doi.org/10.35321/term27-13

Oksana Smirnova FRAME-BASED APPROACH TO THE EU FINANCIAL TERMS WITH THE NOMINAL BASE RISK/RISQUE/RIZIKA IN ENGLISH, FRENCH AND LITHUANIAN (APPLICATION OF CONCEPTUAL FRAMES TO THE ANALYSIS OF ENGLISH, FRENCH AND LITHUANIAN FINANCIAL TERMS IN THE EU WITH THE NOMINAL BASES RISK / RISQUE / RIZIKA): doctoral dissertation. Vilnius: Mykolas Romeris University, 2020, 253 pp.

SIGITA RACKEVIČIENĖ

Mykolo Romerio universitetas

On 4 September 2020, Oksana Smirnova defended her doctoral dissertation at Mykolas Romeris University, entitled Frame–Based Approach to the EU Financial Terms with the Nominal Base ‘Risk/Risque/Rizika’ in English, French and Lithuanian (Application of Conceptual Frames to the Analysis of English, French and Lithuanian Financial Terms in the EU with the Nominal Base “risk / risque / rizika”) (supervisor: Prof. Dr. Sigita Rackevičienė).

In her dissertation, O. Smirnova analyzed terms denoting the concept of RISK in a trilingual parallel corpus of financial texts, using the principles of frame semantics. The aim of the dissertation is to determine which frames conceptualize situations related to risk in the financial domain and to compare the linguistic means used to express these situations in English, French, and Lithuanian. The analysis of the terms was conducted at two contextual levels: microcontext and macrocontext. In the dissertation, microcontext is defined as the closest lexical context of the nominal constituent risk / risque / rizika lexico-

288

Sigita Rackevičienė |

environment, i.e. the dependent components of the head component risk / risque / rizika, with which noun phrases functioning as multi-word terms are formed. Macrocontext is understood in the dissertation as predicate-argument structures formed by verbs, in which one of the arguments is a one-word or multi-word term containing the nominal component risk / risque / rizika.

For the purposes of the study, a parallel corpus of EU financial legislative documents in English, French and Lithuanian was compiled. The corpus contains almost 3,000,000 words. For the microcontext analysis, 203 multi-word financial terms containing the nominal component risk / risque / rizika were selected (70 English terms and their French and Lithuanian equivalents). For the macrocontext analysis, 4,314 English sentences containing 45 verbs, 4,360 French sentences containing 54 verbs, and 2,075 Lithuanian sentences containing 25 verbs were collected and annotated. These verbs form predicate-argument structures in which one of the arguments is a one-word or multi-word term containing the nominal component risk / risque / rizika.

The dissertation consists of an introduction, theoretical and methodological parts, a description of the empirical study, conclusions, a bibliography, appendices, and abstracts in English and Lithuanian. The appendices present the empirical material collected for the study. They include: a list of collocates of the English term risk and statistical data on the use of collocations in the corpus; a list of the English multi-word terms selected for analysis and their French and Lithuanian equivalents; and models of the predicate-argument structures that realise the frames identified in the analysis, together with examples of their expression in the three languages.

The introduction defines the key parameters of the study, as well as the dissertation’s scholarly novelty, value, relevance and practical applicability of its findings. The author emphasises that the dissertation is significant in several respects: 1) the methodology presented for revealing the cognitive structure of a specialised-domain concept makes it possible to understand the relationships between that concept and related concepts within the specialised domain; 2) the models of English, French and Lithuanian multi-word financial terms identified in the dissertation provide useful information for terminology developers and translators, as well as for natural language processing specialists developing automatic term-recognition programmes; 3) the predicate-argument structures characteristic of the financial domain, which include terms, identified in the study, and their systematically organised description are important for translators because they help them formulate coherent passages of text; the semantic-syntactic models may also be used in machine-

Terminologija | 2020 | 27

289

in machine translation programs; 4) the trilingual parallel corpus created for the dissertation is publicly available in the CLARIN repository and can be used for various research purposes; 5) the dissertation’s methodology and results can be used to develop terminological dictionaries based on research into specialised corpora.

The first part of the dissertation presents the theoretical basis of the study, i.e. the principles of the linguistic theories relevant to the research (cognitive linguistics, frame semantics, corpus linguistics, and contrastive linguistics) and their application in terminology research. The second part of the dissertation presents the empirical research methodology: the principles for compiling corpora, selecting the main term under investigation, and analysing the selected terms.

The third part of the dissertation describes the analysis of the empirical data collected. This part consists of two chapters devoted to the analysis of the microcontext and macrocontext of the term risk/risque/rizika.

At the microcontext level, the linguistic analysis focused on the immediate context of the term risk / risque / rizika, i.e. its collocates. The study aimed to classify multi-word terms with different collocates of the main nominal component risk / risque / rizika into semantic categories, based on the principles of frame semantics. The study was based on an analysis of the dependent-component positions opened by the noun risk / risque / rizika, which revealed the number and nature of the dependent components and their semantic relationship to the main nominal component. Based on the results obtained, the most abstract frame from the FrameNet project, representing situations related to risk, was selected for further analysis: RISK_scenario. It was partially modified and adapted to the financial domain. The elements of this frame formed the basis for the semantic classification of multi-word terms with the main component risk / risque / rizika: the terms were grouped into semantic categories according to which frame element their dependent components described. A quantitative analysis of terms using several parameters was conducted, helping to establish their distribution across semantic categories and according to their use in financial texts. This analysis revealed the multifaceted nature of the superordinate concept RISK, its relationship with subordinate concepts, and which objects receive the greatest attention in the financial domain. Following the semantic analysis of multi-word terms, a formal structural analysis of those terms was conducted, revealing the diversity of linguistic expression of universal semantic patterns in the three languages.

290

Sigita Rackevičienė |

The macrocontext-level analysis focused on the positions of single-word terms risk / risque / rizika and multi-word terms with the main component risk / risque / rizika in predicate-argument structures. The study aimed to reveal the relationship between the concepts denoted by the terms and other concepts in the financial domain. The study was based on an analysis of verbs that combine with terms containing the nominal component risk / risque / rizika. These verbs form predicate-argument structures that activate frames representing situations related to risk. Based on the results of the verb analysis, the most suitable frames for the study were selected from the FrameNet project and adapted to the material under investigation. The frames were grouped into categories and stages through which risk progresses from its onset to the point at which specialists take measures to manage it. Two frames were identified as revealing the first stage of risk perception (the emergence and experience of risk: CAUSE_RISK, EXPERIENCE_RISK), and two as revealing the second stage of risk perception (identifying and assessing risk, and managing and controlling it: PERCEIVE and ASSESS_RISK, CHANGE and CONTROL_RISK). After the predicate-argument structures were assigned to the corresponding frames, a three-level analysis was conducted (analysis of frame elements and their interrelations, analysis of the syntactic realisation of frames, and analysis of their lexical realisation), revealing both universal patterns of expression and features characteristic of the languages under investigation.

The fourth part of the dissertation discusses issues concerning the achievement of semantic equivalence in the trilingual parallel corpus used in the study. This part identified and summarised the means used in the languages studied to ensure semantic equivalence, as well as the differences between them encountered throughout the research conducted for the dissertation.

The fifth part of the dissertation presents the final results of the study—the cognitive map of the concept RISK, which shows the relationship between macroframes (expressed through the macrocontext of terms) and the microframe (expressed through the microcontext of terms). The cognitive map reveals the process by which the concept RISK is understood in the financial domain and the interrelationships between participants in situations related to risk.

The dissertation’s conclusions highlight the advantages of applying a methodology based on the principles of frame semantics to the multilingual analysis of financial terminology: this methodology enables semantic-syntactic analysis, providing important information both about the concepts denoted by these terms and their relationships with other concepts, and about the linguistic means of expression used

Terminologija | 2020 | 27

291

to express these concepts and their relationships. A microcontext analysis of the term risk / risque / rizika based on the principles of frame semantics revealed the relationships among specific RISK concepts denoted by multi-word terms, while a macrocontext analysis based on the principles of frame semantics showed which processes and actions are associated with the concept of RISK and which other concepts it interacts with.

The conclusions also discuss the results of the analysis of linguistic means of expression, which make it possible to compare data from the languages under study. A comparative analysis of the formal structure of multi-word terms revealed term-formation trends in the languages examined: the predominant term type by number of components in all three languages is the two-word term; these make up the largest share of the English, French and Lithuanian terms collected for the study, indicating that EU term creators adhere to the fundamental principle of linguistic economy (brevity of terms); nominal modifiers are joined to the head component in different ways—most commonly by direct combination in English and Lithuanian (credit risk, kredito rizika), whereas in French they are joined using the preposition de (risque de crédit). A comparative analysis of the syntactic expression of frames revealed that most semantic-syntactic patterns coincide in the languages under study. The large number of coinciding patterns may be explained by the fact that the research material was collected from parallel EU document texts, in which equivalence between texts is sought at both the semantic and syntactic levels.

The conclusions also note that in the trilingual parallel corpus, semantic equivalence is ensured differently at the microcontext and macrocontext levels. At the microcontext level, semantic equivalence is achieved through identical or very similar linguistic means: most risk-related concepts are expressed by multi-word terms consisting of nominal phrases; moreover, in all three languages, the multi-word terms are identical in their lexical structure (number of components and their meanings), although in rare cases the number of components is greater in one of the languages (Lithuanian). At the macrocontext level, Lithuanian stands out: semantic equivalence is often achieved by using syntactic means different from those in English and French, such as verbal rather than nominal constructions and active rather than passive constructions. This can be explained by the different structures of the languages under study.

292

Sigita Rackevičienė |

Summarizing the research findings, the dissertation author notes that an analysis of a selected group of financial terms based on the principles of frame semantics made it possible to construct a cognitive map reflecting the process by which the concept of RISK is understood in the financial domain. Analysis of predicate-argument structures helped configure the cognitive map’s main conceptual parameters, while analysis of the nouns filling the positions opened by verbs helped reveal the conceptual microenvironment. The cognitive map showed that participants in risk situations are closely interconnected, and their roles in relation to one another may change depending on the situation: the same participants may cause risk in one situation, experience it in another, and participate in risk-management processes in yet another.

O. Smirnova’s dissertation Frame-Based Approach to the EU Financial Terms with the Nominal Base ‘Risk/Risque/Rizika’ in English, French and Lithuanian (Application of conceptual frames to the analysis of English, French and Lithuanian financial terms in the EU with the nominal base risk / risque / rizika) is the first work of this scope in Lithuanian terminology studies to apply frame-semantic methodology to the study of terminological data. Thus, this dissertation brings the work of Lithuanian terminologists and Lithuanian-language terminology data into the global research landscape based on the principles of frame semantics, opening up opportunities for international projects aimed at revealing mechanisms for understanding realities in specialized fields and processes of constructing and conceptualizing domain knowledge. The dissertation was written in English, enabling its findings to be widely disseminated.

Received 2020-11-29

Sigita Rackevičienė

Mykolo Romerio universitetas

Ateities g. 20, LT-08303 Vilnius

Email: [email protected]

Terminologija | 2020 | 27

293