What is the variety of ways one can deal with absorbing words from different languages in a conlang?

Question

The conlang I am working on has very strict rules on word formation:

• They must start and end with a consonant
• They can only have single vowels between consonants (which can be in consonant clusters).
• They can't end in h, w, or y (since they don't make a highly distinct word-ending sound.
• The only use a smallish subset of consonants (missing the "th" sounds, the retroflex sounds, etc.).
• End word with vowel to convert it into a noun/verb/adjective (convert it into speakable form).

But, I want to be able to talk about other languages in the conlang, hence the tone guide. That will make it so you can say the "j" in "measure" even though that is not in the consonant inventory. And other sounds as well. And you should be able to pronounce any sound from any language. But that's about as far as I want to go.

I don't want to allow you to use any word from any language in a sentence necessarily (at least I don't think so as of this time). So if you have a word with the "harsh h" sound like in Hebrew/Arabic, you need to do some slight transforming of it before it fits into the language. Likewise, it only supports 5 vowels in the main language, although you can say 15+ vowel variations to cover all languages when speaking about other languages. However, the "e" sound in English "pet" is, for example, not in the vowel inventory, so you might have to change it to "peyt" (like Spanish accent) or something along those lines.

The question is though, how would you absorb new words into the language? What is the process? Say I give a guide to someone to learn this language. They go out and speak it in front of friends. They are out to a sushi dinner. Now they want to say "maguro" or "sashimi" (well, it turns out these words fit well into the language, sort of). But imagine these words didn't fit in, what would they do? There must be rules of some sort to guide the transformation of any word, is that correct? How do natural languages do it, or a few of them covering the different approaches?

Basically, I am wondering if it would be better to "borrow" the word (morphing it slightly), or just come up with a completely new word and map it to the foreign word.

I would like to control exactly how each word/meaning is sounded out. So sushi would be like "sushim" as the base (start and end with a consonant, single vowels in between), and "sushima" as the food, etc.. (end with vowel to convert into speakable word from base word). But what about more complicated words like "hannukah" with the harsh h (I am pretty sure it has a harsh h, correct me if I'm wrong), or some french word with the guttural r sort of sound, or "crouch" with the compound vowel sound. Basically, wondering what sort of things I should put into place so you can use words from other languages quickly and painlessly, yet they still fit into the language paradigms/rules.

Another example would be proper names like the store Macy's, or Outback, those words don't fit directly into the conlang, so how would they be transformed or redone?

Answer 1

If you want to be strict, coerce them into your phonological rules.

Macy's: that would fit as macys; if /y/ is not a vowel, then macis

Outback: start with a consonant, so use one that is not too pronounced, such as /h/: hotback [*]. Also drop the /u/ from the diphthong. Or use hatback if your /a/ is pronounced more open (in that case hatbeck might work better). Looking at Elopa below, you could also use holdback to approximate the English pronunciation.

It will look strange, but not to speakers of your own language.

Here are some examples of country names from toki pona (leaving out the ma signifier):

• Amerika: Amelika
• Africa: Apika
• Europe: Elopa
• Egypt: Masu - if you can't coerce it, come up with a new word
• Canada: Kanata
• Grenada: Kenata - sometimes you might end up with identical forms for different places/people: that's not a problem, happens in natural languages as well.
• Greece: Elena - you can also use the 'local' name for a country as a starting point.

[*] This is similar to Spanish speakers sometimes adding an initial schwa sound when words start with a consonant when speaking English.

Answer 2

There must be rules of some sort to guide the transformation of any word, is that correct?

If you believe in optimality theory (OT), this comes down to the constraints on valid words in the language. For example, "must begin with a consonant", "can't include /θ/", and "must end in a consonant other than /h/" are all constraints. But the language would also include constraints like, say, "don't insert an /h/" (to keep /h/s from appearing in random places). And the relative ranking of these constraints would determine what happens to foreign words.

For example, maybe "must begin with a consonant" outranks "don't insert an extra /ʔ/". So given the choice between letting a foreign word start with a vowel (violates the first constraint) and inserting an /ʔ/ at the beginning (violates the second constraint), speakers would choose the second option. Why an /ʔ/ rather than any other consonant? Well, maybe "don't insert a /q/" outranks "don't insert an /ʔ/" too, so inserting /ʔ/ is preferable to inserting /q/.

According to optimality theorists, every language includes a vast, near-infinite set of these constraints, with a strict ranking, and speakers will unconsciously modify words to comply with them. One advantage of this explanation is that the same constraints are used for native words as for foreign words, so you don't have to invoke any special machinery to make it work.

If you don't believe in OT (i.e. you prefer rule-based systems over constraint-based systems), you'll generally have a set of repair rules that are invoked whenever a word violates a language's phonotactics. These rules are specific to each individual language, and their goal is to fix all the violations. If your phonotactics don't allow /ħ/, for example, you might have a repair rule that says "replace /ħ/ with /h/". If your phonotactics require consonants at the start of a word, you might have a repair rule that says "insert /ʔ/ before an initial vowel". And so on.

This explanation has the advantage that it clearly explains why all languages seem to have their own idiosyncratic repair rules, which aren't generally necessary for native words. For example, all Swahili words must end in a vowel. When a foreign word doesn't, an /i/ is generally inserted. An optimality theorist would say this is because of some special constraint that's never needed for native words, while a rule-based phonologist would say there's just a repair rule that inserts that /i/. (Māori and Japanese similarly usually require words to end in a vowel; Māori inserts /a/, Japanese inserts /u/. So this doesn't seem to be something universal.)

Answer 3

You can look at how it is done in real languages. What typically happens is that the word is phonologically morphed to match the phonotactics of the adopting language.

Japanese provides some obvious examples. The form of a Japanese syllable is stereotypically (C)V(n): Consonant or Consonant+y plus a vowel, with /n/ as the only final consonant. So what does Japanese do with English?

• ice cream > aisu kurīmu
• mansion > manshon
• pray > purei

Inuktitut (some dialects) also does the same thing:

• coffee > kaapi
• Ottawa > Aatuvaa

To use the last example, Inuktitut only uses the vowels /a, a:/, /u,u:/, and /i,i:/. and there's no /w/. "Ottawa" is normally pronounced [ˈɒt ə wə], so ['a:t u va:] is actually pretty close: if I used it while speaking English, must people would either not hear the difference or assume I was pretending to speak with a comedic German accent by replacing the /w/ with a /v/.