UD for Umbrian 
Introduction
Umbrian is an Indo-European language of the Italic branch. As such it shares a number of characteristics with classical IE languages and especially with Latin. The main similarities between Umbrian and Latin are their declension and conjugation systems. The main difference, beside phonology, is the extendisve use of cliticised postpositions in Umbrian where Latin has plain prepositions.
Tokenization and Word Segmentation
The Iguvine tablets use a word separator to (: in the Umbrian script and ⋅ in the Latin script). We thus follow native word segmentation as much as possible. The main exceptions are :
- when the original segmentation itself is erroneous (e.g. pesni:mu for pesnimu on tablet II face b);
- cliticised adpositions that we decided to separate from their host (e.g. the ubiquitous tutaper is analysed as tuta per).
Morphology
Tags
PUNCTis not used in Umbrian (note that there are word boundaries in the original text but no sentence boundaries).
Features
NOUN,PRON,PROPN,ADJ, andDETare marked withCaseandNumber, andGenderwhen it is known.VERBis marked withVerbForm, andTense,Mood,Person,NumberorCase,Gender,Numberdepending on the finiteness of the form. ** Note that verbs have a future perfect form which comes from a very reduced periphrastic construction. Until we find a better solution, we decided to useAspect=Perin conjunction withTense=Futfor these cases.
Syntax
- Core arguments are identified with case (
NomandAcc) and in absence of case triggering adposition (rupinam-e isobleven if rupinam isAccbecause of the adposition e).
Treebanks
There are 1 Umbrian UD treebanks: