Coping with Confusion in Terminology
This page has not been revised
since March 2000, but
the version on another website has been revised many times
since then, so I strongly recommend that you read
THE
REVISED VERSION.
This web-page contains, with no editing,
Section 2.08-I from my doctoral dissertation.
{ note: "ISM" is an abbreviation for Integrated
Scientific Method, a model
that I developed and then used for the integrative analysis of instruction.
}
Can ISM cope with differences in terminology?
The "Introduction to the ISM
Framework" states that "there is no consistent terminology; instead,
there are important terms...with many conflicting meanings, and meanings
known by many names." This leads to the question, "Can ISM
describe the views of a person who uses different terms than those in ISM,
or who defines the ISM terms differently?"
There are several possible responses
to this challenging question. One approach attempts to solve conflicts
by a simple substitution of one term for another. But a more sophisticated
approach is needed for the really tough inconsistencies; for these conflicts
a "solution" seems impossible, but there may be ways to minimize
the confusion. And in any case, difficulties arising from choices
about terms are not unique to ISM; because of the widespread inconsistency
in terminology, every description of science faces a similar challenge.
Substitution is an easy way to cope with
some differences. For example, the same process of logic is known
by two common names : retroduction and abduction. [1]
There are good reasons to use 'retroduction', [2] but if
someone wants to replace this term with 'abduction' it would cause no significant
change in ISM. Another potential candidate for substitution is 'experimental
system'; in ISM this is defined as "everything involved in an experiment"
for a reason, [3] but sometimes scientists define a system
as only "what is being studied," with "what is done to it,
and the instruments of observation" being external to the system.
In this case, 'experimental system' could be replaced by another term, such
as 'experimental setup' or 'a system and its experimental context'; this
substitution would not alter the hypothetico-deductive logic used in the
ISM framework. Or someone might want to replace theory 'invention'
with 'development' or 'generation'.
But for three ISM terms -- model,
hypothesis, and theory -- the situation is more complex. Part of the
difficulty is inconsistent terminology; these terms and others (such as
principle, law, concept, or conjecture) are often used with similar or overlapping
meanings, or with contrasting meanings that differ from one definer-of-terms
to another. The overall use of these terms lacks both consistency
and precision.
The term 'model' is used in many ways,
so no matter how this term is used there will be conflicts with other definitions.
Often it seems to be a synonym for a theory or sub-theory, or for a certain
type of theory. Or it can refer to an exact application or (more commonly)
a simplified application of a theory for a certain type of system, as when
Giere (1988, 1994) defines a theory in terms of "families of models."
For example, the theory of Newtonian Mechanics includes many models and
families of models, such as gravity-driven motion on an inclined plane with
no friction, or with friction, or with friction plus air resistance.
Giere thus uses the term 'model' in two ways that are different yet compatible :
each of the models (that together comprise the theory) can be useful when
constructing a model (as this term is used in ISM's hypothetico-deductive
box) for certain experimental systems. For example, an 'inclined plane'
family of models is a useful shortcut when applying Newton's theory for
a specific system that is a member of the corresponding family of systems,
while a 'pendulum' family of models is useful for theory application within
another family of systems. By analogy with the distinction between
'domain-theories' and 'system-theories' in Section 2.05, there can be 'domain-models'
(such as the family of inclined plane models that occurs when Newton's theory
is applied, in various ways, within the domain of inclined plane systems)
and 'system-models' (resulting from the application of Newton's theory,
in a certain way, to specific inclined plane systems). And, analogous
to the influence of a domain-theory on system-theories, a general domain-model
(about systems in a domain) will influence the construction of a specific
system-model for one system in its domain.
'Hypothesis' has an even wider variety
of conflicting meanings. Consistent with Giere (1991), ISM defines
a hypothesis as a claim that a theory-based model is similar to a real-world
system "in indicated respects and to an implied degree of accuracy.
(Giere, 1991, p. 27)" Gibbs & Lawson (1992) define it as
"a single proposition intended as a possible explanation...for a specific
effect (p. 143)" in contrast to "a theory...intended to explain
a broader class of phenomena and consisting of multiple hypotheses and/or
general postulates that, taken together, constitute the explanation (p.
149)"; but the authors report with dismay that "a number of textbooks
give examples of hypotheses that clearly are predictions, not hypotheses
(p. 147)," and that most authors "define theories as hypotheses
that have been supported over a long period of time (p. 147)" even
though according to the authors' own definitions "the evidence may
or may not support a theory (p. 148)" -- for example, the Ptolemaic
Theory of earth-centered planetary orbits is still considered to be a theory
even though it now has low status. Darden (1991, p. 17) describes
two meanings for theory -- it can be "[a claim that is] hypothetical,
not yet proven" or "a well-supported, general, explanatory claim"
-- and chooses the latter definition to use in her analysis; and she defines
a hypothesis as "a proposed alternative to a theoretical component.
(p. 18)" Grinnell (1992) usually uses hypothesis in the same
way that ISM uses theory (a word he never uses) -- for example, he says
that scientists "imagine hypotheses to explain these regularities...in
observed natural events (p. 1)," and in criticizing the model of theory
falsification (Popper, 1963), he says that "according to this model,
science progresses through selective falsification of competing hypotheses,
... [and] it takes only one negative result to call a hypothesis into question
(p. 40)" -- but sometimes this word changes meaning and is a prediction :
"[an explicit hypothesis] is the change that I expect to see if I do
[a certain experiment]. (p. 25)"
The chaotic state of terminology is captured
in the paragraph above. In it, hypothesis is defined -- by Giere,
Gibbs & Lawson, Darden, and Grinnell, all authors whose work I respect
-- as a claim about a model for one system, or an explanation for a type
of phenomenon, a sub-theory or theory component, a prediction, a new theory
with low status, a newly proposed theory component, and a theory.
What a wild mix! Generally, however, the difference between hypothesis
and theory tends to be defined along two dimensions -- scope and certainty.
[4] When both factors point toward the same choice
of a term there is general agreement, which includes myself and ISM, about
definitions : if a proposed explanation has narrow scope and low status,
it is a hypothesis; if it has wide scope and high status, it is a theory.
With mixed criteria (narrow scope and high status, or wide scope and low
status) there is less agreement. Another criterion is age; older explanations
tend to be called theories, not hypotheses. For example, the Ptolemaic
Theory remains a theory even though it currently has very low status; once
a theory, always a theory?
In ISM the distinction between a hypothesis
and theory depends only on scope; age doesn't matter; and a hypothesis can
have either low status or high status, as can a theory. In ISM (and
for Giere) the criterion for scope is clear; a hypothesis refers to one
system, while a theory refers to two or more systems. By contrast,
the other authors define the distinction with words like "broader"
or "general."
I do not claim that my use of 'hypothesis'
and 'theory' is the best possible solution for this "terminology problem."
The principle advantages of my definitions are logical simplicity and internal
consistency. The main disadvantage is that, despite agreement with
the use of terms by some philosophers (especially Giere) there is disagreement
with other philosophers and with most scientists.
There are significant logical advantages,
described below, in using the ISM terms. These advantages carry over
to practical concerns. For example, with one non-ISM definition a
low-status theory (hypothesis) may eventually become a theory, but before
this occurs the hypothesis occupies the same status as 'theory' does in
the current ISM. This means that everywhere in the ISM diagram where
"theory" appears (15 places) it would have to be replaced by "theory
or hypothesis"; similar changes would be needed throughout the text
of Chapter 2.
A logical disadvantage of this non-ISM
definition is that if "hypothesis" appears in the "theory"
oval (on the left side of the diagram) rather than in the hypothetico-deductive
box, how does 'hypothesis' enter into 'hypothetico-deductive' reasoning
when a theory (not a hypothesis) is used to construct a model and make predictions?
A logical advantage of the ISM definition
of theory, which avoids the use of status in defining a theory, is that
this lessens the rhetorical value of using the term 'theory' to influence
critical thinking. For example, in the quotations cited above Darden
describes two ways to define a theory; in one a theory is "hypothetical,
not yet proven" while in the other it is "well-supported,"
so the same word can be used to imply low status or high status! I
think it is better to evaluate a proposed explanation (i.e., a theory) based
on merit rather than terminology. We can just say "here is a
proposed explanation; now we can decide how well supported it seems to be,"
instead of trying to short-circuit the critical process by saying "it
is a theory so it is not proven" or "it is a theory so it is well
supported." Another difficulty in drawing a demarcation line
between hypothesis and theory is pointed out by Gibbs & Lawson (1992) :
"Who decides when an hypothesis is supported to the extent that it
gets to be called a theory? (p. 148)"
There are many logical and practical
reasons to use the definitions I have chosen for ISM; these have overcome
the advantages of using a more commonly used definition. This is one
of the few places where I am aware of ISM being normative by "prescribing"
how things should be done in science, by saying "this way is more logical
and practical." This decision was made easier by the fact that
no matter which definitions I use in ISM, they will be inconsistent with
the many other definitions that are commonly used; it is impossible to be
consistent with inconsistency.
FOOTNOTES
1. Almost everyone uses these two terms as synonyms, but as usual there is variation. Ernst McMullin (1992) uses 'abduction' to mean the same thing that 'retroduction' does in ISM, but he uses 'retroduction' to describe a wider-ranging process of thinking that includes "abduction and more."
2. I decided to use 'retroduction' in ISM, for two reasons. First, the prefix 'retro' is a reminder that this logic is oriented backward in time, to explain what already has occurred. Second, if ISM is to be maximally useful for science education, and if part of this usefulness occurs when teachers encourage children to learn a productive form of creative-and-critical thinking, it seems unwise to call this desirable activity 'abduction' -- a term whose primary common meaning is the undesirable activity of kidnapping.
3. Observations are produced by an experiment that involves everything in the experimental system. And to make predictions a model must consider everything, including the instruments of observation.
4. But if a hypothesis is defined as a theory-component (as it is by Gibbs & Lawson, and by Darden), this is not necessarily the same as having narrow scope (which is also used by these authors to define a hypothesis), because a component can have wide scope.
Table of Contents for my dissertation
home-page for my "thinking methods" website
site-map for my "thinking
methods" website