AIM: NLP parser of Old English. Attempts a best guess, rather than certainty.
TAKES: a sentence in Old English
RETURNS: list of words with grammatical tags
Glossaries.
- The files beginning with `data_` are compiled from the main dictionaries of Old English.
Each entry in the JSON array comprises the OE lexeme (that is, the dictionary entry) and its most likely part of speech (POS).
For example, the entry "witscipe" : "NOUN m"
The OE word witscipe 'knowledge, evidence' is the KEY. The part of speech is the VALUE—here a noun of the masculine gender.
Word Lists.
-
The following files were culled from the glossaries, and all nouns put into one files, all adjectives into another, and so on:
- 1. lemmas.txt (all words in dictionary form)
- 2. nouns.txt
- 3. prefixes.txt
- 4. suffixes.txt
- 5. verbs.txt
First, I check to see if an OE word is a value on a list.
Second, I check to see if it is also on another list.
- Some words can be both nouns and adjectives. OE *god* can mean either 'good' or 'God'.
Other words can be nouns or conjunctions! OE *ac* means both 'oak' and 'but'. OE *þa* can be a pronoun,
an adverb, or a conjunction.
Third, I check the OE word for tell-tale signs of its POS (prefixes, suffixes, etc.). Prefixes are listed in a file called prefixes.txt. Suffixes are listed in suffixes.txt.
Fourth, I check the environment of the OE word. Does the word follow a preposition (PRP), for example?