### Is problem of induction actually a problem of language?

Alifar commented on the post "Fire Always Burns?" and he raised a very interesting point. He wrote:

'Fire, as an entity, has a certain property that (we implicitly assume) is inherent in it. And that property is heat. Remove heat from fire, and we will get a flame that does not burn. This will, in a way, do away with the problem of induction, pro tem.
Any phenomenon that we can classify, based on certain visible characteristics, under the broad hypernym of “fire” must necessarily, under certain arbitrarily set yet inflexible linguistic rules, possess the property of burning. That is, that which we call fire must necessarily burn or else we cannot call it by that name. For a layman, the argument runs as follows:

“If it ain’t burnin’, then it ain’t fire.” '

Apparently, it seems to provide a solution to the problem of Induction, but i believe that it only tries to hide the actual issue, like shoving it under the carpet. The problem of induction is not a problem of pure logic... what we are dealing with is not an analytical issue, but an empirical one. And this is precisely what Alifar's solution does: it treats it as an analytical problem.

Let me explain. Consider the question:

"Can there be a square that does not have four sides?"

This is an analytical problem. The answer is simply no, because a square by definition has 'four sides'. If it doesn't have four sides, it wouldn't be a square. This is a logical conclusion from the general rule that it is self-contradictory to deny an analytical statement. [It would be self-contradictory to say 'A square doesn't have four sides'.]

Now, what Alifar is saying is that the question "Does fire always burn?" is logically equivalent to the question "Can there be a square that does not have four sides?". Just like a square by definition has four sides, a fire by definition always burns.
However, the fine point to note is, the definition of a square is independent of experience; it is a purely geometrical issue. This is how we defined a square in the first place.
But the matter of defining a fire is different. Humans first experienced the fire, and then defined it. And the fact that fire burns is a part of the fire's definition only because humans have experienced uptil now that fire burns. But while it is impossible to even imagine a square not having four sides, it is possible to imagine a fire that does not burn.
Say, let me quote Bible. Exodus 3:2. Moses meets messenger of God:

"The Messenger of the LORD appeared to him there as flames of fire coming out of a bush. Moses looked, and although the bush was on fire, it was not burning up."

So, what does it actually mean when a person asks,
"Does fire always burns?"
What he is really asking is:
"Can i ever experience a phenomenon which has all the properties that we normally associate with fire, except that it does not burn?"

Whether we would call that phenomenon 'fire' or not is a secondary issue. I believe that if such a phenomenon is experienced, people would simply revise the definition of fire.

To take another example, let's consider the question:

"Are crows always black?"

The problem of induction states: Just because all the crows we have seen are black doesn't mean that there can't exist a crow somewhere in the world that is of some other colour [say white, or red.]
Alifar would say that being black is part of the a crow's definition. Hence it is impossible to think of a crow being of any other colour. But that only ignores the real question. What the question really means is:

"Is it possible to find a bird that has all the properties we associate with a crow, except that it is black?"

As obvious, this question cannot be shoved underneath the carpet of language. And this is what the genuine problem of Induction is all about.