Exploring
Polybius's syntax
with dependency trees
Neven Jovanović / neven.jovanovic@ffzg.hr
University of Zagreb
Skopje, 22 November 2016
Address of this page:
croala.ffzg.unizg.hr/2016-jovanovic-polybius/
The plan
What?
How? (treebank, database, queries)
To what end?
What?
Polybius in the Perseus Catalog
Textgroups: tlg0543
CITE id: urn:cite:perseus:author.1154
Works Currently Cataloged: Histories
URN: urn:cts:greekLit:tlg0543.tlg001
Polybius in the AGLDT
Currently:
1,001 sentences; 28,271 words
The Ancient Greek and Latin Dependency Treebank (AGLDT) is the earliest treebank for Ancient Greek and Latin. The project started at Tufts University in 2006 and is currently developed and maintained at Leipzig University-Tufts University.
License: CC-BY.
How?
(Prague) syntactic dependency treebank
A treebank is a parsed text corpus that annotates syntactic or semantic sentence structure.
Database and queries
Exploring the treebank: which...?
Which syntactic functions are present in Histories 1?
(A list of syntactic functions in the AGLDT 2 can be found here.)
How many occurrences of each syntactic function is present in Histories 1?
Report from the CroALa Polybius DB.
Exploring the treebank: configuration, complexity, frequency
How are syntactic functions construed, or configured, in Histories 1?
Which configurations are simple, which are complex?
Which configurations are more frequent, which are rare?
Report (for the relation ADV with the complexity of two dependent nodes)
Conclusion
To what end?
Available material, tools, methods
We have (Greek and Latin) texts which we can explore and transform freely, as necessary.
There are (general, industrial quality) software tools that can be adapted for philological research.
There are (general, industrial quality) procedures to control, document, and publish philological research digitally, using the internet.
An inductive Greek syntax?
All this enables us to think about a complete grammar of a text which is described inductively, starting from the actual usages, and which takes into account the frequencies of linguistic phenomena.
This grammar does not have to be a 1000-pages thick printed book.