Brief: Language Acquisition ITS
Jump To: Parent Description
Code: GitHub

  • First of all, there is an immense amount of knowledge that has been written. To expand our minds we ought to learn foreign languages. We choose to learn various languages based on their strengths and weaknesses. Chinese will be useful for picform. Sanskrit is a formal language with a 5000 year history, some of which was oral. Gaelic was oral as well. Hebrew is useful for certain reasons, etc,etc,etc.

    In order to facilitate language learning we have several features. The first is our system models the various languages that we are to learn. It knows our vocabulary and is able to see patterns in the languages and teach us through analogical reasoning and custom Mnemonics. For each language a grammar is used, as well as for spoken languages a pronunciation system (TTS). The user is taught the basic 1400 or so words in each important language and continually refreshed upon them, and advanced vocabulary is developed wherever possible. They are incorporated in communication with the user to preserve use. Ancient and modern texts in the appropriate language are often read by clear to the user. Interesting artifacts in each language are documented and exposed (e.g. in Sanskrit there are poems that read forward as well as backwards, in Chinese there is a poem composed entirely of "Shi", etc.)

    Spoken models are contrasted with the TTS and corrective feedback is given. By knowing what the user knows about the language, custom dictionary entries are generated based on the vocabulary the user already possesses. Furthermore, their language acquisition helps us to formalize the semantics of all the different languages. A model of common myths for each language is created to help expedite learning for large groups.

    Incorporate Justin's writing on this topic when we have internet access. One subsystem of all is termios which is the terminology management server. It is responsible for knowing all the word definitions. Naturally it makes use of many data sources: WordNet, opencyc, Sensus, EuroWordNet, OMCSNet, dictd, etc., as well as many programs: verber, Peterson's WSD, FLogic, Gate. The basic goal is to provide some level of meaning for as much terminology as possible, as well as construct thematic mappings.