If they talk loudly enough, you'll hear the vowels--a, e, i, o, and u, but no consonants, so the words won't make sense. Sound travels very well under water, but some sounds have more trouble than others getting from the air into the water. But why the vowels and not the consonants?

Every spoken sound is actually a combination of different sounds, some low, some high. Even though we don't notice the different sounds, the way they're combined is what gives each spoken sound its own character. In general, consonants contain a lot more high-pitch sounds than vowels. Those are sounds made of faster, smaller sound waves. Compared to consonants vowels are mostly made of low pitches, in other words of larger, slower sound waves.

When the small sound waves hit the uneven surface of the water, they get scattered in all directions like ping-pong balls landing on a rough road. The much larger, lower-pitch waves aren't affected as much by the little water-waves because they hit a much wider area on the water's surface. If we think of a small sound wave as a little ping-pong ball on a rough surface, a larger sound wave is more like a big basketball, which is less affected by little bumps on the road. Unlike balls bouncing on a road, sound waves pass through the water, but like the basketball, the large waves come through with less distortion. That's one reason why when you listen underwater to someone up above, you won't hear the consonants with their high-pitched sounds and short sound waves.