Access to health in an intercultural setting : the role of corpora and images in grasping term variation

Medical concepts can often be lexicalized in severa l ways depending on aspects such as the facet of the concept being unde rlined or the particular communicative setting in which the concept is being used. This feature of terminology is known as terminological variation. I this paper we consider terminological variation as a tool to improve inter linguistic and intercultural communication, a key issue in the pro vision of universal access to health care. To facilitate the identifica tion and analysis of terminological variation, the paper also proposes s ome search strategies to highlight this phenomenon in corpora, the main sour ce of terminological information. Finally, images are proposed as a key issue in the localization process needed to bridge communication gaps between health care providers and lay audiences. The data used in the p aper are taken from an international cooperation project aimed at providin g health providers in Yucatan, Mexico, with materials and training in int ercultural communication for healthcare mainly in Spanish and Mayan, and from a research project on lexical variation .


Introduction
Universal health access is a basic right that however is far from being accomplished due to economic, language and cultural barriers 2 which lead to difficult access to health care and poor communication between patients and health professionals.In many places, such as the Yucatan peninsula in Mexico, it is often the case that mainstream health care based on Western medicine coexists with traditional culture-based practices and this means that interculturality is a key issue in succeeding in the provision of healthcare.Interculturality within health care is defined as the explicit incorporation of the collective cultural load of the patient into his/her relationship with the healthcare provider, who is in turn source, generator and transmitter of a different form of culture (Campos, 2011).In addition, culture encompasses not only the products of a society as embodied in its institutions and objects, be they traditional or Western, but also world views, experiences and behaviour schemata (López-Rodríguez, 2003, p. 152).
Furthermore, low health literacy is often an obstacle reflected in the difficulties the patient faces to understand the medical concepts used by the provider (Brooke Lerner, Jehle, Janicke, & Moscati, 2000).As a result, health authorities and providers are increasingly concerned with health literacy, defined by the American Medical Association (n.d.) as the "ability to obtain, process and understand basic health information and services needed to make appropriate health decisions and follow instructions for treatment".
In this context, translators are essential in the crosslinguistic and crosscultural mediation that often takes place in medical settings where there may be asymmetries of medical knowledge between health professionals and patients, as well as cultural and linguistic differences due to different national identities.According to Angelelli (2008, p. 130) in cross-cultural medical encounters interpreters (and hence translators) occupy the unique position of understanding both interlocutors' points of view and points of reference.The literature on health interpreting includes studies assessing translational accuracy and mistranslations (omissions, substitutions, etc), and their effect in clinical practice (Pöchhacker & Shlesinger, 2007, pp. 2-4).Many of these mistranslations are related to cultural differences between the participants in healthcare communication, patients' low health literacy, and terminological variation.Thus, maintaining accuracy in meaning and promoting intercultural understanding should be a priority for translators and interpreters in medical and health settings, a priority shared by translators in general: "Translators mediate between cultures (including ideologies, moral systems and socio-political structures), seeking to overcome those incompatibilities which stand in the way of transfer of meaning" (Hatim & Mason, 1990, p. 223).
The aim of this paper is to present different forms of terminological variation (register-based and dialectal variation) in medical settings and their relevance in intercultural and interlinguistic communication.We will illustrate ways in which terminological variation can be a useful tool to bring medical knowledge closer to different audiences.Our methodology for retrieving and analysing different forms of denominative variation is based on the use of electronic corpora and search strategies.Finally, images are proposed as a key issue in the localization process needed to bridge communication gaps between health care providers and lay audiences, and we illustrate this with examples from our cooperation project in the Yucatan peninsula.Our data are extracted from the methods and results of research in this cooperation project entailing the production of Mayan and Spanish audiovisual materials for the promotion of health care in the Yucatan peninsula, and in the context of a research project on terminological variation aimed at the study of this phenomenon from a communicative as well as a cognitive perspective.

Culture and medical knowledge in healthcare communication
In interlinguistic healthcare communication there are cultural elements whose understanding is essential for health decisions and treatment.Therefore, the use of translation procedures helping to bridge the gap between source language and target language, while maintaining accuracy, are key.Montalt & González Davies (2007, pp. 180-181) mention the following translation strategies: exoticism, cultural borrowing, calque, transliteration, communicative translation and cultural transplantation.
While these are useful strategies at the microtextual level, translating in an intercultural setting requires going beyond the textual level.Consequently, Montalt & González Davies (2007, p. 184), following Katan (1999) indicate that mutual understanding between different linguistic communities may be hindered by the relevance given by these communities to context vs. text in communication.In this regard, Latin American cultures are closer to so-called High Context Cultures, for which "communication takes place through context rather than through texts" (Katan, 1999).Cultures where text is more important than context in communication are referred to as Low Context Cultures. 3 As part of our cooperation project, we became familiar with the health situation and the reasons for hospital admission of the indigenous populations in the Yucatan peninsula from 2010 to 2012.In this setting, communication takes place between two high context cultures (Mexican and Spanish), however, health information and practices are based on two different medical systems (Western medicine and Traditional medicine 4 ), which are not always successfully integrated.Health resources provided by international organizations such as Unicef and PAHO are neither assimilated nor accepted by the local population.The Yucatan peoples do not accept these resources because they do not represent their physical features and local identities as reflected in customs, clothes, houses, and everyday artifacts.Traditional medicine in Mexico, as in many Latin American countries is based on humoral medicine, defined by Foster (1987, p. 355) as "an ethnomedical system in which local foods and medicines are labeled with "hot" and "cold" markers, and illness is believed to be a disturbance to the temperature equilibrium in the body that is treated through the principle of opposites" (as cited in Worley, 2011, p. 5).
Such folk medicine and practice pose a communication problem for doctors prescribing Western medicine to indigenous peoples since their world views and cognitive frame for understanding disease and healing are different.Smith (2003, p. 3) mentions that in Mexican traditional medicine vitamin C is considered a "cold food", and therefore an inappropriate treatment for a "cold disease" like an upper respiratory tract infection.These cultural differences also pose translation and terminological challenges since there are medical conditions that are endemic to certain geographical areas due to environmental circumstances and/or cultural practices.Such conditions are known as "culture bound syndromes", "culture specific diseases" (O'Neil, 2010) or "syndromes of cultural filiation" (Fagetti, 2004).Culture specific diseases are usually referred to with local names (e.g., ku ben ba, mal viento, tip' te) because of their cultural and environmental specificity, and they are recognized in Traditional medicine but not always in Western medicine.Some of the most reported culture bound syndromes in the area of scope of our project are mal de ojo/ ojo (evil eye or strong eye), aire (group of diseases caused by leaving a hot environment and entering a very cold one), and susto (fright sickness) 5 .
Considering the distance between these two medical systems and the need to acknowledge the cultural idiosyncrasies of the indigenous peoples, it was necessary to find a common framework for communication between doctors and patients in which basic concepts related to disease and their basic conceptual relations (cause, effect, location, etc.) are integrated.In fact, all cultures have systems for classifying diseases on the basis of etiology, signs/symptoms and treatments (Neff, 1996).
The first such common framework is the MEDICAL EVENT proposed by Faber (2002, p. 8) (Figure 1).It constituted the backbone for connecting medical terminology in English and Spanish of varying degrees of specialization in an information resource on cancer called OncoTerm 6 (Faber, López-Rodríguez, & Tercedor, 2001;López-Rodríguez, Faber, & Tercedor, 2006).In the MEDICAL EVENT, the main conceptual categories in a prototypical medical process were identified: RISK FACTOR, BODY_PART, CONDITION, SYMPTOM, DIAGNOSTIC PROCEDURE, INSTRUMENT, TREATMENT, THERAPEUTIC OR HEALING AGENT, SIDE EFFECT, SPECIALISTS and HOSPITAL.It was ascertained that all these categories were valid and applicable to different medical domains (Faber, 1999, p. 99;Faber & Mairal, 1999).(Faber, 2002, p. 8) This figure permits us to relate, define and paraphrase specialized terms on the basis of general language vocabulary.For example, the concept radiation therapy was represented in relation to basic concepts such as TREATMENT, INSTRUMENT, THERAPEUTIC AGENT, RESULT and LOCATION, as shown in Figure 2.This template was the foundation for terminographic definitions.To our mind, this framework can help bridge the gap between the medical knowledge and vocabulary of experts and those of patients, and between users and providers of health services who use different lexicalizations of concepts according to different situations and communication needs because it breaks down concepts into meaning units (see Section 3, Figure 4).As a result, this framework can be applied to integrate the knowledge and practices of different medical traditions, and to explain and define medical conditions to different audiences as in the following example:  The second framework that can help patients and interpreters come to grips with specialized medical objects, entities and procedures, is a better understanding of the Greek and Latin basis of medical terms.In our cooperation project, the population who spoke Mayan and Spanish obtained a better understanding of medical terminology after their etymology had been explained.In this regard, Peckham (1994), studying reactions in 180 orthopedic patients reported that 80% of them thought that there was a difference between a fractured bone and a broken bone.Montalt & González Davies (2007, p. 267) include an appendix of Latin and Greek roots for medical terminology, which are organized thematically.Greek and Latin roots in medical terminology can facilitate classification of terms into main conceptual categories such as these.The above classification constitutes again a framework for the organization of medical knowledge in the same vein as Faber's medical event.

Terminological variation in corpora as a key for use in intercultural communication
The multiple ways to designate a concept through different expressions is often referred to as terminological variation 7 .Thus, term variants are the different lexical manifestations of a term designating a particular concept.Researchers in both translation and terminology have long acknowledged the cognitive and communicative motivation of terminological variation.Although variation has been approached mostly from the cognitive perspective recently (Fernández, 2010;Fernández & Kerremans, 2011;Tercedor, 2011), its communicative aspects have been tackled in many studies as well (Freixa, 2006;Freixa, Fernández, & Cabré , 2008;Tercedor & Méndez, 2000).Medical concepts can be lexicalized in several ways depending on the facet of the concept being underlined, given a particular domain or a particular interest in the sender, something which brings about the multidimensional character of medical terminology, a feature it shares with other specialized domains.Furthermore, a particular term choice can be made on the grounds of the communicative setting in which it is going to be used, entailing geographic features, a marked register, or both.This feature of terminology is key in succeeding in health promotion in an intercultural setting.In this paper, we will therefore concentrate on the communicative motivation of terminological variation and how this feature can be incorporated in corpus linguistics for terminological purposes.
One of the most influential approaches to linguistic variation was the one by Halliday, McIntosh and Strevens (1964, p. 77), who distinguished between linguistic variety according to the use to which language is put (register), and variety according to the characteristics of the users engaged in communication (dialect).This approach to language variation was applied to Translation Studies by many authors (Baker, 1992;Hatim & Mason, 1990; inter alia), and can be applied to terminological variation.The communicative reasons for choosing one term or another are the focus of Sections 3.1 and 3.2.
Language enables us to construe items of world knowledge in different ways (Wierzbicka, 1995) and the way in which it is done depends on many factors intrinsic to the concept and its features or to the way it is perceived.Thus, the distinction between what the word means "objectively" and what it means to its user is not always obvious.In this regard, the cognitive paradigm challenges the objectivist view that distinguishes between the semantics and pragmatics of concepts, the distinction between what the word means and all the encyclopaedic knowledge that you happen to have about the things the word refers to (Lakoff, 1987, p. 138).This consideration of semantics and pragmatics as a continuum is key in the understanding of lexical instances of concepts as equally relevant, their relevance being dependant upon the particular context in which the term is going to be used.
For a long time corpora have proved useful in the identification of terminological variation in medicine (López-Rodríguez, 2001;Tercedor, 1999).The different types of terminological variants, whether registerdependent or not, require different types of corpora.In order to compile register-dependent terminological variants a corpus of texts for a lay audience or a corpus of texts with different levels of specialization is needed, whereas to identify variations within the same level, a corpus containing texts of the same level of expertise can be used (Tercedor & Méndez, 2000).
In a corpus of semi-specialized medical texts, the use of what has been called knowledge probes (Ahmad & Fulford, 1992) and knowledgerich contexts (Meyer, 1994) can help translators detect register variation in texts: one of the variants is the specialized term and the other is a hypernym, which is followed by a paraphrase of the specialized term.Meyer et al. (1999cited in L'Homme & Marshman, 2007, p. 70)  The use of these patterns to detect terminological variants that can be of use in contexts such as the one in Yucatan described above, will be shown in examples taken from the 32 million corpus of the OncoTerm project. 8We will also search the CREA corpus (Reference Corpus for Contemporary Spanish), 9 and Sketch Engine, 10 a Corpus Query System providing access to large corpora and producing concordances, and word sketches (a corpusderived summary of a word's grammatical and collocational behavior).In this paper we exploit these resources either offline (with WordSmith Tools 11 ) or online to extract concordances and word sketches as a way to grasp conceptual relations and term variants (López-Rodríguez, 2007;Tercedor & López, 2008).

Register-based variation
Register is understood as a combination of three variables: field, tenor and mode (Halliday et al., 1964).Halliday (1993, p. 25) defines these variables as follows: (a) field is the social action in which the text is embedded; it includes the subject-matter, as one special manifestation; (b) tenor is the set of role relationships among the relevant participants; it includes level of formality as one particular instance; (c) mode is the channel of communication selected, it includes the medium (spoken or written).
Tenor is normally associated to the personal relation in terms of social distance or formality between the user and his/her audience.This notion has been called "formal tenor" (Hatim, 1997, p. 18) and "personal tenor" (Hatim & Munday, 2004, p. 190).In any case, tenor is a particular privileged category (Hatim, 1997, p. 26) because it interacts with field and mode.The interaction between tenor and field gives rise to technicality: the more formal the occasion, the more technical the use of language (Hatim & Munday, 2004, p. 81).When tenor interacts with mode, tenor determines the purpose for which language is used such as to persuade, to inform, to exhort or to discipline.Hatim and Mason (1990, p. 51) call it functional tenor.
Drawing on this approach, terminological variation due to register mainly encompasses variation reflecting the social relation between the participants (formal register) as well as the knowledge shared by participants (technicality or degree of specialization).For example, a doctor can use the expression chemo in an informal situation with a colleague, low-dose chemotherapy in a research journal, and medication to communicate with his/her patient.
The increasing number of standardized medical vocabularies, such as SNOMED CT®, ICD, CPT®, MeSH, or UMLS-to mention a few-show the complexity and nuances of communication in medical settings, and of variation according to use (register) as shown in the following example involving different types of users such as patients, companies, health organizations, clinicians and information management professionals.One of the ways of acting interculturally when providing health information is through the selection of paraphrases to better target the particular ways of conceiving health by a given cultural and linguistic community.That is why, when recording medical terminology for different purposes, different communicative settings should be considered and marked.
For the purpose of corpus analysis and its use in the improvement of intercultural communication in health settings such as the one described above, the classification of texts according to communicative settings is essential.This can be achieved in two ways: (a) grouping texts in folders according to technicality/degree of specialization (specialized, semispecialized, text for the general public) and formality (formal, neutral, informal); (b) adding in each text a header with tags relating to the pertinent communicative setting(s).
The first approach was followed in OncoTerm, where the texts included only basic tags (for example, author, title, journal and pages in the case of research articles), and the files division into folders was the basis for searches to retrieve terminological variants.For example, the introduction of the knowledge probe called served to generate concordances where the words to the left of called were an explanation or paraphrase of the meaning of the specialized term to the right (bronchi, epidermoid carcinoma, bronchoscopy, and so on).These paraphrases were very indicative of the kind of information to be included in the definition of terms for a lay audience (Pearson, 1998;López-Rodríguez, 2001).Other useful expressions were: is known as / is referred to as / is called / means / is defined as.As to the second approach to classify texts with a view to detect register variation, in the framework of the two research projects described in this paper, a corpus of medical texts for different contexts has been compiled and is now being tagged following the following schema in order to filter texts by keywords, medical domain, audience, and so on.We have used an adaptation of the tags proposed by the Lexicon research group (http://lexicon.ugr.es) for the corpus of EcoLexicon, a terminological knowledge base on the Environment (http://ecolexicon.ugr.es/en/).The annotation is based on XML language, and is the result of selecting from the Dublin Core metadata 12 the most relevant for translation purposes.The domain and keywords tags ensure the comparability of the texts so that the texts under study share similar terminology.These tags also illustrate the overlapping between field, tenor and mode.Functional tenor (interaction tenor-mode) Figure 6: Tags to filter register variation Even when the corpus has not been tagged, when translating medical terms in intercultural contexts it is possible to obtain good examples by consulting on-line corpora such as the CREA (Corpus de Referencia del Español Actual) or the ones available from Sketch Engine (López-Rodríguez & Buendía, 2011) as will be seen in the following section.

Dialectal variation
In terminology, studies on medical variation normally focus on variation according to register.However, in real medical encounters in an intercultural context, health providers and translators should also consider medical variation according to dialect (diatopic, diastratic, and chronolectal variation).All these have an impact on the strategies chosen to adapt the message to particular groups of receivers in a specific geographical, social and temporal context.Within diatopic or geographical variation, corpora are valuable tools to confirm intuitions about the geographical uses of words.The use of tools such as Diatopix 13 or Sketch Engine and its available corpora can provide dialectal variants and confirm "terminological hunches".In the case of the Spanish TenTen corpus of Sketch Engine, only examples of European Spanish can be obtained, although the enormous size of the corpus (more than 2100 million words) constitutes a real advantage.And so, a Google search may indicate that, if we encounter the acronym ETV in Mexico, it is likely to refer to the term Enfermedades Transmitidas por Vector while in Spain, it stands for Enfermedad Tromboembólica Venosa.This can be confirmed by searching the expression ETV in the context of venosa and then, in the context of vector within a collocational span of 15 words in the Spanish TenTen corpus.No results for ETV referring to Enfermedades Transmitidas por Vector were found in the corpus of European Spanish (Figure 7).In any case, if we want to use a corpus including texts in Mexican and European Spanish, the CREA should be the best choice, although it often needs to be complemented with other sources of information and a thorough collocational analysis.
If the corpus is compiled for a special purpose, as in our case, tags referring to geographical variation such as the ones below should be used.Dealing with terminological variation in a multicultural setting also means being aware of variation due to social (diastratic) and geographical (diatopic) differences on the part of speakers.A doctor in Spain or Mexico may use the term metanfetamina (methamphetamine) to refer to an extremely addictive drug.However his/her patients can refer to this drug with one of the street names used in their respective countries, which may not be familiar terms for the doctor.Regarding chronolectal variation, there are not many corpus data shedding some light on either the language of children in clinical settings or the use of child language (i.e., language used by the child and controlled actively by her: Jakobson, 1969) by paediatricians examining them.Expanding previous corpora on child language such as the ones contained in the CHILDES database (http://childes.psy.cmu.edu/) to include healthcare situations in different countries and different dialects could be a promising research area.Such a corpus would be helpful in finding the Mexican equivalent of the children expression pupa found in this example from the Spanish component of the CHILDES project (Figure 10).To facilitate the understanding of medical concepts and their different terminological variants in intercultural and interlinguistic health settings, images can be used as a link between concepts and their lexicalizations.Therefore, we propose that visually representing medical conditions is a basic step into a better intercultural and interlinguistic communication, as well as an important tool for the generation of texts based on controlled languages.Following is an example of a visual representation of a medical symptom for kids (tummy ache / dolor de tripa) and the geographical distribution of the terminological variant chosen.This issue is considered in greater depth under 4.

Images triggering term variation in patient-health provider communication
Images can help bridge communication obstacles in health settings provided they are chosen on the grounds of a particular degree of specialization or a focus on a particular aspect of a medical concept.The use of images to represent general medical concepts is a common communicative tool in hospital and other health facilities.Projects such as the Hablamos Juntos program (http://www.hablamosjuntos.org)have long been working in studying universal signs and icons to direct patients within hospital settings.
Images can be used to bridge communication obstacles and to improve patients' knowledge about disease and health, as in the following information leaflet: The leaflet describes the symptoms of diabetes in plain language and with visual support so that low literacy patients can understand."Necesidad de orinar con frecuencia", is a paraphrase that would rarely be included as a term entry in traditional terminology settings.However, it is a better option to ensure comprehension by lay persons than the medical term polyuria.By the same token, "sed excesiva" is used instead of the medical term polidipsia, and "falta de energía" instead of astenia.This register-based form of variation is key in bringing closer concepts to lay audiences or audiences with limited language proficiency.Such strategy is further enhanced by the use of images that facilitate understanding of medical concepts.Images in this context serve as a sort of interlingua between medical concepts and their lexicalizations, but the interpretation of images, icons and body language by different individuals and cultures is also an important issue to consider, as revealed by our cooperation experience in Yucatan.
Images are thus a thorny issue in any process entailing cultural and linguistic adaptation (localization) for health promotion.In our cooperation project, patients did not have confidence in existing healthcare materials (mainly in printed form) due to the fact that they had been designed with a focus on global use (written in majority languages and lacking cultural referents) and therefore, patients' identities were not reflected.Consequently, we addressed this challenge by designing and developing audiovisual and printed materials in Mayan and Spanish and by assessing the cultural, technical and linguistic aspects that should be considered with a focus on localization, reflected both in language and images.As for language, local cultural concepts were chosen to empower the language and culture.For example, analogies using local concepts such as "me pondré grande como una ceiba" (I will grow tall as a ceiba) were used to illustrate eating habits for children.Images were chosen on the basis of cultural adequacy after anthropologic consultation and field studies to ensure local acceptance.We have briefly illustrated how in multimodal scenarios images play an important role in the access to health serving as a bridge between the concept and its lexicalizations.In this context, their motivated choice relates to the textual configuration and the terminological variants chosen.

Conclusions
In this paper we have illustrated how terminological variation reflects different communicative and social context-bound motivations in designating concepts.The nature of the concept and its salient features are a basic aspect to study terminological variation, although register can determine the choice of one or another lexicalization.Furthermore, the suitability of a particular term choice in a geographical setting can enhance communication between health care providers and patients.In an interlinguistic and intercultural setting, the awareness of these forms of variation is fundamental in succeeding in providing access to health for all.The corpus-based examples and visual resources provided indicate new ways to improve patient-doctor communication and health literacy.Some issues such as the cognitive motivation for a particular terminological choice in a medical encounter or the relation between term variants and images remain to be researched, and experimental methods such as those searching for spontaneous production are a promising venue for researching terminological variation.
use of high-energy rays to damage cancer cells and stop them from growing and dividing.

Figure 2 :
Figure 2: Representation of the concept RADIATION THERAPY

Figure 3 :
Figure 3: Conceptual categories to promote understanding of medical terminology identify three main types of knowledge patterns: • Lexical patterns: is the, is a, such as, and other, known as […] • Grammatical patterns: for example, NOUN + VERB to explain the function of the specialized term • Paralinguistic patterns such as the use of parenthesis or commas.1

Figure 4 :
Figure 4: Different terms and standards for one concept and different users (Source: IMO ® Interface Terminology: http://www.e-imo.com/

Figure 7 :
Figure 7: The acronym ETV in the Spanish TenTen corpus of Sketch Engine

Figure 8 :
Figure 8: Tags to filter dialectal variation

Figure 9 :
Figure 9: Using the corpus to search for instances of dialectal use: the case of cirro

Figure 10 :
Figure 10: A mother and child talking about pain in the CHILDES database 14

Figure 11 :
Figure 11: Visual representation of the concept STOMACH PAIN for children (left) and presence of the terminological variant "dolor de tripa" as retrieved by Diatopix

Figure 13 :
Figure 13: Local features in the design of printed materials (left) and cartoons (right).Source: Animated videos and education materials from the University of Granada cooperation project Procuración de justicia y acceso a la salud por parte de la población indígena de Yucatán: mediación interlingüística e intercultural