Analogical Evaluation

It's fascinating how often religious apologists will use a disanology to support their argument, thereby demoloshing their own argument.  Here's a simple check list which religious apologists appear to be unaware of...

Five steps to evaluate analogical arguments

So how should we evaluate the strength of an analogical argument that is not deductively valid? Here are some relevant considerations :

Truth : First of all we need to check that the two objects being compared are indeed similar in the way assumed. For example, in the argument we just looked at, if the two novels actually have completely different plots, one being an office romance and the other is a horror story, then the argument is obviously unacceptable.

Relevance : Even if two objects are similar, we also need to make sure that those aspects in which they are similar are actually relevant to the conclusion. For example, suppose two books are alike in that their covers are both green. Just because one of them is boring does not mean that the other one is also boring, since the color of a book's cover is completely irelevant to its contents. In other words, in terms of the explicit form of an analogical argument presented above, we need to ensure that having properties Q1, ... Qn increases the probability of an object having property P.

Number : If we discover a lot of shared properties between two objects, and they are all relevant to the conclusion, then the analogical argument is stronger than when we can only identify one or a few shared properties. Suppose we find out that novel X is not just similar to another boring novel Y with a similar plot. We discover that the two novels are written by the same author, and that very few of both novels have been sold. Then we can justifiably be more confident in concluding that X is likely to be boring novel.

Diversity : Here the issue is whether the shared properties are of the same kind or of different types. Suppose we have two Italian restaurants A and B, and A is very good. We then find out that restaurant B uses the same olive oil in cooking as A, and buys meat and vegetables of the same quality from the same supplier. Such information of course increases the probability that B also serves good food. But the information we have so far are all of the same kind having to do with the quality of the raw cooking ingredients. If we are further told that A and B use the same brand of pasta, this will increase our confidence in B further still, but not by much. But if we are told that both restaurants have lots of customers, and that both restaurants have obtained Michelin star awards, then these different aspects of similarities are going to increase our confidence in the conclusion a lot more.

Disanalogy : Even if two objects X and Y are similar in lots of relevant respects, we should also consider whether there are dissimilarities between X and Y which might cast doubt on the conclusion. For example, returning to the restaurant example, if we find out that restaurant B now has a new owner who has just hired a team of very bad cooks, we would think that the food is probably not going to be good anymore despite being the same