Chapter 7
Abduction
Introduction
Nature
Uses
Introduction
Charles Sanders Peirce (1839-1914), the distinguished American philosopher,
introduced abduction into modern logic. Abduction (also called abductive
reasoning
or retroduction) is the process of reasoning that seeks to find the
best explanation
for an event. Abduction is used in a number of fields, some
of the main ones being artificial intelligence, fault diagnosis, belief revision,
and automated planning. Our interest here is with belief revisions since this is
involved in proving facts in a court.
Nature
Introduction
Abduction starts with a collection of data such as a set of facts (or events, or
an observation or a given). This data will be unexpected or anomalous, such
that accepted explanations for it are lacking. Abduction seeks to do two
things. First, it generates hypotheses or explanations
160
for the facts or events.
These explanations can be of two kinds. They can be an established or known
causal rule. For example, after it rains, the grass becomes wet; therefore when
the grass is wet an explanation is that there has been rain. Alternatively, the rule
can be a new causal law; this aspect of abduction can be a source of
discovery in behavioural and physical science. Abduction, it is worth noting, is
the only logical process that does this. It does this by a mixture of inference
161
and creativity.
___________________
160
Sometimes, but only rarely, the expression "explanatory conclusions" is
used instead of "explanations" to describe this part of the abductive process.
161
To illustrate, John
Josephson describes abduction as inference to the best
explanation. See “Inductive Inference in Reasoning and Perception” on
www.cse.ohio-state.edu/lair/ Projects/Abduction/abduction.html.
 |
 |
 |