In contrast, a hypothesis can be proved false for it makes predictions that may or
may not be verified by experiment. To be scientifically useful, hypotheses must
be falsifiable. Thus all scientific theories are constantly in peril of being proven
wrong by new data or observations: experiments are the sword of Damocles for
theories. This is a positive aspect of all scientific investigations for it provides a
natural mechanism for testing and improving scientific knowledge.
An example of a falsifiable hypothesis is Newton’s universal law of gravitation
that predicts the planetary orbits. These predictions are verified, within
experimental accuracy, for all members of the Solar System except Mercury.
The orbit of this planet exhibits a slight deviation from the Newtonian predictions:
Newton’s hypothesis has been falsified. The General Theory of Relativity
also provides predictions for all the planetary orbits, and these agree with
observations, including the case of Mercury. Because of these (and many
other) observations the General Theory of Relativity is now accepted as the
best description of gravitation. Yet this does not mean that future experiments
need to agree with its predictions. Should a discrepancy be confirmed, this
theory will have been falsified and the search for a more accurate description
of gravitational phenomena will begin.
An example of a non-falsifiable hypothesis is the one that claims the Moon is
densely populated with little green men. These beings are then assumed to be
perfectly in tune with human intentions so that whenever we attempt a way of
finding them they would know of this in advance and thwart us. Should anyone on
Earth look at the Moon, the green men have foreknowledge of this and they all
hide. When the Apollo spacecrafts landed on the Moon they all moved to the dark
side (of the Moon) beforehand, and obliterated all traces of their existence. If we
were to observe the whole surface of the Moon simultaneously then, by the time
our equipment is set, they would have dug tunnels deep into the Moon (again
perfectly covering their tracks) and would stay there until we stopped observing. If
we were to obliterate the Moon they would again know of our intentions ahead of
schedule and leave, timing their trip so that we could not see them, etc. It is clear
that by construction the existence of these little green men cannot be disproved,
their existence has to be taken on faith and might be a matter for mystical or
philosophical musing, but it has no place in a scientific discussion.
The experimental verification of hypotheses is paramount, and it is in this
realm that many controversies arise. It is possible for an experiment to provide
an apparent confirmation of a prediction, but this verification is later found to
be the result of poor experimental design, a misinterpretation of the data, the
result of extraneous effects that were not properly taken into account, etc. In
order to minimize these potential problems scientists test predictions using
many different experimental designs. The confirmation of a prediction by a
given experiment will usually set off a flurry of new experiments also aimed at
verifying this prediction, but using many different approaches. In addition the
details of every such experiment are published, and the experimenters open
that would entail testing it in all possible ways under all possible circumstances.