How do people deal with loanwords with highly alien phonemes?

Question

I've been thinking about how a people who speak a language without rhotics would perceive a rhotic sound. Obviously of course, this would depend on exactly which rhotic we're talking about. I thought about making this question about rhotics specifically, but I think it would probably be more beneficial to talk about how languages deal with foreign phonemes in general.

From my experience, they typically just replace it with a native phoneme that overlaps with the foreign phoneme in how its pronounced. For instance, words with the voiceless velar fricative often have that phoneme replaced with the voiceless velar plosive in English. This can be seen in how the word 'loch' as in 'loch ness' is often pronounced, at least in the US.

I've also seen languages that replace an 'f' sound with a 'p' sound. Humans seem to perceive alien fricatives as sounding as plosives. I have noticed that most Americans can't hear the difference between x and k, or a voiceless palatal fricative versus a voiceless post-alveolar palatal fricative (I personally can't understand how someone could fail to hear the difference between those two, but its surprisingly common).

But some rhotics belong to unusual parts of speech. There's the alveolar flap for instance, and the alveolar trill. How do languages deal with phonemes that are part of a manner of articulation their language simply doesn't have? Personally, I suspect the most common replacements for an alveolar flap would be an alveolar approximant or plosive. They might perceive it as being somewhere between say 'l' and 'd'.

And of course there's the extremely rare English 'r' sound. European languages just replace it with their own rhotic, obviously. But what about non indo-european languages that lack a rhotic? The only non IE language I know is Japanese, which does have a rhotic (a lateral alveolar flap, ironnically). The only 'manner of articulation' I've seen any language deal with are nasals, which people seem to like to replace with voiced stops (which makes pretty obvious sense).

Answer 1

Rhotics aren't a special case here. They will still either be deleted or be replaced with the sound that is perceptually (to speakers of the receiving language) closest to the original sound.

English's /r/ [ɹ̠] is similar to /n/ and /l/ among others. For example, in the Algonquian language Innu, English /r/ tends to be replaced by /n/:

Mary → Mânî
Andre → Ântane
Pierre → Pien
Marguerite → Mânikanet
_(source)

Answer 2

They adapt the sounds that are not in their native phonological systems by applying unsuppressed processes of that system, according to the theory of Natural Phonology. Here is a bibliography of some online papers: natural phonology

Answer 3

The lexicalization of loans is a longer process, you have many words which are undiscernible from native words by a layperson because they have been so completely internalized.

In English, you have words like ditty and because which can be traced back to stems of Latin origin, but which have been so completely adapted to the form of modern English that you need specialist knowledge to establish this.

In Swedish, a lot of the old Norse vocabulary has been displaced with German loans which sometimes make little sense in Swedish. Everyday words like anhörig and begära are somewhat opaque (though some morphemes trace back to shared Germanic roots which were also present in older Norse) but make perfect sense in German. Some others which are outright peculiar in Swedish include fogsvans (literally "joint tail", where the German word means "fox tail" -- still a weird term for a saw). Later, French loans have entered the language in the 18th century, originally in forms like lavoir and manquements, but now have modernized spelling and native-like pronounciation lavoar, mankemang.

Finnish is pecular because even more than Swedish, it has replaced much of the original Fenno-Ugric vocabulary with loans of Indo-European origin. A large number of words like pallo (ball) and ranta (beach, cf Swedish strand and German Strand) are clearly of Germanic or even earlier Indo-European origin, and there is a set of trisyllabic words like mehiläinen (bee) and kuningas (king) which are of somewhat later origin, and sometimes attest to forms which are no longer straightforwardly recognizable in the neighboring IE languages.

In all of these languages, you can find recent loans which are used in a form which is much closer to the origin. With reading and writing skills approaching 100% of the population, the process for internalizing these is likely to look quite different -- it is not unlikely that résumé and coup d'état will remain in English in their French form for a very long time. Similarly, obviously English loans like jazz and policy may well have come to stay for the foreseeable future without any further adaptation to the host language.