Announcements

Contribute to the Greek and Latin Treebanks at Perseus!

We are currently looking for advanced students of Greek and Latin to contribute syntactic analyses (via a web-based system) to our existing treebanks. We particularly encourage students at various levels to design research projects around this new tool. We are looking in particular for the following:

Get paid to read Greek! We can have a limited number of research assistantships for advanced students of the languages who can work for the project from their home institutions. We particularly encourage students who can use the analyses that they produce to support research projects of their own.
We also encourage classes of Greek and Latin to contribute as well. Creating the syntactic analyses provides a new way to address the traditional task of parsing Greek and Latin. Your class work can then contribute to a foundational new resource for the study of Greek and Latin - both courses as a whole and individual contributors are acknowledged in the published data.
Students and faculty interested in conducting their own original research based on treebank data will have the option to submit their work for editorial review to have it published as part of the emerging Scaife Digital Library.

To contribute, please contact David Bamman (david.bamman@tufts.edu) or Gregory Crane (gregory.crane@tufts.edu).

The Ancient Greek Dependency Treebank

The Ancient Greek Dependency Treebank is a 192,204-word collection of syntactically parsed Greek sentences. Currently in version 1.1, the treebank is comprised of selections from Hesiod's Works and Days, Homer's Iliad and Odyssey and the works of Aeschylus.

Author Work Word count

Hesiod Works and Days 6,303

Homer Iliad 38,390

Odyssey 99,353

Aeschylus Agamemnon 9,796

Eumenides 6,376

Libation Bearers 6,563

Persians 6,223

Prometheus Bound 7,045

Seven Against Thebes 6,206

Suppliants 5,949

Total192,204

The Latin Dependency Treebank

The Latin Dependency Treebank is a 53,143-word collection of syntactically parsed Latin sentences. Currently in version 1.5, the treebank is comprised of excerpts from eight authors, in the following distribution:

Author Work Word count

Caesar B.G. (Book 2 selections) 1,488

Cicero In Catilinam 1.1-2.11 6,229

Jerome Vulgate: Apocalypse 8,382

Ovid Metamorphoses: Book I 4,789

Petronius Satyricon 26-78 (Cena Trimalchionis) 12,474

Propertius Elegies: Book I 4,857

Sallust Catilina 12,311

VergilAeneid (Book 6 selections) 2,613

Total53,143

Format

The texts in this collection are syntactically annotated under the formalism of a Dependency Grammar (cf. Mel'cuk 1988), in which individual words themselves are linked to their heads. This leads to structures that are familiar to traditional Classical grammars (in which an adjective, for example, "modifies" or "depends on" its head noun), as in the following annotation of ista meam norit gloria canitiem ("that glory will know my old age") from Propertius 1.8.

To represent this structure in a flat text document, we encode it using XML, as illustrated below. Here ista, the first word in the sentence (id="1") modifies the fourth word (id="4"), gloria, via the syntactic relation ATR (attributive). See the README document and syntactic guidelines below for more information on the explicit format and the syntax involved.

	<sentence id="2598662" document_id="Perseus:text:1999.02.0066" subdoc="book=1:poem=8b" span="ista0:canitiem0">
		<word id="1" form="ista" lemma="iste1" postag="p-s---fn-" head="4" relation="ATR" />
		<word id="2" form="meam" lemma="meus1" postag="a-s---fa-" head="5" relation="ATR" />
		<word id="3" form="norit" lemma="nosco1" postag="v3srsa---" head="0" relation="PRED" />
		<word id="4" form="gloria" lemma="gloria1" postag="n-s---fn-" head="3" relation="SBJ" />
		<word id="5" form="canitiem" lemma="canities1" postag="n-s---fa-" head="3" relation="OBJ" />
	</sentence>

Uses

The treebank data provides information about the syntactic relationship of every word in a sentence, along with its morphological analysis and the lemma it's derived from. In addition to providing crucial datasets for NLP tasks such as parsing and grammar induction, treebanks can be used by traditional scholars as well.

Since a treebank provides morphological and lemma information in addition to its syntactic analyses, the text it contains can be subjected to morphological and lemmatized searching (e.g., searching for est as an inflection of the verb edo rather than sum, or searching for ut + a subjunctive).
The syntactic information encoded in a treebank can be used for research in Latin stylistics, linguistics, and lexicography (see Bamman and Crane 2006 and Bamman and Crane 2007).

Student Research

Aside from the scholarly treebank of Aeschylus (described here), each sentence in the Greek and Latin Dependency Treebanks is built from the efforts of two independent annotators that are subsequently reconciled by a third. Many of our annotators are dedicated students, both graduate and undergraduate, and we are committed to engaging students in the act of scholarly research to produce scientific data that can be useful to wider Classical community. We would like to recognize the contribution of the following individuals (and class) to the creation of these treebanks and thank them for their commitment to the advancement of Classical scholarship:

Jennifer Adams, James Artz, Jennifer Curtin, James C. D'Amico, W. B. Dolan, Calliopi Dourou, Scott J. Dube, C. Dan Earley, J. F. Gentile, Francis Hartel, Connor Hayden, Kenny Hickman, Tovah Keynton, Michael Kinney, Florin Leonte, Alex Lessie, Daniel Lim Libatique, Brian Livingston, Viet Luong, Meg Luthin, George Matthews, Molly Miller, Skylar Neil, Robin Ngo, Jessica Nord, Anthony D. Yates and the Tufts University LAT-181 class (Spring 2008).

To contribute, please contact David Bamman (david.bamman@tufts.edu) or Gregory Crane (gregory.crane@tufts.edu).

Downloads

All files

latin-1.5.tar.gz
greek-1.1.tar.gz

Individual files

Greek Latin

treebank-1.1.xml treebank-1.5.xml

treebank-1.5.xsd treebank-1.5.xsd

README.txt README.txt

guidelines.pdf guidelines.pdf

Publications

Bamman, David, Francesco Mambrini and Gregory Crane (2009), "An Ownership Model of Annotation: The Ancient Greek Dependency Treebank," Proceedings of the Eighth International Workshop on Treebanks and Linguistic Theories (TLT8) (Milan), forthcoming [pdf]

Bamman, David and Gregory Crane (2008), "Building a Dynamic Lexicon from a Digital Library," Proceedings of the 8th ACM/IEEE-CS Joint Conference on Digital Libraries (JCDL 2008) (Pittsburgh) [preprint]

Bamman, David and Gregory Crane (2008), "The Logic and Discovery of Textual Allusion," Proceedings of the Second Workshop on Language Technology for Cultural Heritage Data (LaTeCH 2008) (Marrakech) [preprint]

Bamman, David, Marco Passarotti, Gregory Crane and Savina Raynaud (2007), "A Collaborative Model of Treebank Development," Proceedings of the Sixth International Workshop on Treebanks and Linguistic Theories (TLT2007) (Bergen) [pdf]

Bamman, David and Gregory Crane (2007), "The Latin Dependency Treebank in a Cultural Heritage Digital Library," Proceedings of the Workshop on Language Technology for Cultural Heritage Data (LaTeCH 2007) (Prague: Assocation for Computational Linguistics), pp. 33-40 [pdf]

Bamman, David and Gregory Crane (2006), "The Design and Use of a Latin Dependency Treebank," Proceedings of the Fifth International Workshop on Treebanks and Linguistic Theories (TLT 2006) (Prague), pp. 67-78 [pdf]

Bamman, David, Marco Passarotti, Gregory Crane and Savina Raynaud (2007) "Guidelines for the Syntactic Annotation of Latin Treebanks," whitepaper (version 1.3) [pdf]

Author	Work	Word count
Hesiod	Works and Days	6,303
Homer	Iliad	38,390
	Odyssey	99,353
Aeschylus	Agamemnon	9,796
	Eumenides	6,376
	Libation Bearers	6,563
	Persians	6,223
	Prometheus Bound	7,045
	Seven Against Thebes	6,206
	Suppliants	5,949
Total192,204

Greek	Latin
treebank-1.1.xml	treebank-1.5.xml
treebank-1.5.xsd	treebank-1.5.xsd
README.txt	README.txt
guidelines.pdf	guidelines.pdf