UD_Latin-CIRCSE is a repository of treebanks featuring Latin texts natively annotated at the CIRCSE Research Centre in Milan (https://centridiricerca.unicatt.it/circse/en.html) following the Universal Dependencies (UD) (https://universaldependencies.org) annotation scheme. The repository includes prose and poetry texts from different periods.
This treebank repository is a work in progress collective endeavour. Presently, it contains the following annotated texts: Seneca Hercules Furens, Seneca Agamemnon, Tacitus Germania.
Hercules Furens is a tragedy written by Seneca the younger in 1st century CE. The source text was received with tokenisation, and annotation with respect to lemmatisation, POS tagging, and morphological features from the Opera Latina corpus built by the LASLA laboratory in Liège. In few cases, the received annotation with regard to POS tag and morphological annotation was modified during the syntactic annotation; deviations from the received annotation are detailed in the file SenecaYounger_HercF_LASLA_CIRCSE). The syntactic annotation was performed manually at CIRCSE, and follows the UD scheme. The text (7714 tokens, 555 sentences) was enhanced with the annotation of the speakers to whom each sentence is attributed (following Zwierlein 1986). (Zwierlein, O. (1986) L. Annaei Senecae Tragoediae, Incertorum auctorum Hercules (Oetaeus), Octauia. Oxford: Clarendon Press) This annotation, performed manually at CIRCSE, is formatted as a comment in the conllu file following the comment line reporting the text of the sentence, as exemplified in what follows:
#sent_id = Latin_SenecaYounger_HercF_poetry-1
#text = soror Tonantis hoc enim solum mihi nomen relictum est semper alienum Iouem ac templa summi uidua deserui aetheris locumque caelo pulsa paelicibus dedi tellus colenda est paelices caelum tenent
#speaker = Iuno
In cases where more than one speaker utters words in the same sentence, the indication of speakers details the distribution of tokens between the speakers, as exemplified in what follows:
#sent_id = Latin_SenecaYounger_HercF_poetry-291
#text = hic onere uacuam litori puppem applicans repetebat umbras poscit Alcides uiam cedente turba dirus exclamat Charon quo pergis audax
#speaker = Theseus (token 1-16), Charon (token 17-19)
Agamemnon is a tragedy written by Seneca the younger in 1st century CE. The source text was received with tokenisation, and annotation with respect to lemmatisation, POS tagging, and morphological features from the Opera Latina corpus built by the LASLA laboratory in Liège. In few cases, the received annotation with regard to POS tag and morphological annotation was modified during the syntactic annotation; deviations from the received annotation are detailed in the file SenecaYounger_Ag_LASLA_CIRCSE. The syntactic annotation was performed manually at CIRCSE, and follows the UD scheme. The text (5580 tokens, 409 sentences) was enhanced with the annotation of the speakers to whom each sentence is attributed (following Zwierlein 1986). (Zwierlein, O. (1986) L. Annaei Senecae Tragoediae, Incertorum auctorum Hercules (Oetaeus), Octauia. Oxford: Clarendon Press) This annotation, performed manually at the CIRCSE, is formatted as a comment in the conllu file following the comment line reporting the text of the sentence, as exemplified in what follows:
#sent_id = Latin_SenecaYounger_Ag_poetry-1
#text = opaca linquens Ditis inferni loca adsum profundo Tartari emissus specu incertus utras oderim sedes magis fugio Thyestes inferos superos fugo
#speaker = Thyestis umbra
In cases where more than one speaker utters words in the same sentence, the indication of speakers details the distribution of tokens between the speakers, as exemplified in what follows:
#sent_id = Latin_SenecaYounger_Ag_poetry-199
#text = sistito infestum mare uehit ista Danaos classis et Troas uehit nec plura possunt occupat uocem mare
#speaker = Danai (token 1-10), Eurybates (token 11-16)
In cases of reported speech, the character who utters the reported speech is listed as first; the character reporting the speech is enclosed in round brackets, as exemplified in what follows, where the character named Eurybates reports words uttered by the people of Danai:
#sent_id = Latin_SenecaYounger_Ag_poetry-194
#text = nil nobile ausos pontus atque undae ferunt
#speaker = Danai (Eurybates)
Germania is a treatise written by Cornelius Tacitus between 1st and 2nd century CE. The text was received with tokenisation, and annotation with respect to lemmatisation, POS tagging, and morphological features from the Opera Latina corpus built by the LASLA laboratory in Liège. The syntactic annotation was performed manually, and follows the UD scheme. The text consists of 5674 tokens, 299 sentences.
In few cases, the sentence splitting differs from the received one; deviations from the received annotation are detailed in the file Tacitus_Ger_LASLA_CIRCSE.
In adherence to UD guidelines for data split, all three texts within the treebank are maintained together as test set, due to the treebank's size.
Since the UD_Latin-CIRCSE treebank is a work in progress, its structure is subject to changes.
The annotation of Seneca Hercules Furens and Agamemnon has been conducted in the framework of the LiLa: Linking Latin project. LiLa has received funding from the European Research Council (ERC) under the European Union’s Horizon 2020 research and innovation programme – Grant Agreement No. 769994. Warmful thanks to Federica Gamba and Flavio Massimiliano Cecchini for their support and precious advices during the annotation process.