Biology Reference
In-Depth Information
I entirely agree with Darden and Craver. Their insight indicates another function
of experimental discovery:
constraining the direction for construction of
hypotheses to discover mechanisms
.
Experimental discovery has still another function:
organizing data into signifi-
cant phenomena
. To show this, I want to emphasize that “the phenomenon” to be
explained by the description of a mechanism is usually a “significant” phenomenon
possessing a repeatable pattern, structure, or regularity. A significant phenomenon
is intelligible, raising a why question and a motive for its explanation. It is worth
exploring and investigating. But the phenomenon may be hidden and invisible, or
not fully revealed, and the data by themselves are not sufficient to reveal it. An
experimental discovery makes it present, visible, and intelligible. The discoverer
does so by constructing
adequate
and
correct
models to organize fragmentary data,
endow data with significance, and reveal hidden structures, patterns, and
regularities. Therefore,
organizing data into significant phenomena
or
creating
phenomena
in Hacking's sense (Hacking
1983
) is really the first and primary
function of experimental discovery.
As I have argued, Mendel discovered (or created) the significant phenomenon of
trait transmission by implicitly constructing two data models (Fig.
6.3a, b
)as
described by the five generalizations. Observations on the number of offspring
alone would not be enough to constitute a significant phenomenon. This would have
amounted only to a large amount of insignificant data about the transmission of
traits from parents to offspring, data that would have been fragmentary and unin-
telligible. Some hybridists before Mendel had obtained Mendelian ratios (Mayr
1982
, pp. 648-649), but they were not and should not be regarded as discoverers,
for they never proposed any data model or envisaged a mechanism to give signifi-
cance to their data. (This will be discussed further in the next section.) Thus, I
conclude that the relation between experimental discovery and discovery of mech-
anism can be characterized by the three functions of the former: organizing data,
producing a motive to search for a mechanism, and constraining the space of
possible mechanisms.
6 What Counts as an Experimental Discovery
Are experimental discoveries recognized without accompanying theories? The
answer is an emphatic yes. What conditions are necessary to recognize an experi-
mental result as a discovery? In other words, what conditions define an experimen-
tal discovery? In order to synthesize the discussion in the preceding sections, I
suggest the following conditions:
(ED1) An experimenter must explicitly or implicitly propose data models to reveal
significant phenomena.
(ED2) No established theories can explain the new phenomena.
Search WWH ::
Custom Search