Experimental Discovery, Data Models, and Mechanisms in Biology: An Example from Mendel’s Work - Mechanism and Causality in Biology and Economics - page 114

(ED3) The experimenter must envisage searching for underlying mechanisms for
the phenomena, whether or not he or she proposes correct mechanistic
explanations.
One may raise a few questions: How are these conditions to be justified? Why is
a data model not theoretical? Can experimental discoveries really be recognized
without theories?
Because a model of experimental data is the output of an experimental process
(a combination of models of experiments and raw data according to the framework
in Fig. 6.4 ), they are not theoretical. They can be constructed without theories.
However, one can still question whether designing an experiment does or does not
depend on theories—the question of the theory-ladenness of experimentation. It is
the case that one can apply theories when designing experiments, but not all
experiments are designed based on theories. We should distinguish between the
concept of theory and that of background ideas . All experimental designs require
background ideas, but not all background ideas are theoretical. Mendel designed his
experiments with peas on the basis of his ideas about the formation and develop-
ment of hybrids, for which he did not have a complete theory. Furthermore, a data
model and a phenomenon allow a variety of theoretical explanations. Mendel
himself and Mendelian theorists (including Hugo de Vries, William Bateson, and
Thomas Hunt Morgan and his team) proposed different theories to explain the
identical ratios, the data models, and the phenomena of trait transmission (see
Darden 1991 ). 13 In this case and other similar ones, phenomena and data models
were constructed prior to and independent of any theory.
According to the three conditions, Mendel did make an experimental discovery;
it was a key to genetics. Griffith's case also satisfies these conditions, for a data
model can be extracted from Griffith's experimental process, and Griffith envisaged
searching for an underlying mechanism to explain the phenomenon of transforma-
tion . Driesch rejected the possibility of searching for a mechanistic explanation,
although by his experimental process, he did find a new phenomenon that no
established theories could explain. However, his explanation fails to satisfy ED3,
so Driesch did not make an experimental discovery. One may question whether this
judgment is fair to Driesch, because he did really discover something novel.
Answering the question involves a deeper justification for ED3: Why is the
envisaging of mechanisms necessary for recognizing an experimental discovery?
The term “discover” means “not to cover,” that is, “to remove a cover or shelter
over something.” When one removes a cover over something, one makes something
13 David Gooding ( 1990 ) provided another example. The phenomenon that the current in a long
wire can produce a round magnetic field is extracted from the physicist Jean-Baptiste Biot's
experimental process. The contemporary physicists Andr´ Amp`re, Humphry Davy, Michael
Faraday, and Biot himself, respectively, constructed different theories to explain the phenomenon.
So a (significant) experimental phenomenon is always revealed by some certain “adequate” and
“correct” experimental arrangement. If the experimental arrangement or procedure is inadequate
or incorrect, the experimenter may find no significant phenomenon.