12.3.28.1:33: A KWE-STIONABLE (YET INDUBITABLE) CHA-NGE
In "Gaspārdo Līṣā ca", I forgot to mention that Sanskrit ca 'and' (pronounced "cha") is cognate to Latin que 'and'. Both go back to a Proto-Indo-European *kwe. Latin que (pronounced "kwe") preserves this word, whereas Sanskrit ca had undergone several changes:
*kwe > *ke > *ce > ca
*kw simplified to *k and then palatalized to *c to become more like the following palatal vowel *e. Then palatal *e became nonpalatal a.
The shift of *kw to *k must have occurred after original *k became ś:
| Stage 1 (Proto-Indo-European) | Stage 2 | Stage 3 | Stage 4 | Stage 5 (Sanskrit) |
| *k | *ś | *ś | *ś | ś |
| *kw | *kw | *k | *c (before *e) | c (even after *e > a) |
Beekes (1995: 124) reconstructed PIE *k as palatal *ḱ. So it is not surprising that it became palatal ś in Sanskrit in stage 2 before nonpalatal *kw palatalized in stage 4. *kw initially had 'double protection' against palatalization in stage 2: *k was not palatal and medial *-w- was a barrier between nonpalatal *k and palatal *e.
Spanish, French, and Italian underwent similar shifts much later. Palatalized consonants and their descendants are in bold:
| Latin | Spanish | French | Italian |
| c [k] | c [k] ~ [θ] (ceceo and distinción dialects) ~ [s] (seseo dialects) |
c [k] ~ [s], ch [ʃ] | c [k] ~ [tʃ] |
| qu [kw] | cu [kw], qu [k] |
qu [k] | qu [kw], c [k] |
I would predict that in Indo-European in general, palatal *ḱ and plain *k would be more likely to palatalize than 'doubly protected' *kw. This is generally the case (Beekes 1995: 110, Watkins 2000: 46):
| PIE | Old Church Slavonic | Lithuanian | Armenian | Avestan | Old Persian |
| *ḱ | s | š | s | s | θ |
| *kw | k, č, ts | k | kʻ | k, c | k |
But in Greek, it was nonpalatal *kw that palatalized before front vowels, not palatal *ḱ. So *kwe became Greek te 'and', but *ḱe became Greek ke: e.g.,
PIE *ḱer- > Gr keras 'horn' (cf. Skt śṛṅga- 'id.' with palatal ś-)
How did that happen? Here's a wild guess: *ḱ backed to *q to avoid palatalization, fronting to k later:
| Stage 1 (PIE) | Stage 2 | Stage 3 | Stage 4 | Stage 5 (Greek) |
| *ḱ | *k | *q | *q | k |
| *kw | *kw | *k (before *e, *i, *u) *p (before *a, *o) |
*tś (before *e, *i) *k (before *u) *p (before *a, *o) |
t k p |
If Wikipedia is correct, *kw became *p everywhere in stage 3 of Aeolic Greek. (But Beekes 1995: 110 wrote that Aeolic Greek has *kw > *t before *i though not the less palatal vowel *e.)
Mycenaean Greek never went past stage 2; it retained a *k : *kw distinction.
Stage 3 and 4 look implausible to me because they require *q to be more common than *k until the two merge in stage 5. But it is possible for a language to have more q than k. In Arabic, q is slightly more common than k. However, PIE *ḱ was much more common than *kw, so Greek *q would have to be much more common than *k (which would only be before *u in stage 4, and I can hardly find any PIE *kwu in Pokorny!).
This problem gets even more complicated if one reconstructs a three-way distinction between palatal *ḱ, plain *k, and labiovelar *kw in PIE. Would plain *k have merged with palatal *ḱ as *q in stage 3? Or did it become *q only before certain (nonpalatal?) vowels?
Similar questions can be asked about PIE labiovelar *gw and *gwh which also palatalized in Greek unlike PIE palatal *ǵ and *ǵh:
| PIE | Greek | ||
| before *e, *i | before *u | before *a, *o | |
| *ḱ, *ǵ, *ǵh | k, g, kh (before any vowel) | ||
| *kw, *gw, *gwh | t, d, th | k, g, kh | p, b, ph |
I've known about this change for 18 years, but it never bothered me until Sunday. Did some Neogrammarian work out a solution back in the 19th century?
12.3.25.23:59: WHAT WERE THE NAMES OF THE BIBLICAL MAGI?
Gaspard of Gaspard and Lisa (from my last post) is one variant for one of the names of the Biblical Magi. You can make up your own variant by mixing and matching the various parts below:
| G J C K |
a á e o |
s c |
p | a á e |
r | Ø d e s |
Why does French Gaspard have a -d absent in all other versions? Is it by analogy with other -rd names?
The final -s in Latvian Kaspars is presumably the nominative singular ending -s of the first masculine declension.
I was surprised to learn that the name was Mandarinized as 卡斯柏 Kasibo in Taiwan.
Wikipedia's "Biblical Magi" article lists other variations which can't easily fit into the above table:Gondophares (founder of the Indo-Parthian Kingdom; the source of the name Kandahar?)Gushnasaph (in Syrian Christianity)
Karsudan (in Ethiopian Christianity)
The names of the other two - if there were only two others! - have even more variation: e.g., Larvandad and Hormisdas in Syrian Christianity and Hor of Ethiopian Christianity don't match up with English Melchior and Balthazar, though I presume the two Hor-names are related. I have no idea what the originals were.
3.26.00:36: Although Gaspar is thought to be an Indian king in Western Christianity, none of the above names are Indic. The Wikipedia entry "Casper (name)" has yet another non-Indic derivation:
The name Casper and the same sounding name Kasper are derived from Gaspar which in turn is from an ancient Chaldean word, "Gizbar", which according to Strong's Concordance means "Treasurer". The word "Gizbar" [גזבר] appears in the Hebrew version of the Old Testament Book of Ezra (1:8). In fact, the modern Hebrew word for "Treasurer" is still "Gizbar".
If this etymology is correct, how did -i- end up becoming -a-, and why did G- become C/K-? There's nothing un-European about -i- or G-, whereas -zb- looks un-Greek to me, so I can easily imagine it being Hellenized as -sp-.