Pier Paolo Pisani

A computer-controlled language training system for investigating the language skills of young apes

A computer-controlled language training system was designed and constructed to enhance the objectivity and efficiency of inquiry into the language-relevant behaviors of apes. The system allows the S to gain control over the events of the 24-h day in direct correspondence with its competence in using a keyboard on which each key represents a word. Various incentives can be obtained through the selection and depression of appropriate keys in accordance with rules of sentence structure monitored by a computer. The system is flexible and allows for eventual conversation between man and ape, with the computer as the intermediary. A Teletype records all that transpires. Achievements of the chimpanzee S over the course of the first 8 months of the system’s operation attest to the worth of the system and training methods.

The multistore parser for hierarchical syntactic structures

A syntactic parser is described for hierarchical concatenation patterns that are presented to the analyzer in the form of linear strings. Particular emphasis is given to the system of “significant addresses” by means of which processing times for large-scale matching procedures can be substantially reduced. The description makes frequent use of examples taken from the fully operational implementation of the parser in an experimental English sentence analyzer. By structuring an area of the computers central core storage in such a way that the individual locations of bytes and bits come to represent the data involved in the matching procedure, the shifting of information is reduced to a minimum, and the searching of lists is eliminated altogether. The matches are traced by means of binary masks and the state of single bits determines the operational flow of the procedure. The method could be implemented with any interpretive grammar, provided it can be expressed by the functional classification of the items composing the input hierarchical structures.

chapter 6 – Computer Programs

Publisher Summary This chapter presents computer programs associated with the LANA project to perform an objective grammatical analysis of the subjects linguistic input. In the automatic analysis of natural language, a system for the English language was developed. Known as the Multistore Parser, its purpose was to recognize and explain structural patterns in sequences of words, that is, sentences. The recognition of these structural patterns is made by means of a system of rules that operate on a sequence of words whose individual characteristics are pre-established. The individual characteristics are a collection of connectives that represent the possibilities a word has to form a syntactic combination (correlation) with another item. Each word has a pre-established sequence of correlational indices (les), which represent its string of potential links. Each le is identified by the code number of the relation it establishes between two items and by an indication of whether the items to which the le is assigned can be an LH-piece or an RH-piece, because the order in which two items can form a linguistic relation is fixed. To check the correlational compatibility between only two adjacent words, every le of the first word must be tested for a match with every le of the second word. However, the amount of data involved in an analysis of this kind is enormous and one to look for an economical way of handling correlational indices.