Re: Relative frequency of ejectives

David says on the Kalusa comment forum that

>| [k'] is rare, as far as ejectives go. Certainly more common than [q'],
>| but not as common as [p'], [t'], [ts'] or [tS'], as far as I know.
>
>This runs counter to what I thought I knew about the frequencies.
>To my knowledge, [p'] was the really marked one (witness the glottalic
>theory of PIE), [q'] was not all that rare (standard Georgian, for instance,
>has [q'] as its only uvular stop), and [k'] was quite common.
>
>Can anyone corroborate either of these sets of assertions?
>
>
Everything about ejectives I've ever seen confirms what you say, that k'
is by far the most common ejective. p' is certainly a lot less common,
and is often missing from systems that otherwise have one ejective
corresponding to each pulmonic egressive stop. t' and the others are
fairly common, but not as common as k'.