to cohere is to hold together, and explanatory coherence is

1

  Inference to the Best Explanation    Brian
D. Haig University of Canterbury   July 2006  
2

• Overview
• Introduction
• Inference to the Best Explanation
• The Theory of Explanatory Coherence
• Virtues of the Theory
• Conclusion  

3

• Introduction
• Theory appraisal is a neglected feature of
  behavioural research methodology.
• Three major approaches to theory appraisal
• a) hypothetico-deductive method
• b) Bayesianism
• c) inference to the best explanation
• 3. IBE is an appropriate way to evaluate
  explanatory theories in psychology.

4
Inference to the Best
Explanation Theory evaluation in science is
often a matter of IBE. IBE captures the
important idea that much of what we know about
the world is based on considerations of
explanatory worth. IBE involves the process of
judging the best of competing explanatory
theories one accepts a theory when it is
judged to provide a better explanation of the
evidence than its rivals do.
5
The Theory of Explanatory
Coherence The major difficulty with IBE has
been to give it an informative and usable
characterization. Paul Thagard (1992) has
developed a precise formulation of IBE that
involves making judgments of explanatory
coherence. This is the Theory of Explanatory
Coherence (TEC). To cohere is to hold together,
and explanatory coherence is holding together
because of explanatory relations.
6
Criteria of Explanatory
Coherence With TEC, the evaluation of competing
theories in science is decided on the basis of
three main criteria Explanatory breadth.
Captures the idea that a theory is more
explanatorily coherent than its rivals if it
explains a greater range of facts. The
most important criterion.
7
Simplicity. Preference should be
given to theories that offer economical
explanations by making fewer ad hoc assumptions.
Analogy. Explanations are judged more coherent
if they are supported by analogy to theories that
scientists already find credible. Analogy
is an important criterion because it can improve
the explanation offered by a theory.
8
Principles of Explanatory
Coherence The determination of judgments of
explanatory coherence is made through the
operation of seven principles. These principles
specify the relations of local coherence between
a hypothesis and other propositions. Stated
informally, these principles are 1. Symmetry.
Coherence is a symmetric relation, so that if P
coheres with Q, then Q also coheres with P.
Incoherence also is a symmetric relation.
9
2. Explanation. Whenever two
hypotheses work together to explain something,
these hypotheses cohere with each other.
This principle establishes most of the coherence
relations in an explanatory theory. 3. Analogy.
Analogous theories that explain analogous
evidence cohere. This principle is
consistent with the belief that scientists prefer
theories that are analogous to theories that they
already find credible.
10
4. Data priority. Propositions that
describe the results of observation have a degree
of acceptability of their own. They stand
on their own better than propositions whose sole
justification is in terms of what they
explain. 5. Contradiction. Propositions are
incoherent with each other if they are
contradictory. 6. Competition. Hypotheses that
explain the same evidence compete with each other
unless there is reason to believe otherwise.
Hypotheses can compete with each other without
being contradictory.
11
7. Acceptability. An explanatory
hypothesis is accepted if it coheres better
overall than its competitors. According to
TEC, a new theory will replace an older one if
its hypotheses possess greater explanatory
coherence. The decision to accept or
reject a theory as a whole is made on the basis
of local pair-wise coherence relations.
12
The Programme ECHO The seven
principles (and the three embedded criteria)
together provide an informative characterization
of IBE, but they are not sufficient to show how
TEC works. TEC is implemented in a computer
program, ECHO, that straightforwardly applies
connectionist algorithms to the principles of
explanatory coherence (ECHO stands for
Explanatory Coherence by Harmany Optimization).
13

• Applications of TEC
• Through use of ECHO, TEC has been applied to a
  number ofcases of complex scientific reasoning.
  The simulations of scientific cases include
• a) Darwins evolutionary theory (versus
• creationism)
• b) the oxygen theory of combustion
  (versus
• phlogiston)
• c) dinosaur extinction (meteor versus
• volcano).

14

• The Virtues of TEC
• integrated
• theoretically informed
• explanation-focused
• multi-criterial
• coherence discriminating
• precisely formulated
• realizable
• reliable
• Conclusion
• Taken together, these positive features of
• TEC entitle one to regrad it as an
• appropriate method for evaluating the
• comparative worth of causal explanatory