Consideration of Phonemes

This is what the consonant system of Proto-Indo-Euopean looked like (insofar as it can be reconstructed):

(Look at that.  Nine velar stops.  Is it any wonder the descendant languages ran for the far corners of the phonetic space as quickly as they reasonably could?  This is also why PIE sounds pretty ugly to most of us these days, even when we already speak notoriously non-mellifluous languages like English.)

And this is what roots looked like¹, with K for any voiceless stop, G an unaspirated and G^h an aspirated stop, R a sonorant (r, l, m, n, w, y) and H a fricative(either laryngeal or s):

So possible roots: h₂mek, dʰel, ples, ǵwensbʰ (I have no clue whether any of these are actual PIE roots, or if they are what they mean; I was just picking from the possibilities.)

The reason there's only one vowel here is because PIE does a lot of important grammatical stuff with vowel mutation.

So, if I'm doing something PIE inspired, I want to crowd a whole lot of phonemes into a tight space, and I have to be OK with consonant clusters.

1: I got both these charts from Wikipedia, by the way, which is not reliable for many things but tends to be OK on pop culture and reasonably well-established stuff from "boring" sciences (which linguistics most definitely counts as).

links: digg this    del.icio.us    technorati    reddit