When Theories Do Battle
In the middle of the twentieth century, there were two main contenders
to explain the fact that over large scales of space and time, the universe
appeared unchanging. Although the big bang theory,^9 which posited the
creation of the Universe in an enormous explosion, was to emerge trium-
phant, it had a strong rival in the steady state theory. One of the key as-
sumptions of the steady state theory^10 was that one atom of hydrogen was
created de nihilis per 10 billion years in every cubic meter of space. That’s
not a whole lot of creation—but it requires abandoning the principle of
matter-energy conservation that is nominally one of the bedrock princi-
ples of physics. However, there are limits to which scientific principles
can be confirmed by experiment—and in the 1950s (and possibly now as
well), it was impossible to measure with a precision that would invalidate
such a result.
There is an uncertainty (which has nothing to do with the uncertainty
principle) that surrounds any set of hypotheses in physics. The best one
can do with any set of hypotheses is to make deductions and test them by
experiment, and the precision of all experiments is limited. In order to ob-
serve the creation of one atom of hydrogen per 10 billion years per cubic
meter, one can’t just pick a cubic meter and observe it for 10 billion years.
Even granted the fact that it would be hard to find someone or something
willing to sit and watch a cubic meter of space for that long a time, you
might be unfortunate and pick a cubic meter in which nothing happens—
the steady state theory obviously talks about an average rather than an ex-
act occurrence. The steady state theory did not fall by the wayside because
atom creation went unobserved—it fell because in an unchanging uni-
verse, there would have been no cosmic microwave background. Such a
background was predicted by the big bang theory to be a relic of the big
bang—and when it was observed by Arno Penzias and Robert Wilson in
the 1960s, the big bang theory emerged as the clear-cut winner.
Physics is frequently confronted with situations in which it must rely on
statistical methods rather than observations—various theories that have
predicted that protons decay, but with exceedingly long time intervals
before they do, so the solution is to assume that there is a frequency dis-
tribution with which those protons decay and watch a large number of
protons. Many physical theories are confirmed or refuted on the basis of
statistical tests—not unlike theories in the social sciences, except for the
fact that theories in the social science are often accepted or rejected on
the basis of confirmation at the 95 percent level of confidence, whereas
physical theories must meet much more stringent criteria.
38 How Math Explains the World