Library of word classes for the English language

Description

The purpose of this project is to design a CLOS protocol that contains several things:

• A mapping that takes a string and returns a list of words, were each word is written as that string.
• A class hierarchy that determines the category to which a word belongs. Categories are inspired by traditional word classes, also known as part of speech, but it would be advantageous to use a modern categorization scheme, and perhaps even finer categories than that.
• A set of operations in the form of generic functions to determine additional information about words such as tense, mood, person, case, and gender.

While there are already several databases that provide this kind of information, for example Moby Part-of-Speech, the form of these databases is defined in terms of surface syntax encoded as sequences of characters on external storage. The purpose of this project is to define a protocol for accessing this information as an in-memory database with well defined operations.

As part of this project, we suggest providing a set of parsers for some popular existing part-of-speech databases, as well as a version of each of these databases that can be loaded into a Common Lisp system using the Common Lisp function read.

Some interesting extensions of the core protocol can be considered. For example, a function synonyms that takes a word and that returns a list of synonyms of that word, not in the form of strings, but in the form of instances (subclasses) of the class word.