Tuesday, June 06, 2006

Non-Existence vs. Total Undetectability

Consider the Principle of Sufficient Reason:
For every event, there is a cause.
We can paraphrase this claim in another way:
For every event, there may be a detectable cause, and if there is no detectable cause, then there is an undetectable one.
The business of detectable causes is quite uncontroversial, so I'm going to focus on the undetectable ones. Here's the relevant corollary:
Principle of Undetectable Causation:
Every event that has no detectable cause has an undetectable one.
The negation of this can be expressed as:
Anti-Principle of Undetectable Causation:
There are some events that have no cause, detectable or otherwise.
The problem with both the Principle and the Anti-Principle is that neither has any epistemic test. No observation can validate either proposition. Seeing caused events backs up neither claim because such evidence concerns detectable causes, not undetectable ones.

Given that the rational mind treats like as like, the philosopher cannot rationally justify acceptance of either the principle or its negation because there is no evidence to support either one.

We should not be too surprised to find that wholesale acceptance of such principles is irrational. The root of the problem is in the definition of existence itself.

If I say that a unicorn exists in Sherwood Forest, I am creating a model of a set experiences that are consistent with my claim. Seeing a unicorn in the London city center is not relevant to my claim. Seeing a horse in Sherwood Forest is not relevant to my claim. But seeing a horse with a horn on its head in Sherwood Forest is relevant and consistent with it. And, there may be still other experiences that are incompatible with my claim.

More formally:
The claim X exists => There is some set of relevant experiences such that the set is divided into two subsets: the subset of possible experiences compatible with the claim, and the subset of experiences incompatible with the claim.
(Note that there might be some experiences irrelevant to the claim.)

Now, what if we should discover a horse roaming Sherwood with a plastic horn glued to its head? Do unicorns then exist in Sherwood Forest? It depends. If by unicorn, I meant a genetic unicorn that grew a horn by virtue of its DNA, then no. If I meant something that merely looks superficially like a unicorn, then yes.

What we see here is that the meaning of the existence claim is precisely that set of experiences that satisfy the claim. The two are intimately connected. Many existence claims are somewhat imprecise for this very reason. As in the case of the fake unicorn, we do not always anticipate or articulate every possible relevant experience, and even the things we mean to say are often imprecise.

Given the relationship between the meaning of a claim and the experiences that satisfy it, we must conclude that any two existence claims that apply to the same set of experiences, and which subdivide that set in the same way, are actually the very same claim.

This sets us up for the final blow: if an existence claim and its negation are relevant to and consistent with the same set of experiences, then the claim has no truth value at all. If it were true, then its negation would also be true because they both mean the same thing.

We can see that this is true of the Principle of Undetectable Causation, and that the Principle can be neither true nor false. As a proposition, the Principle is meaningless. It literally says that:
P = We have some specific set of relevant experiences, and every element of that set is compatible with this statement.
and its negation says the very same thing. If the Principle means something, it is not as a proposition.

Likewise, when we say that "the existence of God is compatible with every possible experience," we actually mean precisely the same thing as "the non-existence God is compatible with every possible experience." So neither "God exists" nor "God does not exist" are meaningful as propositions.

After subtracting out the naughty bits, what is left of the Principle of Sufficient Reason? This:
For every event, there may be a detectable cause.
Hey, I can agree with that!


Clark Goble said...

Why should metaphysical principles be verifiable?

And what would count as verification? i.e. how to avoid this merely being a positivist critique of metaphysical principles?

Doctor Logic said...

Hi Clark,

I think this post suggests the two reasons why verification is important.

First, how are we to have confidence in the truth of propositions that are not remotely verifiable? I'm using verification quite broadly here. For example, we can verify mentally that 10 + 15 = 25 by assuming the axioms of arithmetic and computing the sum accordingly.

However, how do we mentally verify the Principle of Sufficient Reason as it applies to undetectable causes? What could we possibly experience that would raise our confidence in its truth? Nothing.

Second, do we really know what a proposition is about when it is not verifiable? In the example I cite here, there's a real possibility that non-existence is semantically indistinguishable from in-principle undetectability. I don't think this linguistic confusion is a coincidence, but is a direct result of non-verifiability of both terms.

When we learn the meaning of language, we do so empirically. We associate words with experiences, so the words become symbols for them. The existence of a thing (an object or thought) is a conformation of the experience of that thing to an expectation (more precisely, an experience of an expectation). When we start talking about the existence of things for which there is no expectation, we are certain to run out of linguistic rope. This is why meaning is tied into prediction and expectation.

There are certainly things that we have yet to discover, things for which we currently have no expectation. We can talk about unusual, unexplained experiences, or the possibility of the world being bizarre. However, it's not clear that we can make meaningful statements about the existence of things that we can never experience, not even in principle.

There is a third way of seeing the necessity of verification that I discuss here. Essentially, everything we know is experienced, whether it is a direct sensory experience of a thing, or the experience of the memory or concept of that thing. But what lies behind the experiences? We can never know because only experiences (be they experiences of sensation or reason) are accessible to us. The best we will ever get is an experience of what lies behind the experiences, which metaphysically isn't any more revealing than any other experience we might have.

unenlightened said...

I might wish to talk about the difference between say, male and female human's experiences. Now I know a lot of what is said on the subject is meaningless, but all of it, necessarily? But I cannot in principle experience both.

"However, it's not clear that we can make meaningful statements about the existence of things that we can never experience, not even in principle." Is this a meaningful statement? It seems clear enough that there are things that no one has experienced and we keep discovering new things - in fact the whole world is new every day, no one has read this post yet. But I can talk about the readers of this post and about their possible responses to it though I cannot in principle experience it while I am still writing this.
I know that only virgins can see unicorns anyway and so I could not in principle see one, yet I feel there is still a difference between their existence and non-existence...
Cheers, bob.

Doctor Logic said...

Hi Bob,

You are right that you can talk meaningfully about things not yet experienced, and even about things that will only be directly experienced by others. However, these cases don't contradict my thesis.

Consider first a proposition about things not yet experienced, e.g., "tomorrow I will experience something new when I watch the Indy 500 for the first time." What possible experiences will be compatible with this proposition, and what possible experiences will not be compatible? This statement is normally taken to mean that there will be experiences both new and notable. As long as I have some reasonably fixed meaning for notability, I can easily identify possible experiences that would verify or falsify my proposition.

Similarly, we can fairly ask what experiences would be incompatible with the claim that virgins, and only virgins, can see unicorns? Well, we would expect to see different pattern of behaviors (e.g., startled looks, eyewitness reports, physiological changes, causal links between virginity and sensory capabilities, etc) in virgins vs. non-virgins consistent with our definition of unicorns.

The key to this proposition is that we're talking about things that are exerienced indirectly. Such claims would only become meaningless when you cannot say what personal, indirect experiences would be inconsistent with the claim.

You really have to push propositions pretty hard before they break in the manner I demonstrate in my post. Again, the test is simple: just name those experiences that are implied and denied by your proposition, and you're home safe. But if there are no possible experiences incompatible with your claim, that's a red flag.

Franklin Mason said...

The set of possible experiences that might serve as verification of some existential proposition is infinite. Thus if an existential proposition means that set of possible experiences, that proposition is infinitely complex. But if it is infinitely complex, it cannot be grasped by minds such as ours. We do however grasp many simple existential propositions. Thus . . .

When you say 'detectable', to whom do you mean? Only humans? Surely not. What organs of sense can be put to use? Only those we have? Again, surely not. I don't see clearly what a good account of 'detectable' would come to.

You speak at times of possible experiences. What sense can an empiricist such as yourself make of that talk? Are there possible nonactual things? You seem to assume that there are. But surely such things are undetectable . . .

Doctor Logic said...

Hi Franklin,

Thanks for stopping by.

I responded in a blog post.

c said...

Isn't verifiability principle about meaning kind of difficult to sustain today, after (post)quinean criticism of logical positivism?

Why constrain meaning to empiric verification / falsification (i'm not talking about those supposed a priori propositions) anymore than to eliminate some obviously metaphysical claims?

I think we could easily forget theology and heideggerian metaphysics (for example) on the grounds of purely pragmatic terms, without the needs to postulate some strong meaning criteria. For example, we could say that, even if the theologian's statement that God created universe is meaningful, it is false given the available evidence and because we give much more importance to it than to supposed biblical-based claims, on the grounds of purely pragmatic validation.

Doctor Logic said...


No, I don't think it's too hard to sustain. :)

Verificationism merely holds that the speaker should know of some experiences that should raise or lower confidence in his proposition. If the speaker does not know this, then the speaker does not know the meaning of his utterance. This requirement stands within a free-floating, holistic model of language, or just about any reasonable assessment of meaning.

When we assert meaning, we are not committing ourselves merely to a string of symbols, but to some consequences in experience. If those consequences are purely mental (e.g., in mathematics), then our proposition has only mental meaning. If our proposition has consequences for physical experience, then our proposition is about the physical world. However, we could not be said to know the consequences of a metaphysical proposition beyond its (mental) consistency with other metaphysical propositions. Thus, metaphysical propositions are really mathematics in disguise.

So, I think that verificationism remains a valid way to detect confusion within meaning.

For example, we could say that, even if the theologian's statement that God created universe is meaningful, it is false given the available evidence and because we give much more importance to it than to supposed biblical-based claims, on the grounds of purely pragmatic validation.

This is true, but the theologian hides behind metaphysical claims, not pragmatic ones. I think verification can flush the theologian out of nonsenseland, and back into the pragmatic sphere. Then, absent metaphysical claims, God is just a space alien.