Speakers
Description
Medicine is one of the specialized domains that is of particular interest to different communities of speakers, most of whom cannot be considered experts or semi-experts. Their interest in the domain lies in the fact that a certain level of medical knowledge is needed in everyday life, much like a basic understanding of legal concepts. As a prominent characteristic of the domain, terminological variation has been extensively studied (Bowker & Hawkins, 2006; Lončar & Ostroški Anić, 2013; Tercedor Sánchez & López Rodríguez, 2012, to name a few), focusing mostly on the differences in expertise levels among different speakers, i.e., medical experts and laypeople. The more precise, concise, and systematic the discourse is, the greater the term density and the less term variation. As the degree of specialization decreases, specialized discourse becomes more similar to general discourse in terms of conceptual variation, redundancy, ambiguity,
and the extensive use of synonyms and paraphrases to explain concepts (Cabré Castellví, 1998, in Freixa, 2006). The degree of text specialization causes variation in defining concepts, often referred to as contextual variation (San Martín, 2022), conceptual variation (Freixa & Fernández-Silva, 2017) or vagueness in general language (Geeraerts,
1993). San Martín (2022, p. 2) argues that the context determines the exact meaning of the term in that context, i.e., “the term invokes the same concept, but the activated knowledge differs.” Medical concepts related to diseases, conditions, treatments, procedures, etc., are defined and described differently in different contexts and registers, depending on the intended users. The meaning or concept characteristics remain the same, but different characteristics are highlighted depending on the focus of the communicative setting: the cause of an illness, its symptoms or methods of treatment. In other words, if we regard a disease as a semantic frame with its frame elements (FEs), different elements are in focus depending on the context and the user, which also means that the situation is framed differently by using different terms or term variants. Traditional
terminological or analytical definitions, which consist of the superordinate concept and the defined concept’s delimiting characteristics, will therefore often be replaced with types of definitions that exploit other knowledge patterns, e.g., functional or synonymic (Sierrra et al., 2008). To compare the structural and conceptual differences in the definitions of medical concepts in texts of different registers and levels of expertise, we compared the definitions of terms for 50 diseases found in two corpora of medical texts in Croatian. The aim of this paper is to establish the most common definitional patterns used to define the concepts referring to different diseases in texts for non-experts. The corpora used for the analysis and extraction of patterns had been previously compiled with Sketch Engine tools within the work done by Ostroški Anić & Brač (2022): a scientific corpus of research papers (5,318,395 tokens) and a corpus of texts taken from medical portals with the general public as their intended audience (5,022,639 tokens). All further analysis and definition extraction were also conducted using Sketch Engine tools.
A list of 50 terms for diseases was first established based on concordances of the Croatian term bolest ‘disease’ in the corpus of texts from medical portals. The concordances of each term were then manually inspected, and five definitions per term were selected for annotation. When concordances exceeded 1000 hits, a random sample of 300 was taken. The same list of terms was then queried in the scientific medical corpus using the same procedure. The definitions were then annotated following the Framenet methodology (Ruppenhofer et al., 2016) for the frame elements of the frame Medical_conditions. In addition to annotating FEs, we determined verbal patterns and their lexical markers following the list of markers in Sierra et al. (2010). Table 1 shows two examples of terminological definitions from the two corpora, with the superordinate concept underlined. The definitional patterns established will be used for extracting definitions of other medical concepts to create a dataset of expert and non-expert definitions of medical concepts. The dataset and the typology of definitional patterns will be used for text simplification in order to create popular terminological resources for
non-experts and the general public.
