It may be a pun. Looking up magma in the French wikipedia*, another name for magma in French is groupoïde de Ore†. Here Ore is a Norwegian mathematician, but ore in English is mineral-bearing rock, whereas magma (in both English and French) is molten rock.
Would Bourbaki have based a mathematical term on this pun? I'm not in a good position to judge; maybe somebody else could comment on this.
* Since Bourbaki was a pen name used by a group of French mathematicians, this is the right language to search in.
† There is also a groupoïde de Brandt, also called a groupoid in English, which would explain why Bourbaki felt compelled to coin a new name.
Because that is the standard rule in English. The OED says:
Hence, in modern English, C has
- (1) the ‘hard’ sound
[k] before a, o, u, before a consonant (except h), and when final, as in cab, cot, cut, claw, crow, acme, cycle, sac, tic, epic;
- (2) before e, i, y, it has the ‘soft’ sound
[s]. In all words from Old English or Old French, final c is avoided: the [k] sound being written k or ck, as in beak, meek, oak, book, bark, balk, bank, pack, peck, pick, rock. This is probably due to the claims of derivatives like meeker, oaken, barking, rocky, where c could not be used. Final c however is written in modern words from Latin, Greek, or other languages, and (of late) in the ending -ic, as in sac, tic, epic, critic, music, picnic. In the rare cases in which this c is followed in inflexion by e or i, it is necessary to change it to ck, as in physicking, mimicking, frolicking, trafficker, picnicker. When the [s] sound is final, it must be written ‑ce, as in trace, ice, thrice, and this final e must be retained in composition before a, o, u, as in trace-able, peace-able.
- (3) Ci (rarely ce) preceding another vowel has frequently the sound of
[ʃ], esp. in the endings ‑cious, ‑cial, ‑cion, as atrocious, glacial, coercion (ocean). This sound (which is also taken by t in the same position) has been developed in comparatively modern times by palatalization of [s]. In a few words from foreign languages, c retains the foreign pronunciation, as in It. cicerone [tʃitʃeˈrone].
Which leads us to examples like colicky, havocker, picnicky, plasticky, panicking, picnicking,
panicky, magicked, colicking, picnicked, bivouacking,
colicked, mimicked, frolicked, picnicker, demosaicked,
garlicky, mimicker, havocking, bivouacked, demosaicker,
havocked, panicked, mimicking, frolicking, demosaicking.
Yes, you will sometimes see words like those misspelled without the protective k, but that’s like spelling the plural of bunny as *bunnys instead of as bunnies: it’s just plain wrong. We do not do things that way in English.
Best Answer
OED gives
First use in English was:
They don't give any pronunciation history farther back than 1903.
It's instructive that Late Latin and the Romance languages usually pronounced this word's ending with an /e/ instead of an /i/:
While the Greek underwent a vowel shift of the ending to Modern Greek /-pi'ía/. This reminded me of English's own Great Vowel Shift (GVS). If the first use in English was as far back as 1553, then the vowel may have been subject to the GVS. In particular:
The above would result in an original /-peja/ becoming the modern /pija/. So, it seems possible that the original English borrowing from Latin/Greek had an /e/ and that it has changed to /i/ since then.
This is detective work and reasoning, not citation, so take it for what it's worth.