English /hw-/ > /w-/

[Larry Trask]

Lonhyn Jasinsky asks about the reduction of English /hw-/ to /w-/.
Since I can't find his e-mail address in the posting, I'm replying to
the list.
 
Proto-Indo-European */k/ developed regularly to /h/ (or /x/, if you
prefer) in the Germanic languages, by the First Germanic Consonant
Shift (Grimm's Law).  The resulting fricative was generally pronounced
[h] in syllable-initial position but [x] in syllable-final position,
much as in modern German.
 
The PIE initial cluster */kw-/ accordingly developed in Old English
into /hw-/: hence, for example, Old English <hwaet> `what' and
<hwae:te> `wheat'.
 
The spelling was changed after the Norman conquest, apparently under
Norman influence, to the modern <wh->, and hence orthographic
<what> and <wheat>.
 
The pronunciation remained /hw-/ in England for centuries, at least
for most speakers.  In the south of England, though, the phonetically
natural reduction to /w-/ is attested from the early Middle English
period, but this reduction remained a vulgarism for centuries
afterward; it did not spread widely, and it did not reach educated
speech.  Accordingly, the nearly universal pronunciation /hw-/ was
carried to North America in the 17th century.
 
However, in the 18th century, the innovating pronunciation /w-/ began
spreading rapidly in England, even into educated speech; by 1800, /w-/
was firmly established as the norm in England.  Today, /w-/ is
universal in England and Wales (except in Northumberland), even in the
most careful and prestigious speech, though a few teachers of
elocution and drama still try to to inculcate the /hw-/ pronunciation,
which is still often perceived as elegant.  In Scotland, though, /hw-/
remains universal.
 
The Linguistic Atlas of the Eastern USA, compiled over a generation
ago, shows /hw-/ as normal in most places, with /w-/ the norm in just
three areas, all of them on the east coast: a large area centered on
metropolitan New York, and two smaller ones centered on Boston and
Charleston/Savannah.  This distribution strongly suggests that /w-/
was introduced from England into these port cities and began spreading
out from there.
 
In the last generation, the innovating /w-/ has been spreading across
the USA with astounding speed.  The American linguist William Bright
recently told me (p.c.) that /hw-/ was now confined to "a handful of
old fogies".  I myself (I'm from western New York State) have /hw-/,
like my parents, but my two brothers and my sister (all younger) have
only /w-/.  My mother is acutely conscious of this; she notices the
/w-/ of the young people and regards it as objectionable.
 
There's a brief survey with references in vol. 1, section 3.2.4, of
John Wells (1982), Accents of English, CUP.
 
Useful intros to /h/-dropping generally in English are these two:
 
James Milroy (1992). Linguistic Variation and Change. Blackwell.
[section 5.5]
 
James Milroy (1983). `On the sociolinguistic history of /h/-dropping
in English'. In M. Davenport et al (eds), Current Topics in English
Historical Linguistics, pp. 37-53. Odense University Press.
 
But these two deal with h-dropping in general, and not explicitly with
/hw-/.
 
I might note that the Old English initial clusters /hl-/, /hr-/, and
/hn-/, as in <hlu:d> `loud', <hring> `ring', and <hnutu> `nut', lost
their /h/ completely at a fairly early stage, and hence the reduction
of /hw-/ to /w-/ can be seen as a continuation of a venerable process
of reduction of such clusters.  We are left only with /hj-/, as in
<hue>, but even here there is abundant evidence of the sporadic or
regional loss of /h/ in a number of varieties on both sides of the
Atlantic.  (In my own accent, /hju:/ has become something like /hIw/,
and similarly for other such words, and hence <hue> no longer rhymes
with <you> or <too>.)
 
Larry Trask
COGS
University of Sussex
Brighton BN1 9QH
UK
 
larryt at cogs.susx.ac.uk