-
>
> There was a late ME change of /Er/ to /ar/ finally and
> before consonants, but it didn't take place in all dialects,
> and as a result we have such pairs as <clerk> ~ <clark>,
> <person> ~ <parson>, <university> ~ <varsity>, <certain> ~
> <sartain> (dial.), etc.
>
> In the names in question, however, the variation is much
> older: I believe that there was already variation between
> /er/ and /ar/ in OFr.
>
Oh yes I'm well aware of the change of ME /Er/ to /ar/ (> /Ar/) before
syllable-boundaries, but the variation I'm talking about is chiefly
between /aer/ (I think most often from OF or MF, but also from ME /ar/
before vocalized consonants) and /Er/ before vowels (the latter
apparently of various origin, e.g. < */Ir/ in <merry>, <bury>, < */Er/
before vowel or vocalized consonant in <ferry>, < shortening in
<herring>, otherwise mainly in names it seems). Were e.g. <marry>,
<carry>, <tarry>, <carriage>, <barrow>, <harrow>, <arrow> often
spelled as <merry>, <kerry>, <terry>, <kerriage>, <berrow>, <herrow>,
<errow>? And therefore does the alteration between /Er/ and /aer/ go
back to older ME, and not then a recent development in American
English? (discounting the OE variation between <earh>, <erh>, <arwe>,
etc. in <arrow> and other similar words). You say that there was
variation between /er/ and /ar/ in OF: any typical examples? and
perhaps their descendants in English? The example of the putative
etymon of my name, <Gerard> (and also <Gerald>), produces the pet-form
<Gerry, Jerry> in English, but I've never heard of <Jarry> (even if in
most American English they would be pronounced identically -- but not
in Britain or Trinidad); <Lawrence> produces the pet-form <Larry>, but
I've never heard of <Lerry> (even if this would be the typical
American pronunciation of <Larry> -- they would be kept distinct in
Britain and the Caribbean): the point is, the two pronunciations /Er/
and /aer/ spelled <err> and <arr> seem to have been kept distinct in
NE till recently in American English and till the present in British
and Caribbean English, so it doesn't look to me like <Gerard> or
<Gerald> would be the probable ancestor of <Jarrett(e)>. But you say
the evidence of OF indicates otherwise. Examples?