When does the law change in a system of precedents? The answer depends on how precedent is to be understood. There are three influential models, which are based on the ideas of rules, reasons, and analogies. We argue that these three theories make different assumptions about how the law changes in reaction to novel cases that do not overrule precedent. According to a version of the rule-based approach, the law changes if and only if a new case in a line of precedents reaches a different outcome; whereas, under the reason-based approach, the law changes with every new legal case, in proportion to the number of new features added to the doctrine by each case, regardless of outcome; and, finally, under the analogy-based approach, the law established by the precedent may change with every new case, in direct or inverse proportion to the differences found between cases, depending on the outcome. Although discussion about the nature of precedents often takes a normative perspective, in this paper we tackle it from an empirical perspective: How do people perceive legal change through precedent? Findings support the rule- and, to a very limited extent, the analogy-based approaches. The reason-based approach found no empirical support. We hypothesise that this reflects a conservative tendency to restrict legal change in precedents and to perceive it as non-incremental, thereby enhancing simplicity. Naturally, theorists may argue that the folk idea of when the law changes in a system of precedent is wrong. We suggest that legal theories should address a possible lack of a minimum descriptive requirement and that there are reasons of transparency and publicity to favour legal concepts that are closer to folk concepts.