The answer is . Modern neural networks are incredibly powerful but notorious for not explaining why they made a decision. In high-stakes fields—medicine, finance, law, aviation—regulators demand an audit trail. Expert systems are inherently explainable; they can produce a step-by-step chain of rules that led to a conclusion.

Before probabilistic graphical models became mainstream, expert systems used certainty factors (Shortliffe & Buchanan). The book dedicates an entire chapter to this, explaining how MYCIN combined and propagated certainty through rules. This is a historically important and pedagogically useful section.

The text explores how human knowledge—often informal and experiential—can be codified for a machine. Formal vs. Informal Logic:

IF root-cause = “unforeseeable defect” THEN liability = “act of god” (CF 1.0)