No description available.

The term morphological doubletism refers to a situation in language when there are two (or more) morphemes available for a single cell in an inflectional paradigm of a lexeme. Slavonic languages, with their rich inflectional systems, show particularly high levels of doubletism. In the present dissertation we analyse examples of doubletism in Croatian nominal paradigms. As shown by the dissertation’s subtitle, “a multi-methodological analysis”, we compare and contrast evidence obtained by various methods. First we conduct a corpus study to determine the frequency distributions of the doublet pairs in present-day Croatian. This analysis has shown that the distribution of the doublet pairs is not determined by any intra- or extra-linguistic factor, but that it is not completely random either. These distributions are later used in several additional studies, the purpose of which is to answer the question of how such forms are processed in speakers’ mental grammars. One of the analyses is a computational one, in which we try to reproduce a grammar of a Croatian speaker by using two memory based models (AM and TiMBL). The models were highly successful in producing the desired output without resorting to any rules or generalizations. We also report the results of three questionnaire studies, all of which show that native speakers are extremely sensitive to the language input they receive, in line with usage-based theories of language, as well as that mental grammars are gradient. The speakers’ ratings and production rates closely matched the proportions of the doublet pairs in the corpus. Furthermore, speakers distinguish between several levels of domination of one ending over another. When the domination of one form is weak, speakers resort to a different decision criterion, namely they look at the dominant ending of phonologically similar words.

Examples of the construction are found in Old Ukrainian documents and in modern South Russian dialects, and, although these are few and their usage differs from that found in North Russian sources (where the construction is well documented), it cannot be said that the construction has always been restricted to northern and central dialects. In almost every text both nominative and accusative are used for the direct object of independent infinitives, and often the choice of case seems to be random. From the sixteenth century, however, the accusative gradually becomes more frequent until it displaces the nominative from the written language.

This study examines attitudes towards anglicisms in Russian expressed in print media articles. Accelerated linguistic borrowing from English, a particularly visible aspect of the momentous language changes after the breakdown of the Soviet Union, has engendered a range of reactions. Print media articles spanning two decades and several central outlets are analysed to show how arguments for or against use of anglicisms are constructed, what language ideologies these arguments serve, and whether mediated language attitudes changed during the post-Soviet era. A summary of the history of Russian linguistic borrowing and language attitudes from the Middle Ages to the present day shows that periods of national consolidation provoked demands for the restriction of borrowing. Then, a survey of theories on language ideologies demonstrates that they function through the construction of commonsense argumentation in metadiscourse (talk about talk). This argumentation draws on accepted common knowledge in the Russian linguistic culture. Using critical discourse analytic tools, namely analysis of metaphor scenarios and of argumentation, I examine argumentative strategies in the mediated language debates. Particularly, the critical analysis reveals what strategies render dominant standpoints on anglicisms self-evident and logical to the audience. The results show that the media reaction to anglicisms dramatises language change in discourses of threat, justified by assumed commonsense rational knowledge. Whilst there are few reactions in the 1990s, debates on language intensified in the 2000s after Putin’s policies of state reinforcement came into effect, peaking around times of official language policy measures. Anglicisms and their users are subordinated, cast out as the Other, not belonging to the in-group of sensible speakers. This threat is defused via ridicule and claiming of the moral high ground. This commonsense argumentation ultimately supports notions of Russian as a static, sacred component of Russian nation building, and of speakers as passive. Close textual analysis shows that even articles claiming to support language change and the use of anglicisms use argumentation strategies of negativisation. Overall, a consensus on the character and role of the Russian language exists between all perspectives, emphasising the importance of rules and assigning speakers a passive role throughout.

Dans cette thèse, nous nous intéressons aux trois types d’expressions dont la structure s’écarte du modèle de la proposition transitive canonique, à savoir les constructions : anticausatives, à causalité interne, et impersonnelles en polonais et en russe. Leur point commun est que l’écart par rapport à une phrase transitive porte sur la référence du sujet de la proposition. Dans tous les cas, le référent du sujet est non-agentif. Dans certains cas, le marquage morphologique reflète cette absence, tant sur le prédicat qui est marqué à la troisième personne du singulier neutre que par l’absence d’un sujet nominatif. Ailleurs, la morphologie ne laisse rien, ou presque rien apparaître, le prédicat s’accordant avec l’argument Thème ou le Patient. Ces expressions feront l’objet d’une étude comparative, dans laquelle seront étudiées : la nature de chaque participant non-canonique qui apparaît dans la proposition, les caractéris¬tiques de chaque prédicat présent, et enfin, les propriétés combinatoires entre les participants et les prédicats. Il sera montré que la nature et la structure et enfin la manière dont sont légi¬timés les sujets non-canoniques ne sont pas seulement liées aux propriétés du réfé¬rent de sujet, mais aussi dépendent de la nature du prédicat. Tout comme cela a été proposé pour les constructions anticausatives et à causalité interne (cf. Alexiadou & Anagnostopou¬lou (2003), Alexiadou, Anagnostopoulou & Schäfer (2006), Schäfer (2008)), nous avancerons que les constructions impersonnelles comportent un prédicat Cause qui est responsable de la légitimation thématique des caus(at)eurs nominatifs et obliques. Les différences entre ces structures sont dues à la présence ou à l’absence de la tête Voice. Vues sous cet angle, les constructions impersonnelles sont une sorte de causativisation. Le caus(at)eur indéfini (cf. Kibort (2004)) présent dans les constructions impersonnelles d’une part, et la possible présence des causeurs obliques d’autre part, sont des preuves qui étaient cette hypothèse. / In this dissertation we study three kinds of expressions that deviate from the model of canonical transitive sentences, namely anticausative, internal causation and impersonal constructions in Polish and in Russian. The common point is that the deviation from a transitive sentence has to do with the subject’s reference. In all cases, the referent of the subject is non-agentive. In some cases, the morphological marking reflects this absence, both on the predicates which is frozen in the third person singular (neuter), and by the absence of a nominative subject. In the other two cases, the non-canonicity is not morpologically revealed because the inflected verb agrees with the nominative argument whose referent is the Theme or the Patient.A comparative study of these expressions will be given, in so far as the nature of the non-canonical participants that appear in the sentence, the characteristics of each predicate, and finally, the combinatorial properties between participants and predicates are concerned.It will be shown that the nature, the structure and ultimately the way that non-canonical subjects are licensed are not only related to the properties of the subject’s reference, but also depend on the nature of the predicate. Just as it was proposed for anticausatives and internal causal constructions (Alexiadou & Anagnostopoulou (2003), Alexiadou, Anagnostopoulou & Schäfer (2006), Schäfer (2008)), it will be shown that impersonal constructions include a Cause predicate that is responsible for the thematical licensing of nominative and oblique causers. The differences between these structures are due to the presence or absence of a Voice head. From this perspective, impersonal constructions encode a kind of causativisation. The indefinite causer (analysed in Kibort (2004)), which is present in impersonal constructions on the one hand, and the possible presence of other oblique causer, are thus deemed to be evidence that supports this hypothesis.