Suppose my friend Bassam presents the following hypothesis: "There are undetectable aliens living on Pluto." Suppose I travel to Pluto in search of these aliens. I set up all kinds of instruments to see if there is any evidence of aliens. I return empty-handed. "Bassam," I say. "You said there were aliens on Pluto. I checked. But there aren't any." "Well," Bassam replies, "I said they're
undetectable, didn't I?"
Here I would have a difficult time understanding Bassam's meaning. He's making a positive claim, namely, that there are aliens on Pluto. But at the same time, by saying that these aliens are undetectable, he's ruling out our ability to test his claim. His claim is
unfalsifiable, meaning there's no way, even in theory, that we can prove his theory false via experiment or observation. How useful are such hypotheses?
Consider now the following claim made by Bassam and other Muslims: "The Qur'an has been perfectly preserved." At first, it seems like such a claim is falsifiable. That is, it seems that we can test the claim by doing some historical research. So we do a little research and we learn that Abdullah ibn Masud and Ubayy ibn Ka'b, two of Muhammad's top reciters of the Qur'an, had a different number of Surahs in their versions of the Qur'an. One would think that this falsifies the Muslim claim. "Not so," says Bassam. "Abdullah and Ubayy were simply wrong. The Qur'an we have today is, by definition, the correct one. Hence, everyone who has a different number of chapters must be wrong." So, given such a claim, even if we were to find out that every single one of Muhammad's companions except Zaid ibn Thabit had a different number of Surahs, this wouldn't count as any evidence against the perfect preservation of the Qur'an, since, by definition, only the Qur'an we have today is the correct one.
So we do a little more investigation. We find that there were all kinds of textual variants among early Qur'anic texts. That is, if we were to turn to Surah 2 in the Qur'an of Ubayy ibn Ka'b, this Surah would differ from that in the Qur'an of Ibn Masud, which would differ from that in the Qur'an of Zaid ibn Thabit, etc. These Qur'ans have spelling differences, different words, different phrases, etc. Surely this will count as evidence against the perfect preservation of the Qur'an, won't it? "No, it won't," replies the Muslim. "You see, there were seven [or ten, or twenty] different readings of the Qur'an, and all of them were correct." Here we find that there can be all sorts of differences among copies of the Qur'an, and yet this doesn't at all affect the hypothesis that the Qur'an we have today is a perfect copy of the tablet in heaven.
We dig deeper in search of the truth, and we find Muslim sources reporting that massive sections of the Qur'an have been lost. We find Aisha and Ubayy ibn Ka'b reporting that more than a hundred verses of Surah 33 are missing. Surely this will count as proof that the Qur'an hasn't been perfectly preserved. "Not so," says the Muslim. "Whatever verses are missing from the Qur'an have been abrogated (despite the fact that the Qur'an contains other verses that have been abrogated). You see, Allah changed his mind quite a bit, and he often gave us verses, only to take those verses back." Thus, we find that no matter how much evidence there is that numerous verses are missing from the Qur'an, this will never count as any evidence whatsoever against perfect preservation.
Here non-Muslims are quite confused. Were the missing verses of Surah 33 part of the tablet in heaven? If so, then the Qur'an Muslims have today is very different from the Qur'an in heaven. If not, then why were they revealed as part of the Qur'an? While we're at it, does the perfect Qur'an in heaven contain all seven readings (this would be quite an odd book). If so, then the Qur'an Muslims have today is quite different from the Qur'an in heaven, since Uthman destroyed most (but not all) of the different readings. If not, weren't the variants corruptions of the original, which contained no variants?
In the end, no matter what the evidence says, Muslims will continue to claim that the Qur'an has been perfectly preserved, for they have insulated themselves from the evidence (and this is extremely common in Islam). But non-Muslims are left asking ourselves, "What is the difference between, on the one hand, a perfectly preserved book whose early copies contain different numbers of chapters, different verses, different spellings, different words, different phrases, a different order of chapters, and which, at different times, contained completely different passages (for some were abrogated), and, on the other hand, a book that hasn't been perfectly preserved at all?" As far as evidence is concerned, there is no difference between the Qur'an and a book that hasn't been perfectly preserved, which is why the Muslim claim makes so little sense to anyone who isn't a Muslim.