One of the peculiarities of forming theories is that there is tendency to try to explain everything. For Aristotle, one of the most peculiar aspects of nature was the power of the sun. Where did the heat and light come from? An important observation was that the Sun’s output was known to have been constant for several thousand years, and a quick calculation showed that had it been powered by combustion, such as burning coal, it should have faded. It had not. Now there was a questionable issue here: how far away was the sun? Some time after Aristotle, Aristarchus measured this distance, and was the first to realize how big the solar system really was, and since his measurement was somewhat error-prone, he seriously underestimated the size of the star. Nevertheless, the argument was correct in another sense: if the star was further away, the power had to be correspondingly greater, so qualitatively the argument stood. So, what powered the sun?
There was only one possible explanation that Aristotle could see: the Sun had to be moving, and by moving, it generated a lot of friction, because such friction would be the only physical means of powering the star. The earth did not generate heat, therefore it was not moving. Note that it was not Rumford who established that friction generated heat, in fact the first would be the one who discovered how to start a fire by rubbing one stick in the cavity of another. Aristotle knew that, but somehow in the middle ages the knowledge got overturned by the concept that heat was some subtle fluid called caloric. So, what Aristotle did was to take the only explanation he had that was possible, and also one that helped his theory. It would be too much to expect the ancient Greeks to guess nuclear fusion, but it shows that when developing a theory, every now and again something turns up that should not be explained. There is no fault in admitting you do not know everything.
So, what was the weakness in that theory? The first one might be the phases of the moon. The moon was moving as well, but the phases of the moon were to be explained in terms of reflected sunlight, which is correct, but it meant that the moon was moving approximately as fast, but generating trivial amounts of heat and light. Why was this? Yes, you could find an explanation, but the problem then became, a new explanation was required for one additional fact.
Another interesting fact is that Aristotle and other ancient Greeks considered stars to be other suns, but a long way away. Again, true, but their light was considered to come from the same source: friction. The problem with that is that for those on the equatorial regions their angular velocity was very close to the same as that of the sun, which meant that if they were x times further away, they were going x times faster across whatever was providing the friction, and hence they would emit x times the energy. They should be a lot brighter than they are. A second problem was that those near the poles are travelling much slower, and in principle, the pole star does insignificant travelling. If so, there should be a general dimming from equator to pole, but there was not. Finally, since they have different degrees of brightness, it was argued (correctly) that they were different distances away, but if that were the case, they all had to be travelling on separate disks, all with the same periodic time, but all with different velocities. At the very best, an incredibly complicated scenario. Now the interesting fact is that these difficulties were recognized, but were swept under the carpet. That habit may not have died out just yet.