/hw-/ > /w-/ again

Miguel Carrasquer Vidal mcv at pi.net
Sun Sep 15 09:27:21 UTC 1996

>Mark Hale is quite right to point out that virtually all IEists
>consider that the PIE ancestor of English /hw-/ was a single segment,
>a labialized velar, and not a cluster -- though I have occasionally
>seen it suggested that PIE might actually have had a contrast between
>a /kw/ cluster and a labialized velar.  (Anybody know if this idea is
>still taken seriously?)
I haven't seen Mark Hale's message, so I don't know if he distinguishes
between /kw/ and /k^w/ (palatalized velar).
If /k^w/ counts, there is certainly reason to assume it had a separate,
albeit marginal existence in PIE.  It has distinct, and sometimes rather
bizarre reflexes in at least Indo-Iranian, Greek, Balto-Slavic
and possibly Armenian and Anatolian:
Skt. s'v   (s'va^ "dog")
Av.  sp    (spa^ "dog")
Arm. s^    (s^un "dog")
Grk. pp, tt, kk (hippos/ikkos "horse", kittanos/titanos "chalk")
Lit. s^v   (s^viec^iu` "to shine")
Slv. sv    (svet "light, beside kve^t-/cve^t- "flower")
Gamkrelidze and Ivanov posit a sound law in Anatolian to the effect
that *k^w > s(w) (Hitt. s^uwa- "to fill", Luw. suwanai "dogs").
The case for /kw/ is weaker, but Latin ca^seus "cheese" [Pok.:
"(von *ca^so- aus *kwa^t-so- [cf.] abg. kvasU; das Fehlen des -w-
harrt noch der Erklaerung)" and vapor (*kwapo^s, cf. also Grk. kapnos)
are at least remarkable.  Balto-Slavic /kv/ instead of /k/ (or /s^v/)
seems due to the usual B-S confusion over *k^ vs. *k .
>But I don't think we can really tell whether
>OE /hw-/ was a single segment or a cluster.
It was certainly not deemed worthy of a separate symbol, as were
<w> (wynn), <th> (thorn) and <ae> (ash).  Or <hw> in Gothic, for
that matter.
That leaves the question of why the Norman clerks decided to
change the spelling to <wh> (cf. <sh>, <ch>, <th>).  I assume it
was to reflect the pronunciation [W] (unvoiced labio-velar appr.),
whether it was still /hw/ for the English or not.
Miguel Carrasquer Vidal
mcv at pi.net

More information about the Histling mailing list