Wednesday, March 6, 2013

The Scope of Falsification

Moon cresting over trees
What does it take to falsify a claim?

I think I've been intuitively understanding this up to this point, but until I've addressed recent apologetics, I then started thinking about it more in detail. It may take some more hashing out, so this will be sort of a first draft.

Categorical Differentiation

I yoinked this out of a previous post, because I realized it wasn't fleshed out enough:

When it comes to whether information can occur naturally (assuming the definition doesn't itself require an intelligence), we can ask ourselves a question. Is it possible for information to occur naturally? Assuming we haven't positively demonstrated this yes, there are two other responses - "no" and "I don't know". The "I don't know" is allowing for the possibility, keeping and open mind. The "no" answer is positively asserting a negative, which has a burden of proof, and which cannot be proved - you can't prove a negative. How would you? It could be the natural mechanism that produces information occurs in a way we can't even conceive of yet. You'd have to be omniscient to knowledgeably hold that position.
This comes to down to the null hypothesis. The only thing we can do is apply the null hypothesis - that there is no requirement for information to come from an intelligence, and then work to disprove that (by showing it can happen naturally). So far, all our observations have DNA in nature occurring... naturally (on its own without intelligent interference), even if we aren't exactly sure how abiogenesis might have happened yet. 
Back to nature versus complexity, we can also ask ourselves a question - Is it possible for nature to produce infinite complexity? We're again faced with the same options. If they say no, they have to actually demonstrate that claim, which is impossible. If they insist on this idea that there's a level of complexity that nature cannot produce, it is necessarily a presupposition - held as true without demonstration.
I realized that it sounds as though I'm descending into some solipsistic rabbit hole, where no one can know anything because we can't have absolute knowledge.

After thinking about it some more, I realized that what I was addressing specifically was a  categorical differentiation for different claims made.


I often make this point about the concept of intolerance. There's a cliche of liberals where they're supposed to be tolerant, and conservatives are intolerant, and every time a liberal is criticizing or opposing something bad that someone else is doing, they're violating that idea that liberals are supposed to be tolerant. It's sort of an "intolerance intolerance" conundrum.

It's not a conundrum. The reality, that most liberals cannot seem to articulate, is that there are legitimate and illegitimate things to be tolerant about. It's just understood that when the "tolerance" is discussed, that it's specifically the "legitimate tolerance". For example, intolerance of racism is legitimate. Intolerance of racial diversity is not.

There's more nuance to the discussion that resolves what people already understood intuitively.

I think there's a similar problem when it comes to the Scope of Falsification, and Scope of Demonstration.

Scope of Falsification

So I ask again, what does it take to falsify a claim?

Well, it depends on the claim. Consider these claims.

  1. My dog flies constantly.
  2. My dog can fly.
  3. All dogs fly constantly.
  4. Most dogs can fly.
  5. Some dogs can fly.
You may notice that the claims start with very specific instances, and keep getting broader. How would we falsify these claims? Here's some possibilities (corresponding respectively to the above):
  1. Show a time when the dog isn't flying.
  2. ---
  3. Show a dog that isn't flying.
  4. ---
  5. ---
Note that I blanked out on 2, 4 and 5. There's a falsification issue with those. What does it mean that a dog "can fly"? How would we falsify that?  It'd be difficult.

There's a big difference between saying that my dog flies, and that my dog can fly. If we observe the dog, with footage, for his entire life, from birth to death, and we never observe the dog flying, did we prove that he can't fly? No - we demonstrated he didn't fly. It's possible the dog had some unknown capacity that he simply chose never to utilize, or didn't realize he could utilize. 

The introduction of a simple word "can" into the claim makes it insufferably difficult to falsify. 

One has to carefully evaluate the scope of claim, and be able to discern what it'd take the falsify it - what I'll call the "Scope of Falsification".

For example, it's significantly easier to falsify the claim that humans evolved from earlier primates than it is to falsify the claim that some species have evolved. The second claim is much more broad, and would require an exhaustive amount of refutations accomplish.

This is why the "poking holes in evolution" is an impotent way at falsifying the theory.

Burden of Proof

Here's the good news. We don't have to falsify all broad bizarre claims, because there's a step we've skipped - the burden of proof.

Going back to the claim that "my dog can fly" - we have full evidence that through his entire life, he never actually flew. That wasn't a demonstration of his non-capacity, so much as a non-event.

We don't have to accept the claim as true until disproven. We don't have to accept that the dog can fly merely because weren't able to falsify it. That'd be silly. Intellectual honesty only requires us to consider the possibility if and when I've positively demonstrated that Woofies can fly.

I've struggled in the past to explain why the burden of proof is so important (a concept that Matt Slick at violates left and right, I've noticed). I think I've understood intuitively in the past, but I think the context of this topic helps bring it into focus.

If a claim is unfalsifiable, it's betraying a functional problem of the claim. Broad claims are hard to falsify, and specific claims are easy. If I can actually do a demonstration of Woofies flying, that specific demonstration can be shot down if it turns out the dog is being carried around by fishing line.

There's a utilitarian aspect as well. The point is to make progress in regards to human knowledge. Making a vague claim about dogs somewhere... "out there"... who can fly, doesn't do anyone any good. A specific demonstration gives us something to evaluate, learn and study.

Scope of Demonstration

The burden of proof ties heavily into the Scope of Demonstration - which is, predictably enough, the question of - what would it take to demonstrate a claim?

If we go back to the dog claims (copied for convenience):
  1. My dog flies constantly.
  2. My dog can fly.
  3. All dogs fly constantly.
  4. Most dogs can fly.
  5. Some dogs can fly.
These might be the demonstrations:
  1. A complete birth-to-death record of it always flying (very difficult)
  2. A single demonstration of the dog flying (very easy)
  3. A complete record of all dogs in existence flying all the time (very difficult)
  4. A record of a high percentage of dogs flying at least once (difficult)
  5. A record of a small percentage of dogs flying at least once (somewhat difficult)
You might notice that the difficulty level for demonstration is the compliment of the difficulty level for falsification. 

Maybe instead of calling them "broad claims", it should be "exhaustive" or "absolute" claims. Tentative ones like "some dogs can fly" would very difficult to falsify. Does that mean it's practically unfalsifiable?

Unfalsifiability and Practicality

Listening to what I'm saying, a person might come to a conclusion that the statement "some dogs can fly" would be unfalsifiable. It's the difference between establishing something to an absolute degree, or within practicality.

Let's be reasonable. If we evaluate my dog for 10 minutes, find that he has no apparent mechanism for flying, and he never does fly within that time, it's not unreasonable to come to a conclusion that the claim is false... as long as we understand that conclusion itself isn't absolute - but rather tentative.

Within practical reality, our efforts towards knowledge are more probabilistic than like a mathematical proof. We hold things as true to different degrees of confidence, but never reaching 100%.

Likewise, while it may not break any laws of physics to monitor the entire lives of every dog in existence, it would be practically impossible... so, at the end of the day, essentially unfalsifiable, but we don't need to go that far in the first place.

"Unfalsifiability", generally is achieved when someone has a claim whose falsifiability is pushed to such an extreme that it's practically impossible to falsify. Carl Sagan's Dragon in my Garage comes to mind.

There's a lot of terms and concepts flying around here - absolutes versus practicality, mundane versus extraordinary, demonstration versus falsification. It can get quite fuzzy.

We can throw a party though, because thanks to the burden of proof, we don't even have to worry about any of that until the claim is positively demonstrated. Then, we can sharpen our knives and get to work.

No comments:

Post a Comment