Good and Bad Theories

In some previous videos, I've talked about why a scientific theory is not "just a theory"
and why it means something very different than in everyday speech, such as when some
guy in a bar says, "I have a theory about women."
Philosophers have been debating the essential characteristics of scientific theories for
many years, and it's still an active area of research. However, an important concept
was put forward by Karl Popper, who said that a theory should be falsifiable. In other words,
we may never be able to prove that a theory is true in every possible case, but it should
be possible to prove that it is false.
For example, if a theory says that all swans are white, then the more white swans we see,
the more confidence we have that the theory is true. But we can never prove in the mathematical
sense that it's true, because it's not practical to observe all swans. However, as soon as
we see a single black swan, the theory is falsified.
It's important to understand that saying a theory should be falsifiable is not the same
thing as saying that it should be false. Rather, the theory should go out on a limb and be
vulnerable to being falsified. In other words it should make predictions which can be checked
against observation. If a theory is false, we want to find out as soon as possible so
that we can stop wasting time on it. Conversely, the longer a theory has been tested without
being falsified, the more confidence we have in it.
Suppose I put forward the theory that the universe was created last Thursday, and everyone
had false memories implanted in them to make them thing they were around before then. Similarly,
all the physical evidence pointing to a much older Earth was faked, and photons appearing
to come from distant stars were created in flight, close to the Earth.
To someone naive about science, this might seem like a very powerful theory because nothing
can ever disprove it. But on the other hand, we can never learn anything from it, because
any given set of facts can be explained away within its framework. For this reason it's
actually a pretty useless theory.
However, it's not enough to say that a theory should be falsifiable, since theories are
not tested in isolation. For example, when we use telescopes to test theories about the
solar system, we are making assumptions about the laws of optics. That's why we introduce
a further criterion, commonly known as Occam's Razor. Roughly speaking, if two theories are
equally successful at explaining a set of observations, we should go with the one which
makes fewer assumptions.
Let's take an example. Let's compare the geocentric theory, which says that the Earth is at the
center of the solar system and all other bodies travel around it in perfect circles, with
the heliocentric theory, which says that the sun is at the center. Early astronomers favored
the geocentric theory, mainly on religious grounds. However, as their observations became
increasingly detailed and accurate, they had a problem. The data didn't fit the theory
of the sun and planets going in simple circles around the Earth.
To rescue the theory, astronomers came up with the notion of epicycles. Instead of going
around the Earth, a planet was supposed to orbit around a point which in turn orbited
around the Earth. When this model still proved inadequate, astronomers added epicycles on
top of epicycles, until there were as many as 40 epicycles in the system. In other words,
the theory was getting more and more cumbersome. Nothing new was being learned; instead, the
theory had to be hacked to fit each new data point.
When the heliocentric theory was introduced, and it was shown that the data could be explained
by the Earth and other planets going around the sun in elliptical orbits, vast amounts
of complexity were swept away at a stroke. Astronomy was on a much simpler and more elegant
footing which allowed new facts to be discovered, including the planets Uranus and Neptune.
By the same token, if we very generously treat creationism as a theory that can be compared
with the theory of evolution, we quickly see that creationism is incredibly cumbersome
when faced with explaining the observed facts regarding the diversity of nature and the
age of the Earth. Creationists get increasingly legalistic and talk about hypothetical vapor
barriers, and invent whole pseudo-sciences such as "baraminology", in their desperation
to maintain any shred of credibility for their bible-based worldview. By contrast, the theory
of evolution is much simpler and more powerful, as it relies on a small number of key ideas
while explaining and unifying virtually all the life sciences. As Theodosius Dobzhansky
said, nothing in biology makes sense except in the light of evolution.
In other words, the hallmark of a weak theory that will soon be discarded is that it maximizes
its inputs - the assumptions it makes - while minimizing its outputs - the range of observations
it explains within a coherent framework. Creationism maximizes its inputs and minimizes its outputs,
while evolution does the reverse, constantly being strengthened by observation while opening
up vast new fields of inquiry.
Some creationists wrongly claim that the theory of evolution is not falsifiable. This is a
lie. A single rabbit fossil in a pre-Cambrian layer would falsify the theory immediately.
Likewise, although ignorant creationists think that evolution says a frog can give birth
to a horse, if this actually happened, it would falsify evolution at a stroke. Evolutionary
theory has been subjected to the most rigorous scrutiny for over 150 years, and has passed
every test with flying colors. It is one of the most overwhelmingly well established theories
in all of science.
Now, although the existence of God is generally held to be outside the realm of science, I
think we're justified in considering a similar approach. How does the hypothesis that a god
created the universe and dictates how we should live our lives stack up against the alternative
hypothesis that the universe is governed by impersonal physical forces? Once again, we
see that the God hypothesis becomes more and more cumbersome, and explains less and less
as time goes on.
Even in fields such as morality, which is supposed to be untouchable by science, we
see that the idea of morality being defined by a god immediately runs into problems such
as the problem of evil and the Euthyphro dilemma. Theologians try to work around these problems
with theodicy and increasingly obfuscated definitions of morality that are divorced
from common sense.
Even the concept of God himself is becoming more problematic. As the traditional picture
of an old man with a beard in the sky grows more untenable, theologians paint themselves
into a corner. They can't give up their god entirely, since he is their only job security,
but they are so intent on hanging on to something on which they can hang the label "God", they
invent concepts that are increasingly remote from even being meaningful. The old concept
of God is simply wrong, but "sophisticated" theology is incoherent and vacuous. It's what
a scientist would call not even wrong.
It's time to recognize that naturalism is simply a more elegant and powerful description
of the universe. Perhaps we are approaching a paradigm shift when theism will collapse
under the weight of its own sophistry, and naturalism will emerge as the dominant worldview.
We must certainly hope so. The last time the theistic viewpoint reigned supreme, it was
called the Dark Ages.