punch not < panc

[Nicholas Widdows]

On a non-linguistic forum recently I spotted the throwaway remark that
<punch> the drink comes from an Indian word for 'five' because it has five
ingredients. I immediately barged in and debunked this. But as the _OED_ is
wrong about <zebra>, I thought I'd ask the experts whether they can add
anything to the _OED_ explanation or controvert it.

In summary, it first occurs in English in 1632, and is repeatedly cited,
always with <u> and with variable ingredients; the expression <bowl of
punch> then turns up in Continental languages as <bolleponge>, <palepuntz>,
<bouleponges>, <palepunshen>. Clearly at this date it was pronounced with
[u]. No suitable Indian language has [u] in 'five'.

The first record of an [a] is also the first cited connexion with 'five': in
1698 one Fryer, who had travelled in India in 1672-81, wrote of "_Paunch_
(which is Indostan for Five) from Five Ingredients". It seems obvious to me
that Fryer was indulging in or transmitting a folk-etymology, and all the
years of evidence before him disallow it. The prosaic container <puncheon>
is most likely.

Does anyone have any good reason to disagree?

Secondly, is there a more appropriate list around that delights in
etymological niggles, as this has nothing deep to do with sigmatic aorist
laryngeals? I want to ask someone their opinion of deriving the jive terms
<bug, dig, guy, hepcat, honky> from Wolof; but not here.

Nicholas