Towards a Typology of Lexical Semantic Derivation Models: Linguistic and Psycholinguistic Aspects Типологія моделей семантичної деривації лексики: лінгвістичний та психолінгвістичний аспекти

Objective. The paper focuses on lexical semantic derivation models and the ways they are explicated in experiential vocabulary. The study substantiates linguistic and psycholinguistic features of the models in the contrastive aspect (based on the Ukrainian and Polish languages). Materials and Methods. The material for the analysis is the Ukrainian and Polish adjectives-stimuli that represent the concepts of experiential situation. The psycholinguistic validity of the models is verified based on the results of associative experiments. The results are supposed to help establish types of associations that underlie the experiential vocabulary extensions. Results. The study appeals to the conceptions that uphold the idea of a dynamic conceptualization of the world of discourse (of a certain situation or its fragment). The Role of Personal Internet Orientat ion in the Process . . . 9 © Demenchuk Oleh On that ground, the lexical semantic derivation models are considered as theoretical constructs that represent information on derivational strategies in Ukrainian and Polish. Such an approach has allowed not only to reveal cognitive mechanisms that underlie lexical (experiential) items’ semantic extensions but also to systematize the semantic shifts that occur in the contrasted languages. The extensions are considered within four lexical semantic derivation models: componential (the level of semantic components and their configurations), actantial (the level of predicate-actant structures), topological (the level of image-schema transformations) and constructional (the level of lexical constructions). Conclusions. The results show the relevance of both linguistic and psycholinguistic modelling in analysis of lexical semantic derivation. The choice of a semantic derivation strategy correlates with an informant’s response to a word-stimulus. The similarities and differences in semantic shifts are determined by constraints that set limits on the target situation extensions.

Of paramount importance at present is the issue of combining methodologically adjacent approaches on semantic derivation modelling. The relevance of such an approach is determined by the universal and integral nature of semantic derivation: (a) regular semantic shifts are found in many (if not in all) languages of the world, irrespective of their phylogenic and cultural differences (Wierzbicka, 1992), cf.: © Деменчук Олег «Polysemy is in all probability a semantic universal inherent in the fundamental structure of language» (Ullmann, 1963: 183); (b) semantic shifts are the same at any linguistic level (lexical, wordbuilding, grammatical), as they use common semantic mechanisms (Rakhilina, 2002: 372); (c) semantic shifts are psychologically valid (Radden & Kövecses, 1999: 25), as lexical ambiguity is considered «a central problem of psycholinguistics» (Klepousniotou, Titone & Romero, 2008: 1534, cf.: «From a psychological and anthropological perspective, contiguity (and similarity/contrast) seem to be absolutely fundamental associative relationships» (Koch, 1999: 144). According to the original hypothesis, the modelling of a linguistic item's semantics is realized through a sampling of alternative, hierarchically established semantic dimensions. On that ground, the models of lexical semantic derivation are interpreted as theoretical constructs that represent various ways a certain situation or its fragment is reconceptualized. Within cross-linguistic studies, the models of lexical semantic derivation are supposed to reveal the ways the information on the worldview changes (shifts) is encoded and distributed in lexical semantics of related and non-related languages. Such an approach agrees with the idea of a segmentation of the physical continuum «within which languages may draw either the same or a different number of boundaries and within which they may draw the boundaries at the same or different places» (Lyons, 1968: 58).
The purpose of the paper is to characterize the types of lexical semantic derivation models and to substantiate the psycholinguistic validity of the models in the related (Ukrainian and Polish) languages.
The purpose is fourfold: -to represent a methodological background and method for the analysis of lexical semantic derivation models; -to characterize lexical semantic derivation models and to establish the ways they are explicated in the Ukrainian and Polish languages; -to substantiate the psycholinguistic validity of lexical semantic derivation models in the contrasted languages; -to outline the prospects for the study of lexical semantic derivation models.

Methodological Background and Method
The study of lexical semantic derivation models is presumably based on the following methodological prerequisites: (a) lexical semantic derivation models are models of situations; (b) a model of situation represents the content of a lexical item through a sampling of a multidimensional situation concept that encodes information on the ways a situation is construed; (c) a situation is construed based on the reconceptualization strategies a designator resorts to -identifi cational, evaluative, associative, etc. (those strategies determine the relations between the source and target states of affairs).
Lexical semantic derivation models are considered to correlate with a set of alternative, hierarchically established semantic dimensionsthe aspects that specify the way the information on a situation reconceptualization is distributed in the semantic space of a lexical item. We claim the information to be distributed within four types of lexical semantic derivation models -componential, actantial, topological and constructional.
In order to substantiate the psycholinguistic validity of the models in the contrasted languages, we shall apply to the method of associative experiment. The experiment is supposed to establish (verify) types of semantic associations, as well as to reveal general cognitive mechanisms that underlie the experiential items' semantic extensions.

Results and Discussion
We will make use of the results of the word association tests, carried out by the Department of Computational Linguistics at Jagiellonian University in Kraków (http://www.klk.uj.edu.pl/sssjp/en); see also (Gatkowska, 2017), and of those conducted at Nizhyn Teachers University (http://it-claim.ru/Projects/ASIS/SAS /index.html).
The material for the analysis is the Ukrainian and Polish adjectivesstimuli that represent various experiential situations, cf.: білий 'white', важкий 'heavy', черствий 'stale' vs. biały 'white', ciężki 'heavy', czerstwy 'stale'. The choice is determined by the specifi c feature of an adjective as a semiological subclass of words -adjectives consist of a list of argument positions and reveal dependency that usually holds between the predicate and its actants (Dik, 1992: 21). This suggests the relational (situational) character of the semantic structure of an adjective.
The analysis provides for the evaluation of the responses in the highlight of reconceptualization strategies. The responses are paradigmatic and syntagmatic associations that characterize a situation by a certain semantic feature (paradigmatic associations) (see Table 1): or by an actantial position of a participant (syntagmatic associations) (see Table 2): The interpretation of the associations provides for establishing: (a) shifts that occur at the level of semantic components and their confi gurations (componential model); (b) shifts that occur at the level of predicate-actant structures (actantial model); (c) shifts that occur at the level of image-schema transformations (topological model); (d) shifts that occur at the level of lexical constructions (constructional model).
The componential model provides for the associative shifts that occur between the experiential items' semantic components and their confi gurations.

Fig. 1. The extensions within the componential model
The actantial model is reduced to shift-in-focusing strategies. The strategies provide for the redistributions of attention on the situation participants, when one of the participants is foregrounded and focused upon, and the other is backgrounded and defocused (see Paducheva, 2004;Iriskhanova, 2014). The redistributions of attention may result in: (a) increase or decrease of obligatory participants in the target situation; (b) changes of the participant's referential status or deictic characteristics. These characteristics determine the way a situation concept extends. As for the concept of experiential situation, the extensions occur on a «parameter» scale -from parametric entities to non-parametric ones (see Fig. 2): PARAMETRIC → NON-PARAMETRIC

Fig. 2. The extensions within the actantial model
In differentiating between the parametric and non-parametric entities, we apply to the notion of «organic states», cf.: «The organic states are intrinsic to the meaning of a predicate: the situation is conceptualized as an integral feature of its cardinal participant, and it is not a result of some other event (stone is heavy, grass is green). The derived states, on the contrary, are conceptualized as those obtained from a certain process or activity» (see Kibrik, 1992: 136).
The topological model is reduced to the transformations that occur within basic image-schemas -the recurring patterns of our sensorymotor experience by means of which we make sense of that experience and reason about it (Johnson, 1987: 126). The partial list of schemata includes: PATH (a force vector moving from a source to terminal point), COMPULSION (experience of being moved by external forces), COUNTERFORCE (the head-on meetings of forces), BLOCKAGE (obstacles that block or resist our force), ENABLEMENT (a potential force vector and an absence of barriers or blocking counterforce), REMOVAL OF RESTRAINT (an open way or path, which makes possible an exertion of force), etc.
From this viewpoint, the extension of the experiential situation would be considered in terms of force dynamics interpretations -from a potential (latent) force to a non-potential (actual) one (see Fig. 3): © Demenchuk Oleh POTENTIAL → NON-POTENTIAL
The transformational confi gurations may also be modelled through a sampling of a trajectory-landmark alignment structure (Langacker, 1987: 227). Thus, in respect to the associations важкий (двері) vs. ciężki (głaz), the Trajector and Landmark profi le the experiential situation within the relation: external stimulus (perceptual space) vs. inner reaction (evaluative perception). The Landmark is interpreted as an indicator of a perceptual quality of an object (e.g. an object might look massive and solid), whilst the Trajector -as an (atemporal) feature of evaluative perception (the object is valued as being heavy), cf. важкий «вигляд якого (значні розміри, масивність, густота і т. ін.) свідчить про велику вагу» vs. ciężki «o budowli, stroju itp.: zbyt masywny». The shift of the Trajector relative to the perceptual space provides for the changes in the space evaluation (e.g. getting in touch with an object, the object might prove to be light), cf. the responses важкий (легкий) vs. ciężki (lekki). The Landmark also represents the symbolic space of a «heavy» object, cf. the association ciężki (grzech). The Trajector's shift relative to the space indicates the changes in the degree of evaluation, cf. важкий (обман) «значний за ступенем, силою прояву; великий, значний, який призводить до поганих наслідків».
The constructional model is reduced to the construction content, «associated <…> with semantic structures which refl ect scenes basic to human experience» (Goldberg, 1995: 5).
The semantic structures are argument structures (sort of X causes Y to receive Z) that represent a certain dynamic scene, in which argument roles are interpreted as functions derived from the event, e.g.: someone causing something to move or change state; someone experiencing something; something moving, etc. In terms of Vandeloise's «complex primitives» (Vandeloise, 1987), a scene is viewed as a set of propositions, encoding and distributing information on three experiential dimensions: perception, kinetics and interaction. Among those, a purposive dimension (interaction) is considered the most relevant to human beings (see Correa-Beningfi eld et al., 2005: 355-356).
The analysis of semantic derivation models within the constructional approach proceeds from the assumption that «it is natural for constructions to be associated with a central sense, and with extensions from that sense» (Goldberg, 1995: 203).
From this viewpoint, the extension of the experiential situation would be considered within a «purposive dimension» scale -from a canonical interaction to a non-canonical one (see Fig. 4

Conclusions
The performed study appeals to the conceptions that uphold the idea of a dynamic conceptualization of the world of discourse (of a certain situation or its fragment). The dynamic approach towards the study of experiential vocabulary aims to elaborate a system of lexical semantic derivation models. The models are considered to represent various types of semantic associations that underlie the development of experiential vocabulary's semantic paradigm.
Four types of lexical semantic derivation models are singled out -componential, actantial, topological and constructional. The componential model provides for the associative shifts that occur within the experiential items' semantic components and their confi gurations. The actantial model is reduced to the shift-in-focusing strategies based on the changes of the situation participants. The topological model provides for the image-schema transformations that underlie the extensions at the level of image-schematic concepts. The constructional model is reduced to the extensions, occurring within the experiential items' argument structures.
The results of associative experiments have confi rmed the psycholinguistic validity of the models. It has been established that in responding to a stimulus, informants apply to different semantic derivation strategies. From this viewpoint, the models represent various ways the concept, encoded in the semantics of experiential vocabulary, extends: from typical to non-typical state of affairs (componential model), from parametric to non-parametric entity (actantial model), from potential to non-potential force (topological model), from canonical to non-canonical interaction (constructional model).
The models have proved to be effective means for analyzing semantic derivation strategies in cross-linguistic studies. The analysis has revealed similarities and differences in the extensions of experiential vocabulary in the Ukrainian and Polish languages. It has been established that extensions in Polish are more localized as they reveal some constraints in the target situation extensions -they mostly cover semantic zones that encode information on person's internal (perceptual, cognitive, identifi cational) systems. In the Ukrainian language, the constraints are somewhat loosened, which enables it to extend to the boundaries of both external and internal systems.
It is necessary to carry out further research into semantic derivation modelling in comparative, historical and typological aspects. The choice of the aspects is determined by the tendency of modern semantic studies to a profound analysis of the semantic shifts in related and non-related languages.