Information Technology Reference
In-Depth Information
the birth of Sir Isaac Newton (1642-1727), experiments did not yield the kind of repro-
ducible results we are accustomed to accepting today. Every freshman physics course
has a laboratory experiment on measurement error that is intended to make students
familiar with the fact that experiments are never exactly reproducible; there is always
experimental error. But this pedagogic exercise is quickly forgotten, even when well
learned. What is lost in the student's adjustment to college culture is that this is proba-
bly the most important experiment done during that first year. The implications of the
uncertainty in scientific investigation extend far beyond the physics laboratory and are
worthy of comments regarding their significance.
Most people recognize that they do not completely control their lives; whether it is
the uncertainty in the economy and how it will affect a job, the unexpected illness that
disrupts the planned vacation, or the death of one near and dear, all these things are
beyond one's control. But there is solace in the belief, born of the industrial revolution,
that we can control our destiny if only we had enough money, or sufficient prestige
and, most recently, if we had an adequate amount of information. This is the legacy of
science from the nineteenth and twentieth centuries, that the world can be controlled if
only we more completely understood and could activate the levers of power. But is this
true? Can we transfer the ideas of predictability and controllability from science to our
everyday lives? Do the human sciences of sociology and psychology have laws in the
same way that physics does?
In order to answer these and other similar questions it is necessary to understand how
scientists have traditionally treated variability and uncertainty in the physical sciences.
We begin with a focus on the physical sciences because physics was historically the first
to develop the notion of quantification of physical laws and to construct the underlying
mathematical infrastructure that enabled the physicist to view one physical phenomenon
after another through the same lens and thereby achieve a fundamental level of under-
standing. One example of this unity of perspective is the atomic theory of matter, which
enables us to explain much of what we see in the world, from the sun on our face to the
rain wetting our clothes or the rainbow on the horizon, all from the same point of view.
But the details are not so readily available.
We cannot predict exactly when a rain shower will begin, how long it will last, or how
much ground it will cover. We might understand the basic phenomenon at the micro-
scopic level and predict exactly the size and properties of molecules, but that does not
establish the same level of certainty at the macroscopic level where water molecules
fall as rain. The smallest seemingly unimportant microscopic variation is amplified by
means of nonlinear interactions into a macroscopic uncertainty that is completely unpre-
dictable. Therefore it appears to us that in order to develop defenses against the vagaries
of life it is necessary to understand how science treats uncertainty and explore the limi-
tations of those treatments. Most importantly, it is necessary to understand what science
got right and what it got wrong. For that we go back to the freshman physics laboratory
experiment on measurements.
Each time a measurement is made a certain amount of estimation is required, whether
it is estimating the markings on a ruler or the alignment of a pointer on some gauge. If
one measures a quantity
q
a given number of times,
N
say, then instead of having a
