Please register to attend, preferably the day before the lecture: https://nettskjema.no/a/eveninglecture
To participate by Zoom, please use this link:
https://uio.zoom.us/j/67949975547
About the lecture
How does the Latin language change when we move from the domain of verse to the records of medieval bureaucracy, or from a philosophical treatise to a court room defence speech?
In this talk, I’ll explore Latin use across millennia through the lens of register variation—the systematic ways that language adapts to different communicative situations (Biber, 1988).
Using the method of key feature analysis based on the cohen’s d effect score (Egbert & Biber, 2023) and six large dependency-parsed Universal Dependencies treebanks, I show which grammatical features—parts of speech, morphological features, and syntactic relations—are statistically under- or over-represented in ten different registers of Latin.
Much research has been done on Latin language use in different contexts, but a large-scale, robust computational comparison of grammatical features across time has so far been missing. It is important to test the accuracy of the KFA method on well-known texts and then extend the method to new, unexplored texts.
So far, using the one-million-word corpus, my analysis confirms earlier findings, e.g., the proportionally higher use of proper nouns in late medieval charters to unambiguously identify sellers and locations. I have also found new perspectives, such as there are less distinct register features in classical historiography, where individual authorial style results in fewer grammatical features compared with e.g. scholastic texts. Quite naturally the historical register has a large amount of punctual past tense (perfect). Letters feature a large quantity of adverbs, possibly reflecting a vivid and colloquial tone, and so on.
Sources
Biber, Douglas. 1988. Variation across Speech and Writing. Cambridge: Cambridge University Press.
Egbert, Jesse, and Douglas Biber. 2023. “Key Feature Analysis: A Simple, Yet Powerful Method for Comparing Text Varieties.” Corpora 18 (1): 121–133. https://doi.org/10.3366/cor.
Universal Dependencies for Latin: https://universaldependencies.