Abduction, or "inference to the best explanation" is a form of inference that moves from data describing something to a hypothesis that best explains or accounts for the data.
In engineering design contexts, this process can result in reasoning such as the following:
artefact x must not corrode (premise)
IF an artefact is made of stainless steel,
THEN the artefact will not corrode (rule)
(infer): artefact x should be made of stainless steel (conclusion)
Here, the combination of rule (from the knowledge of the designer) and premise (some element of the design specification) allow a design decision to be made.
Just as deduction may be viewed as an analytical process, abduction, the "reverse deduction", is synthetic - it produces knowledge (in this example, the knowledge that the artefacts should be made of stainless steel has been produced - a design decision). While this shares the same structure as the fallacious "deductive" model of conceptual design, the difference here is that, logically, this is not necessarily a sound argument - if it is also known that the artefact should not conduct electricity, it might be better to abduce that the artefact should be made of a plastic. This is a more plausible "explanation" of the behaviour of the artefact, based on the available evidence - abduction is plausible inference.
댓글 없음:
댓글 쓰기