Janda, Richard D., Brian D. Joseph, & Neil G. Jacobs. 1994. Systematic Hyperforeignisms as Maximally External Evidence for Linguistic Rules. The Reality of Linguistic Rules, edited by Susan Lima, Roberta Corrigan & Gregory K. Iverson, 67-92. Amsterdam: John Benjamins.They trace how common it is for people to substitute more foreign-sounding pronunciations in foreign names and words, even when the source language doesn't have that. So, Beijing and Taj Mahal have (so I'm told) [dž] but many English speakers avoid that utterly normal affricate for [ž]. And famously, English speakers think that French drops final consonants, so people say coup de grace as [gɹa], where French has a final [s] — producing something that would mean 'blow of fat'.
I've long noticed some people using [x] as a pronunciation for [h] as a kind of generalized hyperforeign pronunciation. You get this associated with Germanic, Slavic, Semitic languages and so on. But yesterday, Stephanie Miller used that for the word 'Bahamian'. What's striking there is that I would expect people to treat Carribean English as h-less, and this suggests that [x] has generalized somehow to "non-US English". Or maybe it's some idiosyncrasy.