> I meant 100% literally. You can drop below 100% and still not have a detectable dent in understanding, and I even suggest a few phonemes that would make it ex

> I meant 100% literally. You can drop below 100% and still not have a detectable dent in understanding, and I even suggest a few phonemes that would make it extremely unlikely to come up. Combining /θ/ and /ð/ will mean that it never comes up, not even once, unless you intentionally force the issue by using a sentence like "we saw her teeth" that positions the word in a context where it's not clear whether it's a noun or a verb. In most cases that can't be done, as witnessed by the fact that English nouns regularly zero-derive into verbs - for the zero-derivation to work, you have to be able to tell the difference between a noun and a verb even though they're pronounced identically. On the other hand, you can't actually have meant 100% literally, because that's not an achievable goal. No native speaker can reach a 100% rate of being understood. But a foreign speaker can easily achieve a rate of being understood equal to that of a native speaker while dropping some of the English phonemic inventory. > And the fact also remains that if you apply the drop-phonemes-and-see-if-people-understand-you technique to other languages you can reduce their number of phonemes as well That's true, but the question I saw was just "how many phonemes could you drop from English while still being understood?". And the answer is that English, like every other language, features a lot of redundancy; many things can go wrong before you stop being understood, but every time you do get something wrong, your odds get a bit worse. The amount of slack you have in terms of phonemes depends on how good you are with the rest of the language.