Trekkie With a Phaser

I know this is a stupid question, but what exactly does the word "falsifiable" mean when used in relation to a scientific theory? In the "What is the ulterior motive of evolutionists?" thread it has been said that evolution is a valid theory because it is testable and falsifiable. I understand the testable part, but what exactly does falsifiable mean in this context?

tronvillain

Well, for it to be completely falsifiable there have to be potential observations that would completely destroy it, along the lines of "If p then q, not q, therefore not p." More practically, there have to be potential observations that would decrease the probability of it being true, along the lines of Bayes' Theorem.

tronvillain

Oh, and this is probably better off in S&S.

Trekkie With a Phaser

So, a falsifiable theory is one which can be destroyed, or at least have the probability of it being true drop, under some circumstances. Why, then, is it a good thing for a theory to be falsifiable? Wouldn't you want a theory to be universally true?

(Sorry if these are really stupid questions, but I'm curious and, honestly, a bit confused)

fando

Well, one important criteria for falsifiability is that the theory must be easily testable. If we have a theory that can't be tested, it's not falsifiable. So stating that a theory is testable and falsifiable is sort of redundant. I think that's where your confusion arises.

Corey Hammer

Falsification is not even that important of an issue anymore. Certain fields, like archeology, may generate hypotheses that are not technically falsifiable (e.g., was Cleopatra at Actium or was she really elsewhere when the battle occurred?). However, it is a perfectly testable question.

In the example, all possibilities for where Cleopatra can be are covered (she was at Actium or somewhere else not specified). Falsifiability means that not all possibilities are covered.

Skydancer

What is meant is not that you want a theory to prove false, but that there are predictions that the theory makes that can be proved true or false by subsequent observations. They are not always as obvious as they are in physics, since the deep-time sciences do not allow for human-designed experiments, but in those cases the observations can be on the order of 'we expect to find the following evidence if we look' or 'we do NOT expect to find this no matter what'. For evolution, this is 'we expect to find fossils with characteristics intermediate between whales and some sort of land animal' or 'we do not expect to find human remains in pre-Pliocene rocks'.

In practice, the 'expect to find' can be weak, since fossils may or may not be preserved, but finding a human fossil in ancient rocks would throw a monkey wrench into things - so the theory of human evolution could be disproved by certain observations, even if we don't _expect_ those observations to occur.

Creationists, when pressed, seldom manage to come up with hypothetical observations that could disprove their ... well, not theories, but speculations... since they're apparently always worried that if they commit themselves, we might actually FIND that something. (Rightly so, since it's happened in the past... not that this stops them from claiming just the opposite.) They reserve the option to take any observations at all and claim Goddidit, which is not a scientific attitude.

Brian63

This is how I interpret it, which may not be technically accurate but it has helped me so far until I can read up a bit more on the philosophies of science.

We can come up with all sorts of loony ideas for explaining any event that we wish. For example, why do objects fall to the ground? It may be because of something we call "gravity," or there are invisible spirits pushing down on them until they hit the earth, or any of a countless number of other possibilities. At this point however, all of these are nothing more than speculation. Interesting they may be, but not very useful to us.

If we want to determine which one of these guesses is actually true (if any), we have to be able to rule out the "bad" ones. To do that, each of these different guesses (hypotheses) has to make some sort of prediction. If hypothesis1 is true, then we should observe event1 in a controlled experiment.

If we conduct that controlled experiment and event1 does not occur, then we know that hypothesis1 is not the "right" hypothesis. We have falsified hypothesis1. If we can conduct similar experiments for each of the other hypotheses (they must also include some criteria that we can use to disprove them, aka they must be "falsifiable"), and subsequently disprove many of them, we have narrowed down our bundle of possible guesses as to which one is the "right" one.

Creationism and Intelligent Design are often criticized because they lack the ability to be falsified. In effect, they are nothing more than speculation. They may be right, but it's not useful to us when conducting science (and thus should not be taught in classrooms to children as if they actually are scientific).

If you want to learn more about the workings behind science, a good intro is this article:

<a href="http://www.infidels.org/library/modern/richard_carrier/SciLit.html" target="_blank">Test Your Scientific Literacy! </a>

Brian

[ June 14, 2002: Message edited by: Brian63 ]</p>

Undercurrent

By "falsifiable", they don't mean "falsified", just "capable of being falsified by an observation that may or may not exist."

For example, I could say "Trilobites are extinct." This would be falsifiable, since someone could observe a live trilobite and prove me wrong. It wouldn't be falsified, because no-one has observed a live trilobite.

The falsifiability requirement serves as something of a safety check when we modify theories in light of new data. If, in order to not contradict any data, we have to modify a theory so much that no conceivable data could prove it false, then it's not a good theory. We're not modifying it to make it a better representation of reality, we're modifying it to maintain our belief in spite of its apparent falsity.

This, incedentally, is what the creationists do with their ever-expanding scope of "micro-evloution". They take their statement that "macro-evolution can't occur" and render it less and less falsifyable as new data arrives to maintain belief in spite of apparent falsity.

m.

tronvillain

Objectivist Man:
I suppose you would, but being unfalsifiable doesn't make a theory universally true. Only logical truths are universally true, and theories are never logical truths - if they were no observations or evidence would be required.

Say your theory is something along the lines of "There are magical pixies that always remove any evidence of their existence." Any observation would be consistent with this theory, so no observation will raise the probability of it being true.