|
Morphology without morphemes: complexity, typology and evolution| old_uid | 19587 |
|---|
| title | Morphology without morphemes: complexity, typology and evolution |
|---|
| start_date | 2021/10/19 |
|---|
| schedule | 10h-12h |
|---|
| online | no |
|---|
| location_info | salle Marc Bloch |
|---|
| details | Dans le cadre DILIS |
|---|
| summary | When segmenting inflected words into morphemes, linguists usually rely on conventions inherited from specific grammatical traditions, and draw on complex intuitions informed by diachronic evidence and by their own sense of grammatical elegance. As a result, these analyses are not meant to isolate units which are comparable across languages. How, then, can we ensure the comparability necessary to study the typology of inflectional systems ?
The Word-based approach to morphology provides a useful perspective: It takes paradigms to be structured networks of words which can be insightfully studied by characterizing recurrent, empirical formal relationships between them, without the need to reify traditional, abstract constituents such as stems or morphemes. My project aims to leverage computational tools to study these relationships quantitatively, and address key questions in morphology: How can speakers produce and recognize unknown inflected words ? How does the organization of inflectional exponence vary across languages ? How does analogical change contribute to paradigm evolution ?
I approach these questions by collecting and curating large, unsegmented, inflected lexicons. I then write computational tools which produce comparable descriptions such as (non-morphemic) segmentations and analogical patterns. In turn, these descriptions constitute the basis for quantitative models of paradigm organization and change, and for precise measures of linguistic variation. |
|---|
| responsibles | Coupé |
|---|
| |
|