SEARCHing for Truth

Dateline: November 10, 1999

Rate, according to the Criteria of adequacy, each Hypothesis

Simply listing the evidence for each hypothesis isn't enough in to distinguish which of them is the best and most worthy of belief. There are other factors which we need to consider and which will go a long way in helping us make our decisions.

Testability

The first is Testability. Skepticism is a means of acquiring reliable knowledge, and as such only those claims which can be tested are ones which can dependably labeled as true or false. Because hypotheses cannot be tested in total isolation and only have consequences with regards to a background theory, a new hypothesis must be tested in conjunction with that background theory. For example, if you're trying to assess whether a particular object is capable of levitating, you have to know a little bit about gravity and the normal ways of counteracting it. This means, however, that the hypothesis must add something new which the background theory does not - if we don't learn anything new with the hypothesis, then it is simply not worth our time. Going back to our example, helicopters can "levitate," but since we already know the principles behind it, it can hardly be said to be interesting to the point at hand.

This also means that what we are testing for must be entirely new and that we cannot use the data we are trying to explain as that test. A common situation where people don't follow this principle comes once again from the attempt to use "god" to explain existence. When someone looks at the universe and fails to find a reason for it they too often say that the universe exists because "god" made it. In such a situation we cannot then "test" the hypothesis of a god by pointing to the universe again. The universe was the original data we were trying to explain and so we need something new and different. If you are presented with two hypotheses which allegedly explain events and one is testable while the other is not, then the testable hypothesis should be preferred.

Fruitfulness

The second factor to consider when comparing claims is Fruitfulness. Fruitful hypotheses will make predictions which are not only testable, but which also expand the frontiers of knowledge. When a hypothesis makes an unusual and unexpected prediction which is at first disputed but then later shown to be true, it will be given a lot more attention and credence than a hypothesis which has made verified but uninteresting predictions. Einstein's Theory of Relativity made very dramatic predictions about light bending around strong gravitational fields which no one expected and, at first, no one believed possible. But the predictions could be tested and once they were, their proof made people take a long hard look at what Einstein had to say. The frontiers of knowledge were expanded dramatically - Einstein's theory was very fruitful.

Scope

The third factor is the Scope of a hypothesis, or how much data is can explain. Clearly, the more a hypothesis can explain the more preferable it will be in the long run. Narrowly focused hypotheses are fine but eventually will have to be integrated into larger hypotheses if they are to be useful. A hypothesis which covers a wide range of verified data is also less likely to turn out to be false later on because it will have so much more evidence in its favor. One of the chief reasons Einstein's Theory of Relativity eventually overtook Newton's theories of gravity and motion is because Einstein's theory had much greater scope - Relativity explained everything that Newton's theories did plus much more which Newton's theories could not.

Simplicity

The fourth factor which should be considered is that of Simplicity. If given the choice between a complex and unwieldly hypothesis and a simple hypothesis when both explain the exact same phenomena, then the simpler one is generally preferable. The idea of simplicity is related to that of Scope in that both lead to the goal of achieving a greater unification of knowledge. More scope means that a hypothesis explains more phenomena and more simplicity means that the phenomena are explained in a simpler and easier fashion. Unfortunately, simplicity isn't easily quantifiable like predictive power can be. To a certain degree, simplicity is in the eyes of the beholder, but not entirely. Some elements of simplicity can be described, for example a simpler explanation is one which makes the fewest assumptions. This principle is often expressed via Occam's Razor which stipulates that assumptions should be kept as simple as possible and to a minimum.

Conservatism

The fifth and final factor which should be considered is that of Conservatism. An important condition for something to be regarded as knowledge is its consistency with other known facts. When ideas are inconsistent, either the inconsistency is only apparent or else one of the ideas cannot be true (if either actually are). When we have a system of ideas which are well established and meet all the other above criteria, then it is generally a bad idea to try and adopt a new hypothesis which is not consistent with the current system. So when comparing two hypotheses which purport to explain the same data, then whichever one is most consistent with currently well established should be preferred. This doesn't mean that conservative hypothesis is necessarily more true, but introducing unnecessary inconsistencies is simply a recipie for disaster.

That, of course, assumes that all other elements are equal, which is rarely the case. Conservatism taken to too great a degree can inhibit new discoveries and advancement, so it should be treated with care. Unfortunately, it isn't really possible to know for sure when other factors should take precedence over conservatism - it's a judgment call which can't be made perfectly but which can get better with practice.

All of this, taken together, serves as a formalized method of skeptically approaching claims. Although it may seem like a lot, most of it encompasses things which we all do anyway even if we don't often do them consciously. If we do try to do them deliberately and consistently, however, we are more likely to succeed - in every sense of the word.

Don't miss the other section:

Part 1: Criteria of Adequacy