Or look at the physics involved. The ancient cultures involved - and keep in mind we're talking about a time when the wheel wasn't universal -would have needed to take objects with masses of billions of kilograms, apply a force to them over millions of light years, accelerate them to relativistic speeds, and then gently and neatly arrange them in such a way that they mapped out surface features on the Earth. They would have had to have done this so that the maps would still hold true after thousands and thousands of years in which the Earth's axis processed and the stars themselves moved but that would only be useful during certain seasons. They would also have done this without any other cultures inhabiting comparable latitudes noticing - or at very least without complaining that their maps were being screwed up. Pretty darn impossible.

On the other hand, they could have just picked the patterns to match well-known geography. This isn't as far-fetched as it seems. The problem is that the constellations they're talking about are pretty honkin' bright, as far as constellations go (we're not exactly talking about Camelopardalis here - we're talking Orion, Taurus, etc.). It would be an amazing coincidence that the most obvious groupings in the sky correspond to major landmarks on Earth, at least.

And for what it's worth, I find that a lot of the older constellations bear a pretty strong resemblence to their names. Granted constellations like Auriga and Cancer are a little wonky, but Taurus, Scorpius, Eridanus, Leo, etc. aren't that bad. I would, however, like to know where the ancients got the idea that bears had long tails....