New to Theory Mapping?

Theory Mapping is a new and potentially revolutionary method to improve the quality of theories that society uses. It does this by improving the generation, communication, critique, refinement and selection of theories. It is particularly applicable to areas of inquiry which are not amenable to controlled experiment, where it provides a systematic approach to using empirical evidence and logic in the evaluation of theories.

It consists of drafting Argument Maps for each theory (in which ideas and arguments are represented by boxes connected by arrows) and then measuring how coherently they can explain agreed facts.

Whatbeliefs.com is the home of Theory Mapping. For more information the best place to start is the FAQs, which link to all the various posts on the site.
f

Sunday 28 December 2008

Coherent explanation of the facts

l
In what follows, I explain why the test for the truth of explanation of existing sense experiences should be operationalised as coherent explanation of the facts.

To operationalise the test we need to outline what is meant by existing sense experiences, and what would constitute an acceptable explanation.


‘Existing sense experiences'

I would like to define ‘existing sense experiences’ as meaning publicly verifiable experiences which are reproducible for anyone, and can be stated in the form of ‘facts’. For instance, the statement “Britain is an island” may be verified by anyone who wanted to fly around its coast and have the sense experience of seeing sea all the way around.

If it is not possible for anyone to gain a direct experience of the fact, perhaps because it is about an event in history, then it should be formulated to include sense experience of the means of verification. For instance, “All history books state that X” states that anyone looking in a history book will experience sentences stating that X.

If there is an issue over which there is disagreement, the fact should be stated in a way that all sides can agree to. For instance, on the question of whether UFOs exist or not, the relevant fact is that “Many people think they have seen UFOs with the circumstances documented in books X, Y and Z”.

On any subject, there will be a particular set of facts that the competing theories need to explain. In addition, particular theories may have facts that they see as constituting evidence for their point of view. Alternative theories will also need to explain these facts. For instance, the main fact to be explained on UFOs is the reported sightings. UFO sceptics might want to point to the fact of differences in the descriptions of aliens and their craft. While UFO believers might want to point to facts concerning government secrecy over UFOs. An adequate theory must therefore be able to satisfactorily explain all of these facts.


'Explanation'

The explanation of a fact must meet three conditions:

1. Outline the first-level cause

This is the immediate cause of the fact. For instance, an explanation of “Many people think they have seen UFOs with the circumstances documented in books X, Y and Z” is "Many people have actually seen UFOs".

This first-level cause enables the sense experiences to be predicted in the future in a very rudimentary way, and so ensures that the explanation provides a minimum level of useful information. For instance, explanation A predicts that when a UFO is present, then people will think they see it.

But this is clearly not enough as a test for the truth, since many other explanations could be provided for the same fact. For instance, in the case of UFOs, and an alternative explanation could be "Many people have seen certain things under certain conditions and mistaken them for UFOs". There has to be some justification for why the explanation is true. This leads us to the second condition.


2. Provide a valid deductive argument

There must be a valid argument for the first-level cause using deductive logic. A ‘deductive argument’ is one in which it is impossible for the premises to be true and the conclusion false. For instance, a possible deductive argument (loosely written) for explanation A is:

  1. 1. People who are well-respected members of the community are more trustworthy than government scientists (premise)
  2. 2. Many well-respected members of communities think they have seen real UFOs, while government scientists argue that they were mistaken (premise)
  3. 3. Many people have actually seen UFOs (conclusion).

While an argument for explanation B could be:

  1. 1. If there are different explanations for something, the explanation that is most coherent with current scientific understanding should be preferred (premise)
  2. 2. If people think they have seen UFOs, either they actually saw UFOs or they saw certain things under certain conditions and mistook them for UFOs (premise)
  3. 3. That sightings of UFOs are caused by visual mistakes is more coherent with current scientific understanding than that real UFOs were seen, since science cannot explain how UFOs can move so quickly (premise)
  4. 4. Many people think they have seen UFOs (premise)
  5. 5. If people think they have seen UFOs, the preferred explanation is that they saw certain things under certain conditions and mistook them for UFOs (conclusion from 1, 2 and 3)
  6. 6. Many people have seen certain things under certain conditions and mistaken them for UFOs (conclusion from 4 and 5)

The deductive type of argument must be used because other types contain fallacies which if corrected turn them into deductive arguments. So for instance, an ‘inductive argument’ that ‘A causes B’ based on a statistical correlation between A and B, requires additional premises about there being no third factors and about the direction of causality. While an argument of ‘inference to the best explanation’ that ‘A exists’ because O exists and E is the best explanation of O, requires a deductive argument for why E is the best explanation of O.

However, a deductive argument will only come up with a true conclusion if the premises are true. This leads us to the third condition.


3. Belong to a coherent belief system

As we have seen, some premises may be justified from the fact itself (as for 2 for A, and 4 for B). However, it will always be necessary to bring in some premises that are not contained in the fact itself otherwise it would not be in need of explanation! If each premise needs to be justified by other deductive arguments with their own premises, then there is a danger of an ‘infinite regress’ of premises. Two approaches to this exist in philosophy:

Foundationalism argues that belief systems are like a building in which basic beliefs provide justification for all other beliefs, like the foundation of the building. These basic beliefs do not themselves need justification from other beliefs, either because they are self-evidently true, like the truths of logic and mathematics, or because they are justified directly from sense experiences. However, this view has now been largely discredited within philosophy. Truths which were previously thought self-evident in logic and mathematics have been called into question e.g. by quantum physics. And sense experiences can only yield justified beliefs about the world when combined with background beliefs about how to interpret those experiences e.g. what they mean and how reliable they are, and so the infinite regress is not stopped.

Coherentism argues that there are no basic beliefs, with beliefs being justified by belonging to a coherent belief system. On this view, belief systems are like a spider's web where the strength of the web is provided by its coherent structure, in which each belief or line of web is supported by the others but also plays a role in itself providing support. Beliefs differ in how many other beliefs they support, and hence in their structural importance to the overall system. This is just as sections in a spider’s web nearer the centre are more important structurally than sections nearer the periphery.

There are different views as to how to define 'coherence', but the main view is that it consists of two aspects. Firstly, that the beliefs are logically consistent i.e. they do not contradict themselves. And secondly, that the truth of each belief is justified by others in the system. Since we are insisting on deductive logic to provide justification, then we may say for every belief must be the conclusion of a deductive argument in which the premises are other beliefs in the system.

Not only does coherentism solve the infinite regress problem, but it also seems to be a good description of how our belief systems actually work. For instance, the strength of our belief in logic and mathematics could be said to come from the fact that they are in the centre of all our belief systems, with all the other beliefs depending upon them. Although they are not themselves justified logically, they provide logical justification to all of our other beliefs. Their presence in the system therefore increases its overall coherence and hence is beneficial.

The main criticism is that coherentism is neither a necessary nor sufficient condition for truth. It is not necessary because there is no reason why the truth should appear coherent to our limited human minds, and it is not sufficient because you could have more than one equally coherent belief system. To answer this criticism we have to go back to our definition of truth as the accurate prediction of sense experiences. In order to be able to make predictions about the way that reality interacts with us, we have to have coherent theories that we can practically use. Some level of coherence therefore is necessary for the truth.

In terms of whether one can have more than one equally coherent system, I think there are two concerns here. One is that purely focusing upon the coherence of a belief system within itself allows it to float free from any connection to the sense experiences we want to explain and predict. However, this is not an issue for us because we are insisting that the belief system explain certain facts. Continuing the spider web analogy, these facts can be thought of as fixed points in the environment that the web attaches itself to and gains structural support from - a spider's web without such points to hang from simply collapses. So not only is the belief system explaining the facts, but the fact that it is able to explain them gives it justification.

The second concern is that you could have two equally coherent explanations of the facts. In fact, you could make up an entirely fictitious account and it could be just as coherent as a serious explanation, much like a mentally ill person can create a coherent belief system for why people are trying to kill them. I am not convinced that this problem would ever occur in pratice and believe that experience of such cases is needed in order to formulate an appropriate response. For instance, one possible response is to require that in such a situation the simplist explanation (i.e. with the least number of beliefs) be preferred. This is a popular criterion known as 'Ockham's Razor' after William of Ockham who said "Plurality should not be posited without necessity". I have not yet come across an adequate argument for it, but if experience was to show that deliberately fictional explanations are more complex than serious ones perhaps a case could be made for it. A more promising possibility would be to judge the explanations on their usefulness for making predictions. We said earlier that explanations of facts only need to address the first-level cause. But if they can go beyond this to deeper level causes then the belief system which be much more useful in predicting future sense experiences and guiding action. For instance, for explanation A in the UFO example, a second level of explanation could be about the reasons behind UFOs visiting earth. This would be extremely useful in predicting future sightings and guiding governments in how to respond to them if the sightings were real. How this judgement of predictive usefulness or explanatory depth could be done in an objective manner would need to be worked out through specific cases.

I therefore believe that the coherentist approach is the most defensible, and that the third condition should be that the premises of the explanation belong to a coherent belief system. In the case of most intractable debates (e.g. over religion and the supernatural) these belief systems will each have their own beliefs about truth and how to find it. So the belief system of a believer in UFOs may well include conspiracy theories which are used to support premise 1 for explanation A. Equally, the system of the UFO sceptic may have theories of the scientific method that are used to support premise 1 for explanation B.


Putting it all together

Since the immediate arguments justifying the explanations and the overall coherence of the belief system should all be in terms of valid arguments using deductive logic, we can join the second and third conditions for an explanation together to state that all beliefs within the system including the explanations should be coherent i.e. logically consistent and logically justified. The test of truth may therefore called coherent explanation of the facts.

0 comments:

Post a Comment