“A Logic for Default Reasoning” was written by Raymond Reiter and published in the journal Artificial Intelligence, volume 13, in 1980 (pages 81 to 132), in the same special issue on nonmonotonic reasoning that carried John McCarthy’s circumscription paper. The work addresses a fact about everyday reasoning that classical logic cannot capture: we routinely accept general statements with unstated exceptions, such as “birds fly,” and we conclude that a particular bird flies unless we are told it is a penguin or has a broken wing.
Reiter’s mechanism is the default rule. A default has a prerequisite, a justification, and a consequence: if the prerequisite holds and it is consistent to assume the justification, then the consequence may be concluded. For “birds fly,” knowing that Tweety is a bird lets you conclude Tweety flies, as long as nothing in your knowledge contradicts it. The set of conclusions you reach is called an extension. Crucially, learning that Tweety is a penguin blocks the default and forces you to retract the earlier conclusion, which is why this is “nonmonotonic” reasoning: adding facts can shrink, not just grow, the set of beliefs.
Default logic became one of the standard tools for representing common-sense knowledge and for handling the closed-world and frame problems that bedevil purely classical formalizations. Its prerequisite-justification-consequence shape directly influenced later work on logic programming with negation-as-failure and on knowledge-based systems that must reason with rules-of-thumb.
Why a business reader should care: most real decisions rest on defaults (“assume the shipment is on time unless flagged”), and Reiter’s paper is the foundational account of how to encode such reasonable assumptions in a system while letting it cleanly revise them when an exception appears.