Re: Serious thoughts about objectivity

Sat, 5 Oct 1996 20:07:00 CDT

John McCreery told us:
>Sustaining the difference between observed and composed is where things get

What's worse is that in the "hard(er)" sciences, we pretend that there is no
composition -- no tendency to organize information so that what we even observe
is filtered/composed in some substantial way. Even when we (for example, we
classically trained, 4-fielders) KNOW how culture constructs knowledge even to
the point of affecting the "objective" observations, we often overlook the
subtle ways in which our best efforts fit what we count as observations into our
composed reality.

One good example was explored at length by SJ Gould in an old Natural History
(early 80s sometime) in a piece he titled "Woman's Brains." As I recall, it was
about the famous work of the French anatomist Broca, who used comparative
studies of male and female brains to answer the question, "Why are women less
intelligent than men?" A social construction/composition of observations that
we fully recognize now, but that was invisible to scientists and scholars in his

A second one is from my own research in which I wanted to look at inbreeding
coefficients within a population of endangered monkeys. SO I asked the computer
programmer to give me a paradigm for generating random coefficients for all
possible pairs. Of course, what I meant was all possible *breeding* pairs --
adult males breeding with adult females. What I got was *all* possible pairs --
living and dead, males and males, immatures and ancients, infants and juveniles,
etc., etc. "All possible pairs" meant something different to each of us -- and to
each of us it was intuitively obvious what the term meant.

The problem is -- and John's example shows this exquisitely -- that without the
"paralinguistic" stuff we have no context for organizing the observations. This
is the classical "signal-to-noise" problem. We don't know what is signal and
what is noise until we put it into a context.

So -- now I have to go off and read Sperber, because I haven't seen this one --
how CAN we separate observations and inferences? Especially when what we count
as an observation can be so substantially affected by the framework in which we
hope to draw those inferences?

>In scientific reporting, I like the good, old-fashioned approach of
>separating observations and inferences. (Dan Sperber does a good job of
>describing the difference in _On Anthropological Knowledge_.)

Andrew J. Petto, Editor, National Center for Science Ed.
PO BOX 8880, MADISON WI 53708-8880
voice: 608/259-2926; fax:608/258-2415
NCSE email:

Take Chances! Make Mistakes!! GET MESSY!!!! -- Ms Frizzle