Spanish soft g is [x]!

[RonButters at AOL.COM]

I was wrong in asserting that the the Spanish pronunciation of ANGELES
contains a voiced velar fricative for the sound represented by the spelling
G. The sound typically is the voiceless [x], which will sound like /h/ to
speakers of English.

I still think, though, that perceptually this [x] is, for speakers of
English, closer to /g/ (the so-called "hard G") than to the alveopalatal
affricate (the so-called "soft G" heard at the beginning of the word JUMP).
The perception would be further strengthened by the fact that the preceding
/n/ and the following /E/ are both voiced.

DInIs P wants to make me a bet about this. Unfortunately, that sort of
activity is illegal in North Carolina, but I will be happy to meet him in Las
Vegas (where the medial /g/ IS a voiced fricative); there we can bet legally
and also carry out empirical research on the matter of the perception of
[nxV] sequences by native speakers of English. The loser will have to write
the research-grant proposal.

By the way, isn't it somewhat confusing to speak of "the Spanish soft g"? The
terms "hard G" and "soft G" apply only to English (at least in my
experience).