VinciLingua: who has worked on it

The development of ivi/Vinci and of VinciLingua has only been possible because of interdisciplinary collaboration, over many years. Neither of the two principal researchers could have produced the system on his own. At the same time, development of the system has been greatly facilitated by the participation of graduate and undergraduate research assistants over the years, in the development of the software, the writing of lexicons and grammars, and the conception and implementation of language teaching materials. Our students have come from a variety of areas, ranging from Computer Science, to Linguistics, to Education and to Engineering.

Work at the interface of linguistics and computing science has led us in some surprising directions, as the following examples will illustrate:

  1. Early in the project, we wanted to develop language teaching materials, but one of us (Levison) proposed to generalize the approach beyond canned materials to a full generative environment. Recognizing however that linguists could not be expected to program grammars directly in something like C, we developed a series of metalanguages which follow more closely standard linguistic practice, such as phrase structure rules, morphology rules, and lexical databases.
  2. Impetus for specific developments often came from particular problems to be solved. For example, one of us (Lessard) wanted to develop generation beyond short utterances to more complex texts, including fairy tales, riddles, and limericks. This required the development of a number of features (a more complex semantics, guarded syntax rules, lexical pointers) which have shown their value in other areas.
  3. Grappling with issues of generation has led us to embrace an onomasiological approach to language modelling: we see the goal of modelling not in the movement from form to meaning (parsing), but rather in the embodiment of meaning in form. One of the upshots of this was the production of a monograph with two of our graduate students on the semantic representation of natural language.

These few points just scratch the surface. More details can be found on the VinciProject page page.

One last point: it is clear that development of teaching and learning materials will require more than the two of us. One of the reasons for producing this web page is to encourage others who might be interested in participating in this project to contact us. Contributions are needed in a whole range of areas, including, but not limited to:

  1. development of mini-grammars and exercises in various languages. To date, we have materials in French, English and Spanish, but it would be interesting to see this range expanded, and for more exercises to be developed in the three languages already being worked on.
  2. addition of lexical materials: the richer the lexicons, especially in terms of semantic traits, the more 'delicate' the generation.
  3. exploration of different media and exercise types. We have explored exercises using sound and images, but more remains to be done, and imagination is needed in terms of the kinds of manipulations that learners might profit from in their exploration.