Induction, Reduction and Abduction
Reasoning processes consist of various styles to deal with uncertainty, such as nonmonotonicity, probabilities, ambiguity and vagueness, universality, conclusivity, logical rules governing observations, speculations and subjectivity. In general, logical aspects of these types of reasoning processes are divided into the following three categories: induction, deduction, and abduction. All three stem from the Latin verb “ducere,” meaning to lead, with added prefixes.
The Latin prefix “de” means “from.” Deduction reasoning, also called Reduction reasoning, derives a conclusion from a set of premises or assumptions. In other words, given the truth of the assumptions, a valid deduction guarantees the truth of the conclusion.
The Latin prefix “in” means “toward.” Induction reasoning infers some general principle from a body of knowledge. It is a method of reasoning that involves making generalizations based on specific observations. Inductive reasoning is often used to develop hypotheses.
The Latin prefix “ab” means “away.” Abductive reasoning starts from observations and theories, which might be hypotheses, to make the best guess. It seeks the simplest and most plausible conclusion from a set of available evidence. Abductive reasoning is often used in computer science, artificial intelligence, criminal investigations, medical diagnoses and in diagnostic expert systems where the evidence is incomplete or uncertain. Abductive conclusions do not eliminate uncertainty, which is expressed in terms such as “very typical”, “quite plausible“, “best available” or “most likely.”
