Human experts develop knowledge over time. Some of this is based on their initial training, and so may be obtained from books or training materials, but much is implcit knowledge, which they may not even be able to fully articulate (unknown knowns). Knowledge elicitation attempts to capture this expert knowledge in forms that can be used for hand-crafted expert system rules, or as training data for machine learning. This may involve expert labelling of data.
Used in Chap. 14: pages 217, 218; Chap. 18: pages 271, 272, 281, 283, 284, 289, 291