VINCI is an environment used to simulate language based on linguistic theory. More information on its development and functionality may be found here:

Levison, Michael and Gregory Lessard. \"A System for Natural Language Sentence Generation.\" Computers and the Humanities 26.1 (1992): 43-58. Web.

Program Overview

VINCI is a natural language generation system developed by Michael Levison and Gregory Lessard of Queen's University in Kingston, Ontario (Levison and Lessard 43). It was programmed in C, and the version released in 1992 could run in UNIX, MS-DOS and VM/CMS (43).

Levison and Lessard describe VINCI as

a natural language generation system...which accepts as input a formal description of some subset of natural language, and generates strings in the language. With the help of an attribute grammar formalism, the system can be used to simulate on a computer components of several current linguistic theories. (43)

Their goal was to create a system capable of embodying constructs from a variety of linguistic theories (Levison and Lessard 44). The resulting program is

[o]n the one hand...an attempt to provide an environment for solving a particular range of problems, among them teaching basic concepts of linguistic theory, modelling performance errors and providing a generative computer-aided language learning environment. On the other hand, it attempts to provide the most common devices found in mainstream generative grammar. (44)

Levison and Lessard wanted a system with rules that were as familiar as possible to researchers, that did not require them to learn to program, and that provided a morphology and lexicon rather than require users to construct their own (44). As such, \"VINCI provides the linguist with metalanguages as similar as possible to those used in current linguistic research\" (44).

As of 1992, they were working toward expanding VINCI to include second-language learning, via exercises composed of question and expected answer pairs (Levison and Lessard 55). The prototype looked for common typing errors such as transposition and omitted letters or words, with the intent of identifying errors by type and providing appropriate advice (Levison and Lessard 57). They were also applying VINCI to modelling language errors, or \"malrules\" (Levison and Lessard 57).

The 1992 release of VINCI allowed users to modify the data, including morphology and lexicon, as needed (Levison and Lessard 45). The system could load a language description in under a minute on a SUN network or an IBM 3081G, but was \"considerably slower\" on an IBM PC (286) (45). It was designed to handle a wide range of languages -- though Levison and Lessard mostly used VINCI with French, it was also successfully applied to English, Russian, Italian and basic French phonetics (45).

Even in its earliest version, VINCI was successful as a tool for facilitating linguistic research, and for linguistic students exploring language generation concepts (Levison and Lessard 55).

Note: Vinci continued development into the 2000s and has been maintained online by the developers with comprehensive documentation. It can be found here.

Source

Levison, Michael and Gregory Lessard. \"A System for Natural Language Sentence Generation.\" Computers and the Humanities 26.1 (1992): 43-58. Web.