PhD_thesis

This repository contains all my work for my PhD thesis "Co-occurrence of modal markers in Latin: a quantitative and qualitative analysis". My Phd thesis is part of the wider project "WoPoss. A world of possibilities. Modal pathways over an extra-long period of time: the diachrony of modality in the Latin language", funded by the Swiss National Science Foundation (SNSF n° 176778) for the period spanning from February 2019 to January 2023, and hosted at the University of Neuchâtel, Institut des sciences du langage (ISLa).

Project structure

Notes about SOURCE FILES

• Sallust contains Bellum Iugurtinum; Bellum Catilinae; but also Orationes and Epistulae (not part of the corpus, delete those texts from processing)
• Cicero's Philippicae also contains Pro Milone, Pro Marcello, Pro Ligario, Pro rege Deiotaro. These are not part of the corpus, delete them before text processing

Notes about TXT FILES

• histo_all contains also Bellum Hispanicum, which takes it up to 132k tokens. This probably unbalances the corpus. none of the co-occurrences extracted from this work is modal. probably take it out for the final version of the corpus
• histo_nohisp.txt < txt_files is the file with historiographic texts without Bellum Hispanicum

Notes about quantitative

As can be seen from the name of the output files, all the outputs for quantitative analysis were obtained on the corpus without the Bellum Hispanicum.

Notes about SCRIPTS

The scripts are renamed incrementally: es. extract_sentences_old3.py is more recent than extract_sentences_old2.py.

• extract_sentences_old.py is missing the calculation of percentage in the stats output file
• extract_sentences_old2.py still checks on "dubio" and "certo" as lemmas. This leads to falso co-occurrences with the verb "certo, -are". The adverbial forms of the two adjectives are lemmatised under the lemma of the adjective
• extract_sentences_old3.py is missing the manual check on fortasse: the lemmatisation of fortasse is almost always wrong, so I check on the form instead of the lemma. With this code:
  • 31 new sentences (with fortasse)
  • 9 were already in the previous output file
  • 22 new sentences with fortasse NB. there is one occurrence of fortassis, which is lemmatised as fortasse: so I leave fortasse in the list of lemmas in order to get the occurrences of fortassis

Notes about the annotation

The lemmatisation is systematically wrong for:

• fortasse, usually forio or capital Fortasse
• valeo and all conjugated forms: it lemmatises as the form. So it is great if we want only 3rd person sg. as the WoPoss list says, not great if we are looking for all persons.
• adverbs derived from adjectives are lemmatised with the corresponding adjective e.g. certo ADV is lemmatised as certus (same for dubio)