findings in the acquisition of L2 Spanish tense and aspect. It is hoped that this study will shed light on the second language acquisition of Spanish tense and aspect, and will fill the gaps in the literature on this topic.

METHODOLOGY

Research questions Extended from a previous study on the Spanish tense and aspect with learner written texts (Lu et al., 2015), this study focuses on learners’ oral production. The research questions are as follows:

1. What is the development of past tense morphology and what is the development of lexical aspect classifications of past tense by Chinese-speaking learners of L3 Spanish?

2. Do written and spoken productions of Chinese-speaking learners of L3 Spanish demonstrate the same pattern of past tense marking and lexical aspect of verbs in the acquisition of Spanish tense and aspect?

Research method The research method includes two major parts: construction of spoken corpus as the data source and analysis of oral data.

Data collection: Construction of COATE. For the first part, in the process of constructing the oral corpus, the research team compiled oral data collected from L3 learners of Spanish at four universities (Tamkang University, Providence University, WenZao University, and National Cheng Kung University) in Taiwan, where Mandarin- Chinese is spoken in daily life, and English is learned as a school subject in middle or high schools. All participants signed a consent form to authorize the future usage of compiled spoken data and provided linguistic profiles and language learning experience related to Spanish through questionnaires. Instead of using seat time to determine the participants’ proficiency levels, as has been the case in many other previous studies, a 45-minute Wisconsin Placement Test was administered prior to data collection. Then, the participants recorded a three-minute self- introduction as a warm-up activity in the language laboratories of their institutions during class time. They were given a six-minute preparation section to understand a picture-description task. The speech production that entered the data pool was an oral narrative, in which they described a picture series with key words at the side of each picture illustrating a fairy tale, Little Red Riding Hood (Caperucita Roja), in a period of 18 minutes. Afterwards, the speech data was transcribed with the EXMARaLDA Partitur-Editor (Schmidt, 2011), following the LINDSEI transcription system (Louvain International Database of Spoken English Interlanguage).

Data source: Learner corpus COATE. A total of 15.55 hours of the recorded oral narratives of 71 participants were compiled to construct the Taiwanese Learners’ Oral Corpus of Spanish / Corpus Oral de Aprendices Taiwaneses de Español (COATE). The number of participants, their proficiency level as well as hours and words of recording data are shown in Table 1. Since the participants were college students, the Wisconsin Placement Test, widely used in the US colleges, was administered to identify participants’ Spanish proficiency levels. According to the Spanish placement criteria at Wisconsin test results, participants with scores between 426-548 points (28-48 out of 77 questions correctly answered) were grouped into the beginning level. Participants with scores between 554-618 points (49-58 out of 77 questions correctly answered) were classified to the intermediate level.

Table 1. Data distribution of COATE.

Level Beginning Intermediate Total Participants 44 27 71 Recording (hours) 9.51 6.04 15.55 Transcription (words) 21,887 13,600 35,487

The corpus COATE was constructed from 2013 to 2014. It was combined with the Taiwanese Learners’ Written Corpus of Spanish / Corpus Escrito de Aprendices Taiwaneses de Español (CEATE), which was constructed from 2005 to 2011, to form the two main sub-corpora of the Taiwanese Learners’ Corpus of Spanish / Corpus de Aprendices Taiwaneses de Español (CATE). The COATE, as a sub-corpus, added spoken data into the existing written sub-corpus, CEATE, to expand the scope of the corpus CATE. Based on the construction results of the learners’ oral corpus, COATE, a corpus-based linguistic analysis was conducted afterward.

RLyLA Vol. 14 (2019), 105-115 | 109 Hui-Chuan Lu, An Chung Cheng, and Shen Yun Hung Corpus-based Study of L3 Acquisition on Spanish Past Tense: Evidence from Learners’ Oral Production

Data analysis. For the subsequent data analysis, speech data were annotated with UAM CorpusTool (O’Donnell, 2012) to facilitate the annotation process in two sub-steps. Firstly, Hispanic native speakers who were trained to use the UAM CorpusTool following guidelines of error-correction for CATE marked learner errors with corresponding corrections and those errors were checked and confirmed by our research assistants who had lived and studied in Spain for more than five years. Then, the research team (consisting of 3 trained assistants) annotated the uses of the examined verbs, following guidelines of past tense and aspect for CATE, by contrasting learners’ usage and the corrections as revised by Hispanic native speakers. Furthermore, verbs were annotated according to specific categories including lexical aspects and verbal predicates, as well as tenses.

Finally, a total of 14,043 annotations were made, including the following categories: (1) Tense and aspect use according to a. Hispanic natives and b. Chinese-speaking learners of Spanish: correct and incorrect usages9; (2) the tenses: present, preterit, and imperfect; (3) the lexical aspects: state, activity, accomplishment, achievement, telicity (accomplishment and achievement) and dynamics (activity and telic)10.

To calculate the accuracy rate of learners’ oral productions, we took Hispanic native speakers’ judgments and their revisions for correct uses as a standard reference. The Hispanic native speakers’ judgements formed the baseline to compare and contrast the productions of learners. That is to say, the accuracy rate of imperfect was calculated as: learners’ correct instances of imperfect divided by Hispanic natives’ instances of imperfect. The same calculation was employed to compute the accuracy of learners´ correct uses of preterit11. In order to observe the tendency of learners’ uses of the Spanish past tense and aspect at different developmental stages, we took into account variables such as lexical aspects and language proficiency (beginning and intermediate levels) in the computation.

It should be noted that the two sets of learners’ data (written and oral productions) were elicited with the same tool (a narrative of Little Red Riding Hood with picture prompts) but with two different tasks, written and spoken output.

RESULTS AND DISCUSSION

The data pool was divided into two groups: beginning and intermediate levels, based on the results of the Wisconsin Placement Test. The verbs that appeared in the learner corpora were classified based on their lexical aspects and semantic features as the following: state, activity, telic (accomplishment and achievement) and dynamic (activity, accomplishment and achievement) verbs. The Independent t-test was used to compare the means of two independent groups (beginning and intermediate level) through statistical evidence to verify whether they are significantly different with respect to different lexical aspects. The results of the examined lexical aspects (state, activity, telic and dynamic verbs) that might affect the selection of Spanish past tense morphology (preterit and imperfect) in oral data are shown in Table 2.

Table 2. Language proficiency for different lexical aspects of past tense. Preterit Imperfect Aspect F Value P Value F Value P Value State 0.032 0.958 1.958 0.085 Activity 1.386 0.468 0.494 0.245 Telic V 0.891 0.814 0.178 0.731 Dynamic V 1.950 0.717 0.493 0.313 Average 1.467 0.746 0.875 0.203

According to Table 2, there was no significant difference found by conducting the independent t-test, since the p-values of all relationships were above 0.05 (0.958, 0.468, 0.814, 0.717 and 0.746 for preterit and 0.085, 0.245, 0.731, 0.313 and 0.203 for imperfect). As the difference between learners at the two proficiency levels under consideration did not reach significance for performance accuracy of the preterit and imperfect use in terms of the verbal lexical aspects, it can be concluded that the difference in language proficiency levels (beginning

9 By using the UAM CorpusTool, 3 levels were used to classify the uses of tense and aspect: (1) Classification of tense and aspect when the verb is correctly used by Chinese-speaking learner of Spanish according to Hispanic native, (2) Classification of tense and aspect when the verb is incorrectly used by Chinese-speaking learner of Spanish according to Hispanic native, and (3) Classification of tense and aspect when the verb is incorrectly used by Chinese-speaking learners of Spanish according to learners of Spanish. For example, for the verb “vivió” in the sentence “Ella vivió el bosque”, we annotated (1) incorrect use, (2) incorrect use-imperfect used by Hispanic natives, and (3) incorrect use-preterit used by learners while the verb “estaba” in the sentence “Su madre le dijo su abuela estaba enferma y tenía que visitarla” was annotated with (1) correct use, (2) correct use-imperfect used by Hispanic natives, and (3) correct use-imperfect used by learners. 10 On one hand, verbs used with preterit aspect included (1) stative: sentirse; (2) activity: llevar una cesta; (3) accomplishment: decir, aconsejar, preguntar, ponerse ropa, colocarse gorro, tomar siesta; and (4) achievement: olvidar, aparecer, llegar. On the other hand, verbs used with imperfect aspect included (1) stative: vivir, estar, ser, haber; (2) activity: andar, llamar, ir, cantar, coger flores; and (4) achievement: pedir auxilio. 11 For example, if a student produces a total of 15 correct uses of preterit aspect compared to the total number (19) of preterit aspects confirmed and revised by the Hispanic natives, the result of calculation is 15/[email protected] for this individual student. Then, the average of all students at the same proficiency level were calculated. Furthermore, the same method was applied for calculating averages for different lexical aspects.

RLyLA Vol. 14 (2019), 105-115 | 110 Hui-Chuan Lu, An Chung Cheng, and Shen Yun Hung Corpus-based Study of L3 Acquisition on Spanish Past Tense: Evidence from Learners’ Oral Production

vs. intermediate) did not play any role in significantly affecting the accuracy rate of using Spanish past tense for different lexical aspects in oral data. Nevertheless, we were able to observe learners’ tendency toward oral usage for each level, as shown in Table 3.

Table 3. Distribution of correct uses for different lexical aspects. Lexical Aspect Beginning (%) Intermediate (%) State 34.01 37.04 Activity 43.39 56.57 49.59 58.56

Preterit Dynamic V. 56.75 58.39 Telic V. 59.08 60.43 State 52.22 66.62 Activity 39.87 45.97 49.53 54.28 Dynamic V. 38.81 46.85 Imperfect Telic V. 21.59 24.07

According to the distribution of correct uses for different lexical aspects (stative and dynamic verbs), the results showed the overall accuracy rate12 of preterit usage to be higher than that of imperfect usage by Taiwanese learners of Spanish in spoken production for both levels, 56.57% > 45.97% for the beginning level13, and 58.56% > 54.28% for the intermediate level14. The same tendency was observed in the early stage of Taiwanese learners’ written production, 84.75% > 61.42% for the beginning level (Lu et al., 2015). This has also been found to be the case for English-native speakers of Spanish (cf. Andersen, 1986; Ramsay, 1990; Hasbún, 1995; Salaberry, 2000, 2003, 2008; Salaberry and Ayoun, 2005). Therefore, according to the distributional tendency of accuracy rate, it could be inferred that the preterit use in general is acquired earlier than the imperfect use as a universal phenomenon for both L2 and L3 Spanish learners whose native language is English and Chinese, respectively.

In terms of lexical aspect, as shown in Table 3, learners of beginning and intermediate levels used dynamic verbs (56.75% and 58.56%, activity and telic verbs) more correctly than those of state (34.01% and 37.04%) and 49.53%), and telic (21.59% and 24.07%) verbs. Based on a written task in spoken production. In other words, learners used telic verbs (59.08% and 60.43%), accomplishment and achievement) more correctly(Authors, than 2015) those, the comparisonof activity of verbs the results (43.39% for the currentand 49.59%) study using in oral the perfective aspect (preterit). In the oral production of the imperfective aspect (imperfect), the opposite order was found. The accurate use of the imperfect for beginning anddata andintermediate the results of levels an earlier in studythe stativeusing the verbs written wasdata are52.22% shown inand 66.62%, which were higher than the accuracy rates for the dynamic (38.81% and 46.85%), activity (39.87% and 49.53%), and telic (21.59% and 24.07%) verbs. BasedFigure on 1 .a written task (Lu et al., 2015), the comparison of the results for the current study using oral data and the results of an earlier study using the written data are shown in Figure 1.

100 100

50 50

0 0 Beg. Int. Beg. Int. State 26.21 44.83 State 62.66 82.85 Activity 69.81 64.41 Activity 55.04 83.40

Acc+Ach 84.92 83.87 Acc+Ach 23.63 40.81

Written production: Preterit Written production: Imperfect

100 100

50 50 0 0 Beg. Int. Beg. Int. State 34.01 37.04 State 52.22 66.62 Activity 43.39 49.59 Activity 39.87 49.53 Acc+Ach 59.08 60.43 Acc+Ach 21.59 24.07 Oral production: Preterit Oral production: Imperfect Figure Figure1. Comparisons 1. Comparisons of learner of learner performance performance in writtenwritten and and oral oral production production 12 The accuracy rate was calculated by dividing the total of correct uses by student number.

13 On the one hand, for the preterit usage of beginning learners, the accuracy rate of stative verbs was 34.01% and the dynamic verbs was 56.75%, including verbs of activity (43.39%) and telic verbs (59.08%). On the other hand, for intermediate learners, the accuracy rate of stative verbs was 37.04% and the dynamic verbs was 58.39%, including verbs of activity (49.59%) and telic verbs (60.43%). 27 14 For the imperfect usage of intermediate learners, the accuracy rate of stative verbs was 66.62% and the dynamic verbs was 46.85%, including verbs of activity (49.53%) and telic verbs (24.07%) while for the beginning level, the accuracy rate of stative verbs was 52.22% and the dynamic verbs was 38.81%, including verbs of activity (39.87%) and telic verbs (21.59%).

RLyLA Vol. 14 (2019), 105-115 | 111 Hui-Chuan Lu, An Chung Cheng, and Shen Yun Hung Corpus-based Study of L3 Acquisition on Spanish Past Tense: Evidence from Learners’ Oral Production

As shown in the upper part of the Figure 1, the accuracy rate of preterit uses were higher with telic verbs (accomplishment and achievement) for both levels of learners (beginning, 84.92% and intermediate, 83.87%) than verbs of activity (beginning, 69.81% and intermediate, 64.41%) and stative verbs (beginning, 26.21% and intermediate, 44.83%) in the written production of Taiwanese L3 learners of Spanish. In contrast, the accuracy rate of imperfect uses were higher with stative verbs for both levels of learners (beginning, 62.66% and intermediate, 82.85%) only than telic verbs (beginning, 23.63% and intermediate, 40.81%) (Lu et al., 2015).

As what has been concluded from Table 3, we can observe partial similar tendency for written and oral production: Chinese learners of L3 Spanish used verbs of accomplishment and achievement more correctly than verbs of activity and state in the preterit in both oral and written production. On the other hand, the opposite order was the case with the use of imperfect only in spoken production for different development of proficiency levels. That is, the developmental patterns (state > accomplishment+achievement) were observed only for the earlier stage of development (beginning level) among Taiwanese learners in the use of imperfect of written texts, but not for the latter stage (intermediate level).

The different behavior of learners in terms of written and spoken production in the use of Spanish preterit and imperfect can be explained by the differences of reaction times entailed in the written and oral tasks, that is to say, a spoken narrative is a spontaneous speech that is produced with less reaction time, whereas a written narrative allows for more reaction time to produce the output. Compared to the acquisition pattern of English-speaking learners, the oral production of Chinese-speaking learners of L3 Spanish showed a similar pattern related to developmental stage in terms of verbal lexical aspects. It should be noted that the similarity between learners with two language backgrounds was observed only in the earlier learning stage in the written modality. This might be attributed to the tasks adopted in different studies, where written production was examined in Chinese-speaking learners in Lu et al. (2015) and oral production was examined in this study.

Going one step further, in order to examine the relationship between the oral uses of preterit and imperfect for each proficiency level, a correlation test was conducted, for which the results are shown in Table 4.

Table 4. Relationship between preterit vs. imperfect.

Level Correlation P Value Beginning 0.229 0.135 Intermediate 0.458 0.016

In Table 4, there was no significant difference in the correct uses of preterit and imperfect for learners at the beginning level according to the correlation test (p > 0.05), whereas the correct uses of preterit and imperfect for learners at the intermediate level were positively correlated (p = 0.016). That is to say, the higher is the accuracy rate of preterit use, the higher is that of imperfect uses; but this is not the case for learners of beginning level.

To examine the relationship and determine the difference between lexical and grammatical aspects, paired t tests between lexical aspects within preterit and imperfect for each level of proficiency were conducted. Table 5 shows the results of the paired t-tests with all comparisons turning out to be statistically significant (p < 0.05).

Table 5. Relationship between lexical and grammatical aspects.

Aspect Level Beginning Intermediate Lexical aspect Average±s.d. P Average±s.d. P Preterit State vs. Dyn. -0.22735±.0.46 0.02 -0.21899±0.46 0.022 State vs. Telic -0.25062±.0.46 0.001 -0.23392±0.46 0.014 Activity vs. Telic -0.15738±0.26 0.000 -0.10839±0.22 0.018 Imperfect State vs. Dyn. 0.13408±.0.30 0.006 0.19773±0.27 0.001 State vs. Activity 0.12356±.0.30 0.009 0.17098±0.27 0.004 State vs. Telic 0.30631±.0.41 0.000 0.42551±0.40 0.000

From the upper part of Table 5, the differences between stative and dynamic verbs (verbs of activity and telic verbs), that between stative and telic verbs, and that between verbs of activity and telic verbs for both levels of learners in the preterit aspect were significant (p < 0.05). These results suggested that the specific variable that affects the correct use of preterit is telic verbs. Furthermore, it can be seen that the lexical aspect that really affected the correct uses of the imperfect aspect for both levels was the verbs of state because the relationship between verbs of state and dynamic verbs (activity, accomplishment and achievement), that between verbs of

RLyLA Vol. 14 (2019), 105-115 | 112 Hui-Chuan Lu, An Chung Cheng, and Shen Yun Hung Corpus-based Study of L3 Acquisition on Spanish Past Tense: Evidence from Learners’ Oral Production

state and verbs of activity and that between verbs of stative and telic verbs (accomplishment and achievement) were significantly different (p < 0.05). Therefore, it can be concluded that the telic verbs in preterit aspect and the state verbs in imperfect aspect are the key variable affecting correct selection of Spanish preterit and imperfect aspects. It seems that the lexical aspect does not play a crucial role in determining the selection of two grammatical aspects for these learners.

CONCLUSION

In conclusion, we compiled a L3 learners’ oral corpus of 71 Taiwanese learners of Spanish (15.5 hours and 35,487 words) with annotations of error-correction and lexical aspects in this study in order to investigate the acquisition of Spanish past tense and aspect under the framework of the Lexical Aspect Hypothesis (Andersen, 1991; Andersen & Shirai, 1994). By contrasting written and spoken production, the results showed that the overall accuracy rate of preterit use was higher than that for imperfect use in Chinese-speaking learners of L3 Spanish in both written and spoken production at the earlier stage. However, this pattern appears to only occur in the later stage for the oral production. In addition, learners use accomplishment and achievement verbs more correctly than verbs of activity and state verbs in the perfective aspect (preterit) and the opposite order (state > activity > accomplishment + achievement) characterizes the use of the imperfective aspect (imperfect) in both types of production for learners at earlier stages. This pattern was also observed at a later stage of Taiwanese learners in their spoken production but was not the case for their written production. Furthermore, the oral production of L3 Chinese-speaking learners shared a similar pattern of development with respect to verbal lexical aspects to that of English-speaking learners. However, the findings with the two groups of learners with different language backgrounds only correlate with the results of written production by the beginning learners. Compared to written data, oral data requires more spontaneous reactions to task prompts, which might provide a more authentic representation of acquisition. Finally, according to the analysis of spoken production, the telic verbs in preterit and the state verbs in the imperfect aspect are the key variables related to the correct selection of grammatical aspects. Since Chinese language does not mark past tense and aspect through verbal morphology, it is reasonable that Chinese-speaking learners of L3 Spanish take more time and need more attention and processing time to acquire Spanish past tense and aspect. Future studies could investigate the oral and written production of Chinese-speaking learners at advanced proficiency levels to test the Lexical Aspect Hypothesis.

ACKNOWLEDGMENTS

We are grateful for the grant for the project “Corpus-based study of the L3 acquisition on Spanish past tense” (MOST 101-2410-H-006-088-MY2) provided by the Ministry of Science and Technology of Taiwan.

REFERENCES

Andersen, R. W. (1986). “El Desarrollo de la Morfología Verbal en el Español como Segundo Idioma”, in J. M. Meisel (ed.) Adquisición del Lenguaje – Acquisição da Linguagem. Frankfurt: Klaus-Dieter Vervuert Verlag, 115- 138. https://doi.org/10.31819/9783964562081-011 Andersen, R. W. (1989). “La adquisición de la morfología verbal”, Lingüística, 1: 89-141. Andersen, R. W. (1991). “Developmental Sequences: The Emergence of Aspect Marking in Second Language Acquisition”, in T. Huebner & C. A. Ferguson (eds.) Crosscurrents in Second Language Acquisition and Linguistic Theories. Amsterdam: John Benjamins, 305–324. https://doi.org/10.1075/lald.2.17and Andersen, R. & Shirai, Y. (1994). “Discourse motivations for some cognitive acquisition principles”, Studies in Second Language Acquisition, 16: 133-156. https://doi.org/10.1017/S0272263100012845 Ayoun, D. & Salaberry, M. R. (2005). Tense and Aspect in the Romance Languages: Theoretical and Applied Perspectives. Amsterdam and Philadelphia: John Benjamins. https://doi.org/10.1075/sibil.29 Bardovi-Harlig, K. (2000). Tense and aspect in second language acquisition: Form, meaning, and use. Malden, MA: Blackwell. Bardovi-Harlig, K. (2002). “Analyzing Aspect”, in R. Salaberry & Y. Shirai (eds.) The Acquisition of Tense-Aspect Morphology. Amsterdam: John Benjamins, 129-154. https://doi.org/10.1075/lald.27.08bar Blanco Pena, J. M. (2013). “Escollos lingüísticos de los principiantes chinos de español como lengua extranjera: Causas y sugerencias pedagógica”, Hispania, 96/1: 97-109. https://doi.org/10.1353/hpn.2013.0026 Bonilla, C. (2013). “Tense or aspect? A review of initial past tense marking and task conditions for beginning classroom learners of Spanish”, Hispania, 96: 624-639. https://doi.org/10.1353/hpn.2013.0116

RLyLA Vol. 14 (2019), 105-115 | 113 Hui-Chuan Lu, An Chung Cheng, and Shen Yun Hung Corpus-based Study of L3 Acquisition on Spanish Past Tense: Evidence from Learners’ Oral Production

Cadierno, T. (1995). “Formal instruction from a processing perspective: An investigation into the Spanish past tense”, Modern Language Journal, 79: 179-193. https://doi.org/10.1111/j.1540-4781.1995.tb05430.x Camps, J. (2000). Preterit and imperfect in Spanish: The early stages of development. In Applied Linguisitcs at the End of the Millenium, R. Leow and C. Sanz, (eds), 1-19. Somerville, MA: Cascadilla Press. Camps, J. (2002). “Aspectual distinctions in Spanish as a foreign language: The early stages of oral production”, International Review of Applied Linguistics, 40: 179-210. https://doi.org/10.1515/iral.2002.010 Camps, J. (2005). “The emergence of the Imperfect in Spanish as a foreign language: The association between imperfective morphology and state verbs”, International Review of Applied Linguistics, 43: 163-192. https://doi.org/10.1515/iral.2005.43.3.163 Chin, D. H.-j. (2008). “A cross-linguistic investigation on the acquisition of Spanish aspect. In Selected Proceedings of the 10th Hispanic Linguistics Symposium, ed. J. Bruhn de Garavito and E. Valenzuela, 36-50. Somerville, MA: Cascadilla Proceedings Project. Comrie, B. (1976). Aspect. Cambridge: Cambridge University Press. Cortés, M. (2002). “Dificultades linguísticas de los estudiantes chinos en el aprendizaje del ELE,” Carabela, 52: 77-98. Dalila, A. & Salaberry, M. R. (2005). Tense and Aspect in Romance Languages. Amsterdam: John Benjamins. Dickerson, L. & Dickerson, W. (1977). “Interlanguage Phonology: Current Research and Future Directions”, in P. Corder & E. Roulet (eds.), Interlanguages, Pidgins and Their Relation to Second Language Pedagogy. Librairie Droz, Neufchâtel: Faculté des Lettres and Genève, 18-29. Domínguez, L., Tracy-Venture, N., Arche, M., Mitchell, R. & Myles, F. (2013). “The role of dynamic contrasts in the L2 acquisition of Spanish past tense morphology”, Bilingualism: Language and Cognition, 16/3,558-577. https://doi.org/10.1017/S1366728912000363 Ellis, R. (2008). “Interlanguage variability in narrative discourse: Style shifting in the use of the past tense”, Studies in Second Language Acquisition, 9/1: 1-19. https://doi.org/10.1017/S0272263100006483 Granger, S. (2003). “The international corpus of learner English: A new resource for foreign language learning and teaching and second language acquisition research”, TESOL Quarterly, 37/3: 538-546. https://doi. org/10.2307/3588404 Granger, S. (2009). “The contribution of learner corpora to second language acquisition and foreign language teaching”, Corpora and Language Teaching, 33: 13-32. https://doi.org/10.1075/scl.33.04gra Hasbún, L. (1995). The Role of Lexical Aspect in the Acquisition of the Tense/Aspect System in L2 Spanish (Ph.D. dissertation). Indiana University. Hubert, D. (2013). “The development of speaking and writing proficiencies in the Spanish language, classroom: A case study”, Foreign Language Annals, 46/1: 88-95. https://doi.org/10.1111/flan.12010 Hsieh, H.-H. (2005). A Study of Communication Strategies in Taiwan EFL College Learners’ Spoken Language and Written Language (Master thesis). Retrieved from: http://nccur.lib.nccu.edu.tw/handle/140.119/33399. Last access: June 2019. Larsen-Freeman, D. (2006). “The emergence of complexity, fluency, and accuracy in the oral and written production of five Chinese learners of English”,Applied Linguistics, 27/4: 590-619. https://doi.org/10.1093/applin/ aml029 Liskin-Gasparro, J. (2000). “The acquisition of temporality in Spanish oral narratives: Exploring learners’ perceptions”, Hispania, 83/4: 830-844. https://doi.org/10.2307/346482 Li, P. & Shirai, Y. (2000). The Acquisition of Lexical and Grammatical Aspect. New York: Mouton de Gruyter. https://doi.org/10.1515/9783110800715 Li, W. (2012). “Temporal and aspectual references in Mandarin Chinese”, Journal of Pragmatics, 44: 2045-2066. https://doi.org/10.1016/j.pragma.2012.09.017 Lin, J-W. (2003). “Temporal reference in Mandarin Chinese”, Journal of East Asian Linguistics, 12: 259-311. https://doi.org/10.1023/A:1023665301095 Lin, T.-J. (2005). Corpus de textos escritos por universitarios taiwaneses estudiantes de español. Lingüística en la Red, 3: 1-58. Lopez-Ortega, N. (2000). “Tense, aspect, and narrative structure in Spanish as a second language”, Hispania, 83: 488–502. https://doi.org/10.2307/346038 Lu, H-Ch., Cheng, A. Ch. & Hung, S. Y. (2012). “Corpus escrito de aprendices taiwaneses de español I: Uso y aplicación”, Languages, Literary Studies and International Studies, 9: 45-59. Lu, H-Ch., Cheng, A. Ch. & Hung, S. Y. (2015). “La Adquisición del tiempo-aspecto en L3 para los aprendices Taiwaneses”, Círculo de Linguistica Aplicada a la Comunicación, 63: 200-217. Lu, L.-H. (2016). Corpus-based study of Spanish writing. Foreign Language Studies, 24: 145-169.

RLyLA Vol. 14 (2019), 105-115 | 114 Hui-Chuan Lu, An Chung Cheng, and Shen Yun Hung Corpus-based Study of L3 Acquisition on Spanish Past Tense: Evidence from Learners’ Oral Production

Lubbers-Quesada, M. (2006). “L2 acquisition of Temporal Reference in Spanish and the Interaction of Adverbials, Morphology and Clause Structure”, in N. Sagarra and A. J. Toribio (eds.) Selected Proceedings of the 9th Hispanic Linguistics Symposium. Somerville, MA: Cascadilla Proceedings Project, 157-168. Lubbers-Quesada, M. (2007). “La percepción de las propiedades semánticas y la adquisición de la morfología verbal en el español como L2”, Estudios de Linguística Aplicada, 44: 11–36. Martelle, W. M. (2011). Testing the Aspect Hypothesis in L2 Russian. Dissertation: University of Pittsburgh. Retrieved from http://d-scholarship.pitt.edu/id/eprint/9207. Last access: June 2019. Mendikoetxea, A. (2014). “Corpus-based Research in Second Language Spanish”, in K. L. Geeslin (ed.) The Handbook of Spanish Second Language Acquisition. John Wiley & Sons, Inc., 11-29. https://doi. org/10.1002/9781118584347.ch1 Myles, F. (2005). “Interlanguage corpora and second language acquisition research”, Second Language Research, 21/4: 373-391. https://doi.org/10.1191/0267658305sr252oa O’Donnell, M. (2012). UAM CorpusTool. http://www.wagsoft.com/CorpusTool/index.html. Last access: June 2019. Potowski, K. (2005). “Tense and Aspect in the Oral and Written Narratives of Two-Way Immersion Students”, in D. Eddington (ed.) Selected Proceedings of the 6th Conference on the Acquisition of Spanish and Portuguese as First and Second Languages. Somerville, MA: Cascadilla Proceedings Project, 123-136. Ramsay, V. (1990). Developmental Stages in the Acquisition of the Perfective and the Imperfective Aspects by Classroom L2 Learners of Spanish (Ph. D. dissertation). University of Oregon, Oregon, USA. Ruiz-Debbe, L. D. (2005). “Grammatical development of past tense in learners Spanish as L2: Oral and written productions”, in J. Cohen, K. T. McAlister, K. Rolstad & J. MacSwan (eds.) Proceedings of the 4th International Symposium on Bilingualism. Somerville, MA: Cascadilla Press. Salaberry, R. (1999). “The development of past tense verbal morphology in classroom L2 Spanish”, Applied Linguistics, 20: 151-178. https://doi.org/10.1093/applin/20.2.151 Salaberry, R. (2000). Spanish Past Tense Aspect: L2 Development in a Tutored Setting. Amsterdam: John Benjamins. Salaberry, R. (2003). “Tense aspect in verbal morphology”, Hispania, 86: 559-573. https://doi.org/10.2307/20062909 Salaberry, R. (2008). Marking Past Tense in Second Language Acquisition: A Theoretical Model. New York: Continuum. Salaberry & Ayoun (2005). “The Development of L2 Tense-Aspect in the Romance Languages”, in Tense and Aspect in Romance Languages: Theoretical and Applied Perspectives, 1-33. https://doi.org/10.1075/sibil.29.03sal Salaberry, R. & Shirai, Y. (2002). The L2 Acquisition of Tense Aspect Morphology. Philadelphia: John Benjamins Publishing. https://doi.org/10.1075/lald.27 Schmidt, T. (2011). Partitur-Editor of EXMARaLDA. http://www.exmaralda.org/tool/partitur-editor/ Shirai, Y. & Andersen, R. (1995). The Acquisition of Tense-Aspect Morphology: A Prototype Account, Language, 71/4: 743-762. https://doi.org/10.2307/415743 Skehan, P. (2009). “Modelling Second Language Performance: Integrating Complexity, Accuracy, Fluency, and Lexis.” Applied Linguistics, 30/4: 510-532. https://doi.org/10.1093/applin/amp047 Tarone, E. (1979). “Interlanguage as chameleon”, Language Learning, 29: 181-191. https://doi. org/10.1111/j.1467-1770.1979.tb01058.x Tarone, E. (1985). “Variability in interlanguage use: A study of style-shifting in morphology and syntax”, Language Learning, 35: 373-403. https://doi.org/10.1111/j.1467-1770.1985.tb01083.x Vendler, A. (1967). Linguistics in Philosophy. Ithaca, NY: Cornell University Press. Weisser, M. (2016). Corpus-based Linguistics Links. http://martinweisser.org/corpora_site/CBLLinks.html. Last access: June 2019.

RLyLA Vol. 14 (2019), 105-115 | 115 Revista de Lingüística y Lenguas Aplicadas Vol. 14 año 2019, 117-129 EISSN 1886-6298 https://doi.org/10.4995/rlyla.2019.10007

LA INFLUENCIA DE LA MEMORIA OPERATIVA Y DE LA INTELIGENCIA EMOCIONAL EN LA EXPRESIÓN DE LA EMOCIÓN EN ESPAÑOL COMO SEGUNDA LENGUA

Irini Mavrou Fernando Bustos-López Universidad Antonio de Nebrija, España

Resumen: El presente estudio persiguió examinar la influencia de la capacidad de la memoria operativa y de la inteligencia emocional en la expresión de la emoción en español como segunda lengua (EL2). Para esto, se contó con 41 aprendientes de EL2, de los niveles A2 y B1, que realizaron una tarea de expresión escrita sobre un tema emocional. Se evaluó la calidad global de sus composiciones y se midió la cantidad y la valencia de las palabras emocionales mediante la herramienta emoFinder. Además, se emplearon una medida de autoinforme de inteligencia emocional y tres pruebas para medir la capacidad de memoria operativa de los informantes. Los resultados revelaron indicios del efecto de la inteligencia emocional sobre la producción escrita en EL2, una relación positiva entre la calidad de la producción escrita y la cantidad de vocabulario emocional utilizado, así como una predominancia en el uso de palabras positivas frente a negativas.

Palabras clave: memoria operativa, inteligencia emocional, vocabulario emocional, producción escrita.

THE INFLUENCE OF WORKING MEMORY AND EMOTIONAL INTELLIGENCE ON EMOTIONAL EXPRESSION IN SPANISH AS A SECOND LANGUAGE Abstract: The current study intended to examine the influence of working memory capacity and emotional intelligence on emotional expression in Spanish as a second language (SSL). Forty-one learners of SSL (A2 and B1 levels) carried out a writing task about an emotional topic. The global quality of their texts was assessed by means of an analytical scale, and the web-based search engine emoFinder was used in order to measure the number and valence of emotional words produced by the participants. Moreover, participants completed a self-report measure of emotional intelligence and performed three working memory tasks. The results showed an influence of emotional intelligence on written expression in SSL, a positive correlation between the quality of written expression and the number of emotional words, as well as a greater use of positive words compared to the negative ones.

Keywords: working memory, emotional intelligence, emotional vocabulary, written production.

1. INTRODUCCIÓN

En el aprendizaje de una segunda lengua (L2), en general, y en la producción escrita en dicha lengua, en particular, el aprendiente se apoya en su capacidad cerebral para prestar atención y, así, categorizar, crear, almacenar y recuperar conocimiento. Este aprendizaje también requiere de la dimensión emocional del cerebro. Mora (2014) afirma que la razón no ocurre sin emoción, pues, lo que se percibe en el ambiente se transmite desde áreas cerebrales relacionadas con la emoción hasta áreas que asocian y codifican la información para memorizar y aprender. Por tanto, cognición y emoción no pueden ser vistas como dimensiones separadas, sino como un conjunto de subsistemas interconectados (Dörnyei, 2010). Esto, a su vez, da sustento a la afirmación de Snow (1994:4): “Research that takes person differences seriously also has to recognize the importance of relations among cognitive, conative, and affective functions within and across persons and situations”.

To cite this article: Mavrou, I., and Bustos-López, F. (2019). "The influence of working memory and emotional intelligence on emotional expression in Spanish as a second language". Revista de Lingüística y Lenguas Aplicadas, 14, 117-129. https://doi.org/10.4995/rlyla.2019.10007 Correspondence authors: [email protected], [email protected]

Received: 2018-04-19 Accepted: 2019-03-13 | 117 Irini Mavrou y Fernando Bustos López La influencia de la memoria operativa y de la inteligencia emocional en la expresión de la emoción en español como segunda lengua

En el ámbito de Adquisición de Segundas Lenguas (ASL), factores como la memoria operativa, la inteligencia emocional y el empleo de vocabulario emocional por parte de aprendientes de L2 han sido estudiados desde diferentes perspectivas de análisis y en relación con diferentes variables (principalmente sociodemográficas) y/o destrezas lingüísticas (véase, por ejemplo, Dewaele, 2010a; Lou y Wu, 2012; Abdolrezapour, 2013; Linck, Osthus, Koeth y Bunting, 2014). Ahora bien, apenas encontramos investigaciones en las que dichas variables se aborden de manera conjunta desde la perspectiva de la Lingüística Aplicada, aunque sí existe constancia de estudios centrados en la recuperación de vocabulario emocional en L2 (Anooshian y Hertel, 1994; Altarriba y Bauer, 2004; Ayçiçeği-Dinn y Caldwell-Harris, 2009; Ferré, García, Fraga, Sánchez-Casas y Molero, 2010; Jiménez Catalán y Dewaele, 2017). De ahí el interés del presente estudio que persigue examinar la influencia de la memoria operativa y de la inteligencia emocional en la expresión emocional en español como L2 (EL2), enfocando en una destreza concreta, la expresión escrita.

2. MEMORIA Y EMOCIÓN

En su libro Neuroeducación: Solo se puede aprender aquello que se ama, Mora (2016:91) afirma: “Aprender es, en su esencia, ser capaz de sobrevivir. Y sobrevivir es la ley suprema de todo organismo”. Según el mismo autor, el aprendizaje es un proceso de asociación de sucesos que conduce a la adquisición de nuevo conocimiento (Mora, 2014:164). En la misma línea, Morgado (2014:25) señala: “Aprender significa básicamente adquirir nuevas representaciones neuronales de información y establecer relaciones funcionales entre ellas y las ya existentes en el cerebro”. Para que dicho proceso tenga lugar, el individuo se sirve de un mecanismo especialmente complejo, la memoria.

La memoria comprende tres tipos de almacenes de información: el registro sensorial, mediante el cual se perciben los estímulos visuales y auditivos provenientes del exterior; la memoria operativa, donde se almacena y se procesa temporalmente la información; y la memoria a largo plazo, en la que están almacenados los conocimientos que hemos adquirido a lo largo de nuestra vida (para una revisión véase Mavrou, 2017). Por tanto, aprendizaje y memoria pueden considerarse “como las dos caras de una misma moneda, de tal modo que el uno no puede darse sin el otro, es decir, no puede haber aprendizaje sin memoria ni memoria sin aprendizaje” (Morgado, 2014:25). A su vez, estos procesos modifican el cerebro y sus conexiones, así como la conducta del individuo, y gracias a ellos se pueden transmitir los conocimientos y crear cultura (Mora, 2014:164).

De los almacenes de memoria descritos anteriormente, merece especial atención la memoria operativa (MO). Dicho mecanismo se ocupa del almacenamiento temporal y la actualización de cantidades reducidas de información, las cuales mantiene activas y accesibles para su posterior transformación y manipulación en tiempo real a través del uso de estrategias (Jonides, Lacey y Nee, 2005). Se trata de un tipo de memoria de carácter consciente y prospectivo, que ayuda a mantener un hilo conductor en tareas en curso, entre lo que acaba de suceder, lo que sucede y lo que sucederá en el futuro inmediato (Kane, Conway, Hambrick y Engle, 2007:21). Esta memoria es necesaria para la comprensión y expresión del lenguaje y está implicada en las interacciones comunicativas, sea con hablantes de la misma o de diferentes lenguas (Baddeley, 2003). En última instancia, es el mecanismo que “da un sentido de unidad y continuidad a nuestra experiencia consciente y contribuye a la formación de memorias a largo plazo reforzando las conexiones entre los elementos asociados en ella” (Morgado, 2014:98).

A pesar de la gran importancia del sistema cognitivo –y, para el caso que nos atañe, la MO– en los procesos de aprendizaje, este no puede entenderse en su totalidad sin tomar en cuenta los estados afectivos y emocionales del individuo. De hecho, cognición y emoción se consideran como un “binomio indisoluble y cardinal” para entender la esencia de lo que hay que enseñar y aprender (Mora, 2014:42). Recordamos especialmente aquello que nos emociona, que tuvo un significado particular o impacto considerable sobre nosotros, negativo o positivo; de la misma manera, aprendemos más fácilmente aquello que nos atrae, que despierta nuestra curiosidad e interés, lo que nos agrada y nos motiva.

No obstante, ¿qué se entiende por emoción y cómo esta interactúa con el sistema de memoria? Damasio (2001:781) define la emoción de la siguiente manera: “An emotion [...] is a patterned collection of chemical and neural responses that is produced by the brain when it detects the presence of an emotionally competent stimulus –an object or situation, for example”. De las regiones cerebrales, el sistema límbico es el área que guarda una estrecha relación con los comportamientos emocionales. Dentro de este sistema destaca el complejo amigdalino o amígdala, pieza fundamental en la evolución del ser humano (Damasio, 1994). Esta estructura cerebral está conectada directa e indirectamente con otras áreas cerebrales, como el hipocampo, que se ocupan de los diferentes tipos de memoria. Esta conexión, además, es responsable de la liberación de hormonas y neurotransmisores, como la adrenalina y el cortisol, que aumentan el efecto de la emoción fortificando el recuerdo en la memoria. Cuanta más excitación emocional se produce en la amígdala ante un estímulo que induce algún recuerdo, más

RLyLA Vol. 14 (2019), 117-129 | 118 Irini Mavrou y Fernando Bustos López La influencia de la memoria operativa y de la inteligencia emocional en la expresión de la emoción en español como segunda lengua

detallado y real será este. Por este motivo, aquellos eventos con una intensa carga emocional vividos en primera persona son los más recordados.

La investigación empírica ha mostrado que el entrenamiento en tareas de MO que incluyen estímulos visuales fomenta la capacidad de control emocional y que dicha memoria parece explicar hasta un 50% de la variabilidad en el coeficiente de inteligencia de Spearman (Conway, Kane y Engle, 2003:551). Del mismo modo, se ha sugerido que la realización de tareas de MO que incluyen estímulos con carga emocional va también en beneficio del control emocional (Gutiérrez-Cobo, Cabello y Fernández-Berrocal, 2017a). Este tipo de control constituye un aspecto de la inteligencia emocional, otra noción clave relacionada con la conducta humana, en general, y el aprendizaje de lenguas, en el caso que aquí nos ocupa.

3. INTELIGENCIA EMOCIONAL Y VOCABULARIO EMOCIONAL

El concepto actual de inteligencia emocional (IE), popularizada en 1995 por Goleman (2006) en su libro Inteligencia Emocional, parte de la Teoría de las Inteligencias Múltiples del psicólogo Howard Gardner (1983). Entre las diferentes inteligencias propuestas por Gardner, dos de ellas, la intrapersonal y la interpersonal, constituyeron el punto de partida para que la investigación empírica estudiara la influencia de las emociones en el pensamiento y en el reconocimiento de las emociones en otras personas (Mayer, Salovey, Caruso y Cherkasskiy, 2011:529). Es entonces cuando Salovey y Mayer (1990) propusieron su primera definición de IE, la cual redefinieron unos años más tarde dividiéndola en cuatro subconjuntos de capacidades: “Emotional intelligence is the ability to perceive emotions, to access and generate emotions so as to assist thought, to understand emotions and emotional meanings, and to reflectively regulate emotions so as to promote both better emotion and thought” (Mayer y Salovey, 1997:22). Desde esta perspectiva, la IE se concibe como otro tipo de inteligencia más que implica el razonamiento y procesamiento de información emocional. A partir de esa primera definición, el término fue evolucionando en diferentes modelos de IE como, por ejemplo, los modelos mixtos o de rasgos (Petrides y Furnham, 2003; Bar-On, 2006; Goleman, 2006) que incluyen características personales socioafectivas ajenas al ámbito de la inteligencia, y los modelos integradores o de capacidades (Mayer y Salovey, 1997).

En relación con el vocabulario emocional, los diversos enfoques teóricos contemplan las palabras emocionales como aquellas que se refieren a una emoción, sentimiento o estado de ánimo y sirven para describir y expresar lo que una persona siente (Kovecses, 1995; Altarriba y Bauer, 2004; Pavlenko, 2008). Sin embargo, existe cierta discrepancia sobre si incluir dentro de esta categoría términos que desencadenan una emoción de forma indirecta, es decir, palabras con carga emocional que no hacen referencia a una emoción en sí misma (Pavlenko, 2008:148). En investigaciones recientes con corpus lingüísticos, las palabras emocionales se suelen categorizar en función de su valencia (positiva, negativa o neutra) y activación (baja, intermedia o alta). Las primeras son positivas con activación variada o negativas con alta activación, mientras que las palabras no emocionales se caracterizan como neutras con activación intermedia (Stadthagen-Gonzalez, Imbault, Perez Sanchez y Brysbaert, 2017).

En el ámbito de ASL, tanto las emociones como la IE se han investigado escasamente (Gregersen, MacIntyre, Finegan, Talbot y Claman, 2014). Dewaele y Pavlenko (Dewaele y Pavlenko, 2001-2003; Dewaele, 2008, 2010b; Pavlenko, 2008) han llevado a cabo varios estudios sobre el vocabulario emocional y las diferencias en la expresión de las emociones entre aprendientes de L2 provenientes de diferentes culturas. En estos estudios los autores insisten en la importancia de enseñar a los alumnos léxico para expresar emociones, ya que esto les permitirá no solo comportarse de manera pragmáticamente adecuada en el país de habla de la lengua meta, sino también comprender e identificar las emociones de los hablantes del país extranjero y evitar así los malentendidos culturales.

En cuanto a la IE, las investigaciones empíricas de ASL también son exiguas y se han centrado principalmente en su relación con la competencia escrita (Abdolrezapour, 2013) y la competencia lingüística global en L2 (Zarafshan y Ardeshiri, 2012), el aprendizaje de vocabulario (Pishghadam, 2009; Alavi y Rahimi, 2011) y el aprendizaje y uso de estrategias (Zafari y Biria, 2014; Shakarami y Khajehei, 2015), entre otros. La principal conclusión que se deriva de los resultados de estas investigaciones es que, si bien la IE parece no influir directamente en la competencia global en L2, sí correlaciona positiva y significativamente con el uso de estrategias socioafectivas para afrontar el aprendizaje y lidiar con cuestiones emocionales y afectivas que existen en el entorno escolar, como pueden ser la ansiedad, el miedo a hablar en público o la regulación emocional.

Lo expuesto en las líneas anteriores nos llevaría a considerar que memoria, emoción, IE y léxico emocional no son sino piezas de un mismo conjunto, dependientes las unas de las otras. El presente trabajo persigue aproximarse a estos conceptos y examinar su posible interconexión en el ámbito de EL2.

RLyLA Vol. 14 (2019), 117-129 | 119 Irini Mavrou y Fernando Bustos López La influencia de la memoria operativa y de la inteligencia emocional en la expresión de la emoción en español como segunda lengua

4. METODOLOGÍA

Tomando en cuenta el marco teórico expuesto, el objetivo del presente estudio fue examinar la influencia de la MO y de la IE en la cantidad de palabras emocionales (PE) utilizadas en una tarea escrita sobre un tema emocional, así como en la calidad global de la expresión escrita. En función de este objetivo general, se enunciaron las siguientes preguntas de investigación:

1. ¿Existe relación entre la capacidad de la MO, la cantidad de PE evocadas y producidas en una tarea escrita sobre un tema emocional, así como la calidad global de la expresión escrita?

2. ¿Existe relación entre la IE, la cantidad de PE evocadas y producidas en una tarea escrita sobre un tema emocional, así como la calidad global de la expresión escrita?

3. ¿Existen diferencias en la cantidad de PE y en la calidad global de la expresión escrita en función del nivel de competencia lingüística de los informantes (A2 y B1)?

4.1. Participantes En el estudio participaron 41 estudiantes de EL2, 8 hombres y 33 mujeres, de edades oscilantes entre los 17 y los 29 años (M=20,80, Mediana=20, SD=2,44). De estos, 21 tenían un nivel de competencia lingüística de A2, mientras que los 20 restantes se encontraban en un nivel B1. La mayoría de ellos eran estadounidenses (n=22) y tailandeses (n=11), dos eran de nacionalidad china y los cuatro restantes tenían el ruso, el japonés, el turco y el holandés como lengua materna (LM). Todos los informantes realizaban cursos de verano en la Universidad Antonio de Nebrija y eran estudiantes universitarios en sus correspondientes países, cursando carreras de diversos ámbitos. Mediante la información recabada a través de un cuestionario se constató que la gran mayoría de ellos (n=35) se involucró en el aprendizaje de EL2 a partir de los 12 años. Por lo que respecta a la expresión escrita, un 85,4% afirmaron que dedicaban menos de 3 horas a la semana a la escritura de textos en español, mientras que en su LM la escritura ocupaba más de 3 horas por semana para un porcentaje de 77,5% de ellos.

4.2. Instrumentos 4.2.1. Trait Meta-Mood Scale La inteligencia emocional se evaluó con el Trait Meta-Mood Scale (TMMS; Salovey, Mayer, Goldman, Turvey y Palfai, 1995) que fue administrado en inglés, idioma que todos los informantes conocían. En concreto, se utilizó la versión de 30 ítems que comprende oraciones relacionadas con tres dimensiones del metaconocimiento emocional –atención emocional (13 ítems), claridad de sentimientos (11 ítems) y reparación emocional (6 ítems)– y el informante debe expresar su grado de acuerdo o desacuerdo a través de una escala Likert de 5 opciones.

En cuanto a las tres dimensiones señaladas, la atención emocional está relacionada con el grado de atención que el individuo presta a sus emociones y estados anímicos, la claridad de sentimientos se refiere a la capacidad del individuo de entender y diferenciar entre sus sentimientos y estados de ánimo, mientras que la reparación emocional alude a la capacidad para la autorregulación emocional.

4.2.2. Pruebas de memoria operativa La capacidad de la MO de los informantes se midió con tres pruebas, a saber, Math Span Test (MST), Corsi Block-Tapping Task o Corsi Task y Stroop Test. El MST es una prueba compleja de MO que evalúa la capacidad de procesamiento y almacenamiento simultáneos. La versión utilizada en el presente estudio (Shahnazari- Dorcheh y Roshan, 2012) consistía en 60 operaciones matemáticas sencillas (30 sumas y 30 restas) distribuidas en cinco bloques de 2 a 6 problemas, y se administró mediante el programa de presentación PowerPoint. Los informantes tenían que calcular y decir en voz alta el resultado del problema matemático y, a la vez, intentar memorizar el segundo dígito. Una vez que terminaba cada bloque de operaciones matemáticas, se mostraba una diapositiva con signos de interrogación donde el informante debía recordar y decir en voz alta y en el mismo orden de aparición los segundos dígitos correspondientes a las operaciones matemáticas. La evaluación de la prueba se llevó a cabo considerando como respuestas correctas solo aquellas en las que el informante acertó las operaciones matemáticas y la recuperación de los dígitos correspondientes, siendo 60 la puntuación máxima que se podía obtener.

El Stroop Test (Stroop, 1935) mide la capacidad del individuo de guiar y redirigir la atención hacia un estímulo planteado e inhibir la interferencia de otros estímulos ante tal demanda. La versión utilizada en el presente estudio comprendía tres listados de cien palabras en inglés referentes a colores (brown, red, purple, blue, green; en diez columnas de diez palabras) que los participantes debían leer en un minuto. La primera lista estaba impresa en blanco y negro, la segunda en color, donde el color de la tinta coincidía con la palabra escrita, mientras que la tercera lista, también en color, presentaba las palabras escritas con un color de tinta diferente. Esta prueba, que requiere

RLyLA Vol. 14 (2019), 117-129 | 120 Irini Mavrou y Fernando Bustos López La influencia de la memoria operativa y de la inteligencia emocional en la expresión de la emoción en español como segunda lengua

tres minutos en total (un minuto por página), se administró en formato impreso en tres folios independientes. Para los propósitos del presente estudio, se utilizaron los resultados de los informantes en la tercera lista de la prueba, asignando 1 punto a cada respuesta acertada, esto es, siempre que se leyera la palabra escrita sin interferencias del color de la tinta. La máxima puntuación que se podía obtener era de 100 puntos.

La última prueba de MO, el Corsi Task (Corsi, 1972), evalúa el componente visoespacial de la MO y fue administrada por ordenador utilizando la versión en línea que se puede encontrar en PsyToolKit (Stoet, 2010). Durante su realización, se le presentaba a cada informante nueve cubos dispuestos aleatoriamente en la pantalla y que cambiaban de color, empezando por dos cubos y ascendiendo el número de estos a lo largo de la prueba. El informante, después de leer las instrucciones en inglés y presionar la tecla de “espacio”, escuchaba la palabra “go” y debía clicar sobre los cubos que había visto cambiar de color, en el mismo orden de aparición. Si acertaba, pasaba al siguiente nivel en el que la secuencia comprendía un cubo adicional. En el caso de reproducir la secuencia erróneamente, se generaba otra secuencia con el mismo número de bloques. La prueba finalizaba cuando el informante erraba dos veces en la misma cantidad de bloques que tenía que recordar (por ejemplo, si se equivocaba en una secuencia de seis cubos, se le ofrecía otra oportunidad con otros seis). El número de aciertos definía elblock span.

Es preciso señalar que, por cuestiones de viabilidad relacionadas con el tiempo requerido para la cumplimentación de la totalidad de pruebas administradas en el presente estudio, tanto dentro como fuera del aula, las pruebas de MO se aplicaron solo al grupo de nivel A2, por lo que para contestar la primera pregunta de investigación se tomará como referencia solo a este grupo.

4.2.3. Tarea escrita Para la producción escrita, se utilizó una tarea sobre un tema emocional cuyo objetivo era inducir recuerdos que suscitaran emociones profundas en los informantes (Chapado y López, 2012). Las instrucciones fueron las siguientes: “Todos recordamos una noche especial en nuestra vida. Escribe una redacción de entre 130 y 150 palabras sobre esa noche. Intenta incluir la siguiente información: dónde y cuándo fue, con quién estabas, qué pasó, por qué fue especial”. Si bien las instrucciones estaban en español, se dieron explicaciones también en inglés cuando se consideraron necesarias para la comprensión de la tarea.

4.2.4. Vocabulario emocional Con el objetivo de establecer un criterio homogéneo a la hora de contabilizar de manera idéntica las PE, se optó por la herramienta en línea emoFinder (Fraga, Guasch, Haro, Padrón y Ferré, 2018) y el análisis se llevó a cabo considerando el número total de PE diferentes. En emoFinder pueden encontrarse más de 30.000 palabras evaluadas según dimensiones emocionales (valencia, activación y dominancia), categorías emocionales discretas (alegría, tristeza, miedo, asco e ira) y valores subjetivos (familiaridad, concreción, disponibilidad contextual, edad de adquisición, imaginabilidad, tiempo y experiencia sensorial) provenientes de diez bases de datos diferentes. En el presente estudio, las PE fueron evaluadas según su valencia de acuerdo con el criterio de Guasch, Ferre y Fraga (2016:7), es decir, se consideraron emocionalmente positivas las palabras que obtuvieron una media superior a 5,82 y negativas aquellas con una media inferior a 3,73.

4.2.5. Escala de evaluación de la expresión escrita La evaluación de las producciones escritas de los informantes se llevó a cabo con una adaptación de la escala Writing Assessment Measure (WAM). Dicha matriz había sido utilizada en el estudio de Dunsmuir et al. (2014) para medir la calidad global de la escritura narrativa de niños nativos de inglés de entre 7 y 11 años, mostrando una consistencia interna satisfactoria (α de Cronbach=0,87) y correlaciones inter-ítem significativas al nivel .01 y oscilantes entre 0,63 y 0,85 (validez de contenido). La decisión de utilizar esta escala se tomó después de considerar el nivel de competencia escrita de los informantes, así como el tipo y requisitos de la tarea escrita que realizaron.

Después de traducirse al español, la versión de la WAM utilizada en el presente estudio contó con los criterios de ortografía, puntuación, estructura oracional y gramática, vocabulario, e ideas, puntuándose cada uno de ellos en una escala de 1 a 4 puntos (máxima puntuación: 20 puntos). El parámetro de ortografía consideraba la escritura correcta de palabras, desde palabras comunes con alta frecuencia de aparición hasta palabras complejas con sufijos o formación irregular. En el criterio de puntuación se tomó en cuenta el uso preciso de mayúsculas y signos de puntuación. El tercer parámetro se refería a la estructura global (oraciones simples y compuestas) y a la gramática (precisión en el uso variado de tiempos verbales, concordancia entre sustantivos y determinantes y entre sujeto y verbo, uso de la voz pasiva, etc.). En lo referente al vocabulario, se evaluó el uso de léxico variado y la intención de elaborar una composición más dinámica desde el punto de vista narrativo. Por último, el criterio de ideas trató aspectos relacionados con las descripciones detalladas, la imaginación, la creatividad narrativa, etc.

RLyLA Vol. 14 (2019), 117-129 | 121 Irini Mavrou y Fernando Bustos López La influencia de la memoria operativa y de la inteligencia emocional en la expresión de la emoción en español como segunda lengua

Los dos parámetros que se eliminaron fueron: (a) escritura a mano, que carecía de sentido en el contexto del presente estudio, ya que todos los informantes hacían uso del alfabeto latino (LM o L2); y (b) organización y estructura global, porque las producciones escritas no superaban las 150 palabras y en su totalidad estaban escritas en un único párrafo.

4.3. Procedimiento El procedimiento de recogida de datos comprendió varias etapas. En primer lugar, a todos los informantes se les comunicó el propósito del estudio y se les informó sobre su participación voluntaria y la confidencialidad de sus datos personales y de sus resultados en las pruebas realizadas. Asimismo, se les pidió que completaran y firmaran una hoja de consentimiento donde se explicitaban por escrito todos los detalles que concernían al estudio.

En segundo lugar, los informantes realizaron la tarea escrita dentro del horario lectivo y como actividad de clase. Para el grupo de nivel B1, el docente estuvo al cargo de dar las instrucciones pertinentes que previamente le fueron comunicadas (límite de tiempo e imposibilidad de uso de fuentes de consulta externas), de resolver cualquier duda que surgiera y de repartir la tarea. Una vez finalizada la tarea, se les pidió a los informantes que se tomaran el tiempo oportuno para completar, de forma voluntaria, el TMMS y un cuestionario de perfil lingüístico.

El procedimiento fue ligeramente diferente para los informantes de nivel A2. La tarea formó parte del examen final y, por ende, no tuvieron acceso a ninguna fuente externa de consulta. El TMMS y el cuestionario de perfil lingüístico se les fue proporcionado al final de una clase y se les pidió que lo completaran a su conveniencia, dentro o fuera del horario lectivo, y lo entregaran cuando quisieran, teniendo como fecha límite el último día del curso. Las pruebas de MO, que se administraron solo a este grupo, se llevaron a cabo de manera individual fuera del horario lectivo y después de acordar con los estudiantes sesiones de 20 minutos aproximadamente. Primero, se administraba el Corsi Task, a continuación, el MST y, en último lugar, el Stroop Test; todos los datos se registraron en hojas de respuestas habilitadas para cada prueba.

En la última etapa del estudio, se procedió a la evaluación de las producciones escritas de los informantes mediante la WAM y a la identificación de las PE, que se hizo considerando solo los ítems de contenido referencial y semántico y excluyendo las palabras gramaticales y los nombres propios de persona, ciudades, números, abreviaturas, etc. Posteriormente, dichas palabras se introdujeron en la herramienta emoFinder para su categorización en total de PE, total de PE positivas y total palabras de PE negativas.

5. RESULTADOS

La Tabla 1 recoge los valores descriptivos de las variables del estudio, mientras que la Tabla 2 presenta la matriz de correlaciones inter-ítem entre los diferentes descriptores de la WAM, así como entre la puntuación en esos descriptores y la puntuación final de los informantes en la tarea escrita. La consistencia interna de la escala, que se estimó mediante el coeficiente α de Cronbach, alcanzó un valor de 0,830, muy próximo al valor de α encontrado por Dunsmuir et al. (2014).

Cabe señalar que se registraron un total de 2.441 ítems léxicos (A2: 1.249, B1: 1.192), de los cuales 1.208 resultaron PE (A2: 637, B1: 571). En cuanto a su valencia, 1.155 fueron positivas (A2: 615, B1: 540), mientras que tan solo 53 se clasificaron como negativas (A2: 22, B1: 31).

Tabla 1. Valores descriptivos de las variables del estudio. n Media SD Corsi Task A2 21 5,48 0,93 Stroop Task A2 21 65,67 14,30 Math Span Test A2 21 22,81 9,81 TMMS Atención A2 21 49,67 7,89 B1 20 46,20 8,56 Total 41 47,98 8,31 TMMS Claridad A2 21 38,10 4,85 B1 20 37,35 7,75 Total 41 37,73 6,36 TMMS Reparación A2 21 22,19 4,60 B1 20 22,65 4,59 Total 41 22,41 4,54 Tabla 1, sigue en la página siguiente.

RLyLA Vol. 14 (2019), 117-129 | 122 Irini Mavrou y Fernando Bustos López La influencia de la memoria operativa y de la inteligencia emocional en la expresión de la emoción en español como segunda lengua

Tabla 1, sigue de la página anterior. n Media SD WAM A2 21 9,52 1,78 B1 20 14,10 1,83 Total 41 11,76 2,92 PE A2 21 30,33 7,83 B1 20 28,55 6,63 Total 41 29,46 7,24 PE positivas A2 21 29,29 7,09 B1 20 25,25 8,87 Total 41 27,32 8,17 PE negativas A2 21 1,05 1,63 B1 20 1,55 1,50 Total 41 1,29 1,57

Tabla 2. Matriz de correlaciones inter-ítem entre los descriptores de la escala de evaluación (WAM).

Estructura y Ortografía Puntuación gramática Vocabulario Ideas Ortografía - Puntuación 0,245 - Estructura y gramática 0,429 0,598 - Vocabulario 0,600 0,395 0,508 - Ideas 0,607 0,412 0,512 0,792 - Puntuación final 0,702 0,699 0,810 0,842 0,833

A fin de examinar la posible relación entre la MO, la cantidad de PE y la calidad global de los textos que escribieron los informantes, se efectuó la prueba de correlación de Pearson. Como se señaló anteriormente, dicho análisis se hizo solo con los datos de los informantes del nivel A2 que habían realizado las tres pruebas de MO. De la Tabla 3 se desprende que las únicas correlaciones que se obtuvieron fueron entre la cantidad de PE y PE positivas, por un lado, y la puntuación final en la tarea escrita, por otro (r=0,486, p=0,026, y r=0,472, p=0,031, respectivamente), así como entre las puntuaciones en el Corsi Task y en el MST (r=0,433, p=0,05) aunque el correspondiente a esta correlación valor p fue de 0,05, por lo que los resultados se deberían interpretar con precaución.

Tabla 3. Correlaciones entre la capacidad de MO, la cantidad de PE y la calidad global de la producción escrita (solo nivel A2).

Corsi Task Stroop Task Math Span Test WAM Corsi Task - Stroop Test -0,010 - Math Span Test 0,433* 0,030 - WAM -0,068 0,165 -0,060 - PE 0,011 -0,063 0,086 0,486* PE positivas 0,024 -0,111 0,066 0,472* *p < 0,05

En lo que se refiere a la segunda pregunta de investigación, se volvió a efectuar la prueba de correlación de Pearson entre las variables IE, cantidad de PE y PE positivas y puntuación final en la tarea escrita, tanto con toda la muestra como con cada grupo (A2 y B1) por separado (Tabla 4). Los resultados revelaron una correlación negativa y estadísticamente significativa entre la puntuación total de los informantes del nivel A2 en la tarea escrita y el componente de reparación del TMMS (r=-0,502, p=0,020).

En relación con la tercera pregunta de investigación (Tabla 5), los resultados mostraron que los informantes del nivel B1 produjeron textos más elaborados y correctos en comparación con los estudiantes que poseían un nivel A2 (t=-8,115, p < 0,01). Por otro lado, no se encontraron diferencias significativas en la cantidad de PE en función del nivel de competencia lingüística de los informantes (t=0,785, p=0,437) y, como sería de esperar, tampoco las hubo en lo relativo a las PE positivas (t=0,760, p=0,115).

RLyLA Vol. 14 (2019), 117-129 | 123 Irini Mavrou y Fernando Bustos López La influencia de la memoria operativa y de la inteligencia emocional en la expresión de la emoción en español como segunda lengua

Tabla 4. Correlaciones entre la IE, la cantidad de PE y la calidad global de la producción escrita.

TMMS TMMS TMMS Atención Claridad Reparación WAM WAM (A2 y B1) -0,153 -0,061 0,002 - WAM A2 0,216 -0,209 -0,502* - WAM B1 -0,156 0,094 0,386 - PE (A2 y B1) 0,005 -0,037 -0,175 0,163 PE A2 0,181 -0,097 -0,032 0,486* PE B1 -0,225 -0,008 -0,343 0,372 PE positivas (A2 y B1) 0,025 -0,017 -0,076 -0,065 PE positivas A2 0,185 -0,129 0,032 0,472* PE positivas B1 -0,196 0,017 -0,153 0,028 *p < 0,05

Tabla 5. Comparación de la calidad global de los textos y la cantidad de PE en función del nivel EL2.

Test de Levene para la igualdad de varianzas t-test para la igualdad de medias F Sig. t gl Sig. (bilateral) WAM 0,237 0,629 -8,115 39 p < 0,001 PE 0,273 0,604 0,785 39 0,437 PE positivas 0,095 0,760 0,760 39 0,115

Por último, cabe señalar que no se encontraron correlaciones estadísticamente significativas entre las puntuaciones de los informantes en los diferentes componentes del TMMS y su capacidad de MO (Tabla 6), aunque se apreció una relación positiva entre el componente de reparación del TMMS y los componentes verbal y visoespacial de la MO.

Tabla 6. Correlaciones entre la inteligencia emocional y la capacidad de la memoria operativa (solo nivel A2).

TMMS TMMS TMMS Atención Claridad Reparación Corsi Task -0,188 -0,105 0,282 Stroop Test -0,026 -0,016 0,055 Math Span Test -0,171 -0,177 0,255

6. DISCUSIÓN

El objetivo del presente estudio consistió en examinar la influencia de la MO y de la IE en la producción escrita de estudiantes de EL2 de los niveles A2 y B1, así como en la cantidad de vocabulario emocional utilizado en sus textos. Respecto a la MO, en el grupo de nivel A2, se encontró una correlación significativa, aunque débil, entre las puntuaciones de los informantes en las pruebas Corsi Task y MST; en otras palabras, aquellos informantes con una mayor capacidad de MO verbal presentaron también una mayor capacidad de memoria visoespacial. Este resultado está en consonancia con los obtenidos por Alloway, Gathercole y Pickering (2006), con una muestra de 709 niños de entre 4 y 11 años, que encontraron correlaciones altas entre la MO verbal y la visoespacial, así como por los de Farrand y Jones (1996) quienes llegaron a la conclusión de que ambos tipos de MO comparten los mismos mecanismos de ensayo mental y guardan una relación de equidad dentro de la memoria a corto plazo. En la misma línea, Mammarella y Cornoldi (2005) examinaron el rendimiento de niños con y sin discapacidad en pruebas de MO verbal (forward y backward digit span task) y visoespacial (forward y backward Corsi Task) y observaron que, en ambas pruebas de tipo forward, los niños sin discapacidad obtuvieron resultados más altos, aunque también merece destacar que los niños con discapacidad emplearon de manera eficiente estrategias verbales para realizar la forward Corsi Task. Por tanto, los resultados de la presente investigación podrían explicarse en el marco de un modelo integrador de MO donde se esperaría que ambas capacidades de MO, verbal y visoespacial, estuvieran vinculadas. Esta suposición se sustenta también con los resultados de un estudio con población adulta

RLyLA Vol. 14 (2019), 117-129 | 124 Irini Mavrou y Fernando Bustos López La influencia de la memoria operativa y de la inteligencia emocional en la expresión de la emoción en español como segunda lengua

realizado por Kane et al. (2004; véase también Oberauer, Suß, Schulze, Wilhelm y Wittmann, 2000, y Kane et al., 2007) que arrojaron una varianza compartida de entre 70-80% entre la capacidad verbal y espacial de la MO.

Los resultados del presente estudio también mostraron un mayor uso de PE positivas en comparación con las PE negativas. Dicha prevalencia podría deberse a que el tema de la tarea escrita cubría explícitamente un tema emocional positivo. Investigaciones que examinaron el léxico emocional en relación con la memoria revelaron cómo la valencia de las palabras influye en su posterior recuperación (Ayçiçeği-Dinn y Caldwell-Harris, 2009; Zimmerman y Kelley, 2010; Jiménez Catalán y Dewaele, 2017), además de una predominancia de PE positivas recuperadas. En cuanto a la relación positiva que se obtuvo entre la cantidad de vocabulario emocional producido (PE y PE positivas) y la calidad global de las producciones escritas de los informantes, cabría presuponer que una mayor competencia lingüística suele ir acompañada de una mayor competencia léxica que, a su vez, contribuye positivamente a la calidad de la expresión escrita. De hecho, Dewaele y Pavlenko (2002) concluyeron que, entre otros factores, el nivel de competencia lingüística en L2 determina el porcentaje de PE producidas en dicha lengua.

En lo que se refiere a la IE, los resultados aquí obtenidos contradicen los hallazgos de otras investigaciones de ASL (Abdolrezapour, 2013; Shao, Yu y Ji, 2013; Korpi y Farvardin, 2016) en las que se encontró un vínculo positivo entre la IE y la producción escrita en L2. Por otro lado, en el presente estudio, se reveló una correlación negativa y significativa entre la calidad de los textos que escribió el grupo del nivel A2 y el componente de reparación emocional del TMMS. La reparación emocional implica aquellos intentos para regular o reparar estados de ánimo negativos, poco placenteros o desagradables y/o mantener activos aquellos estados anímicos positivos; además, este componente comprende la activación de estrategias para llevar a cabo dichos intentos de regulación emocional eficazmente (Saloveyet al., 1995:129). Dentro de estas estrategias se encontrarían la supervisión y la atención de las emociones con carácter introspectivo (Bisquerra Alzina, Pérez González y García Navarro, 2015:54).

Por tanto, la correlación negativa hallada podría indicar que aquellos informantes que se autoevaluaron con una alta capacidad de reparación emocional emplearon más recursos cognitivos en las estrategias de regulación emocional ante la realización de la tarea escrita. Esto restaría tiempo y/o atención dedicados a los aspectos lingüísticos de su producción escrita (evaluados a través de los descriptores de la escala WAM) y probablemente también a la revisión de sus textos. Otra interpretación plausible sería que aquellos informantes que se sentían más inseguros o mostraban más ansiedad ante una calidad menor de sus textos necesitaron más estrategias de regulación para paliar los posibles efectos derivados de estos estados anímicos negativos (por ejemplo, nerviosismo) o para mantenerse en un estado de ánimo positivo y calmado.

Lo anterior concuerda con los resultados del estudio de Gutiérrez-Cobo, Cabello y Fernández-Berrocal (2017b; véase también Gutiérrez-Cobo, Cabello y Fernández-Berrocal 2016, 2017a) que arrojaron una correlación negativa entre el componente de reparación del TMMS y el tiempo de reacción en tareas emocionales que exigían recursos cognitivos (concentración y procesamiento de la información). Como señalan los autores, la percepción de una alta capacidad para regular emociones iría acompañada del uso de estrategias menos controladas a la hora de realizar tareas complejas. En cualquier caso, es importante subrayar que el TMMS es una escala de autoevaluación, por lo que en futuras investigaciones sería recomendable la inclusión de medidas más objetivas de la IE, como el Mayer Salovey Caruso Emotional Intelligence Test (Mayer, Salovey, Caruso y Sitarenios, 2003), a fin de poder llegar a conclusiones menos sesgadas sobre su influencia en la producción escrita en L2.

En lo que concierne a la influencia del nivel de competencia lingüística en EL2 en la calidad global de la expresión escrita, el grupo de nivel B1 obtuvo mejores resultados que el grupo de nivel A2, aunque no se apreciaron diferencias significativas en la cantidad de PE y PE positivas utilizadas en sus textos, hallazgo que podría atribuirse a que los informantes del presente estudio se encontraban próximos en el continuum de competencia lingüística (A2-B1). Compartiendo la reflexión de Ferréet al. (2010:780), una diferencia notoria en la cantidad de PE se podría encontrar si se compararan estudiantes de L2 de nivel inicial o intermedio (A1-B1) con aquellos de nivel avanzado (C1-C2), ya que estos últimos han integrado el léxico a nivel conceptual, mientras que en los niveles iniciales todavía predomina la asociación forma-significado entre la L2 y la LM.

Por último, la relación positiva –aunque no significativa– que se obtuvo entre el componente de reparación del TMMS y la capacidad de MO podría indicar que la capacidad para procesar información e inhibir interferencias y/o distracciones es una pieza necesaria en la regulación emocional (Gutiérrez-Cobo et al., 2017a, 2017b; Xiu, Zhou y Jiang, 2016). Como sostienen Xiu et al. (2016), aquellos individuos con mayor capacidad de actualización de la información presente en la MO disponen de más recursos atencionales que pueden emplear en ajustar su estado de ánimo en situaciones que surgen a lo largo del día, en beneficio de su propio bienestar.

RLyLA Vol. 14 (2019), 117-129 | 125 Irini Mavrou y Fernando Bustos López La influencia de la memoria operativa y de la inteligencia emocional en la expresión de la emoción en español como segunda lengua

7. CONCLUSIONES

Los resultados de la presente investigación ponen de manifiesto el papel protagonista que juegan las PE en la producción escrita, especialmente las PE positivas, y marcan un nuevo camino a seguir por docentes y editoriales: se debería crear materiales y enseñar una L2 tratando con detalle el aprendizaje, comprensión y expresión del vocabulario emocional. Igualmente, la correlación hallada entre el componente de reparación del TMMS y la calidad global de la producción escrita acentúa la importancia de enseñar a los estudiantes estrategias de regulación emocional que promuevan un mayor rendimiento en clase, puesto que: “Autorregular el estado emocional [...] ayuda a lograr un óptimo nivel de concentración para el aprendizaje” (Bisquerra Alzina et al., 2015:195).

Con respecto a la capacidad de la MO, Cowan (2015:34) señala que un exceso de información multimodal, es decir, información que requiera de varios componentes de la MO a la vez (por ejemplo, ítems visuales y verbales) puede ir en detrimento del aprendizaje de una L2. Por tanto, el profesor debería delimitar la cantidad y tipología de lo que enseña y concienciar al alumno sobre la importancia del desarrollo de estrategias que le permitan lidiar con input multimodal.

Por último, es oportuno resaltar que los resultados de investigaciones del campo de la Lingüística Aplicada de tipo interdisciplinar arrojan indicios sobre la importancia de considerar no solo aspectos lingüísticos y cognitivos, sino también aquellos que pertenecen al terreno emocional, como lo son la IE y el vocabulario emocional.

FINANCIACIÓN

El presente estudio se enmarca en el proyecto “Emoción, memoria, identidad lingüística y aculturación emocional: Su influencia en el aprendizaje de español como lengua de migración (EMILIA)” [FFI2017-83166-C2-2-R] financiado por el Programa Estatal de investigación I+D+i Orientada a los Retos de la Sociedad del Ministerio de Economía, Industria y Competitividad..

REFERENCIAS

Abdolrezapour, P. (2013). “The relationship between emotional intelligence and EFL learners’ writing performance”. Procedia - Social and Behavioral Sciences, 70: 331-339. https://doi.org/10.1016/j.sbspro.2013.01.070 Alavi, M. y Rahimi, Z. (2011). “On the relationship between emotional intelligence and vocabulary learning among Iranian pre-university EFL learners”. Journal of English Studies, 1/3: 17-25. Alloway, T. P., Gathercole, S. E. y Pickering, S. J. (2006). “Verbal and visuospatial short-term and working memory in children: Are they separable?”. Child Development, 77/6: 1698-1716. https://doi.org/10.1111/j.1467-8624.2006.00968.x Altarriba, J. y Bauer, L. M. (2004). “The distinctiveness of emotion concepts: A comparison between emotion, abstract, and concrete words”. The American Journal of Psychology, 117/3: 389-410. https://doi.org/10.2307/4149007 Anooshian, L. J. y Hertel, P. T. (1994). “Emotionality in free recall: Language specificity in bilingual memory”. Cognition and Emotion, 8/6: 503-514. https://doi.org/10.1080/02699939408408956 Ayçiçeği-Dinn, A. y Caldwell-Harris, C. L. (2009). “Emotion-memory effects in bilingual speakers: A levels-of-processing approach”. Bilingualism: Language and Cognition, 12/3: 291-303. https://doi.org/10.1017/S1366728909990125 Baddeley, A. D. (2003). “Working memory and language: An overview”. Journal of Communication Disorders, 36/3: 189-208. https://doi.org/10.1016/S0021-9924(03)00019-4 Bar-On, R. (2006). “The Bar-On model of emotional-social intelligence (ESI)”. Psicothema, 18/Supl.: 13-25. Bisquerra Alzina, R., Pérez González, J. C. y García Navarro, E. (2015). Inteligencia emocional en educación. Madrid, España: Editorial Síntesis, S. A. Chapado, O. y López, F. (2012). “Las pruebas de expresión e interacción escritas en el DELE”. Ministerio de Educación, Cultura y Deporte. Consultado el 3 de octubre de 2017. http://www.mecd.gob.es/china/ dms/consejerias-exteriores/china/publicaciones-y-materiales/materiales-didacticos/nanjing-2012/ee-en- preparacion-dele/modelos-examen.pdf Conway, A. R. A., Kane, M. J. y Engle, R. W. (2003). “Working memory capacity and its relation to general intelligence”. Trends in Cognitive Sciences, 7/12: 547-552. https://doi.org/10.1016/j.tics.2003.10.005 Corsi, P. M. (1972). Human memory and the medial temporal region of the brain. Tesis doctoral inédita, McGill University. http://trove.nla.gov.au/work/25065049?q&versionId=209435743

RLyLA Vol. 14 (2019), 117-129 | 126 Irini Mavrou y Fernando Bustos López La influencia de la memoria operativa y de la inteligencia emocional en la expresión de la emoción en español como segunda lengua

Cowan, N. (2015). Second language use, theories of working memory and the Vennian mind. Z. (E.) Wen, M. Borges Mota y A. McNeill (eds.). Working memory in second language acquisition and processing. Bristol, UK: Multilingual Matters, 29-40. https://doi.org/10.21832/9781783093595-006 Damasio, A. R. (1994). Descartes’ error. Emotion, reason and the human brain. New York, NY: Avon Books. Damasio, A. R. (2001). “Fundamental feelings”. Nature, 4/6858: 781. https://doi.org/10.1038/35101669 Dewaele, J. M. (2008). “Dynamic emotion concepts of L2 learners and L2 users: A second language acquisition perspective”. Bilingualism: Language and Cognition, 11/2: 173-175. https://doi.org/10.1017/S1366728908003313 Dewaele, J. M. (2010a). Emotion in multiple languages. Basingstoke, UK: Palgrave Macmillan. Dewaele, J. M. (2010b). “Multilingualism and affordances: Variation in self-perceived communicative competence and communicative anxiety in French L1, L2, L3 and L4”. International Review of Applied Linguistics, 48/2-3: 105-129. https://doi.org/10.1515/iral.2010.006 Dewaele, J. M. y Pavlenko, A. (2001-2003). Web questionnaire bilingualism and emotions. London, UK: University of London. Dewaele, J. M. y Pavlenko, A. (2002). “Emotion vocabulary in interlanguage”. Language Learning, 52/2: 263-322. https://doi.org/10.1111/0023-8333.00185 Dörnyei, Z. (2010). The relationship between language aptitude and language learning motivation: Individual differences from a dynamic systems perspective. E. Macaro (ed.).Continuum companion to second language acquisition. London, UK: Continuum, 247-267. Dunsmuir, S., Kyriacou, M., Batuwitage, S., Hinson, E., Ingram, V. y O’Sullivan, S. (2014). “An evaluation of the Writing Assessment Measure (WAM) for children’s narrative writing”. Assessing Writing, 23: 1-18. https:// doi.org/10.1016/j.asw.2014.08.001 Farrand, P. y Jones, D. (1996). “Direction of report in spatial and verbal serial short-term memory”. The Quarterly Journal of Experimental Psychology, 49A/1: 140-158. https://doi.org/10.1080/713755611 Ferré, P., García, T., Fraga, I., Sánchez-Casas, R. y Molero, M. (2010). “Memory for emotional words in bilinguals: Do words have the same emotional intensity in the first and in the second language?”.Cognition and Emotion, 24/5: 760-785. https://doi.org/10.1080/02699930902985779 Fraga, I., Guasch, M., Haro, J., Padrón, I. y Ferré, P. (2018). “EmoFinder: The meeting point of Spanish emotional words”. Behavior Research Methods, 50/1: 84-93. https://doi.org/10.3758/s13428-017-1006-3 Gardner, H. (1983). Frames of mind. The theory of multiple intelligences. New York, NY: Basic Books. Goleman, D. (2006). Inteligencia emocional (63.a ed.). Barcelona, España: Editorial Kairós, S. A. Gregersen, T., MacIntyre, P. D., Finegan, K. H., Talbot, K. R. y Claman, S. L. (2014). “Examining emotional intelligence within the context of positive psychology interventions”. Studies in Second Language Learning and Teaching, 4/2: 327-353. https://doi.org/10.14746/ssllt.2014.4.2.8 Guasch, M., Ferré, P. y Fraga, I. (2016): “Spanish norms for affective and lexico- semantic variables for 1,400 words”. Behavior Research Methods, 48/4: 1358-1369. https://doi.org/10.3758/s13428-015-0684-y Gutierrez-Cobo, M. J., Cabello, R. y Fernandez-Berrocal, P. (2016). “The relationship between emotional intelligence and cool and hot cognitive processes: A systematic review”. Frontiers in Behavioral Neuroscience, 10: 1-13. https://doi.org/10.3389/fnbeh.2016.00101 Gutierrez-Cobo, M. J., Cabello, R. y Fernandez-Berrocal, P. (2017a). “Performance-based ability emotional intelligence benefits working memory capacity during performance on hot tasks”.Scientific Reports, 7/11700: 1-9. https://doi.org/10.1038/s41598-017-12000-7 Gutierrez-Cobo, M. J., Cabello, R. y Fernandez-Berrocal, P. (2017b). “The three models of emotional intelligence and performance in a hot and cool go/no-go task in undergraduate students”. Frontiers in Behavioral Neuroscience, 11: 1-13. https://doi.org/10.3389/fnbeh.2017.00033 Jiménez Catalán, R. M. y Dewaele, J. M. (2017). “Lexical availability of young Spanish EFL learners: Emotion words versus non-emotion words”. Language, Culture and Curriculum, 30/3: 283-299. https://doi.org/10.1080/07908318.2017.1327540 Jonides, J., Lacey, S. C. y Nee, D. E. (2005). “Processes of working memory in mind and brain”. Current Directions in Psychological Science, 14/1: 2-5. https://doi.org/10.1111/j.0963-7214.2005.00323.x Kane, M. J., Conway, A. R. A., Hambrick, D. Z. y Engle, R. W. (2007). Variation in working memory capacity as variation in executive attention and control. A. Conway, C. Jarrold, M. J. Kane, A. Miyake y J. Towse (eds.). Variation in working memory. Oxford, UK: Oxford University Press, 21-48. https://doi.org/10.1093/acprof:oso/9780195168648.003.0002 Kane, M. J., Hambrick, D. Z., Tuholski, S. W., Wilhelm, O., Payne, T. W. y Engle, R. W. (2004). “The generality of working memory capacity: A latent-variable approach to verbal and visuospatial memory span and reasoning”. Journal of Experimental Psychology: General, 133/2: 189-217. https://doi.org/10.1037/0096-3445.133.2.189

RLyLA Vol. 14 (2019), 117-129 | 127 Irini Mavrou y Fernando Bustos López La influencia de la memoria operativa y de la inteligencia emocional en la expresión de la emoción en español como segunda lengua

Korpi, S. y Farvardin, M. T. (2016). “Investigating the relationship between emotional intelligence and writing complexity, accuracy, and fluency among graduate students of TEFL”.Modern Journal of Language Teaching Methods, 6/1: 234-243. Kovecses, Z. (1995). Introduction: Language and emotion concepts. In J. A. Russell, J. M. Fernandez-Dols, A. S. R. Manstead y J. C. Wellenkamp (eds.). Everyday conceptions of emotion: An introduction to the psychology, anthropology and linguistics of emotion. London, UK: Kluwer Academic, 3-15. Linck, J. A., Osthus, P., Koeth, J. T. y Bunting, M. F. (2014). “Working memory and second language comprehension and production: A meta-analysis”. Psychonomic Bulletin & Review, 21/4: 861-883. https://doi.org/10.3758/s13423-013-0565-2 Lou, J. y Wu, R.-l. (2012). “The role of EQ in EFL students’ foreign language learning”. Sino-US English Teaching, 9/11: 1686-1694. Mammarella, I. C. y Cornoldi, C. (2005). “Sequence and space: The critical role of a backward spatial span in the working memory deficit of visuospatial learning disabled children”. Cognitive Neuropsychology, 22/8: 1055- 1068. https://doi.org/10.1080/02643290442000509 Mavrou, I. (2017). Memoria. N. Camuñas Sánchez-Paulete y M. Vaíllo Rodríguez (eds.). Funciones ejecutivas y aprendizaje: Experiencias en educación superior. Navarra, España: Aranzadi, 195-225. Mayer, J. D. y Salovey, P. (1997). What is emotional intelligence? P. Salovey y D. J. Sluyter (eds.). Emotional development and emotional intelligence: Educational implications. New York, NY: Basic Books, 3-31. Mayer, J. D., Salovey, P., Caruso, D. R. y Cherkasskiy, L. (2011). Emotional intelligence. R. J. Sternberg y S. B. Kaufman (eds.). The Cambridge handbook of intelligence. Cambridge, UK: Cambridge University Press, 528-549. https://doi.org/10.1017/CBO9780511977244.027 Mayer, J. D., Salovey, P., Caruso, D. R. y Sitarenios, G. (2003). “Measuring emotional intelligence with the MSCEIT V2.0”. Emotion, 3/1: 97-105. https://doi.org/10.1037/1528-3542.3.1.97 Mora, F. (2014). ¿Cómo funciona el cerebro? (3.a ed.). Madrid, España: Alianza Editorial. Mora, F. (2016). Neuroeducación. Solo se puede aprender aquello que se ama (7.a ed.). Madrid, España: Alianza Editorial. Morgado, I. (2014). Aprender, recordar y olvidar. Claves cerebrales de la memoria y la educación (2.a ed.). Barcelona, España: Ariel. Oberauer, K., Suß, H.-M., Schulze, R., Wilhelm, O. y Wittmann, W. W. (2000). “Working memory capacity – facets of a cognitive ability construct”. Personality and Individual Differences, 29/6: 1017-1045. https://doi.org/10.1016/S0191-8869(99)00251-2 Pavlenko A. (2008). “Emotion and emotion-laden words in the bilingual lexicon”. Bilingualism: Language and Cognition, 11/2: 147-164. https://doi.org/10.1017/S1366728908003283 Petrides, K. V. y Furnham, A. (2003). “Trait emotional intelligence: Behavioural validation in two studies of emotion recognition and reactivity to mood induction”. European Journal of Personality, 17/1: 39-57. https://doi.org/10.1002/per.466 Pishghadam, R. (2009). “A quantitative analysis of the relationship between emotional intelligence and foreign language learning”. Electronic Journal of Foreign Language Teaching, 6/1: 31-41. Salovey, P. y Mayer, J. D. (1990). “Emotional intelligence”. Imagination, Cognition, and Personality, 9/3: 185-211. https://doi.org/10.2190/dugg-p24e-52wk-6cdg Salovey, P., Mayer, J. D., Goldman, S. L., Turvey, C. y Palfai, T. P. (1995). Emotional attention, clarity, and repair: Exploring emotional intelligence using the Trait Meta-Mood Scale. J. W. Pennebaker (ed.). Emotion, disclosure, and health. Washington, DC: American Psychological Association, 125-154. https://doi.org/10.1037/10182-006 Shahnazari-Dorcheh, M. y Roshan, S. (2012). “Developing a non-language related span test for the use in language-specific and cross-language studies”.English Language Teaching, 5/11: 104-110. https://doi.org/10.5539/elt.v5n11p104 Shakarami, A. y Khajehei, H. (2015). “Emotional intelligence and language learning strategies interact in an EFL setting”. International Journal of Applied Linguistics and English Literature, 4/2: 229-237. https://doi.org/10.7575/aiac.ijalel.v.4n.2p.229 Shao, K., Yu, W. y Ji, Z. (2013). “The relationship between EFL students’ emotional intelligence and writing achievement”. Innovation in Language Learning and Teaching, 7/2: 107-124. https://doi.org/10.1080/1750 1229.2012.725730 Snow, R. E. (1994). Abilities in academic tasks. R. J. Sternberg y R. K. Wagner (eds.). Mind in context: Interactionist perspectives on human intelligence. Cambridge: Cambridge University Press, 3-37. Stadthagen-Gonzalez, H., Imbault, C., Perez Sanchez, M. A. y Brysbaert, M. (2017). “Norms of valence and arousal for 14,031 Spanish words”. Behavior Research Methods, 49/1: 111-123. https://doi.org/10.3758/s13428-015-0700-2

RLyLA Vol. 14 (2019), 117-129 | 128 Irini Mavrou y Fernando Bustos López La influencia de la memoria operativa y de la inteligencia emocional en la expresión de la emoción en español como segunda lengua

Stoet, G. (2010). “PsyToolkit: A software package for programming psychological experiments using Linux”. Behavior Research Methods, 42/4: 1096-1104. https://doi.org/10.3758/BRM.42.4.1096 Stroop, J. R. (1935). “Studies of interference in serial verbal reactions”. Journal of Experimental Psychology, 18/6: 643-662. https://doi.org/10.1037/h0054651 Xiu, L., Zhou, R. y Jiang, Y. (2016). “Working memory training improves emotion regulation ability: Evidence from HRV”. Physiology & Behavior, 155: 25-29. https://doi.org/10.1016/j.physbeh.2015.12.004 Zafari, M. y Biria, R. (2014). “The relationship between emotional intelligence and language learning strategy use”. Procedia - Social and Behavioral Sciences, 98: 1966-1974. https://doi.org/10.1016/j.sbspro.2014.03.630 Zarafshan, M. y Ardeshiri, M. (2012). “The relationship between emotional intelligence, language learning strategies and English proficiency among Iranian EFL university students”.Journal of Educational and Instructional Studies in the World, 2/3: 105-114. Zimmerman, C. y Kelley, C. (2010). “Effects of emotionality on memory predictions versus memory performance”. Journal of Memory and Language, 62/3: 240-253. https://doi.org/10.1016/j.jml.2009.11.004

RLyLA Vol. 14 (2019), 117-129 | 129 Revista de Lingüística y Lenguas Aplicadas Vol. 14 año 2019, 131-140 EISSN 1886-6298 https://doi.org/10.4995/rlyla.2019.11080

THE LEXICAL REPRESENTATION OF ENGLISH VERBS OF ACTION. COMPLEX PREDICATES AND STRUCTURES

Ana Elvira Ojanguren López Universidad de La Rioja, Spain

Abstract: This article aims at proposing a lexical representation for a set of English verbs of action. The analysis is carried out on the grounds of Role and Reference Grammar (RRG) and includes aspects like Aktionsart type, macrorole and syntactic function assignment, linking, as well as nexus and juncture. Against this background, the meaning components of the verbs in question are analysed, in such a way that a logical structure based on a lexical representation is defined for each verbal class. Conclusions fall on both the descriptive and the theoretical side. From the descriptive point of view, Fail and Try verbs constitute a unified verbal class as regards their meaning components and grammatical behaviour and, thus, they are represented by means of a unified logical structure. Conversely, Prevent verbs and Forbid verbs require different logical structures that account for their divergent grammatical behaviour, corresponding to the Causative Activity and Causative Achievement Aktionsart types respectively. On the theoretical side, the logical structures of End verbs, Fail verbs, Try verbs and Prevent verbs stick to the canonical representations of RRG, while those of Hinder verbs and Refrain verbs require complex predicates and complex logical structures which allow to incorporate extra meaning components and to combine different Aktionsart types.

Keywords: Lexical Representation, Logical Structures, Verb Classes, Role and Reference Grammar.

1. AIMS, SCOPE AND OUTLINE

The main aim of this article is to propose a lexical representation for some English verbs of action that includes a specification of the internal aspect of the verb and a projection of its argument structure. The discussion is based on the theory of Role of Reference Grammar (hereafter RRG; Foley & Van Valin, 1984; Van Valin & LaPolla, 1997; Van Valin, 2005, 2014), so that it is also an aim of this article to discuss the consequences of the lexical representation of these verbs for the system of Aktionsart types and logical structures posited by RRG.

The article is organised as follows. Section 2 lays the theoretical foundations of the analysis, which include the Aktionsart types, the assignment of thematic roles and semantic macroroles, as well as the nexus and juncture types of RRG. Section 3 makes some methodological remarks concerning the choice of verbs of action. Section 4 offers the lexical representations and logical structures for which no modification of the current framework of RRG is necessary. Section 5 proposes two lexical representations that deviate in some aspects from the canonical representations of RRG. Finally, section 5 draws the conclusions of the research.

2. LEXICAL REPRESENTATION AND LINKING IN RRG

This section summarises the aspects of the theory of RRG that are necessary for addressing the questions raised in this article.

RRG (Foley & Van Valin, 1984; Van Valin & LaPolla, 1997; Van Valin, 2005, 2014) is a theory of language whose main goals are interlinguistic applicability and the association between pragmatics, syntax and semantics. The

To cite this article: Ojanguren, A. E. (2019). "The lexical representation of English verbs of action. Complex predicates and structures" Revista de Lingüística y Lenguas Aplicadas, 14, 131-140. https://doi.org/10.4995/rlyla.2019.11080 Correspondence author: [email protected]

Received: 2018-11-30 Accepted: 2019-03-27 | 131 Ana Elvira Ojanguren López The lexical representation of English verbs of action. Complex predicates and structures

following areas of this theory are reviewed below: lexical representation, including verbal Aktionsart, thematic roles, and logical structures; and the assignment of semantic macroroles and Privileged Syntactic Argument (PSA), as well as the application of the different types of nexus and juncture. Some aspects of the linking between the semantic and the syntactic representation of the clause are also considered. These basic aspects of RRG are discussed in more detail in this section, which is based on the overview of the theory available from http://linguistics.buffalo.edu/people/faculty/vanvalin/rrg/RRG_overview.pdf.

The point of departure of the lexical representation in RRG is the Aktionsart class. The Aktionsart class is the internal aspect of the verb. It is different from the external aspect of the clause, which belongs to the TAM complex of operators (tense-aspect-modality) and expresses epistemic and deontic modality. The RRG typology of Aktionsart consists of four classes: State, Achievement, Accomplishment and Activity. State and Activity are the basic types. States are static, whereas activities are dynamic. The derived types include, in the first place, the Achievement and the Accomplishment. Achievements are punctual changes, whereas accomplishments are durative processes. Van Valin and LaPolla (1997:91-102) also distinguish the class of Active Accomplishments (involving the telic use of Activity verbs) and the causative versions of all Aktionsart classes. Van Valin (2005:32-34) puts forward the Semelfactive, or punctual event, both of the non-causative and the causative type. Aktionsart types are defined with respect to the set of features shown in Figure 1.

State [+static], [-dynamic], [-telic], [-punctual] Leon is a fool. Activity [-static], [+dynamic], [-telic], [-punctual] The children cried. Semelfactive [-static], [± dynamic], [-telic], [+punctual] The light flashed. Achievement [-static], [-dynamic], [+telic], [+punctual] The window shattered. Accomplishment [-static], [-dynamic], [+telic], [-punctual] The snow melted. Active accomplishment: [-static], [+dynamic], [+telic],[-punctual] Paul ran to the store Figure 1. Aktionsart or internal aspect (Van Valin, 2005:33).

The lexical representations in Figure 1 give rise to logical structures. The Aktionsart types with the associated logical structures can be seen in Figure 2. The main distinctions in logical structures arise between stative (predicate´) and dynamic (do´) Aktionsart types; and between non-causative and causative Aktionsart types. The variables x, y and z stand for verbal arguments (semantically compulsory). The labels INGR(essive), SEM(e)L(factive), BECOME and CAUSE represent, respectively, ingressives, semelfactives, accomplishments and causatives.

Aktionsart type Logical Structure STATE predicate´ (x) or (x, y) ACTIVITY do´ (x, [predicate´ (x) or (x, y)]) ACHIEVEMENT INGR predicate´ (x) or (x, y), or INGR do´ (x, [predicate´ (x) or (x, y)]) SEMELFACTIVE SEML predicate´ (x) or (x, y), or SEML do´ (x, [predicate´ (x) or (x, y)]) ACCOMPLISHMENT BECOME predicate´ (x) or (x, y), or BECOME do´ (x, [predicate´ (x) or (x, y)]) ACTIVE ACCOMPLISHMENT do´ (x, [predicate1´ (x, (y))]) & BECOME predicate2´ (z, x) or (y) CAUSATIVE α CAUSE β, where α, β are LSs of any type Figure 2. Aktionsart types and logical structures in RRG.

Given the logical structures in Figure 2, the first step towards the syntactic representation of the clause is the assignment of semantic macroroles. Macroroles are generalised semantic roles, called Actor and Undergoer, on which the semantic interpretation of verbal arguments depends. Macroroles make grammatical generalisations across the argument structure of verbs. In a transitive predication, the x argument of the verb is the Actor and the y argument bears the macrorole Undergoer. In an intransitive predication, the semantic properties of the verb determine whether the only argument functions as an Actor or as an Undergoer. The maximal number of macroroles that a verb can take is two. In ditransitives, such as like to present someone with something or to give something to someone, the third argument is called the non-macrorole oblique core argument.

The syntactic representation of the clause comprises PSA, direct core arguments and oblique core arguments. It must be noted in this respect the syntactic functions subject and object are not considered universal in RRG. For this reason, this theory posits the concept of PSA. The PSA is a construction-specific function resulting from the neutralisation of thematic roles and pragmatic functions for syntactic purposes. This means that the PSA is the controller of the agreement between the finite verb and the first argument; and/or the noun phrase around which the complexity of the construction revolves, that is to say, the noun phrase that is omitted, shared, raised etc. Any

RLyLA Vol. 14 (2019), 131-140 | 132 Ana Elvira Ojanguren López The lexical representation of English verbs of action. Complex predicates and structures

argument in the sentence different from the PSA is either a direct core argument or an oblique core argument. For instance, in someone gives something to someone else, someone gets PSA, something is a direct core argument and to someone else is analysed as an oblique core argument because it is governed by preposition. As has been said above, someone gets the macrorole Actor and something the Undergoer.

Linking is the correspondence both from semantics to syntax (production) and from syntax to semantics (comprehension). The linking syntax-semantics is governed by the Completeness Constraint, which stipulates that all the arguments in the semantic representation of the sentence must be realized in the syntax, and, conversely, that all the elements of the syntactic expression must be linked to some argument in the semantic representation in order to be interpreted. The main elements of linking are verb agreement, case assignment and prepositional government. The treatment of constructions is in accordance with these aspects.

To recapitulate, the main descriptive device of the linking between semantics and syntax is the logical structure. The logical structure is a layered representation that originates in the lexical representation of the verb and is expanded by means of the macroroles and syntactic functions. The resulting structure is a tree-diagram representation of the clause that displays the argument projection and the operator projection (semantic and morphological features like external aspect, tense, modality, etc.). All the elements involved in the realization of the arguments, like agreement, case assignment and prepositional government, together with the relevant construction, constitute the linking algorithm of RRG.

The linking algorithm operates clause by clause. The layered structure of the clause is a hierarchical structure that can be broken down into several semantic layers defined by the scope of operators (especially the TAM -tense-aspect-modality- complex; Foley & Van Valin, 1984:210). The following layers are distinguished in the layered structure of the clause: the core, including the verbal nucleus, its arguments and its argument-adjuncts, as in eat cake and go to the mountains respectively; the clause, which comprises the core and the periphery (as in play football on Sunday), and the sentence, which consists of one or more clauses, as in I always read a novel before going to bed.

The treatment of complex sentences in RRG is based on the hierarchical structure of the clause just described. RRG distinguishes the type of unit (juncture) from the type of relation (nexus) involved in the complex sentence. These aspects are considered independently. The default choice for the formation of complex sentences is the combination of nuclei with nuclei, cores with cores, clauses with clauses, and sentences with sentences. These are the differentlevels of juncture. According to the degree of complexity of the combining units, several levels of juncture arise: nuclear juncture, core juncture, clausal juncture, and sentential juncture. Nuclear junctures, to begin with, are complex constructions that contain several nuclei. For example, in John forced open the can, two nuclei, force and open, appear in a single core. Core junctures are made up of two or more cores, as in I had Fred to force the can open. In this type of juncture, a core argument is shared by two cores (the noun phrase Fred). In English, intransitive verbs only can be found in the linked predication of a nuclear juncture (Van Valin & LaPolla, 1997:445). A clause juncture can be identified in more complex structures like John met Mary yesterday and Jim met her too. Further differences between the levels of juncture are due to complementisers (to, from, etc.). Thus, nuclear junctures do not take complementisers, whereas core junctures may include them. Consequently, the two nuclei can be adjacent in a nuclear juncture, but not in a core juncture.

The syntactic and semantic relations between the units in a juncture, called nexus, are coordination, subordination and cosubordination. Subordination can belong to two types: if the subordinate clause is an argument, as in That they got married surprised everyone; and peripheral subordination, if the subordinate clause is a periphery, as in The news was everywhere before they noticed. Daughter subordination (involving linked predications as arguments of the matrix predication) and peripheral subordination (consisting of a linked predication as periphery of a matrix predication) can take place at the levels of the nucleus, the core, and the clause. However, subordination requires the possibility of clefting and passivisation. For example, Mary criticised Jim´s resigning his position is an instance of subordination because It was Jims´s resigning his position that Mary criticised (cleft) and That Jim resigned his position was criticised by Mary are possible (passive; Van Valin & LaPolla, 1997:445). The third nexus type of RRG is called cosubordination, or dependent coordination. In cosubordination, the dependence results from the scope of the operators. In other words, the units must share at least one operator at the relevant level of juncture. For example, in Jack sat drinking beer the operator of imperfect aspect has scope over both nuclei, considering that a paraphrase like Jack sat and drank beer is possible.

3. SOME METHODOLOGICAL REMARKS

The selection of verbs for this article draws on Faber and Mairal’s (1999) lexical domains of English. These authors provide a list of thirteen lexical domains, including Action. Within this domain, the following subdomains belong in a group that can be described as verbs of inaction: To not to do something; To cause somebody not to do

RLyLA Vol. 14 (2019), 131-140 | 133 Ana Elvira Ojanguren López The lexical representation of English verbs of action. Complex predicates and structures

something; To stop doing something; and To make an effort in order to be able to do something. Two subdomains not included in Faber and Mairal (1999:279-293) are taken into consideration in this work so that the group of verbs of inaction is coherent and exhaustive: To make it difficult for someone to do something; and To refrain oneself from doing something. This is shown in Figure 3.

To stop doing something [end, finish] relinquish, cease, stop, desist; To make an effort in order to be able to do something try,[ attempt] strive, struggle, endeavour; To not do something [fail, neglect] desist, omit, give up, abandon; To cause somebody not to do something [prevent] forbid, prohibit, restrain (someone from), detain, constrain, impede; To make it difficult for someone to do somethinghinder [ ] hamper, impede; To refrain oneself from doing something [refrain] restrain, abstain, forbear. Figure 3. The classes of verbs of inaction.

The subgroups of verbs of inaction in Figure 3 are considered, in line with Levin (1993), verbal classes. This means that each set constitutes a unified and consistent inventory that is justified on the basis of the semantics and the syntax of the verbs that belong to the class. As in Levin (1993:10), the semantics of a given verb determines the range of expressions in which it is found, while the syntactic configuration of the class of verbs is a criterion of class membership. Put differently, the number and form of the arguments of a verb are determined by the meaning components of the verb in question; and the verbs that belong to the same class on the grounds of their grammatical behaviour are expected to share meaning components. The grammatical behaviour of verbs in this framework include argument realization, diathesis alternations and morphological properties.

According to two authoritative sets of dictionaries (https://en.oxforddictionaries.com; http://dictionary.cambridge.org), the verbs in Figure 3 share a meaning component that can be described as the non-happening of an event, either because the action referred to by the verb ceases, as in We stopped the collaboration; or because it is never completely accomplished, as in They tried to break into the apartment. Consequently, these verbs call for a secondary predication expressing the action that finished or never occurred.

The sentences in which verbs of inaction appear show a considerable degree of semantic and syntactic integration. The integration of events is reflected in the syntax, so that these verbs take a secondary predication of the form of a linked clause with a finite form of the verb and introduced by a conjunction; or a linked core with a non-finite form of the verb (infinitive or gerund) without subordinator marker; or a gerund governed by a preposition. For instance, Prevent verbs admit two of these configurations: The law prohibits drivers from using mobile phones vs. The law prohibits that drivers use mobile phones.

Regarding the morpho-syntactic alternations presented by the different argument realizations, but for Fail verbs and Hinder verbs these classes are found in the Nominalisation alternation, which can be identified in pairs like The board attempted to approve the new policy vs. The board attempted the approval of the new policy. Of the alternations proposed by Levin (1993), Try verbs appear in the To/for alternation, thus The country struggled to survive vs. The country struggled for survival; and the Understood reflexive object alternation is found withRefrain verbs, thus He refrained from bursting out vs. He refrained himself from bursting out.

With these premises, the research method consists of three main steps. In the first place, the meaning components of the verb are analysed with a view to defining a lexical representation. Secondly, the following parts of the semantic representation and the syntactic representation are dealt with: Aktionsart type, macrorole and syntactic function assignment, nexus and juncture and linking. Finally, the layered representation (based on the lexical representation) of each verbal class is provided. Section 4 concentrates on the verbal classes whose lexical representation can be proposed without deviation from the canonical model of RRG. Section 5 proposes two modifications of the canonical model of lexical representation of RRG that are required by the data and, moreover, are compatible with other modifications that have already been proposed.

4. THE REPRESENTATION OF END VERBS, FAIL VERBS, TRY VERBS AND PREVENT VERBS

The Aktionsart class of End verbs is the Achievement, which corresponds to an ingressive and telic event. Thus, the lexical representation shows that the ongoing activity has a punctual endpoint. End verbs take one macrorole only. In an expression like John finally stopped playing the drum, the x argument plays the thematic role Effector and receives the macrorole Actor.

RLyLA Vol. 14 (2019), 131-140 | 134 Ana Elvira Ojanguren López The lexical representation of English verbs of action. Complex predicates and structures

End verbs ACHIEVEMENT INGR do´ (x, [stop´ (x, y)]) Figure 4. The logical structure of End verbs.

The y argument in the logical structure in Figure 4 is a linked predication. In an expression such as She stopped working, the juncture takes place at the nuclear level because there is no complementiser between the two nuclei and the second verb is intransitive (Van Valin & LaPolla, 1997:460). The nexus type is cosubordination given that one argument that receives the macrorole Actor is shared by the matrix predication and the linked predication. The complex sentence, therefore, is an example of nuclear cosubordination. In expressions such as She ceased to work, the juncture takes place at core level because the two nuclei are not adjacent. The nexus relation is also cosubordination because the Actor is not the controller of the pivot missing in the linked predication (or the dependent clause has its own subject, in traditional terminology. The complex construction is an instance of core cosubordination.

Try verbs and Fail verbs can be represented by means of an Accomplishment logical structure which expresses that the first participant is not successful in doing something, as in The party failed to attract new voters. Although the presupposition involved in this type of expression is that the action does not take place, there is a change because the main participant realizes that they will not be able to perform the intended action and give up. Considering that failing and trying are durative, the BECOME component of the logical structure of the Accomplishment stands for both the feature of change and the component of duration. The x argument plays the thematic role Experiencer and receives the macrorole Undergoer, while the second is typically a linked predication.

Try verbs, Fail verbs ACCOMPLISHMENT BECOME (NOT successful´ (x, y)) Figure 5. The logical structure of Fail verbs.

The logical structure in Figure 5 represents expressions like The visitors tried to take pictures of the hall. In such expressions, the juncture takes place at core level because the two nuclei are not adjacent but separated by the complementiser to. The nexus relation is cosubordination because deontic modals have scope over the two cores (as, for instance, in The visitors should not try to take pictures of the hall). The complex sentence, therefore, is an instance of core cosubordination.

Prevent verbs express an event in which someone precludes the happening of an action, in such a way that both the verb in the matrix predication and the verb in the linked predication are related to each other by causation. The Aktionsart type is the Causative Activity because the event is durative, so that both expressions in pairs like the following are possible: A software issue prevented customers from updating their profiles and A software issue was preventing customers from updating their profiles. The logical structure of the linked predication of Prevent verbs contains the lexical representation NOT do´ (y, z). The x argument of the Activity, which plays the thematic role Agent and gets the macrorole Actor, impedes that the y argument of the linked predication performs an activity. This is also the y argument and the Undergoer of the matrix clause. When it gets PSA, passives on the Undergoer result like They were prevented from joining the club.

CAUSATIVE ACTIVITY [do´ (x, [predicate´ (x, y)])] CAUSE [NOT do´ (y, [predicate´ (y, z)]] Figure 6. The logical structure of Prevent verbs.

As can be seen in the logical structure in Figure 6, the juncture takes place at core level because the two nuclei are separated by the complementiser from. The nexus type is coordination. Although the Undergoer of the matrix clause is also the Actor of the linked core (in traditional terminology, the dependent clause has its own subject), the nexus cannot be cosubordination because this type of nexus requires the same subject in the matrix clause as in the linked core; and operators with scope over the two cores. It cannot be subordination, either, because it is a requirement of subordination that the linked predication is a macrorole argument of the matrix predication. In this respect, it must be taken into account that passives like the following cannot be made: *That customers updated their profiles was prevented by a software issue/*For customers to update their profiles was prevented by a software issue. The complex sentence, consequently, is an instance of core coordination.

RLyLA Vol. 14 (2019), 131-140 | 135 Ana Elvira Ojanguren López The lexical representation of English verbs of action. Complex predicates and structures

While the logical structure in Figure 8 is valid for verbs such as prevent, it cannot be applied to verbs like forbid or prohibit. Unlike prevent, forbid and prohibit do not presuppose that the forbidden action does not take place. Thus, for instance, expressions like The directors prevented the members from objecting to the proposal, which presuppose that the action was not accomplished; as opposed to others like The directors prohibited the members to object to the proposal, which do not presuppose that the action was not accomplished. Furthermore, Prevent verbs, unlike Forbid verbs, take the complementiser from.

To represent Forbid verbs, the Causative Achievement logical structure has been selected. It describes a process initiated and terminated by someone as a result of which someone else is no longer allowed to do something. The component INGR in the logical structure in Figure 9 indicates that there is telicity and the change is punctual. This logical structure is a simplified version of the representation ofpromise (Van Valin & LaPolla 1997:551): [do´ (w, [express.(a).to.(b).in.language.(c)´ (w, x)])]. As in this segment of the logical structure of promise, the logical structure of forbid specifies that the verb is a speech verb and that the activity relates an addresser to an addressee, so that the first argument of [do´ (x, [predicate´ (x, y)])] performs the thematic role of Actor and the second argument is the Recipient. The remaining element in the complementation pattern is the Theme, which is performed by the second argument of INGR (NOT allowed´ (y, z)).

Forbid verbs CAUSATIVE ACHIEVEMENT [do´ (x, [predicate´ (x, y)])] CAUSE [INGR (NOT allowed´ (y, z))] Figure 7. The logical structure of Forbid verbs.

Both the x, the y and the z argument can be macrorole arguments and get PSA status, depending on the nexus relations and juncture levels. As regards the assignment of macrorole, this constitutes a case of competition between two arguments for receiving the status of macrorole: the Patient and the Theme.

If the x argument in the matrix clause is the PSA of the construction, an active expression results such as The directors prohibited that the members objected to the proposal. It has to be noted that this construction is infrequent in English and usually found with fixed expressions likeGod forbid that… or Heavens forbid that…. Nevertheless, the nexus relation is subordination because the linked predication is Undergoer of the matrix predication. The juncture takes place at core level because the clause is an argument of the verb. There is a complementiser and the linked predication contains a finite verb, so that the linked unit is a clause. The construction, consequently, is clausal subordination (note that the linked predication is a clause, thus the name, and that it belongs in the core of the matrix predication).

If the y argument is the PSA, the construction in point is The directors forbid the members to object. A passive can be formed on the argument y, which is shared by the matrix predication and the linked predication: The members were forbidden to object by the directors. Given that, as has been remarked above, only macrorole arguments can be PSAs (in traditional terminology, only arguments with macrorole can become subjects of the active or the corresponding passive), the Patient (the members) receives the macrorole Undergoer and then the y argument in the logical representation in Figure 7 is assigned PSA. It has to be taken into account that the y argument is shared by the matrix and the linked predication, but it is the second argument of the matrix predication (less marked as Undergoer) and the first argument of the linked predication (more marked as Undergoer). If the y argument receives the macrorole Undergoer and is the PSA, the juncture takes place at core level because the complementiser to separates the two nuclei. The nexus relation is coordination because a deontic modal operator like must does not have scope over the two cores (Van Valin & LaPolla 1997:460). Therefore, the resulting construction is an instance of core coordination.

When the z argument in the linked predication is the PSA of the construction, the resulting construction is a passive like That the members objected was prohibited by the directors. The juncture involves a clause because the linked predication is introduced by a complementiser and contains a finite verb. The linked clause is the Undergoer of the matrix clause and, consequently, the nexus relation is subordination. Put another way, the linked predication is an argument in the core of the matrix predication. The term clausal subordination makes reference to the linked predication rather than to the matrix predication, given that a clause is inserted into a core.

To summarise, this section has discussed the lexical representation of End verbs, Fail verbs, Try verbs and Prevent verbs. The similarity of the meaning components and especially the coincidence of the grammatical behaviour of Fail verbs and Try verbs advises to provide them with a unified logical structure. For the opposite reason, that is, the divergence of the grammatical behaviour of Prevent verbs with respect to Forbid verbs, two different logical structures have been proposed for these classes.

RLyLA Vol. 14 (2019), 131-140 | 136 Ana Elvira Ojanguren López The lexical representation of English verbs of action. Complex predicates and structures

5. THE REPRESENTATION OF HINDER VERBS AND REFRAIN VERBS

It has not been necessary to deviate from the canonical lexical representation of RRG so as to define the logical structures discussed in section 4. The verbs discussed in the following section require a less canonical representation, although it is convergent and compatible with similar solutions adopted within the framework of RRG. Two questions are addressed in this section, namely the representation of complex predicates, with respect to Hinder verbs; and complex logical structures, with respect to Refrain verbs.

Complex predicates may require complex representations involving predicates with further lexical decomposition. In general, RRG does not favour this solution, although lexical decomposition below the level of the predicate has already been used. The Lexematic-Functional approach has analysed several classes of English and Old English verbs, including verbs of warning (González Orta, 2002), verbs of running (Cortés Rodríguez & Torres Medina, 2003), verbs of writing (Cortés Rodríguez & Martín Díaz, 2003), verbs of smell perception and emission (González Orta, 2003), verbs of speech (González Orta, 2004), remember verbs (González Orta, 2005), verbs of sound (Cortés Rodríguez & González Orta, 2006), verbs of feeling (García Pacheco, 2013), and verbs of existence (García Pacheco, 2013); as well as some specific verbs like(ge)séon and (ge)lócian (Sosa Acevedo, 2007), and some constructions, including the resultative (González Orta, 2006) and the conative (Sosa Acevedo, 2009).

González Orta (2006) proposes a lexical template for the class of verbs of speech. A lexical template is an enriched version of the lexical representation of RRG that includes syntactic and semantic information within the same format, which is based on RRG logical structures. According to González Orta (2006), the resultative construction, which describes the state resulting from an action (Levin, 1993:101), can subsume subconstructions with verbs of speech. Verbs of speech, in this approach, code events that consist of certain subevents. With these subevents, the constructional template appears in a set of construction-based templates that represent the related constructions. This can be seen, with respect to verbs of command, in Figure 8.

Command verbs [do´ (x, [use´ (x, voice/words)] CAUSE [do´ (x, [express.instructions.(a).to.(b). in.language.(c)´ (x, y)])] CAUSE [do´ (y, z)], where y = b , z = a. Figure 8. Constructional templates with command verbs (González Orta, 2006).

The lexical decomposition of do´ as express.instructions.(a).to.(b). in.language.(c)´ explicates this predicate as the action of a an addresser (a) who gives instructions to an addressee (b) in a certain language (c). This is lexical decomposition below the level of the predicate and with a detail comparable to the description of clausal semantics. It draws on Van Valin & LaPolla (1997:551), who propose a logical structure for the verb to promise that is also based on a verb of speech decomposed lexically as express.(a).to.(b).in.language.(c)´. The logical structure in Figure 9 shows that the speaker expresses an obligation to someone and this causes that they become obligated (BECOME obligated) to do something.

Promise CAUSATIVE ACCOMPLISHMENT [do´ (w, [express.(a).to.(b).in.language.(c)´ (w, x)])] CAUSE [BECOME obligated´ (w...)] Figure 9. The logical structure of to promise.

A solution similar to the ones just reviewed, which rely on lexical decomposition below the predicate level is adopted for representing Hinder verbs. The logical structure of Hinder verbs is related to the one for Fail verbs presented in Figure 7. However, Hinder verbs are transitive. While They failed is acceptable, They hindered is not. Like Fail verbs, Hinder verbs are considered durative because they can be found in expressions such as The meeting was failing for the position of some of the delegates and The position of some of the delegates was hindering the meeting. On the other hand, Fail verbs are usually telic (the state of failure reaches its logical end) whereas Hinder verbs tend to be atelic (the hindrance can go on while the hindered action is not implemented). For these reasons, the logical structure of Hinder verbs is causative, expressing that someone or something makes an action become difficult or impossible; and atelic, meaning that the activity of hindering does not have clear-cut temporal boundaries. The Aktionsart proposed for Hinder verbs is the Causative Activity type, in such a way that the first argument of the Activity, which typically gets the thematic role Agent and receives the semantic macrorole Actor, causes the first argument of the linked clause not to do something.

RLyLA Vol. 14 (2019), 131-140 | 137 Ana Elvira Ojanguren López The lexical representation of English verbs of action. Complex predicates and structures

Hinder verbs appear in two syntactic configurations: a simplex clause and a complex sentence. In the simplex clause, the thematic role Patient is played by a noun phrase that frequently entails a predication, as in The arrest of the two women hindered the success of the talks (Two women were arrested, which hindered the success of the talks). In the complex configuration, the maximum of arguments with realization is two, but the passive is possible on the argument that receives the thematic role Patient (They were hindered from gathering by the speed of the wind). It is not possible to assign PSA to the Theme if it is a verbal predication: *That the emergency services arrived was hindered by the speed of the wind/*For the emergency services to arrive was hindered by the speed of the wind. A nominalization is necessary, in which case we cannot speak of a complex sentence: The arrival of the emergency services was hindered by the speed of the wind. In the complex configuration, the juncture takes place at core level because the two nuclei are not adjacent, but separated by a complementiser. As for the nexus relation, the dependent clause (in traditional terminology) has its own subject and, consequently, the nexus cannot be cosubordination, which requires the same subject in the matrix clause and the linked clause. Moreover, modal operators do not have scope over the two cores, which is a requisite of cosubordination. The nexus is one of coordination because the linked predication cannot become the PSA in passives based on the Theme (if such a passive was possible, the resulting nexus would be subordination). The complex sentence, then, is an instance of core coordination. This can be seen in Figure 10.

Hinder verbs CAUSATIVE ACTIVITY do´ (x, [predicate´ (x, y) CAUSE [NOT do.sucessfully´ (y, z)] Figure 10. The logical structure of Hinder verbs.

It must be taken into account that the complex predicate do.sucessfully has been incorporated into the lexical representation of Hinder verbs. This has been done in order to explain the fact that a hindered action is not an action that never takes place, but an action that does not take place as planned or that does not take place successfully enough. This means that this logical structure accounts for the unspecific version of the verb in the linked clause (it is unclear whether the action referred to by the verb took place or not); or for the imperfective version of the verb in the linked core (the action referred to by the verb may have taken place despite the difficulties). That is to say, given expressions like Huge debts hindered them from buying a house, the logical structure in Figure 10 is accurate if the reading is unspecific or imperfective. If the reading is perfective (i.e.they never bought the house), the expression is synonymous with Huge debts prevented them from buying a house and, in consequence, the logical structure of Prevent verbs (presented in Figure 6) is more accurate.

As has been said above, this section discusses complex logical structures with respect to Refrain verbs. Complex logical structures have been used so far in RRG to account for two different phenomena: the causative version of the Aktionsart types presented in Figure 2 and the Active Accomplishment or telic version of Activity verbs, as in eat pizza (Activity) vs. eat the pizza (Active Accomplishment) or run in the park (Activity) vs. run to the park (Active Accomplishment). According to Van Valin (2014), the lexical representation of accomplishments must consist an activity and a resulting state in order to account for the processual part of durative telic events. As Van Valin (2014) remarks, running, writing and eating are incremental processes measured out by the incremental path or theme. The process should belong in the Active Accomplishment logical structure, but as simultaneous with the Activity. The logical structures of verbs of creation and consumption include an incremental Theme, while the ones of verbs of motion include an incremental Path. As can be seen in Figure 11, an activity and a process result in a change of location.

Sally ran two miles to the park [do´(x, [run´ (x)]) PROC cover.path.distance´(x, (y))] & INGR be-at´(z, x) ‘x runs and simultaneously effects a process of covering a path of distancey , both of ⋀ which terminate, and this leads to the result that x is located at z´

Sally ran two miles in only twenty minutes [do´(x, [run´ (x)]) PROC cover.path.distance´(x, (y))] & INGR be-at´(path.endpoint, x) ‘x runs and simultaneously effects a process of covering distancey , both of which ⋀ terminate, and this leads to the result that x is located at the endpoint of a path of length y’ Figure 11: Incremental paths in active accomplishments of motion (Van Valin, 2014).

Two aspects of the representation in Figure 11 deserve some comment. Firstly, complex lexical representations are used to express incrementality: cover.path.distance´ and be-at´(path.endpoint, x). Secondly, the Activity is

RLyLA Vol. 14 (2019), 131-140 | 138 Ana Elvira Ojanguren López The lexical representation of English verbs of action. Complex predicates and structures

parallel to the Accomplishment, which is represented by means of the symbol . A similar solution is proposed for Refrain verbs. ⋀ Refrain verbs are different from the other verbs discussed here because they are basically stative, but also convey a meaning of action. Refrain verbs are verbs of volition that make reference to an act of will that excludes a certain action on the part of the main participant. There is no interaction with another participant. Refrain verbs are either syntactically intransitive (He refrained from taking vengeance) or reflexive (He refrained himself from taking vengeance), but cannot be used causatively: *He refrained her from taking vengeance. An argument in favour of a State Aktionsart for Refrain verbs is that they do not easily admit progressive tenses: *He was refraining himself from taking vengeance. This also happens to other verbs of volition and preference, such as to want, to like, etc.: *I am wanting..., *I am liking... Refrain verbs are semantically complex and this complexity is also present in their syntax. Refraining presupposes that the action from which someone refrains does not take place. At the same time, it is necessary to want something in order to be able to refrain from it. Furthermore, Refrain verbs are semantically transitive. That is to say, expressions like ?They refrained are possible but unacceptable without a specific context. This suggest that the logical structure of Refrain verbs is a compound one, with a first part of volition and a second part of inaction. A compound logical structure is the solution adopted given the problem of representing causativity when there is one participant only. Refrain verbs can be interpreted as verbs of obligation, in such a way that the target of obligation is oneself, thus the explicit or implicit reflexivity. In other words, the existence of a single participant excludes a causative structure, while a compound logical structure suitably displays the volition component and the inaction component when the main participant is the same.

In the stative part of the logical structure of Refrain verbs, the first argument receives the thematic role of Wanter and gets the Macrorole Undergoer. In the active part of the logical structure of these verbs, the first argument realizes the thematic role Effector and is assigned the semantic macrorole Actor. It is also the Undergoer of the stative part of the logical structure. This conflict is solved by including the negation NOT before the active predicate do´, which indicates that the Actor does not perform any action and that the Undergoer (which certainly undergoes a given state) has preference over the Actor in the interpretation. If the expression is reflexive, the first argument of the stative part of the logical structure is realized twice, but both realizations correspond to the thematic role Wanter and the semantic macrorole Undergoer. The second argument of the stative part of the logical structure of Refrain verbs is Desire. This argument does not get a semantic macrorole. For this reason, when it is realized verbally, it gives rise to a non-macrorole core linked by means of a complementiser; and if it is realized nominally, it gives rise to a non-macrorole oblique constituent governed by a preposition. When the thematic role Desire is realized by a verbal argument, the complex structure is core juncture because the two nuclei are not adjacent to each other. In traditional terminology, the dependent clause shares the subject with the main clause. This indicates that the nexus relation is cosubordination, for which it is a condition that the matrix clause and the linked core share this argument. The logical structure of Refrain verbs can be seen in Figure 12.

STATE & ACTIVITY [want´ (x, y)] [NOT do´ (x, [predicate´ (x, y)]]

Figure 12. The⋀ logical structure of Refrain verbs.

The representations in Figure 11 and Figure 12 are based on resources available from the theory of RRG itself that, furthermore, are convergent with the solutions proposed in other problematic areas like speech act verbs (to promise, for instance) and the active version of durative processes (active accomplishments). Complex predicates like do.sucessfully´ and complex logical structures containing [NOT do´ (x, [predicate´ (x, y)]] allow us, respectively, to incorporate meaning components and to represent hybrid verbs that combine two Aktionsart types. ⋀

6. CONCLUSION

This article has analysed the meaning components of a set of English verbs of action in order to propose a lexical representation for each class. Then, their grammatical behaviour has been discussed on the basis of RRG, including macrorole and syntactic function assignment, nexus and juncture and linking. The verbal classes under analysis have been represented in terms of logical structures based on different types of Aktionsart: Achievement (End verbs), Accomplishment (Try verbs and Fail verbs), Causative Activity (Prevent verbs, Hinder verbs), Causative Achievement (Forbid verbs) and State and Activity (Refrain verbs). Several syntactic constructions have been identified: cosubordination (End verbs, Try verbs and Fail verbs), subordination (Forbid verbs) and coordination (Prevent verbs). Given these results, conclusions can be drawn from the descriptive and the theoretical perspectives.

RLyLA Vol. 14 (2019), 131-140 | 139 Ana Elvira Ojanguren López The lexical representation of English verbs of action. Complex predicates and structures

From the descriptive point of view, the meaning components and the grammatical behaviour of Fail verbs and Try verbs indicate that they constitute a unified class and that, consequently, they should be represented by means of a unified logical structure. On the other hand, the divergent grammatical behaviour ofPrevent verbs with respect to Forbid verbs calls for two separate logical structures: Forbid verbs correspond to the Causative Achievement Aktionsart, whereas Prevent verbs represent the Causative Activity Aktionsart.

On the theoretical side, it has not been necessary to deviate from the canonical lexical representation of RRG to propose logical structures for End verbs, Fail verbs, Try verbs and Prevent verbs. The representation of Hinder verbs and Refrain verbs, on the other hand, needs additional resources: complex predicates for Hinder verbs, and complex logical structures for Refrain verbs. Complex predicates and complex logical structures have already been used by RRG to deal with, respectively, speech act verbs (to promise, for instance) and the active version of durative processes (active accomplishments). These resources have the advantage of allowing us to incorporate extra meaning components and to represent combined Aktionsart types. It remains for future research to determine if these procedures can be generalised in the system of RRG lexical representation.

ACKNOWLEDGEMENT

This research has been funded through the grant FFI2107-83360P, which is gratefully acknowledged.

REFERENCES

Cortés Rodríguez, F. & M. A. Martín Díaz. (2003). “The meaning-syntax interface of writing verbs: templates, constructions and linking rules within a lexical grammar of Old English verbal predicates”, Revista Canaria de Estudios Ingleses, 47: 13-35. Cortés Rodríguez, F. & M. González Orta. (2006). “Anglo-Saxon verbs of sound: Semantic architecture, lexical representation and constructions”, Studia Anglica Poznaniensia, 42: 249-284. Faber, P. & R. Mairal. (1999). Constructing a Lexicon of English Verbs. Berlin: Mouton. https://doi. org/10.1515/9783110800623 Foley, W. & R. Van Valin. (1984). Functional Syntax and Universal Grammar. Cambridge: Cambridge University Press. García Pacheco, C. L. (2013). Los verbos de sentimiento en inglés antiguo: Arquitectura léxica e interfaz semántica- gramática. PhD Dissertation. Universidad de La Laguna. García Pacheco, L. M. (2013). El dominio verbal de la existencia en anglosajón. Análisis semántico-sintáctico. PhD Dissertation. Universidad de La Laguna. González Orta, M. (2002). “Linking syntax and semantics in Old English verbs of warning”, Estudios Ingleses de la Universidad Complutense, 10: 157-182. González Orta, M. (2003). “The Old English verbs of smell perception and emission: Analysis of the interface of their semantic and syntactic representation”, SELIM, 12: 33-48. González Orta, M. (2004). Diccionario sintáctico del léxico verbal del inglés antiguo: verbos de habla, Servicio de Publicaciones de la Universidad de La Laguna. González Orta, M. (2005). “The interrelation of semantic structure and syntactic variation in Old English verb classes: Catalogue of syntactico-semantic constructions”, Revista Alicantina de Estudios Ingleses, 18: 111- 128. https://doi.org/10.14198/raei.2005.18.05 González Orta, M. (2006). “The resultative construction in Old English: Towards a semantic network of verb classes”, Studia Neophilologica, 78: 123-137. https://doi.org/10.1080/00393270601022282 Koopman, W. F. (1992). “The study of Old English syntax and the Toronto Dictionary of Old English”, Neophilologus, 76/4: 605-615. https://doi.org/10.1007/BF00209877 Levin, B. (1993). English Verb Classes and Alternations. Chicago: University of Chicago Press. Sosa Acevedo, E. (2007). “The semantic representation of Anglo-Saxon (ge)séon and (ge)lócian: syntactic evidence for meaning decomposition”, RÆL. Revista Electrónica de Lingüística Aplicada, 6: 92-107. Sosa Acevedo, E. (2009). “Lexical classes and the conative construction in Old English”, Studia Anglica Posnaniensia, 45: 69-90. https://doi.org/10.2478/v10121-009-0005-6 Van Valin, R. (2005). Exploring the Syntax-Semantics Interface. Cambridge: Cambridge University Press. https://doi.org/10.1017/CBO9780511610578 Van Valin, R. (2014). Some questions concerning accomplishments. Lecture delivered at the 2014 Symposium on Verbs, Clauses and Constructions, held at the University of La Rioja. Van Valin, R. & R. LaPolla. (1997). Syntax: Structure, meaning and function. Cambridge: Cambridge University Press. https://doi.org/10.1017/CBO9781139166799

RLyLA Vol. 14 (2019), 131-140 | 140 Revista de Lingüística y Lenguas Aplicadas Vol. 14 año 2019, 141-151 EISSN 1886-6298 https://doi.org/10.4995/rlyla.2019.9564

METAPHORS OF VICTORY AND DEFEAT IN SPORTS HEADLINES IN ENGLISH AND SPANISH

Sara Quintero Ramírez Universidad de Guadalajara, Mexico

Abstract: Metaphor is one of the most frequently used resources in the specialized language of sports (cf. Segrave, 2000; Herráez Pindado, 2004; Segura Soto, 2009; Medina Montero, 2015). The focus of this study is on how victory and defeat are expressed through metaphors in sports headlines. The data collection consists of 100 sports headlines in English and 100 in Spanish. Based on our findings, we argue that there is a diversity of metaphors that take advantage of mutual semantic fields to present victory and defeat in the two corpora, the semantic fields that were identified in the study are: a) war, b) laws, c) cleaning, d) royalty, e) life and death, f) space and g) pain. Finally, when a team’s nickname is an animal-related name, journalists map the properties of the animal onto the team (Silaški, 2009) to make the headline more attractive for the audience.

Keywords: sports headlines, metaphors, victory, defeat, semantic fields.

1. INTRODUCTION

Sports discourse is a specialized language that has precise and specific characteristics that differ from those of other specialized discourses such as in politics, religion and advertising (Groppaldi, 2009:107). Sports discourse is characterized, among other traits, by the recurrent use of rhetorical devices such as metaphors, metonymies, hyperboles and similes (cf. Nomdedeu Rull, 2004; Saiz Noeda, 2010a; Corado Valenzuela, 2011; Suárez Ramírez & Suarez Muñoz, 2016; Quintero Ramírez & Álvarez Amaral, 2016; etc.).

A large number of studies have focused on the use of metaphor in sports discourse (cf. Segrave, 2000; Herráez Pindado, 2004; Silaški, 2009; Segura Soto, 2009; Medina Montero, 2007, 2009, 2015). However, there are few studies that have centred on victory and defeat metaphors (for example, Silaški, 2009). Moreover, we have found no research focusing specifically on the metaphors of victory and defeat in sports discourse from a contrastive perspective between English and Spanish.

Therefore, the aim of this research is to determine the main linguistic characteristics of victory and defeat metaphors used in sports headlines in different newspapers published on the Web in English and Spanish. More specifically, we intend to identify similarities and differences between the metaphors used in both languages. We believe that even if each language has a particular way of presenting victory and defeat in its sports’ headlines, both languages also have some common traits. In fact, according to previous studies (Segura Soto, 2009; Saiz Noeda, 2010a; Mapelli, 2010; Medina Cano, 2010; Quintero Ramírez, Valenzuela Indart & Castañeda Hernández, 2015), one shared trait consists of considering sports events as a battle, a war, and a fight.

The results of this study characterize sports discourse. In order to reach the stated objective of the research, this paper is organized as follows. First, a literature review section is presented. In this section, three significant topics are examined, media headlines, sports discourse and metaphors in sports discourse. Next, there is a methodology section in which two main themes are explained, a) how the corpus of the study was set up, and b) the procedure that was followed in order to analyze the corpus. Then, both corpora are examined and classified into headlines that refer to a) victory and b) to defeat. Finally, conclusions are drawn.

To cite this article: Quintero Ramírez, S. (2019). "Metaphors of victory and defeat in sports headlines in English and Spanish". Revista de Lingüística y Lenguas Aplicadas, 14, 141-151. https://doi.org/10.4995/rlyla.2019.9564 Correspondence author: [email protected]

Received: 2018-02-08 Accepted: 2019-04-08 | 141 Sara Quintero Ramírez Metaphors of victory and defeat in sports headlines in English and Spanish

2. LITERATURE REVIEW

2.1. Media headlines Headlines are considered as brief texts that summarize the content of the article they precede. Runjić-Stoilova & Galić (2013:275) assert that “headlines are specific types of texts in which one or more words announce the following text. The headline takes the central place in the text: it is graphically separated from the text body and often classified in the group ofsmall texts”. Furthermore, for Pou Amérigo (2001:145), print and online headlines constitute the first contact between the reader and the news; therefore, Dor (2003:696) defines headlines as ‘negotiators’.

Ifantidou (2004:699) emphasises that headlines have two main purposes: a) to summarize the information presented below them and b) to attract the reader’s attention so that he/ she feels interested in reading the entire article. Salaverría (2005:80) states that online headlines have another function, which is a hypertextual one, since very often the headline is at the same time the hyperlink on which the readers can click in order to gain access to the whole article.

Despite these three functions, there are many readers that feel satisfied with the information they read in the headlines (Dor, 2003:695). Moreover, there are also readers that claim not to have enough time to read the full-texts in the media and prefer to only read headlines to be informed. Nevertheless, Ifantidou (2004) warns us that one of the negative consequences of simply reading a headline is the misinterpretation of the text:

Reading headlines as self-contained texts can have a twofold pragmatic effect. Firstly, headlines may receive ‘distorted’ interpretations, because unless the full-story is also read, there is always the risk of retrieving a ‘wrong’ interpretation. Secondly, headlines may not be strictly read for their newsworthiness, but for a rough approximation to information, because unless the accompanying news report is consulted, accurate newsworthy information cannot be safely obtained (Ifantidou, 2004:702).

In spite of Ifantidou’s warning, nowadays, there are more headline readers than full-text readers (Nadal Palazón: 2011:20). Consequently, in this study, we have decided to analyze newspaper headlines due to the current relevant status of this text genre and for the reason that “[n]ewspaper headlines are an interesting field to research, given the specific kind of linguistic phenomena that can be observed in this particular register” (Bucaria, 2004:280). Furthermore, our main interest does not focus on newspaper headlines in general, but on sports headlines found in the electronic version of the newspapers, in particular, those headlines that present a sports event in terms of victory or defeat through metaphorical terms.

Even though online headlines do not have the same restrictions of space as print headlines, they continue to be written in a very concise way in order to attract the reader’s attention (cf. Quintero Ramírez, 2013). However, the main difference between print and online headlines is context, a print headline is surrounded by photos and outlines that help the reader understand the setting; whereas in the online version, the headline is somewhat distant from the context because very often the reader sees the headline on a search results page or in a social media post (Lieb, 2016).

2.2. Sports discourse Sport is a cultural manifestation that possesses its own rituals (Galindo Cáceres, 2010:53). These sport rituals are closely linked to the process of communication established by the participants in the sports practices, i.e., athletes, trainers, journalists, fans, etc. Doing sport generates different communication situations such as a training session, a conversation among the players of a team, a narration of a competition, etc. The different participants in these situations develop specific terminology and expressions (Kowalikowa, 2009:64).

In sports discourse, analysts study, among other aspects, how sports events are presented by specialized journalists and how these events are understood by an audience that is interested in obtaining that kind of information. For sports fans, this kind of discourse is easy to understand. However, for a person who is not familiar with the expressions and the jargon used by the journalists, sports discourse is difficult to comprehend and sometimes even mysterious (Alasalmi, 2014:7).

Since the last century, but particularly in the last few decades, the media focuses on sports events. Society as a whole is informed about worldwide sports events through the media. Marín Montín (2000:241) claims that there is no other human activity that generates more expectancy in society than sport does. This fact means that the sports journalist is considered a link between the audience and the sports events.

Sports discourse in the media has been studied from different linguistic angles, since this specialized discourse displays a very precise way of narrating and describing sports events. According to Alvar Ezquerra (2009:159), it

RLyLA Vol. 14 (2019), 141-151 | 142 Sara Quintero Ramírez Metaphors of victory and defeat in sports headlines in English and Spanish

is interesting to study the lexicon employed in sports discourse, since journalists must have a variety of linguistic skills when producing their texts in order to attract, entertain and interest their audience in what they have to say (Curvadic García & Vargas Castro, 2010:218).

Mapelli (2004:171) asserts that there are a lot of rhetorical figures that are usually produced in sports discourse. Furthermore, these figures are not static because the sports journalists conceive new rhetorical expressions in order to create more creative texts. Indeed, the language of sports journalism is filled with metaphors, metonymies, hyperboles and similes (Quintero Ramírez, et al., 2015; Quintero Ramírez & Álvarez Amaral, 2016). As mentioned above, this study focuses on those metaphors that are used to express victory or defeat in a sports event in newspaper headlines in English and Spanish.

2.3. Metaphors in sports discourse “In the cognitive linguistic view, metaphor is defined as one conceptual domain in terms of another conceptual domain” (Kövecses, 2002:4). Corado Valenzuela (2011:28) states that a metaphor has two main parts: the tenor and the vehicle. The tenor is the referent to which features are attributed. The vehicle is the object whose features are borrowed. Lakoff & Johnson (1995:39) assert that metaphors are not exclusive to literary language, but they are very frequently used in common life situations. Actually, some examples of metaphors in the current discourse are: a) an argument is war, b) love is a journey, c) theories are buildings, d) ideas are food, etc. (Kövecses, 2002:5). Moreover, for Segrave (2000:48), life itself is seen as a sport.

Sports discourse is filled with metaphors. Certainly, these rhetorical figures are considered as a necessary linguistic tool for describing, narrating and commenting on sports because there is a strong association between sports and a diversity of semantic fields. This association permits a rich production of metaphors by sports journalists. In order to present the different situations that occur in sports events, metaphors associated with war (Kövecses, 2002), transportation, religion (Lewis, 2013), love, literature, etc., are employed (cf. Medina Montero, 2009; Segura Soto, 2009; Saiz Noeda, 2010a; Mapelli, 2010; Medina Cano, 2010; Quintero Ramírez et al., 2015).

Furthermore, Kövecses (2002: 18) claims that specialized linguistic terms produced in sports are very frequently used to explain abstract concepts. Indeed, according to Lakoff (1992), for some time, people have understood war more clearly in terms of sports, i.e., as a competitive game, since in war there must be a winner, a loser, a beginning and an end, as happens in sports events. Moreover, “[t]his metaphor is taken very seriously. There is a long tradition in the West of training military officers in team sports and chess. The military is trained to win” (Lakoff, 1992:472).

As stated before, in this study we focus on victory and defeat metaphors produced in sports newspaper headlines. According to Silaški (2009:62), sports headline authors consider the communicative situation in order to produce the metaphors for victory or defeat according to “certain (defining) properties of the characters” involved in the body of the article. Headline authors play with the nicknames of the teams to produce very creative sports headlines and make them attractive and interesting for their audience.

In her study about topic-triggered metaphors in Serbian newspaper headlines, Silaški (2009) observes that there is a strong link between sports and war, because of the main goal of many sports events which consists of “defeating the opponent”. Indeed, in some examples of her corpus victory is conceived as eating, fragmenting, taming, trapping and even killing the adversary (Silaški, 2009:63-64). These metaphors stress “the confrontational character of sport competition” (Silaški, 2009:64).

3. METHODOLOGY

3.1. Corpus formation The corpus of this study consists of 100 English-language sports headlines and 100 Spanish-language sports headlines found on different newspapers in their online versions. The headlines were published from June to August 2017. The criterion we followed to select the headlines was that the headline refers to a sports event in terms of victory or defeat. We have selected the 200 headlines randomly. This means that we did not look for the most popular or the most read newspaper. In other words, the selection was arbitrary, based only on the above- mentioned criterion.

We constructed the corpus in English from the following newspapers: The Independent, Newham Recorder and The Guardian from Great Britain; Chicago Tribune, The Register Guard, Houma Courier, Minot Daily News, The Washington Post and New York Post from U.S.A.; The Australian and The West Australian from Australia; Hindustan Times and The Tribune from India; Toronto Sun from Canada; and New Zealand Herald from New Zealand.

RLyLA Vol. 14 (2019), 141-151 | 143 Sara Quintero Ramírez Metaphors of victory and defeat in sports headlines in English and Spanish

The newspapers that were considered to build the corpus in Spanish are: La República and El Universo from Ecuador; Medio tiempo, Excélsior, El Diario de Coahuila, Esto and Récord from Mexico; Clarín, Diario Hoy and Olé from Argentina; La Nación from Chile, El Nuevo Herald from U.S.A; and Marca, El Español and El Confidencial from Spain.

We decided to constitute our corpora from the online version of newspapers, since nowadays in digital journalism it is very important to attract and keep the reader’s attention so that he/ she feels the necessity to continue reading the body of the article. Therefore, “the digital headline should be sufficiently attractive” (Suárez Ramírez & Suárez Muñoz, 2016:83) and the headline author should exploit all the creative linguistic tools at his/ her disposal.

Moreover, we built our two corpora from newspapers from different countries because our intention is to show how metaphors are used in different parts of the world in order to express victory and defeat in sports events. Therefore, we expect our results to be wide-ranging regarding the phenomenon of the metaphors used in sports headlines.

3.2. Analysis procedure For the analysis, we considered the ideas presented in the theoretical framework and the previous research on metaphors in sports discourse, in general (Lakoff, 1992; Segrave, 2000; Kövecses, 2002; Medina Montero, 2009; Mapelli, 2010; Medina Cano, 2010; Lewis, 2013; Quintero Ramírez et al., 2015), and in sports headlines, in particular (Silaški, 2009; Quintero Ramírez, 2013; Suárez Ramírez & Suárez Muñoz, 2016).

Once the two corpora were constituted, the sports headlines of each language were examined in detail. The sports headlines from both corpora were classified into headlines that refer to a) victory and b) to defeat.

First of all, victory headlines were analyzed in order to identify the metaphors used in both languages. Once we listed the metaphors of victory in each language, we detected the characteristics that are common to both languages and those characteristics that belong only to one language.

Furthermore, defeat headlines were studied in the same way. Based on these results, we established similarities and differences between the victory and defeat metaphors used by sports journalists in newspaper headlines in English and Spanish. Finally, our results are discussed in terms of current theoretical inquiry.

4. RESULTS AND DISCUSSION

In the headlines in English, there are 68 metaphors; whereas in Spanish, there are 51. The rest of the headlines present victory and defeat in no metaphorical terms. In our both corpora, the metaphors related to victory are more abundant than those related to defeat. Indeed, in English, we register 55 victory metaphors and only 13 defeat metaphors; whereas in Spanish, we find 36 victory metaphors and 15 defeat metaphors. Graphic 1 shows the frequencies of victory and defeat metaphors in our materials in English and Spanish.

80 70 60 50 40 30 20 10 0 English Spanish Defeat metaphors 13 15 Victory metaphors 55 36

Graphic 1. Frequencies of victory and defeat metaphors in both corpora.

RLyLA Vol. 14 (2019), 141-151 | 144 Sara Quintero Ramírez Metaphors of victory and defeat in sports headlines in English and Spanish

4.1. Common characteristics in English-language and Spanish-language corpora 4.1.1. Victory metaphors As it has been mentioned above, in both corpora, headlines tend to present sports news more in terms of victory than defeat. This is, to a certain point, logical, since the central character of the news is very often the winning team or athlete. As in previous studies, in our corpora, sports headlines present victory in a variety of metaphorical terms and through different linguistic forms, however most times victory is announced through a verb1 in English and Spanish.

Lakoff (1992:472) says that war is understood in terms of sports. In our corpora, the contrary is true. On the one hand, we find 26/68 (38.24%) metaphors that belong to the semantic field of war in English. On the other hand, we register 14/51 (27.45%) in Spanish. When the headline focuses on a clear victory, verbs that refer to war are used by headline writers in both languages (Alin & Tochon, 2004:43; Gómez Torrego, 2010:152).

Examples (1-6) display the verbs crush, the Spanish-language equivalent aplastar, overwhelm, the Spanish equivalent masacrar, humble and the Spanish equivalent humillar respectively to present the superiority of one protagonist over another. These metaphors can be considered cases of hyperbole too because of the exaggeration expressed through them (Gómez Torrego, 2010:157).

(1) Penguins crush Predators 6-0 to take 3-2 lead in Stanley Cup Final [Chicago Tribune, 08/06/2017]

(2) Francia aplastó a Holanda y la dejó muy complicada [Clarín, 31/08/2017]

(3) Rafael Nadal overwhelms Dominic Thiem to reach French Open final [The Guardian, 09/06/2017]

(4) Yasiel Puig masacra a Miami con su poder y los Marlins inician mal la segunda mitad [El Nuevo Herald, 14/07/2017]

(5) New Zealand humble USA to win 35th America’s Cup [The Australian, 16/06/2017]

(6) Brasil ‘B’ humilla a Australia en partido amistoso [Récord, 13/06/2017]

Furthermore, examples (7) and (8) show the verbs sink and the Spanish-language equivalent hundir correspondingly and compare a football game and a swimming competition with a fleet. These metaphors not only compare sports events to war, but more especially to a nautical battle (Mapelli, 2010:167).

(7) Football: Arsenal sink Chelsea in FA Cup final [New Zealand Herald, 28/05/2017]

(8) El ruso Chupkov hunde la flota japonesa en los 200 braza [Marca, 28/07/2017]

Furthermore, in examples (9) and (10) headlines display the verbs punish and the Spanish-language equivalent castigar to present how a team defeated the other team in a resounding way. These metaphors belong to the use of legal language (Gómez Torrego, 2010:156). We only identify 2/68 (2.94%) victory metaphors belonging to this semantic field in our materials in English and 1/51 (1.96%) metaphor in Spanish.

(9) India punish England to win by 35 runs in Women’s World Cup encounter [The Independent, 24/06/2017]

(10) Valencia Basket empata la final de la ACB y castiga a un pobre Madrid en el segundo partido [El Español, 11/06/2017]

Moreover, from the semantic field of cleaning and maintenance, in some sports such as football (11), baseball (12) and tennis (13-14) the verb sweep in English and the Spanish-language equivalent barrer2 are used to express that in a series of certain number of games (two, three or more) one team or athlete triumphs over the other one showing a huge superiority and winning the series without letting the contender(s) do anything (Saiz Noeda, 2010b:230). In English, there are 5/68 (7.35%) metaphors that belong to this semantic field; whereas in Spanish, we register 6/51 (11.76%).

(11) Mexico looks to sweep Americans [The Register Guard, 16/06/2017]

(12) Max Scherzer, Ryan Zimmerman lead Nationals to sweep of Giants [The Washington Post, 01/06/2017]

1 Although other lexical categories such as nouns and participles are not at all excluded, as is seen in some examples below. 2 In example (13), the verb barrer is followed by the preposition a; whereas in example (14) the verb is followed by the preposition con.

RLyLA Vol. 14 (2019), 141-151 | 145 Sara Quintero Ramírez Metaphors of victory and defeat in sports headlines in English and Spanish

(13) Final de campeones en Roland Garros: Nadal barrió a Thiem y Wawrinka frenó a Murray [Clarín, 09/06/2017]

(14) Nadal barre con Wawrinka y se corona campeón en Roland Garros [La Nación, 11/06/2017]

Sometimes, a specific victory means to win a championship or a tournament title. This victory is often described in terms related to royalty, so that the attainment of the title is seen as a crowning (Quintero Ramírez, 2013:178). 4/68 (5.88%) metaphors in our materials in English and 4/51 (7.84%) metaphors in our corpus in Spanish are related to royalty.

In (15) the participle crowned is employed to explain that Golden State Warriors obtained the NBA championship title. In (16) the verb coronarse is used to indicate that a Venezuelan baseball player won a title in a specific competition: The Home Run Derby. In (17) the significant actions of a player are presented through the verb lift3; moreover, the team’s attainment of the championship title is stated through the noun crown.

(15) GSW crowned NBA champions [New Zealand Herald, 13/06/2017]

(16) El venezolano Frank Díaz se corona en el Home Run Derby 2017 [Marca, 17/06/2017]

(17) Cristiano Ronaldo lifts Real Madrid to second successive Champions League crown [Hindustan Times, 04/06/2017]

As mentioned above, Segrave (2000:48) asserts that life is considered as a sport. In our study, the contrary is true, i.e., victory is conceived in terms of life. Indeed, sports journalists take their metaphors from the semantic field of life and death (Mapelli, 2010:169). Hence, the fact that a team or an athlete wins means also that this team or athlete is still alive, as in example (18) or sale vivo in a tournament, as in example (19). In English, we register 3/68 (4.41%) metaphors related to the semantic field of life and death; whereas in Spanish, we only identify 1/51 (1.96%).

(18) Egge still alive in state tennis consolation bracket [Minot Daily News, 02/06/2017]

(19) El Barça B sale vivo de Cartagonova y da un paso de gigante en la eliminatoria [Marca, 04/06/2017]

Furthermore, when a sports club’s nickname is an animal-related name, then the properties of the animal are mapped onto the team and these characteristics influence the headline author to present victory or defeat of the club in very creative terms (Silaški, 2009:62). On the one hand, there are 9/68 (13.24%) animal-related metaphors in the materials in English. On the other hand, there are 3/51 (5.88%) animal metaphors in Spanish.

In the examples below (20-23), verbs such as fly, volar, rugirand cantar are used by the headlines authors because they associate the characteristics of animals used as nicknames of the respective teams to announce victory of that specific team. In other terms, these are metonymy-based metaphors triggered by animal-related and bird-related nicknames of sports clubs (Silaški, 2009).

(20) Cardinals fly past Eagles, 48-34 [Houma Courier, 08/06/2017]

(21) Águilas FCA vuelan con campeonato [El Diario, 03/06/2017]

(22) Rugieron las Panteras [Olé, 06/06/2017]

(23) ¡Canta el Gallo! Querétaro se lleva la SuperCopa MX [Esto, 16/07/2017]

Additionally, the characteristics of the animal-related name of the team influences also the characteristics attributed to the athletes that play for that team. Therefore, a player can find his wings if he plays well for a team whose nickname is Eagles, as in example (24). The association presented in this headline is based on the metaphor victory is flying (Silaški, 2009:64).

(24) Lewis Jetta finally finds his wings as Eagles beat Geelong at home [The West Australian, 15/06/2017]

3 In accordance with conventional orientational metaphors up is good (Alin & Tochon, 2004: 43; Silaški, 2009: 64).

RLyLA Vol. 14 (2019), 141-151 | 146 Sara Quintero Ramírez Metaphors of victory and defeat in sports headlines in English and Spanish

4.1.2. Defeat metaphors As stated before, the headlines from our corpora present the sports events focusing more on victory than defeat. Nonetheless, our sports headlines present defeat in different metaphorical terms and varied lexical categories. Certainly, when the headline focuses on the defeat of a team or an athlete, there are two common verbs in both corpora: fall / caer (25-29) and suffer / sufrir (31-31). There are 6/68 (8.82%) headlines with the verb fall and 2/68 (2.94%) with the verb suffer in English; whereas there are 3/51 (5.88%) headlines with the verb caer and only 1/51 (1.96%) with the verb sufrir in Spanish.

Concerning the verbs fall / caer, “[n]egatively evaluated concepts […] are most frequently spatially oriented downward and are conceptualised as DOWN” (Silaški. 2009:65). Indeed, downward movement produces a negative idea such as sadness, unhappiness, sorrow, and, in our specific examples, defeat (Alin & Tochon, 2004:43). Consequently, the notion conveyed in headlines (25-29) is based on the metaphor defeat is a fall.

Moreover, in example (27) the metaphor alluded to is preceded by another one: Kerber loses her wings. In this specific case, the metaphor does not refer just to the defeat of a team whose nickname is a bird-related name, but also to an athlete’s defeat, more particularly, to the best tennis woman player according to the Women’s Tennis Association (WTA) rankings. Thus, the concept of going down and falling is elaborated through the metaphor of losing one’s wings to express the idea of an important athlete’s defeat.

(25) Hockey: Black Sticks fall to Belgium [New Zealand Herald, 26/06/2017]

(26) Wallabies fall to defeat as spirited Scotland gain degree of revenge [The Guardian, 17/06/2017]

(27) Kerber loses her wings, falls in first round again [The Tribune, 28/05/2017]

(28) El Athletic cae ante el Valladolid [Marca, 22/07/2017]

(29) España cae ante Inglaterra y deja en el aire su clasificación para cuartos [Marca, 23/07/2017]

Regarding the verbs suffer/ sufrir, they both display a negative meaning, since suffering refers to feel pain or undergo a punishment. This negative meaning is reinforced by the noun phrase that follows the verbs. In both examples the noun phrase represents a metaphor to refer defeat: in (30) capital punishment and in (31) descalabro.

(30) Tottenham suffer capital punishment as West Ham dethrone ‘kings of London’ [The Guardian, 06/05/2017]

(31) Chile sufre descalabro ante Rumania previo a Confederaciones [Excélsior, 13/06/2017]

Furthermore, in both languages, in order to present defeat, headline authors use the past participles of the verbs that they employ to present victory. Indeed, we observed 5/68 (7.35%) past participles in our headlines in English, and 7/51 (13.73%) in Spanish. In examples (32-36) the past participles of the verbs sweep, sink, knock out, golear, humillar, aplastar and noquear are used to show a clear defeat. In English, the first two participles, swept and sunk, are preceded by the verb get and be respectively to emphasize the passive voice; whereas in Spanish the participles are preceded by noun phrases.

In example (36), the writer not only uses the past participle of the verb aplastar, but also that of the verb noquear which is a verb employed in the discipline of boxing to indicate that a boxer has knocked out his/ her adversary. Moreover, the writer adds a metaphor that belongs to the bullfighting discipline,una cornada de varias trayectorias, to emphasize the idea of the defeat suffered by Barcelona.

(32) Giants get swept, team bus breaks down, and Madison Bumgarner isn’t there to fix any of it [The Washington Post, 08/05/2017]

(33) West Ham are sunk by dramatic late penalty at Southampton [Newham Recorder, 19/08/2017]

(34) US Open live: Konta knocked out by Krunic on day one [The Guardian, 28/08/2017]

(35) Chivas, goleado y ‘humillado’ por Santos en amistoso [Medio tiempo, 02/07/2017]

(36) Un Barcelona aplastado, noqueado y con una cornada de varias trayectorias [El Confidencial, 17/08/2017]

RLyLA Vol. 14 (2019), 141-151 | 147 Sara Quintero Ramírez Metaphors of victory and defeat in sports headlines in English and Spanish

4.2. Differences between English-language and Spanish-language corpora 4.2.1. Victory metaphors Besides characteristics that coincide in both languages, there are a few features that are slightly different between the two corpora. The main difference we observe is that in Spanish, there is a more varied list of verbs that is used to express the victory of one team or one athlete over the adversary, such as desarbolar (37), frenar (38) and asustar (39). We did not find any verbs with similar meanings in the corpus in English.

(37) El City de Guardiola desarbola a un Madrid indolente en defensa [La República, 17/07/2017]

(38) El casi cuarentón Tommy Haas frena el retorno de Federer en Stuttgart [Marca, 14/06/2017]

(39) La nueva generación alemana ‘asusta’ a los ilustres campeones del mundo [Medio tiempo, 30/06/2017]

Moreover, in English, related to the semantic field of the jewelry, journalists use the verbshine to describe a solid performance of a team whose nickname is Diamonds (40). As Silaški (2009:62), explains, in this particular case, the properties of the diamond are mapped onto the team to convey its performance in metaphorical terms. Furthermore, once again from the semantic field of the war, the home run of a baseball player is described in terms of an explosion (41). Since the explosion of a missile, a rocket or a projectile is considered lethal for the adversary, the writer uses the adjective dangerous to describe how the explosion of the alluded player makes his team lethal for its adversaries.

(40) Geelong thump GWS, Pies give Dees finals scare, Diamondsshine and more: sportwatch [The Guardian, 26/08/2017]

(41) Gary Sanchez explosion makes Yankees all the more dangerous [New York Post, 02/06/2017]

In our corpus in English, there are 6/68 (8.82%) victory metaphors that do not coincide with those in Spanish. Moreover, in our materials in Spanish, we register 7/51 (13.73%) victory metaphors that do not have a similar meaning in our headlines in English.

4.2.2. Defeat metaphors In Spanish, as a victory is often considered in terms of a crowning to express the attainment of a title or a championship, a defeat that means the fact of losing a title is also considered in terms related to royalty. In examples (42-43), the fact that tennis players Djokovic and Murray are overtaken by other players is expressed through the metaphor perder su/ la corona4.

(42) Djokovic pierde su corona ante Thiem y cae al tercer puesto del ranking mundial [Marca, 07/06/2017]

(43) Murray pierde la corona de Queen’s ante el australiano Thompson [Marca, 20/06/2017]

Furthermore, in Spanish, verbs such as hundir are preceded by the personal pronoun se to illustrate the defeat of the subject presented in first position of the headline, as in examples (44-45). The alluded verb highlights the negative idea of defeat, since it refers to a downward movement which is negatively evaluated (Silaški. 2009:65; Alin & Tochon, 2004:43).

(44) Santos y Chivas se hunden en el fondo de la general [Esto, 23/08/2017]

(45) Uruguay se hunde ante Italia en Niza y sufre su quinta derrota seguida [Marca, 07/06/2017]

As it can be seen, in our corpus in Spanish, we only identify 4/51 (7.84%) defeat metaphors that do not coincide with our defeat metaphors registered in English.

Finally, we offer a synthesis of quantitative data. Certainly, Table 1 shows the frequency of all the metaphors related to victory or defeat that we registered in both corpora. As it can be seen, the metaphors related to war are the most frequent ones in both corpora. Moreover, in English the animal-related metaphor is the second most common one; whereas in Spanish the variety of metaphorical verbs constitute the second most frequent strategy.

4 Even if we have not seen any similar example in our corpus of English, we are aware that this metaphor is not exclusive of the Spanish-language, since we have seen this metaphor in English-language in other materials.

RLyLA Vol. 14 (2019), 141-151 | 148 Sara Quintero Ramírez Metaphors of victory and defeat in sports headlines in English and Spanish

Table 1. Frequency of metaphors in both corpora.

Type of metaphor English Spanish War 26/68 38.24% 14/51 27.45% Laws 2/68 2.94 1/51 1.96% Cleaning 5/68 7.35% 6/51 11.76% Royalty 4/68 5.88% 4/51 7.84% Life 3/68 4.41% 1/51 1.96% Animal-related 9/68 13.24% 3/51 5.88% Spatial verb 6/68 8.82% 3/51 5.88% Negative-meaning verb 2/68 2.94% 1/51 1.96% Past participles 5/68 7.35% 7/51 13.73% Others 6/68 8.82% 11/51 21.57%

4.3. Discussion In the corpora of our study, victory seems to be a protagonist of the sports news, and headline authors show their creativity when they write about the triumphs of the different teams and athletes in a numerous variety of metaphorical terms, especially through the lexical category of verbs, as has been described in the above paragraphs. However, even if defeat is not usually the central character of the sports news published in the different newspapers in English and Spanish, headline authors present it in a diversity of metaphorical terms, in particular through past participles and verb phrases, as has been argued.

Although some differences between the two corpora are illustrated in the above paragraphs, the differences are minimal, since sports journalists in English and Spanish operate under the same semantic principles to express metaphors in their headlines. Indeed, many terms that refer to war are commonly used in headlines from both languages to present victory and defeat (Alin & Tochon, 2004:43; Gómez Torrego, 2010:152). Other semantic fields that are exploited by sports journalists in their headlines are: laws, cleaning and maintenance, royalty, life and death. Based on our findings, we argue that there is a diversity of metaphors that take advantage of mutual semantic fields in the two languages.

Furthermore, orientational metaphors (Alin & Tochon, 2004:43; Silaški. 2009:65) play an important role when headlines refer to defeat, since the verbs that imply a downward movement are used. Additionally, sports reporters take advantage of the teams’ nicknames. When a team’s nickname is an animal or an object-related name, journalists map the properties of the animal or the object onto the team and the headline is presented alluding these characteristics (Silaški, 2009:62). Finally, there are also similar terms that are used in the two languages to describe equivalent sports realities such as victory and defeat.

5. CONCLUSIONS

The main aim of this study was to identify how victory and defeat are expressed through metaphors in sports headlines in English and Spanish. After having studied the corpora, we have detected some relevant features that can be summarized in the following aspects. In both corpora, victory is presented in a variety of metaphorical terms and mostly through a verb such as crush / aplastar, punish / castigar, humble / humillar, sink / hundir, overwhelm / masacrar, etc. Most of the verbs belong to the semantic field of the war.

Moreover, victory metaphors expressed through verbs such as fly, volar, cantarand rugir are used when a sports club’s nickname is an animal-related name. The headline author maps the properties of the animal to express the team’s victory, as in Silaški’s (2009) study. Furthermore, when victory means to win a championship, metaphor based on royalty, i.e., crowning, is observed in both languages. Finally, in some sports, victory is considered in terms of giving life to a team or an athlete.

In the case of defeat, in both corpora, there are two frequent verbs that are used to express defeat: fall and suffer. On the one hand, the verbs fall/ caer are based on the spatial conception of down is bad; therefore, defeat is a fall. On the other hand, the verbs suffer/ sufrir denote a negative meaning which is emphasized by noun phrases that denote defeat. In both corpora, some defeats are expressed through the past participles of some of the verbs that are employed to present victory such as sunk, knocked out, noqueado, aplastado, humillado, etc.

RLyLA Vol. 14 (2019), 141-151 | 149 Sara Quintero Ramírez Metaphors of victory and defeat in sports headlines in English and Spanish

Although there are previous studies that focus on metaphors of sports discourse, those studies concentrate more particularly on football discourse (cf. Segura Soto, 2009; Saiz Noeda, 2010a; Medina Montero, 2007, 2009; among others). Moreover, the few studies that analyze the use of metaphors in sports discourse from a contrastive perspective compare Spanish and Italian (cf. Gómez Torrego, 2010; Mapelli, 2010; Medina Montero, 2015) or French and Vietnamese (cf. Xuân Thu, 2011). Therefore, the originality of this study lies in the contrastive perspective of analyzing victory and defeat newspaper headlines in English and Spanish.

Finally, the findings of this research present a broad overview of the victory and defeat metaphors that sports headlines display on different newspapers published in English and Spanish languages. We are conscious that our results are not categorical nor conclusive. Therefore, further research should be carried out to examine other rhetorical figures of sports headlines in printed and online media in English, Spanish and other languages.

REFERENCES

Alasalmi, T. (2014). Problems of translating the laws of Rugby Union from English into Finnish. M. A. Thesis. Jyväskylä: University of Jyväskylä. Alin, C. & Tochon, F. (2004). “La construction identitaire de l’enseignant: le rôle des métaphores”. Recherche & Formation, 47: 41-54. https://doi.org/10.3406/refor.2004.1926 Alvar Ezquerra, M. (2009). El deporte y los términos deportivos. In Hernán Gómez-Prieto, B. (Ed.) Il linguaggio dello sport, la comunicazione e la scuola. 149-166. Milan: Edizioni Universitarie di Lettere Economia Diritto. Bucaria, C. (2004). “Lexical and syntactic ambiguity as a source of humor: The case of newspaper”. International Journal of Humor Research, 17/3: 279–309. https://doi.org/10.1515/humr.2004.013 Corado Valenzuela, H. (2011). Análisis retórico de los signos sonoros y lingüísticos de la narración radiofónica de futbol en emisoras de frecuencia modulada. Tesis. Guatemala: Universidad de San Carlos de Guatemala. Cuvardic García, D. & Vargas Castro, E. (2010). Recursos lingüísticos en la titulación periodística costarricense: el caso de La Nación y el Diario La Extra. Filología y Lingüística, 36/1: 207-232. https://doi.org/10.15517/rfl. v36i1.1132 Dor, D. (2003). “On newspaper headlines as relevance optimizers”. Journal of Pragmatics, 35: 695-721. https://doi.org/10.1016/S0378-2166(02)00134-0 Galindo Cáceres, J. (2010). Comunicología, comunicometodología y deporte. Crítica epistemológica de las Ciencias y las ingenierías del deporte. In Martínez, S. (Coord.). Fútbol-espectáculo, Cultura y Sociedad. 53-67. Mexico City: Afínita. Gómez Torrego, L. (2010). “Aspectos semánticos del lenguaje del fútbol en España”. Monográficos MarcoELE: Lenguas de Especialidad y su Enseñanza, 11: 150-158. Groppaldi, A. (2009). Il lessico degli appassionati di pallavolo: lingua speciale e gergo per tifosi. In Hernán Gómez- Prieto, B. (Ed.). Il linguaggio dello sport, la comunicazione e la scuola. 107-120. Milan: Edizioni Universitarie di Lettere Economia Diritto. Herráez Pindado, A. J. (2004). “La metáfora en la formación del vocabulario del ciclismo en francés”. Ibérica, 7: 107-123. Ifantidou, E. (2009). “Newspaper headlines and relevance: Ad hoc concepts in ad hoc contexts”. Journal of Pragmatics, 41/4: 699–720. https://doi.org/10.1016/j.pragma.2008.10.016 Kowalikowa, J. (2009). “Language of sport in the context of communication and culture”. Studies in physical culture and tourism, 16/1: 63-67. Kövecses, Z. (2002). Metaphor. A practical introduction. Oxford: Oxford University Press. Lakoff, G. (1992). Metaphor and war: The metaphor system used to justify war in the Gulf. In Pütz, M. (Ed.)Thirty Years of Linguistic Evolution, Philadelphia/ Amsterdam: John Benjamins. https://doi.org/10.1075/z.61.36lak Lakoff, G. & Johnson, M. (1995).Metáforas de la vida cotidiana. Colección Teorema. Madrid: Cátedra. Lewis, T. V. (2013). “Religious Rhetoric in Southern College Football: New Uses for Religious Metaphors”. Southern Communication Journal, 78/3: 202-2014. https://doi.org/10.1080/1041794X.2013.770064 Lieb. T. (2016). Editing for the Digital Age. Towson: Towson University. Mapelli, G. (2004). Locuciones del lenguaje del fútbol. In Atti del XXI Congresso dell’AISPI, Letteratura della memoria. La memoria delle lingue: la didattica e lo studio delle lingue della penisola iberica in Italia. 171- 181. Messina: Lippolis. Mapelli, G. (2010). “Ámbitos semánticos del lenguaje del fútbol en Italia”. Monográficos MarcoELE: Lenguas de Especialidad y su Enseñanza, 11: 159-172. Marín Montín, J. M. (2000). “La crónica deportiva: José A. Sánchez Araujo”. Ámbitos, 5/1: 241-257. https://doi. org/10.1007/978-94-011-4265-6_20

RLyLA Vol. 14 (2019), 141-151 | 150 Sara Quintero Ramírez Metaphors of victory and defeat in sports headlines in English and Spanish

Medina Cano, F. (2010). Los narradores deportivos y sus epopeyas cotidianas. In Martínez, S. (Coord.), Fútbol- espectáculo, cultura y sociedad. 157-207. Mexico City: Afínita. Medina Montero, J. F. (2007). La metáfora en el léxico futbolístico: el caso de los participantes en español, y sus posibles equivalentes en italiano. In Luque Toro, L. (Coord.). Léxico español actual: Actas del I Congreso internacional de léxico español actual. 197-240. Venice: Università Ca’Foscari di Venezia. Medina Montero, J. F. (2009). La metáfora en el léxico futbolístico el caso de la actividad deportiva en español, y algunas propuestas de traducción al italiano. In L. Luque Toro (Coord.), Léxico español actual II. (pp. 155- 202). Venice: Università Ca’Foscari di Venezia. Medina Montero, J. F. (2015). “La metáfora en el lenguaje futbolístico: el caso del tiempo, el lugar y los útiles deportivos en español, y propuestas de traducción al italiano”. Rivista Internazionale di Tecnica della Traduzione, 17: 137-155. Nadal Palazón, J. G. (2011). El Discurso ajeno en los titulares periodísticos. PhD thesis. Salamanca: Universidad de Salamanca. Nomdedeu Rull, A. (2004). Terminología del fútbol y los diccionarios: elaboración de un diccionario de especialidad para el gran público. Tesis doctoral. Barcelona: Universidad Autónoma de Barcelona. Pou Amérigo, M. J. (2001). “Los titulares de prensa y los nuevos servicios de información por correo electrónico y teléfono móvil”. Estudios sobre el Mensaje Periodístico, 7/1): 145-157. Quintero Ramírez, S. (2013). “Análisis sintáctico de titulares deportivos en la versión electrónica de seis periódicos mexicanos”. Lengua y Habla, 17: 165-182. Quintero Ramírez, S., Valenzuela Indart, N. & Castañeda Hernández, R. C. (2015). “Figuras retóricas en crónicas futbolísticas y beisboleras”. RAEL, Revista Electrónica de Lingüística Aplicada, 14/1: 118-130. Quintero Ramírez, S. & Álvarez Amaral, D. (2016). “Figuras retóricas en notas periodísticas deportivas”. Verbum et Lingua, 8: 75-91. Runjić-Stoilova, A. & Galić, J. (2013). The representation and reception of paraphrase in newspaper headlines. In Kišiček, G. & Žagar, I. Ž. (Eds.) What do we know about the world? Rhetorical and Argumentative Perspectives. 459-482. Ljubljana: Educational Research Institute/ Digital Library. Saiz Noeda, B. (2010a). “Notas sobre la retórica del lenguaje futbolístico”. Monográficos MarcoELE: Lenguas de Especialidad y su Enseñanza, 11: 196-227. Saiz Noeda, B. (2010b). “Expresiones con valor retórico. Glosario”. Monográficos MarcoELE: Lenguas de Especialidad y su Enseñanza, 11: 228-236. Salaverría, R. (2005). Redacción periodística en Internet. Pamplona: EUNSA. Schirato, T. (2013). Sports discourse. London/ New Delhi/ New York/ Sydney: Bloomsbury Academic. Segrave, J. O. (2000). The sports metaphor in American cultural discourse. Culture, Sport, Society, 3/1: 48-60. https://doi.org/10.1080/14610980008721862 Segura Soto, G. A. (2009). “Prepara, apunta, dispara… fusila al portero, la metáfora bélica en el fútbol”. Káñina, Revista de Artes y Letras, 33: 67-74. Silaški, N. (2009). Topic-triggered metaphors in newspaper headlines. Professional communication and translation studies, 2/1-2: 59-66. Smith, A. (2015). “Figuras literarias en los titulares de la prensa costarricense”. Revista de Lenguas Modernas, 22: 35-42. https://doi.org/10.15517/rlm.v0i22.19664 Suárez-Ramírez, S. & Suárez-Muñoz, A. (2016). “La retórica del titular deportivo en la prensa española”. Documentación de las Ciencias de la Información, 39: 83-118. https://doi.org/10.5209/DCIN.54410 Xuân Thu, Phạm Thị. (2011). Étude contrastive de l’utilisation de la métaphore et de la métonymie dans le langage du sport en francais et en Vietnamien. Mémoire de Master. Da Nang: École de Langues Étrangères, Université de Danang.

RLyLA Vol. 14 (2019), 141-151 | 151 Revista de Lingüística y Lenguas Aplicadas Vol. 14 año 2019, 153-172 EISSN 1886-6298 https://doi.org/10.4995/rlyla.2019.10527

REMEDYING FOREIGN LANGUAGE ANXIETY THROUGH CLIL? A MIXED-METHODS STUDY WITH PUPILS, TEACHERS AND PARENTS

Mathea Simons Claudio Vanhees Tom Smits Karen Van De Putte University of Antwerp, Belgium

Abstract: Foreign Language Anxiety (FLA) is an important challenge for language learning. In Content and Language Integrated Learning (CLIL), subjects are taught in a language that is not the learners’ mother tongue; a context that could reduce FLA. This study analyzes whether CLIL can positively influence FLA and which characteristics determine its presence in a CLIL context. Data were collected from 225 pupils, their parents and teachers in Flemish-speaking Belgium. Quantitative and qualitative methods were applied, with pre- and post-measurement. Results indicate that CLIL can indeed positively influence FLA. Pupils experienced growth in their self-confidence to use the foreign language. Their teachers observed more active participation, especially from more silent and less proficient pupils. The parents also noticed an increase in communicative attitudes. We found eight variables to have an influence. The foreign language used in CLIL appeared to have the most important influence besides the pupils’ interest in language learning and their personality traits, extraversion and agreeableness.

Keywords: CLIL, foreign language anxiety (FLA), mixed-methods, pupils, teachers, parents.

1. INTRODUCTION

Research has shown that success (or failure) in foreign language learning in an educational context is largely related to affective factors. Foreign language anxiety (FLA) is such an affective factor and can be held responsible for its debilitating effect on language learning. Researchers and stakeholders are seeking approaches that can help reduce FLA in the classroom. Content Language Integrated Learning (CLIL) is an educational approach in which subjects are taught in a foreign language, and that could help reduce FLA. Therefore, the aim of the present study is to analyze the potential influence of CLIL on FLA and to provide more insights into characteristics that influence the presence of FLA in the CLIL classroom.

2. THEORETICAL FRAMEWORK

A large and growing body of literature has demonstrated that affective factors significantly influence the processes involved in foreign language learning. Both negative and positive learner emotions can be considered “the fuel for foreign language learning” (Dewaele et al., 2017). The domain of language learning and emotions is marked by several key concepts such as motivation, willingness to communicate, self-efficacy and (foreign) language anxiety (Simons & Smits, 2018). In the last decade we have witnessed a growing interest in the influence of positive emotions on language learning (Dewaele, 2005, 2015; Gregersen et al., 2014; 2016; MacIntyre & Gregersen, 2016; Oxford, 2015; Ross & Starcke, 2016; Seligman & Csikszentmihalyi, 2000), At the same time, we observe a renewed and global interest in the emotion and psychological construct of Foreign Language

To cite this article: Simons, M., Vanhees, C., Smits, T., and Van De Putte, K. (2019). "Remedying foreign language anxiety through CLIL? A mixed-methods study with pupils, teachers and parents". Revista de Lingüística y Lenguas Aplicadas, 14, 153-172. https://doi.org/10.4995/rlyla.2019.10527 Correspondence authors: [email protected], [email protected], [email protected], [email protected]

Received: 2018-07-13 Accepted: 2019-04-08 | 153 Mathea Simons, Claudio Vanhees, Tom Smits and Karen Van De Putte Remedying foreign language anxiety through CLIL? A mixed-methods study with pupils, teachers and parents

Anxiety (FLA) (Young, 2018). FLA is “a distinct complex of self-perceptions, beliefs, feelings, and behaviors related to classroom language learning arising from the uniqueness of the language learning process” (Horwitz et al., 1986:128). The construct emerged in the mid-1980s when it first sparked the interest of teachers and researchers, continued to do so in the 90s, and provoked a growing interest in the XXI Century (Young, 2018). The influence of FLA on language learning has been extensively studied. For reviews we refer to Dewaele and MacIntyre (2014); Gardner and MacIntyre (1993) and Horwitz (2001, 2010).

FLA has been shown to have a weakening effect on language learning (Aida, 1994; Dewaele & Al-Saraj, 2015; Horwitz, 2010, Pérez-Paredes & Martínez-Sánchez, 2000; Thomas & Sylvén, 2015), for both proficient and less proficient pupils (Gregersen & Horwitz, 2002). In some cases a negative influence on acquisition and retention (Horwitz, 2000, 2001; MacIntyre, 1995) or on motivation (MacIntyre, 2002; Yamashiro & McLaughlin, 2001) has been demonstrated as well. As, depending on the language, most studies mention occurences of 1 to 3 or 1 to 4 pupils (Simons & Decoo, 2009; Pihko, 2007), teachers and stakeholders are seeking approaches or specific contexts that reduce anxiety (Lasagabaster, 2009).

Content and Language Integrated Learning (CLIL) is an educational approach in which a subject (content) is taught and learned in a language distinct from the learners’ mother tongue, with a dual focus on content and language. Both aspects are equally important and are to be integrated in each other (Coyle, Hood, & Marsh, 2010; Strobbe et al., 2012). The concept originated in the 1990s and spread rapidly. Dalton-Puffer and Smit (2007:8) list arguments that support CLIL education: CLIL “creates conditions for naturalistic language learning, (…) it revolves around the purpose and the meaning of language use in the classroom; (…) attending CLIL classes means a substantial increase in the amount of target language exposure”. They also point to the potential positive influence on affection: reducing target language anxiety and increasing learners’ motivation. This positive impact of CLIL on affective variables is also mentioned by other researchers. Muñoz characterizes CLIL as “a relatively anxiety-free environment” (2002:36). Lasagabaster (2009) adds that besides enhancing motivation, improving language attitudes and fostering multicultural openness, CLIL may reduce anxiety. Thompson and Sylvén (2015) explain that the theoretical underpinning for lower anxiety in CLIL lies in the focus on communicating content rather than on language form.

Research into the link between CLIL and FLA is recent and scarce. Macintyre, Baker, Clement and Donovan (2003) applied a survey with a qualitative section to study the effects of language, sex and grade on FLA, willingness to communicate (WTC), and perceived communicative competence in CLIL and traditional classes in a Canadian junior high French late immersion programme.They found that students’ L2 WTC, perceived competence, and frequency of communication in French increased from grades 7 to 8 and was maintained between grades 8 and 9, despite a drop in motivation between grades 7 and 8 and a steady level of FLA across the three grades. Boys’ overall WTC and FLA levels remained constant across the three grade levels, whereas girls showed an increase in WTC and a decrease in FLA from grade 8 to 9.

Pihko (2007) investigated English Language Anxiety in CLIL and traditional classes in Finnish comprehensive schools through a survey. She found that a considerable number of pupils in both settings experienced FLA, yet CLIL pupils to a lesser extent. She doubts if effects are due to CLIL or pupils’ higher personal motivation.

Doiz, Lasagabaster and Sierra (2014) compared pupils of grades 7 and 9 in CLIL and non-CLIL in the Basque Autonomous Community in Spain. In grade 7, CLIL pupils showed a higher motivation than non-CLIL pupils, but also experienced significantly more anxiety, which might be explained by the more demanding CLIL approach. However, no significant differences in anxiety were found in grade 9, so they hypothesize that CLIL pupils became gradually accustomed to English as the subject language. Lasagabaster and Doiz (2017) subsequently studied both age groups during a longer period. Results for non-CLIL pupils did not reveal important changes despite higher anxiety levels in grade 9 compared to grade 7. CLIL pupils reported more anxiety in higher grades, leading the authors to consider age as a possible influencing variable.

Thompson and Sylvén (2015) studied Swedish CLIL and non-CLIL pupils in grade 10. CLIL pupils exhibited significantly less anxiety, increased self-confidence, more affinity with English classes, a more relaxed attitude towards English, increased confidence with native speakers and less fear of ambiguity in English. Strikingly, these significant differences were found at the start of high school, when pupils initiated CLIL, which made the authors hypothesize that the differences cannot be attributed to CLIL, but rather to pupil profiles of CLIL learners, whose lower anxiety may be innate. Möller (2016) compared different types of anxiety in CLIL and non-CLIL in German secondary education, bearing in mind the approach used in primary education. No significant differences were detected between pupils as regards exam anxiety. As for fear of success (i.e. the belief that success will lead to negative consequences such as jealousy or bullying), however, CLIL pupils exhibited significantly more anxiety, which the author assigned to the selectivity of German CLIL education. In one of the most recent studies Somers and Llinares (2018) compared motivation and feelings of anxiety in students enrolled in two bilingual tracks of differing intensity (Bilingual Education programme of the Community of Madrid). They observed significantly more anxiety in high-intensity students and considered it a wake-up call for CLIL educators who should counterbalance this with greater immersive language support.

RLyLA Vol. 14 (2019), 153-172 | 154 Mathea Simons, Claudio Vanhees, Tom Smits and Karen Van De Putte Remedying foreign language anxiety through CLIL? A mixed-methods study with pupils, teachers and parents

3. RESEARCH QUESTIONS

Results from the above studies demonstrate the relevance of exploring the phenomenon of FLA in a CLIL context more profoundly. In previous studies, FLA was generally measured at a specific moment without considering the evolution or checking the specific influence of CLIL (e.g. through pre- and post-measurement). Moreover, only pupils were investigated, while other stakeholders, namely (language and subject) teachers and parents, could also provide crucial information on its effects. Research on the subject has been mainly quantitative in nature whereas qualitative data could provide deeper insights into the phenomenon. Furthermore, no prior study has investigated the underlying factors that teachers might have an influence on. Therefore, this study set out to answer the two following research questions:

RQ1: How do pupils, their parents and teachers (CLIL teachers and language teachers), perceive the (potential) influence of CLIL on FLA?

RQ2: Which underlying characteristics (e.g. gender, home language, personality, CLIL language, motivation for CLIL) influence the presence of FLA in the CLIL classroom?

4. METHODOLOGY

4.1. Research context This study is carried out in Flanders, the Dutch-speaking part of Belgium. Flemish pupils generally start to learn French in primary education by the age of 10. At 14 they initiate English classes and subsequently add German or Spanish by the age of 15. Although French is the first foreign language in the curriculum and one of Belgium’s official languages, contact with English is much more frequent and attitudes towards it much more positive (Housen, Janssens, & Pierrard, 2002).

CLIL is relatively new in Flanders (Strobbe et al., 2012): French, English or German can be used as CLIL languages, but are also still taught as foreign languages. CLIL education encompases a maximum of 20% of the teaching time, and is generally organized for one subject per grade, which results in 1 or 2 CLIL periods a week. Flemish pupils participate voluntarily in CLIL and a parallel programme in Dutch is always offered. If pupils – even during the school year – are unwilling to continue CLIL, they switch to the parallel programme. CLIL teachers sign up voluntarily, but have to master the CLIL language at C1 level. In order to answer the research questions, a study was conducted on two campuses of a large scale secondary school (1247 students). The study was carried out across three grades of secondary education, with three subjects and two foreign languages (see Table 1):

Table 1. Context of the study Grade Age span CLIL subject CLIL language 10th 14-17 Geography English 11th 15-17 Art History English 12th 17-18 Economics French

4.2. Respondents and Research Instruments Three respondent groups participated in the study: pupils, their parents and their (CLIL and language) teachers. The mixed-method approach consisted of a quantitative (survey) as well as a qualitative (semi-structured interviews) strand involving pre- and post-measurements. All research questions were answered by the three respondent groups. The following table shows the number of respondents per group and the instruments for data collection:

Table 2. Research design. Respondent group(s) Method Pupils (N= 225) Quantitative (Survey) Pre measurement: September Post measurement: May (N= 14) Qualitative (Interview): February Teachers (N= 8) • CLIL-teachers Qualitative (Interview): February • Language teachers Qualitative (Interview): February Parents (N= 130) Quantitative (Survey) Pre measurement: September Post measurement: May

RLyLA Vol. 14 (2019), 153-172 | 155 Mathea Simons, Claudio Vanhees, Tom Smits and Karen Van De Putte Remedying foreign language anxiety through CLIL? A mixed-methods study with pupils, teachers and parents

We decided to collect both quantitative and qualitative data from the pupils. Given the small number of teachers (N=8), they were interviewed. The parents answered a survey (twice) due to their large number (N=130) and time constraints.

4.2.1. Respondent group 1 – Pupils The participating pupils first completed a web survey and then a selection of them took part in semi-structured interviews with the researchers.

a. Web survey. The web survey was administered with the EvaSys-tool at the beginning (September, pre- measurement) and the end of the school year (May, post-measurement). Thanks to teacher support, a 100% response rate was realized in both measurements and all items were completed by all pupils. However, the number of participants differs between pre- and post-measurement due to illness, pupils changing subjects or schools and other reasons. 187 pupils completed both surveys. Table 3 provides an overview in which gender, but also home languages, nationalities and academic age are taken into account.

Table 3. Profile of the pupils - Survey

Language spoken Gender at home Nationality Age Total Female Male 1 2 or more 1 2 Required age Younger Older Context N % % % % % % % % % 10th Geography - English 133 71.4 28.6 90.2 9.8 95.5 4.5 88.7 6.0 5.3

11th Art History - English 43 72.1 27.9 88.4 11.6 95.3 4.7 88.4 7.0 4.7

12th Economics - French 11 45.5 54.5 81.8 18.2 100 0.0 72.7 0 27.3

Total 187 70.1 29.9 89.3 10.7 95.7 4.3 83.3 4.3 12,4

As Table 3 shows, more girls than boys participated in the survey, 1 pupil out of 10 has more than one home language, and 4.3% more than one nationality. The overlap between nationality and other home languages is complete: pupils with another nationality also speak another home language. Most have the stipulated academic age: 4.3% are ahead of their age, 12.4% lag behind.

The web survey (see Appendix 1) consisted of (1) items to measure FLA, in order to compose the dependent variable FLA (RQ1), and (2) items to identify independent variables that possibly influence FLA (RQ2).

Foreign language anxiety as a dependent variable

To measure the dependent variable FLA, we used a translated version (Dutch) of the Foreign Language Classroom Anxiety Scale (FLCAS), which has been frequently used in empirical research (33 items, scored on a 5-point Likert scale) (Horwitz, 1986). Despite its extensive use, some researchers have criticised the validity of the scale. Sparks and Ganschow (2007) and Sparks and Patton (2013) state that FLCAS does not measure an anxiety distinct to FL learning, but rather individual differences in students’ skills and/or self-perceptions of their language learning skills. Other researchers (e.g. Skehan, 1989; Trang, 2012) observe the possibility that some FLCAS concepts (e.g. test anxiety) refer to a common type of fear, not exclusive to foreign language learning. Despite the development of other scales (e.g. MacIntyre & Gardner, 1994; Woodrow, 2006) we opted for the FLCAS because we are convinced that the items best reflect the concept we want to measure, because it is still the most widely used instrument, because it provides a relatively clear theoretical framework (Trang, 2012), and because some authors attribute differences between factor analysis results to the role of culture and ethnicity in shaping students’ beliefs (Yang, 2012). To avoid using an instrument with insufficiently proven validity, we decided to apply EFA ourselves and to retain only those items referring univoquely to the concept of FLA as defined in the first section.

Independent variables

In addition to (traditional) personal characteristics (gender, age, nationality, home language), the CLIL language, personality, motivation for CLIL, interest in the subject, interest in language learning and extracurricular language learning were used as independent variables. Personality, motivation and extracurricular language learning were measured with (validated) instruments or items.

RLyLA Vol. 14 (2019), 153-172 | 156 Mathea Simons, Claudio Vanhees, Tom Smits and Karen Van De Putte Remedying foreign language anxiety through CLIL? A mixed-methods study with pupils, teachers and parents

The pupils’ personality was determined with the Big Five, “a hierarchical model of personality traits (…) which suggests most individual differences in human personality can be classified into five broad, empirically derived domains” (Gosling et al., 2003:507): (1) Emotional stability, (2) Agreeableness, (3) Conscientiousness, (4) Extraversion and (5) Openness to experience. Each bipolar factor (e.g. Extraversion vs. Introversion) summarizes several specific facets (e.g. sociability), which subsume even more specific traits (e.g. talkative, outgoing). For reviews, see John and Srivastava (1999), and McCrae and Costa (1999). We used the short, free version of the test (20 items).

We assumed that also motivation for CLIL influences (possible) feelings of anxiety a pupil develops. The construct of motivation in foreign language learning has been extensively studied both within the SLA field and the broader educational field, with several key publications (e.g. ardner, 2001; Dörnyei & Schmidt, 2001). In this study, we decided to use a modified version of Vansteenkiste and Soenens’ (2015) scale. Four motivational types are measured, subdivided in two categories: (a) external motivation, linked to obligation, pressure and stress, labelled as external regulation (punishment, reward expectation) and introjected regulation (ego-involvement, shame, guilt), and (b) internal motivation, with a lot of psychological freedom for the individual, called autonomous motivation (personal relevance) and willingness (pleasure, passion, interest).

Besides motivation for CLIL, interest and motivation for the subject and for language learning may also play a role. Nine items measured these underlying characteristics.

b. Semi-structured interviews. In-depth midterm interviews were conducted on 14 pupils (February, 6 months of CLIL experience) to deepen the quantitative results from the web survey. Their grade, gender and CLIL language were taken into account for the selection (see Table 4): 9 girls and 5 boys were interviewed; 7 had languages as their major subject; 11 exclusively spoke Dutch, while 3 spoke an additional language. Based on the pre- measurement results, 6 pupils were affected by FLA. The interviews (Appendix 2) comprised four main questions. The first three aimed at answering and deepening answers to RQ1. The last question focused on RQ2.

Table 4. Profile of the Learners – Interview (Results section uses learners’ code names).

Languages Code Name Gender Age spoken at home Grade Language Major FLA profile CLIL language L1 Male 18 1 12th Yes French L2 Female 17 1 12th No FLA French L3 Female 17 2 12th No French L4 Male 17 1 12th No FLA French L5 Male 18 2 12th Yes French L6 Female 18 1 12th Yes FLA French L7 Male 18 1 12th No French L8 Male 18 1 12th No French L9 Female 15 1 10th Yes English L10 Female 16 1 10th No FLA English L11 Female 15 1 10th Yes English L12 Female 16 1 10th Yes English L13 Female 15 2 10th No FLA English L14 Female 16 1 10th Yes English

4.2.2. Respondent group 2 – Teachers For the full picture to emerge, the pupils’ teachers (N = 8) were equally involved in the study: five ‘regular’ language and three CLIL teachers. Table 5 gives more details on their profile. We organized the interviews (again at midterm) with all teachers from both groups, each built around four questions (Appendix 3). The first three questions focused on RQ1: how do teachers perceive FLA in regular language and CLIL classes for CLIL, but also non-CLIL pupils. Moreover, the questions focused on the evolution of anxiety after six months of CLIL education. The last question enquired about causes of language anxiety (RQ2).

RLyLA Vol. 14 (2019), 153-172 | 157 Mathea Simons, Claudio Vanhees, Tom Smits and Karen Van De Putte Remedying foreign language anxiety through CLIL? A mixed-methods study with pupils, teachers and parents

Table 5. Profile of the Teachers (Results section uses subject and language teachers’ code names).

Code Name Gender Type Mother tongue Teaching Experience CLIL Experience Ts1 Male Subject Teacher Dutch 5-10 years 1 year (Geography) Ts2 Male Subject Teacher (Art Dutch 5-10 years 2 years History) Ts3 Female Subject Teacher French 5-10 years 1 year (Economics) Tl4 Female Language teacher Dutch > 20 years 1 year (English) Tl5 Female Language teacher Dutch > 20 years 2 years (English) Tl6 Female Language teacher Dutch 5-10 years 1 year (English) Tl7 Female Language Teacher Dutch 5-10 years 2 years (English) Tl8 Male Language teacher French 1 year 1 year (French)

4.2.3. Respondent group 3 – Parents. All of the participating pupils’ parents completed a survey on paper at the beginning and the end of the school year (Appendix 4). 130 parents returned both questionnaires (69.5% response rate), displaying strong parental involvement (see Table 6).

We decided to use a limited set of open questions so as to receive as much data as possible. The first questions dealt with pupils’ attitudes towards CLIL and their parents’ perceptions. The third focused on language anxiety (RQ1) and tried to map possible causes (RQ2). The last questions enquired about comprehension during CLIL lessons, the self-confidence to participate and the difficulty of the subject matter.

Table 6. Profile of the Parents.

Parents’ Response Rate (N= 130) Parents’ Response Rate (%) Child’s CLIL subject Child’s CLIL language

94 (out of 133) 70,7% Geography English

29 (out of 43) 67,4% Art History English

7 (out of 11) 63,6% Economics French

4.3. Data Analysis 4.3.1. Web survey analysis (pupils) A first group of variables was measured with one or a very limited set of items (e.g. CLIL language, interest in the subject). In this case, data reduction was unnecessary and these characteristics were included in the analysis as separate variables. A second set was measured through a series of items, making it necessary to reduce the data through exploratory factor analysis (EFA). As recommended in the literature, we used Principal Components Analysis with Varimax-rotation (Nunnally, 1978; Kline, 1994) and only included items with loadings higher than 0.30. For a factor to be retained, its eigenvalue had to be > 1 (Kaiser, 1960) and the reliability of the new scale (Cronbach’s alpha) > 0.70. Based on the new scales, new (composed) variables were calculated.

a. Dependent variable: FLA

EFA (KMO: 0.950; Bartlett 0.000) resulted in a three factor solution, which explained 51.8% of the variance (see Table 7). All item loadings were higher than 0.30, but several items showed cross loadings on more than one component. If the difference in loadings was inferior to 0.20, the item was removed.

RLyLA Vol. 14 (2019), 153-172 | 158 Mathea Simons, Claudio Vanhees, Tom Smits and Karen Van De Putte Remedying foreign language anxiety through CLIL? A mixed-methods study with pupils, teachers and parents

Table 7. Solution EFA FLCAS.

Explained Cronbach’s Number of Items Factor variance alpha items FLA1 – Foreign language anxiety 25.0% α = 0.923 10 R1, 3, 7, 9, R11, 13, R14, R18, 23, R24 FLA2 – Didactic process anxiety 15.7% α = 0.788 6 10, 12, 15, 16, 19, 29 (evaluation, error correction) FLA3 – Negative feelings towards 11.0% α = 0.703 3 5, 6, 17 CLIL context Note: R=Reversed item

By means of sum scores, new variables were created from the items. In order to analyze the influence of independent variables (RQ2), we tried to compose one single dependent variable. Therefore, we analysed the correlations between the three newly created variables and performed an additional EFA:

N = 187 Variable ‘FLA global’ FLA1 0.810 FLA2 0.864 FLA3 0.701

As the Cronbach’s alpha of the three variables shows a global score of 0.706, we decided to create a new variable ‘FLA global’ to be used as the central dependent variable in the explanatory analysis (RQ2).

b. Independent variables

Several (independent) variables were targeted with multiple items. To verify whether the underlying constructs were measured, we conducted exploratory factor analyses (EFA).

1. Personality. The respondents’ personality was measured through 20 items. EFA (KMO: 0.716; Bartlett 0.000) indicated a solution of 5 components (see Table 8), together explaining 54.9% of the variance.

Table 8. Solution EFA Personality – Big Five.

Factor Cronbach’s alpha Number of items Items Personality 1 – Extraversion α = 0.735 3 5, 11, 19 Personality 2 – Conscientiousness α = 0.708 4 R3, 7, 12, R18 Personality 3 – Agreeableness α = 0.702 3 6, 15, 17 Personality 4 – Emotional stability α = 0.560 3 1, R2, 8 Personality 5 – Openness to experience α = 0.573 2 4, R10 Note: R = Reversed item

In view of the low validity of Personality 4 and 5 (α < 0.600), these variables were discarded from further analyses.

2. Motivation for CLIL. EFA of the motivation-related items (KMO = 0.708; Bartlett’s test 0.000) led to a two- factor solution, together explaining 60.1% of the variance (see Table 9).

Table 9. Solution EFA Motivation.

Factor Cronbach’s alpha Number of items Items Internal motivation α = 0.761 5 a, c, e, g, h External motivation α = 0.720 3 b, d, f

RLyLA Vol. 14 (2019), 153-172 | 159 Mathea Simons, Claudio Vanhees, Tom Smits and Karen Van De Putte Remedying foreign language anxiety through CLIL? A mixed-methods study with pupils, teachers and parents

3. Extracurricular language learning. This variable was measured through 7 items. EFA (KMO: 0.697; Bartlett’s test: 0.000) resulted in a two-factor solution, with 49.8% of the variance explained (see Table 10).

Table 10. Solution EFA Extracurricular language learning.

Factor Cronbach’s alpha Number of items Items

Seizing opportunities for language learning autonomously α = 0.569 3 5, 6, 7

External motivation α = 0.488 2 4, 8

Given the low reliability of the new scales (α < 0.700), we decided to leave these variables out of the subsequent analysis.

The following table gives an overview of the dependent and 11 independent variables that were taken into account in the analysis.

Table 11. Overview of dependent and independent variables.

Variable Name Number of items Cronbach’s alpha Categories Dependent variable FLA global Language Anxiety 10 0.788 Independent variable 1. Gender 1 0 = male 1 = female 2. Age 1 3. Home language 1 0 = language of the region 1 = other 4. CLIL language 1 0 = English 1 = French 5. Interest in the subject 1 0 = no interest or neutral 1 = interest 6. Interest in language 1 0 = no interest or neutral learning 1 = interest 7. Personality 1 Extraversion 3 0.735 8. Personality 2 Conscientiousness 4 0.708 9. Personality 3 Agreebleness 3 0.702 10. Motivation 1 Internal 5 0.761 11. Motivation 2 External 3 0.720

4.3.2. Interviews (pupils and teachers) and parents’ survey analysis The pupil and teacher interviews were audio-taped and transcribed verbatim, and the open questions of the parents’ interview were categorized. The first coding phase consisted of an explorative analysis of the transcripts by one researcher. Both research questions were used as a framework for the initial coding scheme, which was refined and extended based on the data. Then, the data were re-coded based on the extended framework and analyzed using NVivo10 software. Several strategies assured the validity structure (Johnson, 1997) of the qualitative data analysis. Investigator triangulation and peer review: two researchers monitored the data collection, after which one researcher performed an independent coding of the data in two steps (intra-rater reliability). Inconsistencies in coding between both phases were discussed with the other researchers. The use of low inference descriptors: the respondents’ actual language and personal meanings were described verbatim and reported with direct quotations (see Results).

The following table gives an overview of the qualitative data coding scheme, containing four main categories: the FLA currently experienced (pupils) or perceived (teachers, parents), the development of FLA, causes of FLA and causes of (language) confidence. Text fragments were coded: generally to one; to two or more when appropriate. The table shows the number of pupils and teachers (N) mentioning the subcategory as well as the number of times it was mentioned during the interviews (N of refs).

RLyLA Vol. 14 (2019), 153-172 | 160 Mathea Simons, Claudio Vanhees, Tom Smits and Karen Van De Putte Remedying foreign language anxiety through CLIL? A mixed-methods study with pupils, teachers and parents

Table 12. Coding scheme and results of qualitative data analysis.

Pupils Teachers N Pupils N of refs N Teachers N of refs Main category Subcategory (N= 14) Pupils (N= 8) Teachers Current FLA - Speaking 6 8 2 6 - Listening 5 6 1 1 - Reading 3 4 - Writing 1 1 FLA evolution - Speaking 6 12 8 22 - Listening 4 5 - Reading 2 2 - Writing 3 4 - Evolution of results 1 1 - Evolution of subject and language course 11 19 FLA causes - To feel judged 1 2 - Classmates, atmosphere 2 6 - Teacher 2 8 - Fear of errors 1 1 6 14 - Individual language skills 1 1 4 4 - LSP in the CLIL language - Personality 4 6 8 16 - Nature of the foreign language 5 8 2 2 - Type of activity, exercise, task 2 2 3 6 - Lack of preparation, time 2 2 2 2 - Main subject 2 2

5. RESULTS

This section will successively answer the research questions underlying the study. First it describes FLA in the CLIL context and how pupils, their teachers and parents perceive the (potential) influence of CLIL on FLA. Then the second research question will be addressed, in search of subjacent causes explaining differences in perceived FLA. The data from the (web) survey with the pupils (pre- and post- measurement) are used as a starting point. Subsequenty, we deepen our findings on the basis of the data from the pupil and teacher interviews and the parents’ survey. For the interpretation of the descriptive statistics we consider mean scores > 3.5 (on 5) high, mean scores below 2.5 low. In case the standard deviation (SD) is > 1, the group is regarded as heterogeneous.

5.1. The impact of CLIL on FLA (RQ1) At the start of the CLIL programme, the global FLA of the pupils (N = 225) is low to neutral, although standard deviations show important in-group differences: M = 2.68 (on a scale of 5), SD = 1.04. 24.6% of the pupils have a score of more than 3.5 (to 5) on the factor Foreign Language Anxiety. This percentage confirms earlier research. At the end of the school year, global FLA stays at the same level, although internal group differences remain important (M = 2.62; SD = 1.00).

Table 13 shows mean scores and standard deviations on the three factors distinguished in the FLA scale, based on the pre- and post-measurement, complemented with the global FLA variable. Paired samples tests were used to estimate differences between pre- and post-measurements.

Table 13. Descriptive analysis and analyses of variance.

N= 187 Mpre SDpre Mpost SDpost t df Sig 1 FL anxiety 2.67 1.03 2.62 1.00 1,15 186 0.253 2 Didactic process anxiety 2.16 0.85 2.02 0.81 2.72 186 0.007** 3 Negative feelings towards CLIL context 2.03 0.84 2.19 1.02 -2.56 186 0.011* 4 FLA global 2.41 0.84 2.32 0.81 2.45 186 0.015* * = p < 0.05; ** = p < 0.01; *** = p < 0.001

Pre- and post measurement results indicate that pupils experience low anxiety feelings, neither for the foreign language, neither for the FLT/CLIL pedagogical approach nor for the specific CLIL context (M < 2.67 of 5). The first variable ‘Foreign language anxiety’ obtains the highest score and SDs are also important here. Paired sample

RLyLA Vol. 14 (2019), 153-172 | 161 Mathea Simons, Claudio Vanhees, Tom Smits and Karen Van De Putte Remedying foreign language anxiety through CLIL? A mixed-methods study with pupils, teachers and parents

t-tests show that one year of CLIL education does not significantly decrease this variable. Nevertheless, we see (small yet) significant differences for the two other variables and for the general FLA global variable. Despite low initial anxiety feelings, from the statistically significant results we infer that CLIL education can help to reduce these feelings even more.

The variable ‘Didactic process anxiety’ shows the most significant decrease. After one year of CLIL, pupils are less anxious about pedagogical aspects of language learning: e.g. not always understanding the teacher, the teacher who corrects every error, making mistakes. Consequently, the anxiety to learn a subject in a foreign language diminishes.

We also see a low but significant increase over time for the third variable ‘Negative feelings towards CLIL context’, which stands for not feeling like going to class, doing other things, not willing to take more CLIL education (see Table 7). Hence, it shows that after one year of CLIL education, the novelty element has disappeared, the focus is (again) more on learning success and the enthusiasm to engage in CLIL has slightly decreased.

As described before, the FLCAS is subject to criticism. Therefore, we included a control item: “After one year of CLIL education, I experience less fear when I speak or write [the CLIL language]”. The pupils confirmed this (M = 3.21), which indicates an increase in self-confidence in the FL. 44% of them even obtain a score of > 4 on a total of 5, but there are important in-group differences (SD = 1.15).

46 pupils (= 24.5%) show a high score on initial language anxiety (> 3.5), 25 of them even higher than 4 on a total of 5. For this group, we separately analyzed the influence of CLIL education on feelings of anxiety (see Table 14).

Table 14. Descriptive analysis and analysis of variance of pupils with high levels of anxiety.

N= 46 Mpre SDpre Mpost SDpost t df Sig 1 FL Anxiety 4.06 0.43 3.90 0.54 1.80 45 0.078 2 Didactic process anxiety 2.80 0.92 2.67 0.74 1.21 45 0.234 3 Negative feeling towards CLIL context 2.28 1.03 2.57 1.16 -2.42 45 0.020* 4. FLA global 3.43 0.55 3.28 0.50 1.78 45 0.082 * = p < 0.05; ** = p <0.01; *** = p < 0.001

The anxiety these pupils develop towards the FL appears to have diminished after one year of CLIL education. The decrease is not significant (p = 0.078), but may indicate a tendency. However, the difference with the overall pupil group is clear: for the entire group, a decrease of only 0.05 (to 5) could be found, whereas within the group of anxious pupils, the decrease of FLA amounts to 0.16 (to 5).

For the latter group, one year of CLIL did not have an impact on their anxiety related to didactic aspects. Their negative feelings towards CLIL in general also increase slightly but significantly. Consequently, a year of CLIL education did not positively impact their motivation to enroll in (even) more CLIL classes.

When pupils reflect on language anxiety in a CLIL context, they mainly focus on speaking and listening (see Table 12). Three pupils also describe their anxiety when reading and one when writing in the FL: “I sometimes feel scared of writing in French. It’s so difficult…” (L6). Despite low anxiety feelings among most, many undergo a positive evolution: “Those two extra hours in the FL boost my confidence.”(L1). Most pupils (N= 11, out of 14) notice an improvement for the CLIL subject and language classes: “I think it’s thanks to CLIL, my friends that don’t take CLIL have more difficulties with understanding and speaking” (L3). Almost half of them mention progress for speaking. Most do not notice substantial differences, but three explicitly mention less stress. Their writing skills also evolve, especially the use of specific FL jargon: “I’m more confident to write specific words as I’ve seen them written on the blackboard during CLIL, even without knowing their meaning...” (L14).

When (language and CLIL) teachers mention anxiety in a CLIL context, they mainly refer to oral skills (see Table 12) – to no surprise. The occurrence of FLA has mainly been described for productive skills, especially for speaking (e.g. Aida, 1994). Both CLIL and language teachers observe a rise in self-confidence: I“ notice most pupils are less afraid to express themselves” (Tl7). One also refers to FLA in listening: at the beginning of CLIL the pupils were keyed up when listening to a text, but now he observes more self-confidence. Teachers perceive a positive evolution in pupils’ oral skills, especially in quiet or weaker pupils: “Some pupils, particularly those initially afraid to speak, show less fear and more fluency when given a turn” (Tl7), or as Ts1 states: “One boy with particular difficulties in English still doesn’t talk much, but says he now enjoys discussions because he understands more thanks to CLIL”.

RLyLA Vol. 14 (2019), 153-172 | 162 Mathea Simons, Claudio Vanhees, Tom Smits and Karen Van De Putte Remedying foreign language anxiety through CLIL? A mixed-methods study with pupils, teachers and parents

A vast majority of the parents (N = 94; 72.3%) state their child is enthusiastic about CLIL, whereas 10% (N = 13) say they regret the choice. One out of ten (n = 15) notices their child experiences stress. No other attitudes or feelings are mentioned.

Almost half of the parents never hear about CLIL (46.9%), which is possibly connected with the pupils’ age (puberty). If so, pupils often comment positively (33.8%): they learn a lot, it goes ‘amazingly’ well, and it is fun to learn a language in another way. Some mention quieter lessons and more concentration. Negative feedback (13%) is often about feelings of frustration: “She doesn’t understand enough, which makes her feel bad”, or refers to additional efforts and pressure due to CLIL.

Almost half of them (48 %) think their children feel more at ease in the CLIL language, and show more self- confidence and understanding. Most parents (84%) perceive a positive influence of CLIL on asking questions in class, specifically due to more exemplary questions and the lack of negative comments. 9 out of 10 positively evaluate the comprehension of the subject matter.

Yet not all are convinced. An important group (40.1%) does not see a noticeable improvement on (self-) confidence, largely because their child already felt self-confident in the FL:She “ already dared to express herself quite well”. Others think one or two CLIL periods a week are insufficient for an effect. Almost half of those who do not see any improvement, still mention a considerable fear to communicate: “She still doesn’t speak spontaneously because she thinks her English is not good enough”.

5.2. Elements influencing FLA in a CLIL context (RQ2) Based on exploratory factor analysis (EFA) we could discern one dependent variable, ‘FLA global’, and eleven independent variables (see Table 11). These independent variables were incorporated in the analysis and their influence on the dependent variable was examined by multiple linear regression analyses. We ascertained its significance (significant limit p < 0.05) and examined the effect size using the Beta coefficient, that indicates which share of the total variance within the dependent variable can be attributed to the independent variable (i.e. contribution on a total of 5). It also marks the direction of the effect on language anxiety (positive or negative).

A model composed of all independent variables explains FLA significantly (F(11,186) = 7.08, p < 0.001, R² = 0.27). It accounts for 27% of the variance within the dependent variable. The collinearity between independent variables is correct (VIF < 1.93; Tolerance > 0.56). Table 15 shows the influence of the independent variables on FLA in CLIL.

Table 15. The influence of independent variables on the dependent variable FLA.

FLA – Foreign Language Anxiety Independent variables (dependent variable) β T p 1. Gender (1 = female) 0.242 1.97 0.050* 2. Age -0.272 -2.77 0.006** 3. Home language (1= other) -0.061 -0.370 0.712 4. CLIL language (1= French) 0.895 2.99 0.003** 5. Interest in the subject (1= is interested) 0.058 0.493 0.623 6. Interest in language learning (1= is interested) -0.373 -3.22 0.002** 7. Personality 1 – Extraversion 0.276 4.99 0.000*** 8. Personality 2 – Conscientiousness 0.045 0.499 0.618 9. Personality 3 – Agreeableness 0.177 2.03 0.044* 10. Motivation 1 – Internal -0.149 -2.25 0.025* 11. Motivation 2 – External 0.163 3.13 0.002** * = p < 0.05; ** = p < 0.01; *** = p < 0.001

FLA in CLIL is significantly influenced by eight variables, namely: Gender, Age, CLIL language, Interest in language learning, Extraversion, Agreeableness, Internal and External Motivation. The variable CLIL language has the largest Beta coefficient (β =0.895). Thus, the anxiety feelings a pupil develops, depend mostly on the CLIL instructional language, which in our analysis constitutes CLIL education in French.

Interest in language learning has a β-value of -0.373. Pupils that find the language course interesting, will experience less FLA in the CLIL context. Furthermore, we observe three variables with similar β-values.

RLyLA Vol. 14 (2019), 153-172 | 163 Mathea Simons, Claudio Vanhees, Tom Smits and Karen Van De Putte Remedying foreign language anxiety through CLIL? A mixed-methods study with pupils, teachers and parents

Personality-Extraversion has a Beta coefficient of β = 0.276 and displays the highest significance (p < 0.001). Its direction may, however, surprise: the more pupils believe to be extravert, the more FLA they experience. Age (β = -0.272) and Gender (girl) (β = 0.242) also exert a significant influence. Female pupils experience more FLA, as Donovan and MacIntyre (2004) also observed, whereas FLA is inversely proportional to pupils’ age (15-18 years). Finally, three variables have a smaller, yet significant impact, the first of which isPersonality - Agreeableness (β = 0.177). The more a pupil’s susceptibility to interests of others, the higher anxiety may be. This constitutes the drive experienced to live up to the expectations of others (teacher, peers). If pupils are motivated by external factors (parents’ expectations, obligation), FLA increases (β = 0.164), while it decreases when they choose CLIL based on internal motivation (β = -0.149).

The qualitative data on the pupils confirm this analysis. The participants mention seven causes for FLA (see Table 12). In CLIL, the language itself is the main cause for FLA: “I’m not afraid, but if geography were taught in another language I wouldn’t feel at ease” (L11). Pupils also believe personality traits influence FLA: “Well, X never talks. But that’s his character” (L14). They also mention 15 elements that could reduce FLA in CLIL (see Table 12) and teachers are clearly key. They should encourage the use of FL (L13: “Don’t be afraid to speak English in class. When you need help, I’ll be there”), allow mistakes and correct errors discreetly (L4: “He corrects our mistakes without subtracting points”), be available for questions and offer supporting materials, use the mother tongue adequately and translate if needed, but also separate subject and language contents in evaluation (L8: “We have to study vocabulary and course units, but she separates them well”), and give clear instructions and feedback. Besides teachers, 11 out of 14 pupils signal the importance of motivation for CLIL. In interesting lessons the will to communicate increases: “When it’s interesting, you naturally participate more” (L1). The influence of peers and class atmosphere on well-being and self-confidence is mentioned by the same number of pupils. Speaking in a tolerant group, helps to reduce FLA in CLIL considerably: “There’s an open atmosphere in class. Mistakes are not an issue” (L12). Pupils seem to be concerned about their classmates’ opinion: “If they laughed, I wouldn’t dare to raise my hand again” (L11).

Teachers indicate nine possible FLA causes. All (N = 8) consider pupil personality the main cause: “A few perfectionist pupils don’t express themselves easily” (Tl6). Another important cause is the fear to make mistakes: “Some pupils are just afraid to commit errors” (Tl8). All teachers point to group atmosphere and peers as key elements. FLA diminishes in a safe environment: “When pupils feel comfortable, they express themselves more easily” (Ts2). Six also emphasize the importance of adequate feedback and evaluation: “Feedback should focus on what they’ve just learned and they shouldn’t be interrupted. Otherwise they won’t dare to speak anymore” (Tl7). They also notice their own impact on FLA: “We must encourage them and give everyone a turn as often as possible. We create the learning environment” (Ts1).

Most parents are positive about the influence of CLIL on self-confidence and (possible) feelings of anxiety. Parents notice a greater WTC at home as well as abroad (e.g. during holidays). Only a small group of parents (14%) answer the open question and provide possible causes for their child’s increased self-confidence. Most of them (n = 12) refer to the teacher’s proficiency and pronunciation in the CLIL language as key elements for success. At the same time, they are convinced that the teacher’s attitude and pedagogical approach play a role in the emotions pupils develop towards the FL: “He corrects mistakes, but doesn’t include them in the evaluation”. Some parents (n = 9) attribute increased self-confidence simply to more frequent contact with the FL: “The more you engage with the language, the more you dare to use it”, or to larger vocabulary: “He’s more willing to express himself because he knows more words”. When their child experiences anxiety, parents mention the fear of making mistakes (n = 4) and peer pressure in class (n = 3): “My daughter experiences stress while answering, and is always a bit afraid to make mistakes”; “She feels as if other pupils might laugh at her when she makes a mistake”.

6. CONCLUSION AND DISCUSSION

As in CLIL education more attention is paid to content (subject) than to language, researchers suspect that this approach might positively influence FLA. In the present study we used a mixed-methods approach to operationalize this hypothesis in a sample of 225 pupils, their parents (N = 130) and their teachers (N = 8; CLIL and regular teachers) to ascertain if CLIL affects FLA and what aspects impact on its occurrence.

Most pupils exhibited little anxiety at the onset of their CLIL programme. In spite of this low initial FLA, it diminished even further – thanks to the CLIL experience –, particularly the fear of pedagogical-didactic aspects of the approach (e.g. fear of mistakes and study failure). Yet a quarter of the pupils suffered from high initial FLA. This number corroborates earlier research on FLA incidence (Simons & Decoo, 2009; Pihko, 2007), also in CLIL settings (Thompson & Lee, 2013). Among the latter feelings of anxiety equally decreased, albeit not significantly. Hence, FLA appears to be relatively stable. CLIL education can reduce FLA, but from this study its potential only appears as a tendency without statistical significance. Qualitative data revealed that pupils experienced growing confidence to use the FL, both in CLIL and language classes, which was confirmed by teachers of both courses.

RLyLA Vol. 14 (2019), 153-172 | 164 Mathea Simons, Claudio Vanhees, Tom Smits and Karen Van De Putte Remedying foreign language anxiety through CLIL? A mixed-methods study with pupils, teachers and parents

Despite its low perceptibility, they noticed that (especially weaker and quieter) pupils more often dared to speak in class. Parents equally commented (relatively) positively on the decrease in communicative timidity. Nearly half of them reported that their child felt more comfortable in the CLIL language and dared to use it more.

An explanatory model composed of 11 variables allowed us to pinpoint eight variables that significantly account for FLA in a CLIL context. The language CLIL is taught in seems to have the biggest impact. English and French were the instruction languages in this study. Pupils taking French CLIL classes risk seeing FLA rise with 0.9 to 5 compared to pupils taking English CLIL classes. The qualitative data underpin this observation. The morphological complexity of a language and its linguistic familiarity might be contributing factors to FLA. This result confirms previous research by Dewaele (2010), who found that in a FL from a familiar linguistic family anxiety tends to be significantly lower. In this study the pupils’ mother tongue was Dutch, a Germanic language such as English, which made English CLIL classes less anxiety provoking. Given this fact, stakeholders might consider initially organizing CLIL classes in an instruction language linguistically closer to the pupils’ mother tongue and gradually moving towards less familiar languages.

A second variable with a considerable effect is pupil interest in the CLIL language. If present, it causes FLA to subside. It constitutes a more global attitude of pupils towards the FL: if they are interested in the CLIL language (i.e. read in the FL at home, try to use it when the opportunity arises, watch series in the FL etc.), anxiety feelings decrease.

Furthermore, pupil personality (i.e. extraversion and agreeableness) turned out to be of influence. Rather surprisingly, more extraverted pupils displayed higher FLA levels, which might be explained by Baker and MacIntyre’s (2003) findings that immersion pupils’ communication anxiety correlated most strongly with WTC. FLA reduces with age or combined with internal motivation, whereas external motivation (obligation, expectations) causes it to increase. Female CLIL pupils are more susceptible to FLA than males. The qualitative data on pupils and teachers confirm the effects of CLIL language and personal traits. As regards FLA and developing language confidence, all pupils mention the teacher, whose deliberate choice of a didactic-pedagogical approach and sympathetic attitude benefit pupils’ assuredness. Peer group and class atmosphere also affect communicative courage, which is confirmed by teachers. Parents perceive teachers as key figures but also underline the importance of habituation: the more contact with L2 (in language and CLIL classes), the faster language confidence can grow. These findings, on the basis of the qualitative data, offer an interesting option for further research.

A few limitations of the present study, however, are noteworthy. First, pupils’ L2 proficiency was not measured. A decrease in anxiety should lead to an improvement in L2 proficiency. Moreover, it could be interesting to administer measures of L1 skill and/or FL aptitude (eg. The MLAT) to participants prior to the study, especially to investigate the role of confounding variables in FLA. Data could be collected from the pupil files on their performance on previously-administered measures of language-related skills.

Furthermore, elements such as teacher characteristics, pedagogical approach, pupil group characteristics, pupil’s relative standing in the group of learners, class atmosphere and the scope of the CLIL programme have not (yet) been included in the study. Besides such a cross-sectional study, a longitudinal study of CLIL pupils from grades 7 to 12 on the evolution of FLA over time, would be particularly interesting as this age group knows “a bump in anxiety during the teens and twenties that drops off in older groups” (Dewaele & MacIntyre, 2014). It would be equally of interest to include larger respondent groups for each CLIL language, as it constitutes the variable of most important consequence for FLA in the present study. Models that equally include those aspects, will manage to shed more light on the decisive factors for the prevention of FLA in CLIL.

REFERENCES

Aida, Y. (1994). "Examination of Horwitz, Horwitz, and Cope’s construct of foreign language anxiety: The case of students of Japanese". The Modern Language Journal, 78/2: 155-168. https://doi.org/10.1111/j.1540-4781.1994.tb02026.x Baker, S.C., & MacIntyre, P.D. (2003). "The role of gender and immersion in communication and second language orientations". Language Learning, 53: 33-64. https://doi.org/10.1111/0023-8333.00224 Coyle, D., Hood, P., & Marsh, D. (2010). CLIL: Content and language integrated learning. Cambridge, UK: Cambridge University Press. Dalton-Puffer, C., & Smit, U. (2007). “Introduction”, in C. Dalton-Puffer and U. Smit (Eds.) Empirical Perspectives on CLIL Classroom Discourse (pp. 7-23). Frankfurt: Peter Lang. https://doi.org/10.3726/978-3-653-01829-5 Dewaele, J.-M. (2005). "Investigating the psychological and emotional dimensions in instructed language learning: Obstacles and possibilities". Modern Language Journal, 89/3: 367-380. https://doi. org/10.1111/j.1540-4781.2005.00311.x

RLyLA Vol. 14 (2019), 153-172 | 165 Mathea Simons, Claudio Vanhees, Tom Smits and Karen Van De Putte Remedying foreign language anxiety through CLIL? A mixed-methods study with pupils, teachers and parents

Dewaele, J.-M. (2015). "On emotions in foreign language learning and use". The Language Teacher, 39/3: 13-15. Dewaele, J.-M. (2010). "Multilingualism and affordances". International Review of Applied Linguistics, 48: 105-129. Dewaele, J.-M., & Al-Saraj, T. M. (2015). Foreign language classroom anxiety of Arab learners of English: The effect of personality, linguistic and sociobiographical variables.Studies in Second Language Learning and Teaching, 5: 205-228. https://doi.org/10.14746/ssllt.2015.5.2.2 Dewaele. J.-M., & MacIntyre, P.D. (2014). Foreign Language Enjoyment and Foreign Language Classroom Anxiety. Unpublished manuscript. Dewaele, J.-M., Witney, J., Saito, K. & Dewaele, L. (2017). "Foreign language enjoyment and anxiety in the FL classroom: the effect of teacher and learner variables". Language Teaching Research, 22/6: 676–697. https://doi.org/10.1177/1362168817692161 Doiz, A., Lasagabaster, D., & Sierra, J. M. (2014). "CLIL and motivation: The effect of individual and contextual variables". Language Learning Journal, 42/2: 209-224. https://doi.org/10.1080/09571736.2014.889508 Donovan, L.A., & MacIntyre, P.D. (2004). "Age and sex differences in willingness to communicate, communication apprehension, and self‐perceived competence". Communication Research Reports, 21/4: 420-427. https://doi.org/10.1080/08824090409360006 Dörnyei, Z., & Schmidt, R. (Eds.). (2001). Motivation and second language acquisition. Honolulu, HI: University of Hawaiï. Gardner, R. C. (2001). "Integrative motivation and second language learning: practical issues". Journal of Foreign Language Education and Research, 9: 71-91. Gardner, R.C., & MacIntyre, P.D. (1993). "On the measurement of affective variables in second language learning". Language Learning, 43/2: 157-194. https://doi.org/10.1111/j.1467-1770.1992.tb00714.x Gosling, S.D., Rentfrow, P.J., & Swan, W.B. (2003). "A very brief measure of the big five personality domains". Journal of Research in Personality, 37: 504-528. https://doi.org/10.1016/S0092-6566(03)00046-1 Gregersen, T., & Horwitz, E.K. (2002). "Language learning and perfectionism: Anxious and non-anxious language learners’ reactions to their own oral performance". Modern Language Journal, 86/4: 562-570. https://doi.org/10.1111/1540-4781.00161 Gregersen, T., MacIntyre, P. D., Finegan, K. H., Talbot, K., & Claman, S. (2014). "Examining emotional intelligence within the context of positive psychology interventions". Studies in Second Language Learning and Teaching, 4: 327-353. https://doi.org/10.14746/ssllt.2014.4.2.8 Gregersen, T., MacIntyre, P. D., & Mercer, S. (2016). Introduction. In P. D. Mac-Intyre, T. Gregersen and S. Mercer (Eds.), Positive psychology in SLA (pp. 1-9). Clevedon, UK: Multilingual Matters. https://doi.org/10.21832/9781783095360-002 Horst, M., Cobb, T., & Meara, P. (1998). "Beyond a Clockwork Orange: Acquiring second language vocabulary through reading". Reading in a Foreign Language, 11/2: 207-223. Horwitz, E.K., Horwitz, M.B., & Cope, J. (1986). "Foreign language classroom anxiety". The Modern Language Journal, 70/2: 125-132. https://doi.org/10.1111/j.1540-4781.1986.tb05256.x Horwitz, E.K. (2000). "It ain’t over ‘til it’s over: On foreign language anxiety, first language deficits, and the confounding of variables". The Modern Language Journal, 84/2: 256-259. https://doi.org/10.1111/0026-7902.00067 Horwitz, E.K. (2001). "Language anxiety and achievement". Annual review of applied linguistics, 21: 112-126. https://doi.org/10.1017/S0267190501000071 Horwitz, E.K. (2010). "Foreign and second language anxiety". Language Teaching, 43: 154-167. https://doi.org/10.1017/S026144480999036X Housen, A., Janssens, S., & Pierrard, M. (2002). Le français face à l’anglais dans les écoles secondaires en Flandre. Louvain-la-Neuve: Duculot. John, O.P., & Srivastava, S. (1999). The Big Five Trait Taxonomy, in L.A. Pervin and O.P. John (Eds.) Handbook of Personality. Theory and Research (pp. 102-138). New York: The Guilford Press. Johnson, R. (1997). "Examining the validity structure of qualitative research". Education, 118/2: 282-292. Kaiser, H.F. (1960). "The application of electronic computers to factor analysis". Educational and Psychological Measurements, 20: 141-151. https://doi.org/10.1177/001316446002000116 Kline, P. (1994). An easy guide to factor analysis. Newbury Park: Sage. Lasagabaster, D. (2009). "The implementation of CLIL and attitudes towards trilingualism". International Journal of Applied Linguistics, 157: 23-43. https://doi.org/10.2143/ITL.157.0.2042586 Lasagabaster, D., & Doiz, A. (2017). "A longitudinal study on the impact of CLIL on affective factors".Applied Linguistics, 38: 688-712. MacIntyre, P.D. (1995). "On seeing the forest and the trees: A rejoinder to Sparks and Ganschow". The Modern Language Journal, 79/2: 245-248. https://doi.org/10.1111/j.1540-4781.1995.tb05438.x

RLyLA Vol. 14 (2019), 153-172 | 166 Mathea Simons, Claudio Vanhees, Tom Smits and Karen Van De Putte Remedying foreign language anxiety through CLIL? A mixed-methods study with pupils, teachers and parents

MacIntyre, P.D. (2002). Motivation, anxiety and emotion in second language acquisition, in P. Robinson (ed.) Individual differences and instructed language learning. Amsterdam/Philadelphia: John Benjamins, 45-68. https://doi.org/10.1075/lllt.2.05mac MacIntyre, P.D., Baker, S.C., Clèment, R., & Donovan, L.A. (2003). "Sex and age differences on willingness to communicate, anxiety, perceived competence, and L2 motivation among junior high school French immersion students". Language Learning, 53: 137-165. https://doi.org/10.1111/1467-9922.00226 Maclntyre, P. D., & Gardner, R.C. (1994). "The subtle effects of language anxiety on cognitive processing in the second language". Language Learning, 44: 283-305. https://doi.org/10.1111/j.1467-1770.1994.tb01103.x MacIntyre, P. D., & Gregersen, T. (2016). Positive psychology in SLA. Clevedon, UK: Multilingual Matters. https://doi.org/10.21832/9781783095360 McCrae, R.R., & Costa, P.T. (1999). The five-factor theory of personality, in O.P. John, R.W. Robins and L.A. Pervin (eds.) Handbook of Personality. Theory and Research. New York: The Guilford Press, 159-181. Möller, V. (2016). A question of cause and effect: How educational settings at the primary and secondary level interact with intelligence, motivation and anxiety. Paper presented at the Focus on the Learner Conference. October, 2016. Konin, Poland. Muñoz, C. (2002). Relevance and potential of CLIL. In Marsh, D. (ed.) CLIL/EMILE – The European dimension. Actions, trends and foresight potential. Strasbourg: European Commission. Nunnally, J. (1978). Psychometric theory (2nd ed.). New York: McGraw Hill. Oxford, R. L. (2015). "Emotion as the amplifier and the primary motive: Some theories of emotion with relevance to language learning". Studies in Second Language Learning and Teaching, 5: 371-393. https://doi.org/10.14746/ssllt.2015.5.3.2 Pérez-Paredes, P. F., & Martínez-Sánchez, F. (2000). A Spanish version of the foreign language classroom anxiety scale: Revisiting Aida’s factor analysis. Revista Española de Linguistica Aplicada, 14: 337-352. Pihko, M.-K. (2007). Foreign Language Anxiety in Content Language Integrated Learning (CLIL) and in Traditional Classes. From Brawn to Brain. Proceedings of the ViKiPeda 2007 Conference. Helsinki: University of Helsinki, 129-141. Ross, A. S., & Stracke, E. (2016). "Learner perceptions and experiences of pride in second language education". Australian Review of Applied Linguistics, 39/3: 272-291. https://doi.org/10.1075/aral.39.3.04ros Seligman, M. E. P., & Csikszentmihalyi, M. (2000). "Positive psychology: An introduction". American Psychologist, 55/1: 5-14. https://doi.org/10.1037/0003-066X.55.1.5 Simons, M. & Decoo, W. (2009). "Communicatiedurf versus taalangst". Levende Talen Tijdschrift, 10/2: 3-13. Simons, M. & Smits, T. F. H. (2018). Language Education and Emotions. Proceedings of the Third International Conference on Language Education and Testing. Antwerp: University of Antwerp. Skehan, P. (1989). "Language testing part II". Language Teaching, 22/1: 1-13. https://doi.org/10.1017/ S0261444800005346 Somers, T. & Llinares, A. (2018). "Students’ motivation for content and language integrated learning and the role of programme intensity". International Journal of Bilingual Education and Bilingualism. https://doi.org/10.1 080/13670050.2018.1517722 Sparks, R.L., & Ganschow, L. (2007). Is the foreign language classroom anxiety scale measuring anxiety or language skills? Foreign Language Annals, 40/2: 260-287. https://doi.org/10.1111/j.1944-9720.2007.tb03201.x Sparks, R. L., & Patton, J. (2013). "Relationship of L1 skills and L2 aptitude to L2 anxiety on the Foreign Language Classroom Anxiety Scale". Language Learning, 63/: 870-895. https://doi.org/10.1111/lang.12025 Strobbe, L., Sercu, L., Strobbe J., & Welcomme A. (2012). Je vak in een vreemde taal? Wegwijzer voor de CLIL- onderwijspraktijk. Leuven: Acco. Thompson, A.S., & Lee, J. (2013). "Anxiety and EFL: Does multilingualism matter?" International Journal of Bilingual Education and Bilingualism, 16/6: 730-749. https://doi.org/10.1080/13670050.2012.713322 Thompson, A.S. & Sylvén, K. (2015). "Does English make you nervous? Anxiety profiles of CLIL and non-CLIL students in Sweden". Apples - Journal of Applied Language Studies, 2. https://doi.org/10.17011/apples/ urn.201512093950 Trang, T. (2012). "A review of Horwitz, Horwitz and Cope’s theory of foreign language anxiety and the challenges to the theory". English Language Teaching: 5/1: 69. https://doi.org/10.5539/elt.v5n1p69 Vansteenkiste, M., & Soenens, B. (2015). Vitamines voor groei: ontwikkeling voeden vanuit zelf-determinatietheorie. Leuven: Acco. Woodrow, L. (2006). "Anxiety and speaking English as a second language". RELC Journal, 37/3: 308-328. https://doi.org/10.1177/0033688206071315 Yamashiro, A.D., & McLaughlin, J. (2001). "Relationships among attitudes, motivation, anxiety, and English language proficiency in Japanese college students". Second language research in Japan, 112-126.

RLyLA Vol. 14 (2019), 153-172 | 167 Mathea Simons, Claudio Vanhees, Tom Smits and Karen Van De Putte Remedying foreign language anxiety through CLIL? A mixed-methods study with pupils, teachers and parents

Yang, H.C. (2012). "Language Anxiety, Acculturation, and L2 Self". Electronic Journal of Foreign Language Teaching 9/2: 183-193. Young, D. (2018). Turning a Hazy View into Focus. Trends and Turns in Thirty-two Years of Language Anxiety Research, in M. Simons, T. Smits and P. Castaño-Sequeros (Eds.) What a Feeling! Language Education and Emotions. Book of Abstracts (pp. 157). Antwerp: University of Antwerp.

RLyLA Vol. 14 (2019), 153-172 | 168 Mathea Simons, Claudio Vanhees, Tom Smits and Karen Van De Putte Remedying foreign language anxiety through CLIL? A mixed-methods study with pupils, teachers and parents

APPENDICES

APPENDIX 1 – SURVEY PUPILS

A. Items to measure the dependent variable Foreign language anxiety

The following items refer to the subject courses [i.e. geography, art history, economics] taught and learnt in a foreign language [i.e. English, French]. Please indicate what applies: 1 (strongly disagree)–2 (disagree)–3 (neutral)–4 (agree)–5 (strongly agree).

1. I always feel quite sure of myself when I am speaking in my foreign language class 2. I don’t worry about making mistakes in language lessons. 3. I tremble when I know that I’m going to be called on in language lessons. 4. It frightens me when I don’t understand what the teacher is saying in the foreign language. 5. It would bother me to take more classes in a foreign language. 6. During class, I find myself thinking about things that have nothing to do with the course. 7. I keep thinking that the other students are better at languages than I am. 8. I am usually at ease during tests in [FL] for the [subject] class. 9. I start to panic when I have to speak [FL] in [subject] classes. 10. I worry about the consequences of missing my class of [subject]. 11. I don’t understand why some people get so upset about speaking [FL] in [subject] classes. 12. Sometimes I can get so nervous that I forget things I know. 13. It embarrasses me to volunteer answers in [FL]. 14. I would not be nervous speaking [FL] with native speakers. 15. I get upset when I don’t understand what the teacher is correcting. 16. Even if I am well prepared for language class, I feel anxious about it. 17. I often feel like not going to my [subject] class as it is taught in [FL]. 18. I feel confident when I speak [FL] in [subject] classes. 19. I am afraid that my teacher is ready to correct every mistake I make. 20. I can feel my heart pounding when I’m going to be called on. 21. The more I study for a language test, the more confused I get. 22. I don’t feel any pressure to prepare very well for [subject] in [FL]. 23. I always feel that the other students speak [FL] better than I do. 24. I feel very self‐conscious about speaking [FL] in front of other students. 25. [Subject] classes move so quickly I worry about getting left behind. 26. I feel more tense and nervous in my [subject] class than in my other classes. 27. I get nervous and confused when I am speaking in my [subject] class. 28. When I’m on my way to the [subject] class, I feel very sure and relaxed. 29. I get nervous when I don’t understand every word the language teacher says. 30. I feel overwhelmed by the number of rules one has to learn to speak [FL]. 31. I am afraid that the other students will laugh at me when I speak [FL]. 32. I would probably feel more comfortable around native speakers than with students when I have to speak [FL]. 33. I get nervous when the [subject] teacher asks questions that I haven’t prepared in advance.

Control item in post-measurement:

34. After one year of CLIL education I am less anxious when speaking or writing [FL].

RLyLA Vol. 14 (2019), 153-172 | 169 Mathea Simons, Claudio Vanhees, Tom Smits and Karen Van De Putte Remedying foreign language anxiety through CLIL? A mixed-methods study with pupils, teachers and parents

B. Items to identify independent variables:

1. Personal data

Name; Gender; Nationality; Home language

2. Items measuring personality traits

Please rate each statement according to how well it describes you: 1 (strongly disagree) –5 (strongly agree).

1. I am relaxed most of the time. 2. I worry about things. 3. My bedroom is often a mess. 4. I have a vivid imagination. 5. I don’t mind being the center of attention. 6. I take time for others. 7. I follow a schedule. 8. I rarely get irritated. 9. I put the others first. 10. I have difficulty understanding abstract ideas. 11. I keep in the background. 12. I am always prepared. 13. I lift a conversation to another level. 14. I talk to a lot of different people at parties. 15. I make people feel at ease. 16. I complain easily. 17. I feel other people’s emotions. 18. I shirk my duties. 19. I don’t like to draw attention to myself. 20. I am quick at understanding things.

3. Items related to motivation for CLIL

Please indicate to which extent the following propositions apply: 1 (strongly disagree)–5 (strongly agree).

I opted for CLIL because

A. I like it. B. Others (parents, teachers etc.) expect it of me. C. I can learn a lot from it. D. I would feel guilty if I didn’t. E. I realize why it is useful F. Others forced me to do this. G. I want to prove myself H. I find it interesting.

RLyLA Vol. 14 (2019), 153-172 | 170 Mathea Simons, Claudio Vanhees, Tom Smits and Karen Van De Putte Remedying foreign language anxiety through CLIL? A mixed-methods study with pupils, teachers and parents

4. Motivation for the subject and the language

Please indicate what applies: 1 (strongly disagree)–5 (strongly agree).

1. I find the subject interesting. 2. I find the language class interesting. 3. I read [FL] books, magazines, newspapers etc. in my spare time. 4. I go on holiday to an [FL]-speaking country. 5. I try to speak [FL] when the opportunity presents itself (e.g. on holiday abroad)). 6. I watch [FL] series or programmes (e.g. on television or on my computer). 7. I try to understand the lyrics of [FL] songs. 8. I correspond in [FL] with friends abroad, acquaintances etc. (e.g. Facebook, Twitter etc.). 9. I have participated in an [FL] summer course.

APPENDIX 2 - GUIDELINES FOR THE INTERVIEW, PUPILS

Main questions Subquestions (RQ1) You are taking (subject) in (CLIL - when speaking (CLIL language)? (Afraid of making mistakes?) language). Do you feel confident/ - when writing? comfortable: - when hearing? - when reading? (RQ1) Do you feel more self-confident - Do you feel that you can easily express yourself in (CLIL language)? (E.g. than last year, when you didn’t take in a class conversation; Are you more at ease asking questions?) (subject) in (CLIL language) yet? - Do you feel that you understand better? (E.g., when the teacher explains something or when reading something) - Do you feel that you can write more fluently? (E.g., when writing an essay or writing down answers.) (RQ1) Do you enjoy taking (subject) - Would you opt for it again? more now that it is being taught in - Would you like to take more courses in another language? Which (CLIL language), or less? subjects? Which languages? (RQ2) From the previous questions - Your own personality? (Are you someone who doubts sometimes or we can infer that you feel secure/ comes across unsure of him/herself? Or are you rather self-assured and insecure during (subject) classes in don’t you care so much what other people think of you?) (CLIL language)… What influences this - The language itself? (Is (CLIL language) your forte, do you consider it an feeling of (in)security? enjoyable/interesting/difficult language?) - Your interest in (CLIL language) or in (subject)? - Your classmates? (Do you help each other if you don’t understand something?) - The teacher/support? (Does (s)he encourage you? Do you feel at ease addressing him/her? - Your motivation? Why did you opt for the CLIL programme? (personal interest, friends, parents, teachers)

RLyLA Vol. 14 (2019), 153-172 | 171 Mathea Simons, Claudio Vanhees, Tom Smits and Karen Van De Putte Remedying foreign language anxiety through CLIL? A mixed-methods study with pupils, teachers and parents

APPENDIX 3 - GUIDELINES FOR THE INTERVIEW, TEACHERS

Main questions Subquestions (RQ1) Would you say that the students - How can you tell? E.g. asking and answering questions, expressing their are at ease in your classes? (dis)likes, disclosures, formulating reasons, a point of view or opinion, actively taking part in a conversation,. - Do you notice a difference between the skills?

(RQ1) Do you notice a difference with - On which point do you perceive an evolution? the beginning of the school year? - Do you notice a difference between the skills?

(RQ1) Do you notice a difference - Do you notice a difference between the skills? with students not receiving CLIL education? (RQ2) What might influence students’ - What causes of feelings of anxiety or insecurity do you discern? language confidence? - What causes of language confidence do you discern?

APPENDIX 4 – SURVEY, PARENTS

1. What is your daughter/son’s attitude towards the CLIL classes? My daughter/son* (several answers possible)

• is enthusiastic about the CLIL classes. • regrets his/her decision to take up CLIL. • experiences more stress because of the CLIL classes. • other:

2. When my daughter/son talks about the CLIL classes at home, this feedback is

• positive, because (explain briefly): • negative, because (explain briefly): • my son/daughter never talks about the CLIL classes at home.

3. Do you feel that your daughter/son feels more at ease with (CLIL language)? (e.g. dares to express themselves more, understands better) YES/NO

If yes: does that hold true for all skills? (speaking, writing, listening, reading) If not: what causes this and how can it be improved?

4. My daughter/son indicates understanding the teacher WELL/NOT WELL and being able to follow the teaching content WELL/NOT WELL.

If well: has this improved compared to the beginning of the school year and why do you think? If not well: what causes this and how can it be improved?

5. My daughter/son indicates (s)he DOES/DOES NOT manage to ask questions during CLIL lessons.

If so: has this improved compared to the beginning of the school year and why do you think? If not: what causes this and how can it be improved?

6. My daughter/son indicates that studying the learning content is EASIER/MORE DIFFICULT.

If easier: has this improved compared to the beginning of the school year and why do you think? If more difficult: what causes this and how can it be improved?

RLyLA Vol. 14 (2019), 153-172 | 172 polipapers.upv.es/index.php/rdlyla