Introduction
This paper describes an "intelligent" tutoring system, (IKITS), for learning Chinese. An "intelligent" tutoring system teaches in an "intelligent" manner with help from stored expert knowledge and assumptions about the student's knowledge and understanding.
IKITS is constructed for students in Chinese at senior high school- and university level.
The purpose of IKITS is to support students in their studies in Chinese. The system is not intended to replace teacher, but to function as a supporting supplement to students who want to conduct independent studies in Chinese.
I chose to do a tutoring system for Chinese, because I think that the form of education will be frequently used in the future.
The subset of Chinese I have represented in the system is representative for the grammar rules and grammatical structures and words a beginner is learning in Chinese.
The system consists of grammar rules in Chinese to read about, word
exercises, grammar tests and grammar
exercises and a lexicon.
The system is implemented in HyperCard and HyperProlog in system 7.0 on a SE/30 Macintosh-computer.
This paper consists originally of a description of second language learning, tutoring systems, tutoring systems for second language learning, a general description about Chinese, the subset of Chinese represented in IKITS, a general description of IKITS and a detailed description of the system and it's parts and functions. The English version of this paper only covers the parts the subset of Chinese represented in IKITS, a general description of IKITS and a detailed description of the system and it's parts and functions.
Comments
A complete tutoring system for language learning in Chinese should except for the parts that are in this system also consist of character analysis and speach.
Language learning is individual since we all learn in different ways. Therefore the tutoring should also be individual and it's very difficult to construct a system that fits all learning types.
It seems to be an impossible task to replace (language-) teachers with tutoring systems. Partly because pedagogical knowledge is very difficult to represent and partly because of the advantage a teacher has, over a computer, in his intuitive ability to sense what a student might learn and in which way he might learn, compared to the pedagogical theories and inference algorithms that a computer might possess. I think that a system of this type is functioning best as a tool for exercises in education, for both teacher and students.
The construction of this system has been full of problems. I've had problems with the memory in HyperProlog because that the system often has been sending big pieces of information between HyperCard and HyperProlog and between HyperProlog and HyperCard. The connection between HyperCard and HyperProlog hasn't always worked without problems. To be able to take full advantage of a link like this between two applications they have to have similar functions defined in approximately the same way. One example is that HyperProlog is interrupting an execution that fails, and that HyperCard on the other hand can't always report problems in the executions. This is a problem in searching for errors in functions in programming.
If I get an opportunity to develop the system further, I will first develop the existing parts but also think about how character analysis and pronounciation would be implemented in a tutoring system.
Description of Interaktivt Kinesiskt ITS (IKITS) Interactive Chinese Intelligent Tutoring System
General description
This system is made for students who study Chinese at senior High School- or university level. This system is made for independent studies for learning Chinese. The idea of the system is to learn Chinese grammar and in Chinese words. Character analysis and fonetic studies is not parts of this system. It would take another direction of didactic domain knowledge which would be analysis of speach and characters. The purpose of this system is for the student to decide himself what to study, how to study and in which order to conduct his studies. The knowledge a student needs to be able to work with the system is general knowledge of how a Macintosh-computer works and general knowledge about English grammar (Swedish).
The system is individual and discovery-based. Individual in the way that the student gets tips from the system, based on the student's results, and the system is discovery-based in the way that the student discovers how the Chinese grammar works, through reading about it or through doing exercises in the system.
The system is passive so that the student can conduct independent and optional studies without being interrupted by the system. There is however one exception to the passive behaviour, when the system gives the student a message, but it's not an interruption, but a tips, which doesn't have to affect the student.
The jugdement of the students results is done in a passive diagnosis, when the system, without interference from the student, is doing an analysis.
Registration of the student's activities is partly done by a belief model, in which the results are represented from analysis of purely syntactical constructions, not semantical or pragmatical. The grammar rules the student will learn, are semantically limited. Besides that there is a registration of the student's results, there is also registration of the student's knowledge in the form of his vocabulary, in an overlay model.
The construction of the system is made with the idea that it's supposed to be a general system. It will be possible to expand and continue developing parts of the system that already exists.
98-03-03 Sofie Johansson Kokkinakis sofie.johansson.kokkinakis@svenska.gu.se
