One goal of information science research is the improvement of database query languages, making them simpler to use, easierto learn, and less prone to user error. Most non-technical endusers cannot use existing large online bibliographic database systems without help from specially trained intermediaries or experts in the use of the online systems, and even trained intermediaries are not equally proficient in all of the many bibliographic systems available. The use of artificial intelligence techniques to produce computer assistants or intermediaries will make it possible for end users and experts to retrieve information directly and efficiently by assisting with query construction. The specific emphasis is on directing the user's uncontrolled vocabulary queries toward controlled indexing vocabulary for a particular database. A computer intermediary to assist users with query construction will be implemented to validate this principle on the Chemical Abstracts database. The results obtained in this study will be applicable to query construction for information retrieval, particularly by non- technical personnel and casual users, on a wide range of database systems.
