Tuesday, May 8, 2012

Determining Likely Explanations based on Precedent Probability

Topic

When investigating into a phenomenon, on a basic level, we determine what possible explanations are likely, and thus are worth spending time exploring, based on past precedence.

Introduction

When discussing with theists their supposed evidence for the existence of God, we often run into a problem, one that strangles their ability to demonstrate their claims.

An example:

The theist sets out to prove that there's a god. The presented evidence is to drop to one's knees and, with a humble heart, ask God to reveal himself to you, and he will.

There's typically two problems with this approach to evidence.
  1. They're doing it backwards. They start with a conclusion and try to find anything that might support it. This is opposed to the scientific approach, which is to follow the evidence to a conclusion.
  2. Given the supposed evidence, there's multiple explanations for the phenomenon, some more likely than others, and they choose the least likely explanation.

Details

For the sake of discussion, let's assume that the aforementioned phenomenon was consistently repeatable. A person follows the instructions, knees crunching into the ground with the crack of hands clasping together, resisting the urge to make hand-farts, and within moments, the person stands up, and walks away a devout believer.

We have observed a phenomenon. The next step is to ask the question, "What's going on?"

Precedent Probability

Precedent probability is a fancy term that I just made up to refer to the concept that we can figure out probable explanations for something based on how often it has occurred in the past, or how common the pattern is already established.

Example of Precedent Probability

I return home one day and find that the bags of garbage I had outside of my house are torn open and their contents spread about.

I live in Maine. I think about what the possible explanations for this phenomenon are.
  • It could be raccoons. Their trash-picking behavior is very precedented. Their existence in my area is very  precedented.
  • It could be bears. Their trash-picking behavior is very precedented. Their existence in my area is precedented, but not terribly common.
  • It could be cats. Their trash-picking behavior is not that precedented, particularly if the garbage is sealed up. Their existence in my area is very precedented.
I stand in my driveway, glaring towards the wreckage, mentally building a list of potential possibilities to be angry with. For each possibility, I can assess how likely it is. For instance, since bears aren't common in my area, that's not a likely candidate. Since the garbage bags were sealed up, and cats attacking sealed containers of garbage isn't common, that's not likely either.

The raccoons, however, seem like a very likely candidate. Not only do they live in my area a lot, but they go after garbage bags frequently as well. It would be quite rational of me to conclude that's the most likely candidate.

What if I thought of another possibility?
  • It could be penguins.
It's possible. It could be that some penguin someone had, as an exotic pet, escaped and had a desire for my garbage. There's a couple problems with that. Not only is their garbage-picking behavior unprecedented, but their existence is very unprecedented in my area. Heck, they live at south pole.

I would be out of my mind to think that penguins were the most likely candidate.

It could be aliens, but that's even more unprecedented. It'd be taken into a psychiatric ward if I was running around hunting down an alien for tearing into my garbage bags, and rightly so. The only context for which that wouldn't be insane is if garbage-picking aliens were precedented.

The moral of the story is that when we're evaluating possibilities, we can, on a basic level, determine the probabilities of those possibilities. Of course, the evidence should lead us, but barring further evidence, that's a common starting point for investigation and hypothesis generation.

This is not to say that I've concluded it's a raccoon, so much as that's the first thing I'd check out.

Precedent Probability on the Example Evidence

Going back to the presented evidence for the existence of a god, we have a couple of possible explanations:
  1. The claim is true, and one is actually in telepathic communication with a supernatural ghost thing.
  2. You are experiencing a combination of psychological factors and are mistaken in what you think you are experiencing.
Telepathy and ghosts (and gods) aren't just uncommon. They are entirely unprecedented, in terms of objective demonstrable reality.

Conversely, psychological factors, such as biases, perceptual errors, memory errors, priming, etc, are incredibly precedented.

It, without further evidence, would be absurd to think that #2 was more likely than #1.

Disclaimer and Practical Epistemology

As I eluded to before, this probabilistic assessment is not in itself a conclusion. It may in fact turn out that a penguin did tear into my garbage backs, and I did have telepathic communication with a ghost.

It's possible, and further evidence could reveal those to be what actually happened.

We only have so much time in a day, so much money, and so much time. That's the point of assessing probabilities - for time management. If I start on the most probable possibilities, I'm statistically more likely to have my efforts pay off then if I start with the least likely possibilities.

This isn't unusual. Research grants are given to those scientists who can write a grant proposal that can demonstrate that there's merit to the research project the scientists wish to engage.

There has to be a semblance of veracity to a claim or possibility before it's even worth spending resources to investigate.

Extended Examples

Phenomenon: The Bible predicts many things that we are seeing now in the modern era, such as the existence of particles in Physics (ignore for a moment that the Greeks hypothesized this before the Bible).

Possibilities:
  1. Some people were ahead of their time, their data was written loosely into the book and the main knowledge was lost, and now we're rediscovering it in the modern era.
  2. Divine inspiration.
#1 is more precedented, and therefore, the more likely explanation.

Phenomenon: A person is diagnosed with cancer. People pray for the cancer to get the hell out, and the cancer goes into remission.

Possibilities:
  1. Cancer has been known to go into remission without prayer or intervention. Prayer is common, especially since most of the (U.S.) population is religious. The two events align, and is a coincidence.
  2. Divine intervention.
#1 is more precedented, and therefore, the more likely explanation.

Discussion

The main problem apologists have is that they simply skip the other possibilities and head straight for the most unlikely one. Why would one ignore the possibility that the cancer simply went into remission? "No, that can't be it!", they say with a high, shrill voice.

This is one reason why the scientific method doesn't operate on single pieces of evidence. In order to distinguish which is true, #1 or #2 from above, we need additional data. We need to accumulate large quantities of data in order to converge on an answer. Apologists frequently don't understand this, either.

It'd be like trying to conclude what a 500 piece puzzle was a picture of using only one puzzle piece. No, you need to put together a significant portion of the pieces before you can safely make a conclusion.

Rebuts

It may seem, given the assessments above, that it would be impossible to prove a god, because we keep going with the natural explanation. That's where following up with additional confirming evidence helps us come to an accurate conclusion. If the #2's are correct, the evidence should ultimately indicate that.

Once that happens, and a previously unprecedented claim is demonstrated to be true - it's now precedented, and the probability analysis changes.