Morphological process feeding in the formation of old english nouns: Zero-derivation, affixation and compounding

Resumo

This doctoral thesis describes in an exhaustive manner the relations established between the derivative processes intervening in nominal formations in Old English, by analysing their interaction, combinability and recursivity. The aim of this work is twofold. On the descriptive side, it contributes to fill the gap on derivational studies for this stage of the language as stated by Kastovsky (1992). On the theoretical domain, the study contributes to the current debate on the position of morphology within linguistics in general and to the discussion on word formation in particular. In the first phase of the research the morphological process that gives way to each of the analysed predicates have been identified. The predicates, along with all their semantic and inflectional properties had previously been filed in the database Nerthus (www.nerthusproject.com). All in all, 16,694 predicates have been studied, of which 2,848 are underived, and 13,670 are complex. Within the complex nouns, 4,084 are derived (1,025 prefixed and 3,052 suffixed), 4,347 are compound and 1,239 are the result of zero-derivation. The second phase of the research accounts for the establishment of the relations between the different predicates, and follows from the principles of analysis that guide the research, among which recategorization, relative ordering of processes, transparency, compositionality, directionality and gradation must be included (Torre Alonso et al. 2008). The main conclusions obtained from this word include the non-existence of a relative ordering of processes, as all processes feed each other in pre-terminal and terminal operations. The study also concludes that recursive lexical creation is richer and much more frequent than non-recursive word formation. In fact, a total of 131 different word structures have been identified containing up to six levels of complexity, according to the number of processes involved. The vast majority of the analysed predicates are three or four-level structures.