Facts, Truth, and
Theory
Objective truth as meant by philosophers is impossible to
know since as human beings we can only observe the world through our senses.
Since science cares primarily about discovering ways to predict outcomes in the
world, the philosophical meaning of objective truth is a useless waste of
time. Thus I will be using objective truth, as it is understood by scientists,
to mean repeatable observations which can be replicated by anyone else with
similar equipment and skills. It is important to point out that only
observations and experimental results are considered scientific facts. Models of
reality and the rules, principles, and ideas they are composed of are
scientific Theories. There is one other type of truth in science and that is a
deductive truth. Deductive truths are reached by following logical reasoning
from a set of assumptions (known as axioms in mathematics, or premises in
philosophy). Deductive truths are certainties if the assumptions they are based
on can be demonstrated to be true.
To illustrate these different
types of truth consider the Theory of Evolution. Evolutionary change in
response to natural selection is a deductive truth given the assumptions of:
variation of traits being present, this variation being inheritable, and
reproductive success and survival being in some way dependent on these traits
(all of which have be observed to be true). The observation, that all known
living things share a set of characteristics, is a scientific fact. But the
inference from these two truths that all living things are descended from a
single common ancestor is a theory, the current best model to explain the
observations.
Fitting Data to
Models
The most frequent scientific activity (described by Kuhn as
‘normal science’) is to fit new data to existing models of reality. A scientist
uses their current model of reality to make a prediction or to ask a question
for an experiment/situation which has yet to be examined, then examines that
experiment or situation to determine if the prediction is realized and to
determine a reasonable answer to the question. These experiments and
observations are often complex and difficult to perform or acquire. In addition,
they are performed by human researchers frequently using often novel equipment
or novel applications of existing technology. Both humans and novel techniques
are potential sources or error. Humans make mistakes, equipment or reagents can
be faulty or contaminated, technologies can fail. The experiment itself may be
poorly designed thus not observing what it was meant to observe. Experimental
errors are so common scientists are frequently skeptical if an experiment
appears to have succeeded the first time it is attempted. Even if the utmost
care is taken to avoid errors there is always the possibility the ones results
were just the result of chance rather than a real replicable result. Using the
widely accepted threshold of statistical significance of p < 0.05, it is
expected 5% of results will be simply due to chance.
Models of reality are once again
important for filtering out these errors. Results and observations which
contradict expectations based on the current core model of reality in the respective
field will be considered likely to be the result of experimental error; and
every attempt will be made to correct these errors or correct the experimental
design to make the results fit the predictions or expected range of answers
from the core model. A recent example of this situation was the report of neutrinos
travelling faster than the speed of light.
Even despite these filters many
published scientific results are false. As such, the quality and
believablity of a field of study is directly related to the specificity and the
breath of its core models of reality.
Through
such work the scope and utility of the model of reality is expanded and
verified. Eventually a result which is unexpected or contradictory to a
prediction made by the model of reality is discovered, and resists all attempts
to ‘correct’ it. Such results will lead to modifications to the model of
reality or on occasion shifts to a new different model.
Part 1
Part 2
Part 4
Part 1
Part 2
Part 4
