This paper presents the results of a design study funded by the British Library Research and Development Department, aimed at developing an interactive information retrieval system that identifies structural representations of anomalous states of knowledge (ASKs) underlying information needs. The system attempts to resolve these anomalies through various retrieval strategies applied to a database of documents in compatible structural formats. Part I discusses the background and theory of the project, while Part II (next issue) presents the methods, results, and conclusions. The project's basic premises include the idea that information needs are not precisely specifiable, that problem statements can be elicited from users to derive representations of ASKs, that ASKs exist in classes, and that all elements of information retrieval systems should be based on the user's ASK. A freeform interview technique was developed for eliciting problem statements, and a statistical word co-occurrence analysis was used to derive network representations of problem statements and abstracts. Structural characteristics of these representations were used to classify ASKs, and both ASKs and information structures were evaluated by users and authors. Results indicate that interviewing is a satisfactory technique for eliciting problem statements from which ASKs can be determined. The statistical analysis produces structures generally appropriate for both documents and problem statements. ASKs represented in this way can be usefully classified according to their structural characteristics. Of thirty-five subjects, only two had ASKs for which traditional 'best match' retrieval would be intuitively appropriate. The results suggest that at least some of the project's premises are reasonable, and that an ASK-based information retrieval system is at least feasible. The paper discusses the theoretical basis of the ASK information retrieval system, emphasizing the need to move beyond the traditional 'best-match' principle. It argues that the 'best-match' principle is based on two assumptions that are not always valid: that users can specify their information needs precisely, and that information needs are functionally equivalent to document texts. The paper suggests that an ASK-based approach, which considers the user's state of knowledge as anomalous, is more appropriate for information retrieval. It also discusses the design of an interactive system that incorporates the user's ASK, allowing for iteration and interaction in retrieval. The system is designed to be highly iterative and interactive, with the user's evaluation of retrieval results feeding back into the system to modify the retrieval strategy and problem structure. The paper concludes that an ASK-based information retrieval system is a feasible approach to improving information retrieval.This paper presents the results of a design study funded by the British Library Research and Development Department, aimed at developing an interactive information retrieval system that identifies structural representations of anomalous states of knowledge (ASKs) underlying information needs. The system attempts to resolve these anomalies through various retrieval strategies applied to a database of documents in compatible structural formats. Part I discusses the background and theory of the project, while Part II (next issue) presents the methods, results, and conclusions. The project's basic premises include the idea that information needs are not precisely specifiable, that problem statements can be elicited from users to derive representations of ASKs, that ASKs exist in classes, and that all elements of information retrieval systems should be based on the user's ASK. A freeform interview technique was developed for eliciting problem statements, and a statistical word co-occurrence analysis was used to derive network representations of problem statements and abstracts. Structural characteristics of these representations were used to classify ASKs, and both ASKs and information structures were evaluated by users and authors. Results indicate that interviewing is a satisfactory technique for eliciting problem statements from which ASKs can be determined. The statistical analysis produces structures generally appropriate for both documents and problem statements. ASKs represented in this way can be usefully classified according to their structural characteristics. Of thirty-five subjects, only two had ASKs for which traditional 'best match' retrieval would be intuitively appropriate. The results suggest that at least some of the project's premises are reasonable, and that an ASK-based information retrieval system is at least feasible. The paper discusses the theoretical basis of the ASK information retrieval system, emphasizing the need to move beyond the traditional 'best-match' principle. It argues that the 'best-match' principle is based on two assumptions that are not always valid: that users can specify their information needs precisely, and that information needs are functionally equivalent to document texts. The paper suggests that an ASK-based approach, which considers the user's state of knowledge as anomalous, is more appropriate for information retrieval. It also discusses the design of an interactive system that incorporates the user's ASK, allowing for iteration and interaction in retrieval. The system is designed to be highly iterative and interactive, with the user's evaluation of retrieval results feeding back into the system to modify the retrieval strategy and problem structure. The paper concludes that an ASK-based information retrieval system is a feasible approach to improving information retrieval.