The BG Language Creation Guide
Calques and Borrowings
When ever two languages get next to each other, they mix. this has to do partly with the fact that bilingualism is more or less the natural state of the human mind: or so it would seem if you look at statistics about how many people are bilingual (perhaps 66% of the world's people).
People use whatever is available to them when it's time to communicate.
Except in the cases of language shift or language death, this mixing doesn't damage either language. Remember that all languages are changing all the time. Adding words or structures from one language doesn't damage the other. Spanglish (a mixture of English and spanish) may be frowned upon by prescriptivists, but there's nothing inherently wrong with it.
Languages are eager to take words from one another. there are two basic ways this is done:
The most direct way to do this is by borrowing.
When a word is borrowed, it's pronunciation almost always has to change to fit in with the phonology of the new language.
Spanish and Japanese, for example, both borrowed the word 'strike' from English, and to keep our discussion lively, we'll also assume that Liqupa borrowed it, too. Here's how that was done:
- The first thing we have to do is see if all of the phonemes in the word we want to borrow are also phonemes in the new language.
'strike' has 4 consonants and one vowel phoneme:
/s/ /t/ /r/ /k/ and /aj/
Spanish has all of these. the /r/ phoneme is realized phonetically as an alveolar tap, however. this is the first place where the word's pronunciation will change slightly.
- On to syllable structure:
Spanish does not allow three consonants in a row at the beginning of a syllable. Two are ok, and you can have three if they are not word initial. Spanish will not allow syllables to end in /k/. So, as it stands, the word /strajk/ is unpronouncible. Its syllable structure is CCCVC. It has to be changed.
To fix the first problem, Spanish adds a word-initial /e/ segment. the technical term for this is epenthesis. It then deletes the final /k/. You end up with
(As it happens, many Spanish speakers can, in fact, easily pronounce this final /k/ even though it's technically not allowed: I've heard plenty of spanish speakers pronounce the word with the final /k/... but few without that epenthetic /e/.)
- Morphology: Don't forget also to asign the word to a gender. All spanish words are either masculine or feminine. The choice is arbitrary. I believe 'strike' is masculine.
- Here we go to Japanese:
- Phonemes: All of the phonemes in the word
/s/ /t/ /r/ /k/ /aj/
are also in Japanese. Once again, the /r/ is realized differently though.
- Syllable Structure: The japanese syllable structure is different from that of Spanish.
It looks like this:
where G is a glide and N is a nasal. Lots of problems await the Japanese speaker trying to say a CCCVC word like /strajk/.
Japanese, like Spanish, starts with epenthesis to solve the problem. All of those word-initial consonants have to be separated by vowels. and, instead of deleting the final /k/ phoneme, Japanese adds a word-final vowel, creating another syllable. Which vowels should be added is difficult to say: you can't predict which ones a language will add exactly, but you can account for it afterward. Here's what the word looks like after all this epenthesis is done:
The glorious thing about this word is that Japanese actually borrowed it twice:
once for baseball and once for labor. In doing so, it changed the english homophones into two distinct words. Clever, I think.
- Morphology: If this were a verb, there'd be more to do in Japanese, figuring out which verb category it goes in. Luckily, it's a noun. there are no noun classes, so our work here is done.
Finally, we arrive at the mythical isle of Kasinicinala where the pine-tree worshipping synesthetes have discovered a cultural passion for baseball and thus join the torrent of languages which have borrowed the word 'strike'.
- The phonemes first once again:
/s/ /t/ /r/ /k/ /aj/
- /s/ is fine: Liqupa has an /s/ phoneme.
It is represented by allophone [S] though word initially, so to pronounce this word with a good Liqupa accent, you need to begin with [S] not [s].
- /t/: Liqupa has no /t/: no voiceless alveolar stop. We look then, for something similar and find a /d/ (voiced alveolar stop).
Differing only in voicing, these segments are close enough. Note that we could as easily have picked /s/ which differs only in manner or /p/ which differs only in place. The choice is somewhat arbitrary... an artistic rather than scientific choice at that point.
- /r/: Once again, we find no /r/ and must go sniffing about for a decent substitute.
The other available liquid is /l/.
- /aj/: This one, thank goodness, is fine! Liqupa has /a/ and /i/ and will allow these to occur together. We'll keep writing the vowel as /aj/ since the real pronunciation difference between [ai] and [aj] is minimal.
- /k/: There is no /k/ phoneme, but we know that [k] is an allophone of /q/.
For now, let's leave the [k] alone an dsee how that goes.
So, our word so far in Liqupa is:
- syllable structure: Alas, the maximum syllable in Luqupa is CVC.
Having worked hard to determine which segments to use, we are now going to have to get started either epenthesizing vowels or deleting consonants.
We can do anything we like here: languages solve this problem in all kinds of ways. Why not do one of each? I arrived at this solution by asking my 8 year old which she preferred. She told me to epenthesize a /u/ after the [S] and delete the voiced stop: (Ok, not in those words exactly...)
this gives us:
Almost done. Although Liqupa does allow a consonant at the end of a word, that consonant has to be a [q]. I could
Any of these choices is fine: I'm partial to the one with [ka] so that's the one I'll pick. and,
- (a) epenthesize a vowel at the end of the word,
Since [k] is the allophone of /q/ that occurs before /a/, I'd epenthesize [a] and get this:
- (b) Delete the offending consonant and get this:
- (c) Use a word-final [q] and get this:
- Morphology: I'll make it a [na] class noun just for kicks.
The English word 'strike' /strajk/ thus becomes:
depending on the requirements of the morphology and phonology of the language which borrowed it.
same word: three different languages with three different phonological systems taking it in.
- (el) [estrajk]
- [Sulajka] (na class)
On to calques
Sometimes, a language will choose to avoid the phonological adaptation route entirely and calque a word. that means, translating its individual morphemes.
If Spanish borrowed the word 'skyscraper' [skajskrepR] (R = syllabic r), it might look like this:
It didn't borrow the word though: it calqued it.
The verb 'to scrape' is 'rascar' and sky = cielo.
Now, in order to use the verb, one has to conjugate it: make it agree with the subject of the sentence. We get:
rasca (it scrapes)
The morphemes should appear in Spanish, not English word order. Perversely, English makes its compounds backward with the object of the verb coming before the verb even though english is an SVO language:
- can-open-er (the object of the opener is the can)
Spanish (also an SVO language) turns these around and makes its compounds verb then noun.
And then pluralizes the object. So, a skyscraper becomes
It is also possible to calque half of a word and borrow the the other half.
Liqupa might do this with the word 'strikezone'. The Liqupa word 'kana' means 'place' which is close to 'zone' in meaning. So, I borrow 'strike' (as shown above) and calque the 'zone' part to get: