What Is Induction ? Does It Work? Essay, Research Paper
In this essay I will begin by explaining what induction actually is. Then I will highlight the difference between induction as a form of reasoning and deduction in an attempt to clarify the main ‘Problem with induction’. Next I will go through a number of different philosophers attempts to deal with this problem as well as pointing out a number of extra problems with induction. Finally concluding with my opinion on how to deal with these daunting problems of induction.
Induction is a form of reasoning on which most if not all of science is based. Everyday life is also heavily dependent on it. It works by drawing a universal theory from a finite number of empirical observations.
A simple example of induction could be demonstrated by asking a person whether the sun will rise tomorrow morning. That person will, more then likely, answer ‘Yes’. If you ask them how they know this to be true however they will probably say something like ‘Because the sun rises every morning’. By ‘every morning’ the subject means every morning that he has observed. Although this example appears to be true and self evident, it is not logically correct to go from a finite number of observations in the past to a definite prediction of the future. This is the infamous ‘Problem with induction’.
It is important at this point to explain the difference between induction and deduction. Deduction is an argument, where, if the premises are correct then the conclusion is necessarily correct. I.e. If P then Q. P. Therefore Q. For example:
If a brick is thrown at the window then the window will break.
A brick has been thrown at the window.
Therefore the window is necessarily broken.
An inductive argument on the other hand could have completely correct premises and still arrive at an incorrect conclusion. Inductive arguments have two features: true premises provide support for the conclusion but do not guarantee it; and the conclusion contains information that is not in the premises. For example:
Every swan that I have ever seen was white.
Therefore all swans are white.
The conclusion contains the new information that ‘all swans’ not just the ones that I have seen are white. The premises of the argument are correct i.e. ‘every swan that I have ever seen was white’, however, the conclusion is false because there are in existence black swans which I have never seen.
Although induction is just as important in everyday life as it is to science, science is much more worried with the problem of induction. This is because man has developed a method of dealing with inductive reasoning, which provides false conclusions. Instead of losing his mind when the known world throws up unknown or unexpected outcomes, man applies the emotion of surprise. A computer, on the other hand, will ‘crash’ or breakdown when confronted with conclusions that cannot, according to its in-built logic, be correct.
Bertrand Russell recognises the problems with induction but wants to hold on to it never the less, as it is so useful in every day life. He attempts to deal with it by using probability. For example:
The more times A is associated with B the more probable it is that A is associated with B.
The more A continues to be associated with B the more certain it is that A will always be associated with B.
Experience therefore, does not disprove a theory according to Russel i.e. just because it fails to rain on one particular dark cloudy day, doesn’t diminish our belief that it will rain on the next similarly dark cloudy day.
There are, unfortunately, a lot of problems with this solution.
1. Even you could say something was 95% probable, it still doesn’t explain how we can predict the future from observations in the past.
2. How many observations are necessary for the probability to be high enough? The probability of a universal statement being true could never be any percentage except zero because a finite number divided by infinity will always be zero.
3.It assumes that observation is a secure method and that human observation has direct access to external world properties. However it is known that the eye can be tricked and the brain is subject to influence by such things as culture and expectations.
4.The biggest problem with this argument is probably that it is a circular one i.e. It uses induction to solve the problem of induction. It states that induction has worked in the past and therefore it will work again in the future. As induction is not a logically correct argument then neither is Bertrand’s argument.
Hume recognises these problems and comes to the conclusion that it does not work and should be discarded totally. In Humes opinion our beliefs and scientific theories are simple a psychological habit which we cant help but acquire due to the repetition of the same results. Bertrand feels this is too harsh a criticism as it is still a useful form of reasoning regardless of the problems associated with it.
Karl Poppers approach to the problem of induction is to offer the scientific world an alternative to induction – Falsificationism. Popper didn’t believe that science started with observation, rather that scientists made bold conjectures and then tried to falsify them. Every time a theory stands up to a test (isn’t falsified) the stronger that theory becomes. Thus theories are never necessarily right, they just aren’t wrong, or at least have not been proven wrong so far. This theory uses deduction instead of induction. For example:
This A is not B
Therefore not all A’s are B’s
It makes life a lot easier for the scientist as one refuting test disproves the theory as opposed to numerous tests attempting but never quite managing to prove it true.
Inductivists argue that Popper’s theory just refuses to recognise the problem of induction, for if scientific theories are simply bold conjectures then there is no rational basis for believing predictions of the future. Clearly however, it is rational to at least attempt to predict the future for example: If a dog starts growling at you it would be rational not to go any closer to it because the obvious prediction if you do is that the dog will bite you.
So does induction work? Although philosophers have highlighted the problems with induction and the fact that there is absolutely no logical reason to rely on it at all. They do as a rule see that it is still a very useful method of arriving at conclusions. In fact without it we would be completely lost in this world of ours.
Perhaps the best way to deal with the problem of universal statements would be to use a mixture of falsification and inductive reasoning. We could accept tried and tested universal statements until they are falsified. A case of the scientist being innocent until proven guilty so to speak
A.F.Chalmers, What Is This Thing Called Science?, (Milton Keynes, 1982).
B.Russell, The Problems Of Philosophy, (Oxford, 1971).
K.Popper, The Logic Of Scientific Discovery, (London: Hutchinson, 1968).