Building on Taylor’s (1968) levels of need and Wersig’s (1979) “problematic situ- ation,” Belkin (1977, 1978, 1980) developed the “anomalous state of knowledge”
(ASK) hypothesis. When encountering a problematic situation, users cannot solve the problem by applying existing knowledge, and their anomalous state creates cog- nitive uncertainty that prohibits them from adequately expressing their information need. They need additional information to clarify their thoughts. The driving force of information retrieval is the users’ problem that leads to recognition of their inad- equate knowledge to specify their information need. Simultaneously, users evaluate the information retrieved from an IR system related to the problematic situation, and that might also determine their Anomalous State of Knowledge. In other words, Belkin identified the ASK underlying users’ information needs. Information needs and information retrieval are dynamic, and they change along a user’s cognitive structure. Figure 1.1 presents a cognitive communication system for information retrieval (Belkin, 1980, p. 135).
Belkin, Oddy, and Brooks (1982a, 1982b) further applied ASK to information retrieval system design. Based on the assumptions that (1) users cannot specify their information needs and (2) there are classes of ASKs that an IR system needs to be built on, Belkin, Oddy, and Brooks (1982a, 1982b) developed an ASK-based information retrieval system. The system was designed to ask a user to describe the ASK instead of specifying the need as a request to the system. The system was based on the cognitive viewpoint that human interaction is mediated by people’s state of knowledge. In addition, the researchers considered the IR situation as a recipient- controlled communication system, as suggested by Paisley and Parker (1965). An ASK-based information retrieval system design was suggested as follows (Belkin, Oddy, & Brooks, 1982a):
1. User’s problem statement
2. Structural analysis of problem statement
3. Choice of retrieval strategy according to type of ASK
4. Abstract presented to user simultaneously with explanation of why text was chosen
5. Structured dialog between system and user to infer user’s evaluation of a. Method of choice
b. Suitability of document to problem c. Whether need has changed
6. Modifications according to evaluation or finish 7. Return to 2 or 3 as necessary (p. 69).
Figure 1.1. Belkin’s cognitive communication system for information retrieval.
From “Anomalous states of knowledge as a basis for information retrieval” by N. J.
Belkin, 1980. Canadian Journal of Information Science, 5, p. 135. Copyright 1980 by University of Toronto Press. Used with copyright permission.
GENERATOR’S IMAGE OF THE WORLD
USER’S IMAGE OF THE WORLD
TEXT REQUEST
CONCEPTUAL STATE
OF KNOWLEDGE INFORMATION ANOMALOUS STATE
OF KNOWLEDGE
CONCEPTUAL STATE OF KNOWLEDGE belief, intent,
knowledge of user transformations
realization of need Linguistic,
pragmatic transformation
User-Orented IR Research Approaches
ASK hypothesis has been widely cited as a theoretical basis for research on informa- tion retrieval, especially interactive IR research. It sets the foundation for the major interactive information retrieval models, such as Ingwersen’s (1992) cognitive model, Ingwersen and Järvelin’s (2005) integrated IS&R research framework, Belkin’s (1996) episode model, and Saracevic’s (1996b, 1997) stratified model. A predomi- nant characteristic of research in information retrieval in recent years has been the adoption of a “user-centered” approach to the design of IR systems. Belkin’s ASK hypothesis was credited for the shift in emphasis. According to Lima and Raghavan (2004), this shift began primarily after Belkin proposed his ASK hypothesis.
Research has been conducted in terms of how to represent ASK. By analyzing the problem statements of users, Oddy, Palmquist, and Crawford (1986) were able to represent the anomalous state of knowledge for the creation a top-down, across- domain IR system. Cole et al. (2005) investigated the anomalous state of knowledge hypothesis in a real-life problem situation by studying how history and psychology undergraduates sought information for a course essay. They identified the importance of Belkin, Oddy, and Brooks’ ASK papers. First, the ASK hypothesis creates a frame- work for dealing with a user’s unknown information need without asking the user to specify it. Second, an interoperable structural code can be built into IR systems to connect the user’s ASK to the relevant documents in the IR system. Wu (2005) examined elicitation during retrieval interaction as microlevel information-seeking (MLIS). She enhanced the dialogue structure derived from the ASK hypothesis by identifying the differences between patrons’ and intermediaries’ elicitation behaviors in terms of frequency and timeframe, and further revealed that intermediary elici- tation is preplanned and patron elicitation is situational. Liddy, Oddy, and Bishop (1988) applied the conceptual framework of ASK to problem statements, and they identified four major components: user traits, subject traits, information traits, and position in the problem-solving process.
Belkin (1993) extended the research of ASK to information-seeking strategies, and further developed the episode model of interaction with text in which interaction is the central process. Belkin and his associates conducted a series of studies to establish relationships between information-seeking strategies and users’ informa- tion-seeking goals and problematic situations. For example, Belkin, Cool, Stein, and Theil (1995) developed four dimensions of information-seeking strategies: method of interaction, goal of interaction, mode of retrieval, and resource considered. Xie (1997, 2000, 2002) identified the patterns between users’ interactive intentions and information-seeking strategies and further explored their shifts in information- seeking strategies. Lin and Belkin (2005) modeled multiple information-seeking episodes by analyzing eight types of information problems and their relationships with multiple searching episodes.