Previous Lecture Complete and continue  

  Does the quantity of alien sightings affect the likelihood that aliens exist?

Question:

Does the quantity (in and of itself) of alien sightings affect the likelihood that aliens exist?

Put simply, my friend asserts the following claim: “If a million people around the world claim to have seen an alien at some time in their life, then the mere fact that such a large number have reported seeing one makes it more probable that aliens do in fact exist.”

- Don La Mere

This is a very rich question! It does push us to think about what it means to have a good reason to believe something, and under what conditions we would agree that evidence increases the likelihood of a belief being true.

There are a couple of ways to hit this.

The first would be give a quick “yes” to the question — if a million people around the world claim to have seen an alien, all other things being equal, that does count as evidence in favor of the hypothesis that aliens exist.

But a lot turns on that “all other things being equal” part. A critical perspective on the question will push us to ask deeper questions about the case.

For example, we need to clarify what claims are counting as evidence in this case, and what claims are counting as the hypothesis.

The evidence here is the statement that “a million people claim to have seen an alien at some time in their life”.

By itself, this statement is too vague or ambiguous about what’s actually being claimed.

Are they saying they saw something mysterious in the sky that they couldn’t explain, but which they interpret as an alien craft of some kind?

Are they saying they saw physical creatures of some kind, which they interpret as aliens?

Is there consistency across these claims, are is there quite a bit of variety in what people report seeing?

So, there are issues about what the nature of the evidence is that we’re supposed to be considering. If the body of evidence includes the motley assortment of mysterious sky sightings or ambiguous figures in corn fields that make up a lot of modern day reports, that’s quite different from a million people having a very specific, common experience of encounters with physical beings. What you can infer from these bodies of evidence is quite different.

So as it stands, there’s actually no way to assess this reasoning, because the nature of the evidence is unclear.

But let’s proceed with the case, because it raises some important general issues in logic of scientific reasoning.

Let’s assume that evidence is just that “many people report experiences which they interpret as contact with aliens”. That’ll do for our purposes.

I take it that the hypothesis is easier to clarify. It’s the claim that in at least some of these cases, beings from another planet actually interacted with these people.

Now, let’s reconstruct the reasoning that might be offered for the conclusion that aliens exist.

This case involves reasoning from observations or experiences, to a hypothesis about the cause of those experiences.

The most natural argument form is to treat it as a hypothetical argument. Let’s start with this basic form:

1. If H then O
2. O
Therefore, H

In other words,

1. If (aliens exist and were attempting to contact human beings,) then (we would expect to see people reporting experiences which they interpret as contact with aliens)
2. People are reporting experiences which they interpret as contact with aliens.
Therefore, aliens exist and are attempting to contact human beings.

This is the argument form that philosophers of science call “hypothetico-deductive confirmation”.

It’s important to note that the argument form is deductively invalid — it’s guilty of the formal fallacy of affirming the consequent.

But proponents of this argument don’t treat it as a deductive argument, they treat it as a type of inductive argument. So that objection doesn’t carry much weight.

Supporters want to claim that if a hypothesis (H) makes a prediction (O), and the prediction is confirmed, this counts as evidence in favor of the hypothesis.

So the claim isn’t that the evidence O guarantees the truth of H. The claim is merely that evidence O increases the probability that H is true (it counts “in favor” of H).

To reflect this inductivist reading of the argument, we should rewrite it as

1. If H then O
2. O
Therefore, P(H|O) is greater than P(H)

Or,

1. If the hypothesis true, then we would expect to see O
2. We do see O.
Therefore, the probability that H is true, given the evidence O, is higher than the prior probability that H is true.

The problem with this argument form is that, in this simple form at least, we have no reason to think it’s valid or strong. How is the conclusion supposed to follow from the premises? How much does the probability increase? How are we supposed to calculate it?

This is a general problem for inductivist arguments like this, it has nothing particularly to do with the aliens case.

One of the logical issues that arguments like this face is that in most cases, different incompatible hypotheses can generate the same observable predictions. When that happens, the evidence doesn’t allow us to distinguish between them.

Example:

1. If John has the flu, then he’ll have a fever.
2. If John has an infection, then he’ll have a fever.
3. John has a fever.
Therefore, …?

The observation in this case doesn’t tell us which of the competing hypotheses is true.

The 20th century philosopher of science Karl Popper is famous for arguing there is now way of avoiding this logical problem. He rejected inductive methodologies in science, and argued in favor of deductive methods that focus on trying to falsify hypotheses rather than confirm them.

So, a simple model of hypothetic-deductive falsification is this:

1. If H then O
2. Not-O
Therefore, not-H

Or

1. If the hypothesis H is true, then we expect to see observation O.
2. We don’t see O.
Therefore, H is false.

Now, in real science, when we design experiments to test a hypothesis, we usually have some alternative hypothesis in mind to act as competitor. Sometimes this is a specific alternative, and sometimes it’s just a “null” hypothesis which is defined to be logically incompatible with the test hypothesis.

So ideally what we’re looking for is to set up an argument with premises that look like this:

1. If H1 then O
2. If H2 then not-O

This is the logical conditions for what is sometimes called a “strong test”.

Then we test for O. If O is observed, this counts against H2. If O is not observed, this counts against H1.

Note that it would be hasty to conclude that if O is observed and we reject H2, that we should therefore accept H1. That runs us back into the fallacy of affirming the consequent.

No. What we CAN say is that if O is observed, then this bit of evidence strongly supports H1 over H2.

That is, we can make the comparative claim, that H1 is more likely true than H2, given evidence O. But that’s NOT the same as saying that H1 is true, or even likely true.

This is the basic logic of comparative hypothesis testing that is built into standard approaches to hypothesis testing that you might learn in a statistics class in school.

It’s basically Popperian falsificationism, wrapped in a statistical formalism that (unfortunately) tends to obscure the simple logic underlying it.

Now, getting back to our alien example, the lesson we can draw from this is that we need to consider alternative hypotheses, and then whether any of these alternatives could explain the reports of alien visitors.

These are the key variables:

H1 = Aliens exist and are attempting to contact human beings.
H2 = Aliens do not exist.
O = Many people report experiences which they interpret as contact with aliens.

So the critical question can be rephrased like so. Is the situation more like this? …

1. If H1 then O
2. If H2 then not-O
3. O

… in which case we could infer that the evidence strongly supports the alien hypothesis.

Or is the situation more like this? …

1. If H1 then O
2. If H2 then O
3. O

… in which case the evidence doesn’t support one hypothesis over the other.

It’s that second premise — if H2 then O — that is crucial. If there were no aliens, would we still expect to see people reporting experiences which they interpret as contact with aliens?

This is where it matters what kind of experiences we’re talking about. Ambiguous lights in the night sky? Shadowy figures in the cornfield? Experiences of paralysis in your bed surrounded by alien-looking figures?

You can’t answer these questions without getting into the details of the cases. Alternative hypotheses will have be detailed and specific, and their plausibility will rely on background knowledge about such cases.

For example, it will matter to your estimation of the probabilities if you’re familiar with the skeptical literature on UFOs, close encounters, alien abductions, etc. Knowing about cases of fraud, knowing about the psychology of perception, knowing about sleep paralysis and other psychological phenomena that can generate hallucination, etc. All of that will matter to your assessment.

On the other hand, knowing about cases of government cover-ups can influence you the other way. It will also matter if it’s happened to you; if you’re the one who has an experience that you interpret as alien contact. That will affect the probabilities that you assign those these alternatives.

Now, returning to the original question — does the quantity of observations, the sheer numbers, by itself, count in favor of the hypothesis?

I would say that this depends on the details of the alternative hypotheses. If the alternatives appeal to principles of human psychology, or statistical facts, that are universal, or cultural facts that are widely shared (e.g. widely shared experiences of aliens in movies and television), then we might well expect that in a population of 7 billion people, millions of people may have experiences which they interpret as evidence of alien contact.

And if that’s the case, then even millions of reports may not count in favor of the alien hypothesis, because we’re still dealing with a situation like this …

1. If H1 then O
2. If H2 then O
3. O

… where the evidence doesn’t strongly support one hypothesis over another.

Now, if we can get an alien ship landing on the lawn of the White House in Washington, or blowing up New York and Paris, that’ll be a different situation.

I think I’ll stop here, but I hope this is stimulating or helpful in some way!

0 comments