Skip to content

Latest commit

 

History

History
18 lines (9 loc) · 1.62 KB

README.md

File metadata and controls

18 lines (9 loc) · 1.62 KB

Irish Sentence Bank

This is a collection of approximately 4,500 sentences in Irish, tokenized and lemmatized, with English translations. Most of the sentences have been hand-picked from published language-learning texts and would serve well as example sentences in a dictionary. Many of the sentences do in fact occur as example sentences in the Pota Focal House Glossary, but many others don't. The tokenization and lemmatization has been done by hand and is believed to be to practically 100% correct.

License

Available under the Open Database License.

Format

Each sentence is enclosed in a <sentence> tag, consisting of an <original>, a <translation> and a number of <flex> elements. The <original> tag contains the sentence in Irish. It consists of a sequence of <token> tags interlaced with whitespace and punctuation. Each <token> tag has an index number in its @slot attribute, and its corersponding lemma is given in a <flex> tag with the same @slot number.

Notes

  • The relation between tokens and lemmas is one-to-many: in some cases a single token corresponds to more than one lemma, for example the token "ón" ('from the') corresponds to the lemmas "ó" ('from') and "an" ('the').

  • In some cases of indeterminacy I had to take an executive decision as to what is a lemma and what isn't, whether a token is a word form of another lemma or whether it is a lemma in its right, and so on. My decisions may have been arbitrary but they are consistent: I have tried to apply the same thinking to all sentences.