monotonic reasoning

Terms from Artificial Intelligence: humans at the heart of algorithms

Page numbers are for draft copy at present; they will be replaced with correct numbers when final book is formatted. Chapter numbers are correct and will not change now.

In monotonic reasoning new information can add new inferences, but never invalidate old ones. For example, RDF uses and open world model, that is it assumes that there may always be more assertions than are currently available. This is a safe conservative position, but makes it hard, for example, to make strong negative assertions such as "there are no pink elephants". In contrast non-monotonic reasoning may make use of default rules, or assume if soemting is not stated it is not true, but may have to retract inferences as new information emerges.

Used in Chap. 3: page 29

Also known as monotonic