© Roger M Tagg 2009 revised 2010
Welcome to FROLIO – a new attempt to merge philosophy and the "semantic web" . This website is under continuing development.
It may sound an innocuous concept, but Randomness is in fact a contentious issue, even today.
As the Wikipedia page (which itself is subject to dispute) says, there are common meanings which do not equate with the formal definitions as used in subjects like Statistics, where its main application lies. The OED definition quoted includes the following shades of meaning of random.
while the statistical use has meanings like:
Some religious fundamentalists adhere to the view that everything that happens is ordered by the design of an omnipotent God, including what we decide to do next. Therefore randomness does not really exist, but things may appear that way because we have not cracked the code. Some people believe in magical signs and codes (e.g. derived from the Bible) that will tell us what is destined to happen. Several sects have predicted the end of the world - fortunately for us, incorrectly - several times. Earlier religions used divination, augury etc; some people today still read palms, tea leaves or horoscopes - although in most cases, light-heartedly.
The Wikipedia page contrasts the religious views of Martin Luther and C.S. Lewis on free will. Luther allowed very little free will, but Lewis said that God gave us free will.
The Wikipedia page also quotes Donald Knuth (a famous computer scientist) as saying "God exerts dynamic control over the world without violating any laws of physics, suggesting that what appears to be random to humans may not, in fact, be so random".
A philosophical position independent of religion, known as Determinism, holds that randomness cannot exist, and that all we have is unpredictability (because we do not, or cannot, know enough. The Wikipedia page, in explaining this, says "the universe is fully governed by causal laws resulting in only one possible state at any point in time".
A major activity in statistics is sampling, i.e. observing a smaller - but still representative - set of things in order to make decisions about the whole population. Sampling is needed because it would take too long, would cost too much, or would otherwise be impracticable to observe every member of the population. For this to work, selection of which members of the population to sample has to be unbiased (i.e., each member should have an equal chance of being selected). This is where randomness comes in - we select the members to be sampled at random.
In practice this is achieved by using a series of random numbers that indicate which member we should select next. How do we get these random numbers? Well, one can get a computer program to supply them. Other options are to take the next N digits of the constant π (pi, the ratio of the circumference of a circle to its diameter). One can even use the middle digits of a series of telephone numbers in a phone directory. In my first job, we used to have books with tables of random numbers - we just made a chance decision as to where in the tables to start using the numbers.
Of course one can then say that these numbers are no longer random - they are the particular ones in the RAND books, or the ones produced by a certain program etc. Purists therefore call them pseudo-random numbers (see this web article)
Gamblers sometimes use random numbers - giving rise to what mathematicians sometimes call Monte Carlo methods.
Drawing conclusions from samples can be thought of as a special case of Induction. Scientists and mathematicians often use this approach to detecting patterns, and thus to suggesting hypotheses. Of course if the sample is limited in some ways, one may arrive at the wrong conclusion (e.g. that all swans are white, if no Australian swans were in the sample).
It may vary with the granularity we apply in our observations. For example, Ohm's law (that the current in an electrical conductor is proportional to the voltage difference) may be deduced by induction from many observations of conductors, but the process at the microscopic level (i.e. movements of electrons within atoms) may to all intents and purposes be regarded as random.
Statistics and induction alike can never account for all the variation one can observe in a physical process or phenomenon. Any explanation of the observed facts usually includes a measure of noise. As is recognised in FROLIO and many other models, any thing that we can observe is only that which we can differentiate from a background. A physical background is never completely "quiet" - just as a piece of paper is never perfectly white. The noise can come from many sources, but most of the time, as long as we can distinguish the "signal", we regard the noise as random. Sometimes, statisticians call this the "error" element - the sort of thing preceded in everyday speech by the phrase "plus or minus".
Chaos theory is the study of systems that show extreme sensitivity to small changes or initial conditions, and which may display behaviour which might be classed as random, but which is more correctly termed chaos. Weather, especially when it gets turbulent, is an area that chaos theory often addresses. From my youth, I remember seeing the film of the Tacoma suspension bridge violently oscillating in a storm.
The Wikipedia page has a paragraph that distinguishes random from chaotic data.
Talking about chaos is not such a problem for religious people, since at least it is mentioned in the Bible, although before the 6-day creation started. However they would presumably say that it does not exist now.
I (currently!) take the view that all physical things are governed by some natural behaviour patterns (probably never fully understood, but hopefully expressible through FROLIO relationships), and that all living things have some level of autonomous behaviour patterns (again not fully understood) which lead them to develop in certain ways. Plonked together in our universe, all these things get in each other's way and interact. This interaction process is extremely unpredictable and may be chaotic.
I see randomness as a concept that describes that part of what we observe in which we can detect no pattern. Maybe as science advances, randomness is on the retreat, but I think it will never be eliminated, especially given Quantum Mechanics, Heisenberg's Uncertainty Principle etc. Even a statistical view of what happens at the sub-atomic level is not agreed on; there are at least 2 different statistical models (Bose-Einstein, Fermi-Dirac etc).
With abstract things the position is not quite the same, as they are essentially man-made and in some cases have clear definitions. Also, man-made physical things may behave less randomly than natural things.
In the end, I think it may be best to treat randomness as a relative concept, with degrees of randomness - rather than "yes, this is random" or "no, this isn't random".
Index to more of these diatribes
Some of these links may be under construction – or re-construction.
This version updated on 3rd February 2010
If you have constructive suggestions or comments, please contact the author rogertag@tpg.com.au .