Follow up to: The Metaphysics Dilemma

My main conclusion by the end of “The Metaphysics Dilemma” was that naturalism as a position should not be the metaphysical claim that only matter exists, it should be the epistemic claim that we have no reason to adopt supernatural positions (positions that involve irreducibly mental entities) because such positions are incoherent.

But if the concept of souls and Gods are incoherent, why would people ever endorse them as answers to a question? Who would ever answer questions of consciousness with that of a soul, answer questions of morality with that of divine commands, or answer questions of how the universe came to be with that of God?

Well, because answers such as these make the question go away.

What is an Explanation?

Once upon a time… say 1667… scientists were attempting to explain the entire combustion process, including flames and rust. The scientists came up with this idea called phlogiston, which was an odorless, invisible substance that was ejected from objects as they burned.

Substances that were easily flammable were said to be rich in phlogiston, and the fact that things would only burn for a certain amount of time before the fire was extinguished was a clear indication that all the phlogiston had been released. The fact that you could “snuff” a flame by enclosing it was an indication that certain amounts of air could only contain certain amounts of phlogiston.

Even the act of respiration (breathing in and breathing out) in animals (such as humans) was considered to be how our bodies removed ejected phlogiston.

The End of Phlogiston

This theory, as compelling as it sounds, surprisingly turned out to be false. It was eventually found that certain materials gained weight after burning while other materials lost weight after burning. And while phlogiston could explain either event (by being said to have either a positive or negative weight), it could not explain both without contradiction.

And since scientists don’t believe that phlogiston works in mysterious ways, they moved on to other theories of combustion. There was no need for an a-phlogistonist movement with books, blogs, and conferences. Luckily, there also aren’t any fringe groups that stubbornly hold onto phlogiston, like those who stubbornly reject biological evolution and common descent.

Rooted in Prediction

But there’s another key lesson: phlogiston could go either way. In fact, phlogiston made no predictions about how combustion would work that weren’t already known — instead, phlogiston was always retrofitted and updated to explain new observations after the fact, and thus was only disprovable when forced to explain logical contradictions.

At least as I argue it, the foundation of knowledge is based upon making predictions that anticipate what you will experience next (see the series starting with “The Origin of Truth” if you don’t agree with me on this).

Dinosaur Comics by Ryan North www.qwantz.com (click to enlarge)

The Anatomy of an Explanation

When we seek an explanation, what is it that we’re looking for? Surely we’re not just looking to satisfy our feeling of curiosity. What does it mean to actually explain something?

Say that we have an event to explain: there is one cookie missing from the cookie jar.

A few hypotheses could come to mind:

• The Fairy Hypothesis: There is a cookie stealing fairy that took the cookie from the cookie jar.
• The Little Jane Hypothesis: A little kid named Jane who lives in the house took the cookie from the cookie jar.
• The Roger Ripley Hypothesis: A next-door neighbor named Roger Ripley took the cookie from the cookie jar.
• The CIA Action Hypothesis: the CIA sent in an agent with a cloaking device who took the cookie undetected.
• The Cookogliston Hypothesis: The cookie jar contains a substance called cookogliston that when ejected causes cookies to disappear.

Which one of these five hypotheses the most reasonable? Which one is most likely to be true? Which one is the best explanation?

What’s Going On?

Explanations seem to be one part causal knowledge and one part historical knowledge: we want to know not just which events in the past, if they happened, would have been capable of causing the event of a missing cookie to logically follow; but we also want to know which causal events in the past actually happened.

Having an explanation is of instrumental importance to us because it lets us understand why events unfold the way they do: if a certain event is observed, we can use causal and historical reasoning to determine that another specific event will occur afterward.

So for each hypothesis, to determine if it is a viable explanation, we need to establish (1) if the suggested event would actually result in a missing cookie and (2) if the suggested event actually took place. The explanation that is most likely to be causal and historical is the one that is most likely to be true.

But how do we test and rank our five explanations above?

Explanatory Virtues

One approach to finding the best explanation is to test it against certain explanatory virtues, where the virtues have been properties of successful explanations in the past and are properties of what an ideal and good explanation looks like, just as durability and lightness are properties of successful hammers in the past and are properties of what ideal and good hammers look like. But what are these virtues?

Luke Muehlhauser in his essay “God Did It is a Terrible Explanation” surveys the philosophical literature and finds six key virtues for explanations to be judged upon, to which I add one more (coherence):

• Testability: A better explanation renders specific predictions that can be falsified. More specifically, the better explanation will make only true predictions, and make some predictions that are novel — not just retrodicting what we already know, but predicting things we observe after the prediction is made.
• Scope: A better explanation explains a wider variety of phenomena.
• Precision: A better explanation explains phenomena with greater precision.
• Simplicity: A better explanation makes use of fewer claims, and especially makes use of fewer unsupported or ad-hoc claims.
• Analogy: A better explanation is consistent with knowledge we already have independent of the explanation, not contradicting previous information without also disproving it.
• Coherence: A better explanation is consistent within itself, not entailing contradictions.
• Past success: A better explanation fits within a tradition or trend of past explanatory success (For example, is from the field of astronomy, not astrology).

Why These Virtues

Why these virtues and not others virtues? Are these virtues arbitrary?

Simply because untestable, imprecise, ad-hoc, overly complicated, and/or self-contradictory explanations are suspicious, and much less likely to be causal or historical.

Additionally, if an explanation seems otherwise plausible but requires us to throw out all we know of atomic theory as useless and false without reason to do so, then the explanation is probably not correct.

Defeating Phlogiston

So based on testing for causal and historical nature, and by using the seven explanatory virtues, we see that phlogiston is not a successful explanation. It is not ideally testable — phlogiston only makes retrodictions, not novel predictions that have been verified. Additionally, it is not precise — we know nothing about what phlogiston is or what phlogiston itself is made of.

Modern theories of combustion succeed much further in all of these virtues — successfully making novel predictions and unify a wide variety of phenomena with the same explanation.

Judging Between Cookie Jar Hypotheses

So now we can use casual/historical reasoning and the seven virtues to judge between the following cookie jar hypotheses:

• The Fairy Hypothesis: There is a cookie stealing fairy that took the cookie from the cookie jar.
• The Little Jane Hypothesis: A little kid named Jane who lives in the house took the cookie from the cookie jar.
• The Roger Ripley Hypothesis: A next-door neighbor named Roger Ripley took the cookie from the cookie jar.
• The CIA Action Hypothesis: the CIA sent in an agent with a cloaking device who took the cookie undetected.
• The Cookogliston Hypothesis: The cookie jar contains a substance called cookogliston that when ejected causes cookies to disappear.

Right away, we worry that some of these hypotheses are completely untestable — if the CIA agent was completely undetectable with a cloaking device, how could we ever establish that he was there? Likewise, how could we verify the existence of a fairy? On the other hand, questioning Little Jane or Roger Ripley is definitely testable — we can see if they crack under interrogation, or if they left any evidence at the scene of the crime.

What about precision? Several of these hypotheses are hopelessly vague — how does the CIA’s cloaking device work, and how do we even know the CIA has one? What is a fairy and what reason do we have to believe that fairies exist, independent of the fact that their existence explains the cookie? And what the hell is cookogliston?!

On the same vein we have problems with simplicity — fairies, cloaking devices, and cookogliston are all being proposed ad-hoc precisely because they make the explanation work, and not because they’ve been verified with any independent evidence.

Only the Little Jane and Roger Ripley hypotheses succeed on the virtues, and could be further decided between by investigating their means and motive. Maybe the cookie jar is too high for Little Jane to reach without considerable effort, or maybe Roger Ripley was verified to have been on vacation in Pittsburgh during the established time of theft.

All of these factors would mean that while both Little Jane and Roger Ripley could have caused the theft of the cookie, one of the two is more likely to have been the one who actually did it in history.

A certain conspiracy theory says that contrails are explained by governments who want to spread mind controlling gasses into the atmosphere. Is it a good explanation?

Making the Question Go Away

Hypotheses like Cookogliston or the cookie-stealing fairy do definitely propose entities that, should they exist, would never fail to cause the event in question. But the question is also one-part historical — do these entities actually exist? Do we have any independent reason to believe in fairies or cookogliston?

Perhaps we could verify it by capturing a fairy or by doing controlled tests and determining that the jar looses weight prior to a cookie spontaneously vanishing, but until then cookogliston and cookie-stealing fairies are just ad-hoc entities invented solely because of their immense causal power.

They just make the question go away — we ask where the cookie went, and then invent an entity that, if it existed, would never fail to explain exactly where the cookie went. But we don’t actually answer the question, because we have no idea if this entity existed.

We should never invent an entity solely because that entity would explain the phenomena — instead, we should seek to explain things only with what we’ve already verified to exist. Otherwise, we discover the tooth fairy to be the best explanation for why teeth are replaced with money when placed beneath your pillow, or justify the appeal to poltergeists to explain the movement of furniture.

Is Casper the best explanation for friendliness?

What Does the Supernatural Explain?

This is exactly what I think is happening with supernatural explanations. What caused the universe to exist? Well, a God of course! But why suspect a God? Simply because a God has all the properties that we would expect to create the universe — such as omnipotence. If God existed, the universe would exist. But does God exist?

It’s like saying the universe was made by the thing that makes the universe — it’s completely incoherent and imprecise. What is the nature of a God? How does a God exist? How is God able to make the universe exist? Why is God omnipotent? We simply don’t know, and probably can’t ever know — even in principle. Such knowledge is impossible.

In fact, the existence of a God explains too much. Like phlogiston, which could explain either burning objects loosing mass or gaining mass and therefore actually explains either, the existence of a God could explain any universe. And if a God could be invoked to explain absolutely everything and never rules anything out, how is it actually explaining something, as in making one outcome appear more likely than the other?

Additionally, theories of God make no novel predictions, are not precise, simple, or testable, and come from a tradition (religion) that has had poor past success. Also, some theories of God contradict our background evidence.

And don’t forget all the other reasons I don’t believe in the supernatural, with a variety of other problems involved when invoking God or prayer.

From John Danaher in http://commonsenseatheism.com/?p=9984 (click to enlarge)

What About Souls?

Souls also don’t explain anything — we ask a question about how consciousness arises, simply respond “Souls!” or “Dualism!” and then the question appears to go away. And this is because souls are defined such that they automatically give rise to consciousness. If only we could simply define brains to automatically give rise to consciousness and make the question go away — but we know better.

Instead, we must question — what is it about knowing we have a soul that makes our consciousness any more understood? Can we use souls to solve any problems in psychology — such as explain mental disorders, why our senses trick us, or how we come to lose our senses? Can souls explain why we’re prone to certain biases and fallacies, but not others? Can souls explain why certain people are psychopaths and certain people aren’t? Can souls explain multiple personality disorder?

Even if we have a soul, we know nothing more about how we think than we did before. In this sense, souls are empty as a concept. Likewise, knowing that God created the universe tells us nothing more about why the universe has certain features and not others, or what features we can expect to find in our further explanations. God does nothing to unify quantum mechanics with relativity, and does nothing to explain why the universe appears to be expanding in an accelerating manner.

Perhaps we now know why Pierre-Simon Laplace is said to have responded “I had no need of that hypothesis” when asking why he did not write about God in his works on astronomy.

What if We Have No Other Explanation?

Perhaps the supernatural could be defended by appealing to the fact that we have no other explanation — the so-called God of the Gaps. But it’s obvious that just because we don’t currently have an explanation that we should adopt a horrendously unlikely and implausible one.

Even if we ruled out both Roger Ripley and Little Jane, we still would have no reason to invoke the CIA hypothesis or appeal to fairies or cookogliston. We would just have no current explanation. But this wouldn’t mean that we should be content with not knowing — instead we should hunt for more evidence until another explanation becomes likely.

Only if we turn up an indication that the CIA had cloaking technology and wanted the cookie for national security, or captured a fairy and were able to verify it’s ability and desire to swipe cookies, or verify in controlled experiments that jars really precipitated the spontaneous disappearance of cookies, would we be justified in believing in these strange hypotheses.

Likewise, we would need some sort of justification for the existence of god(s) beyond them/him/her/it being convenient explanations for a wide variety of phenomena. We could invent any entity and say it explains everything, but that wouldn’t mean such an entity exists, for all the reasons I’ve already argued.

Perhaps we do have this independent explanation in the form of an amazingly inerrant holy book of some sort, but I doubt so. And even if we did, this is not how apologists are reasoning — they start with arguments to the best explanation first and then move on to using the existence of a God to demonstrate inerrancy of holy scriptures, not the other way around.

If Calvin has no explanation for the laws of mathematics, should he use Hobbes's explanation involving what he calls "imaginary numbers"?

Why Invoke and Defend These Entities?

There are simply too many unanswered and unanswerable questions to be made in response to supernatural concepts. Irreducibly mental entities just don’t seem to be testable, even in principle. And even if they were testable, we have no idea how they actually explain anything — the existence of irreducible mental entities seem to add nothing to our knowledge of what they allegedly explain.

However, one remaining lingering question is the opening question of this essay — if these concepts are incoherent and don’t actually explain anything, why do people invoke them? I cannot actually answer this question because I don’t have any evidence or analysis, psychological or otherwise, that I can use to back up my claim, and I don’t just want to speculate baselessly.

Though, if I were to speculate baselessly, I imagine it would probably have something to do with these fake-explanations seeming like they explain what is going on because of their immense casual power, and people ignoring the historical component of explanations, and/or not noticing that this causal power is completely contrived.

Now, even after these entities are invoked as explanations, and then debunked, why would people continue to defend them insistently — either claiming that they must still have explanatory merit, or that they exist for other reasons? Honestly, it further mystifies me. I may not understand my ally here.

Perhaps some believer can help me out, and figure how it is that supernatural entities explain anything, let alone constitute the best explanation?

Followed up in: Proving God Through Cosmology?

Should Calvin stick to his theory of addition in light of the contradictory evidence?

Be Sociable, Share!

•