Oldest words in English?

[Alice Faber]

Joel S. Berson wrote:
> At 3/1/2009 02:04 PM, Baker, John wrote:
>> I'm doubtful that tin has been known long enough to be from a
>> pre-Indo-European substrate.
>
> Can't one determine from geological and archeological dating what
> substrate tin comes from?
>

I don't know about tin specifically, but I've dealt with similar issues
in Semitic/Afro-Asiatic etymologies. And the answer is "not
necessarily", which, in fact, reduces to "no".

First, a lot of early evidence for metal usage comes from archaeological
finds of smelting detritus. However, there's some clear stratological
evidence for earlier use of some of these metals, with the presumption
that they come from nodules in meteorites.

Second, identifying an archaeologically-identifiable cultural group with
a specific language group is totally non-trivial.

Third, identifying a pervasive loan-word isn't always easy, in the
absence of a clear source (where you can demonstrate both that the word
existed in that source language and a persuasive contact situation in
which the transmission would have occurred).

Fourth, arguments that lexical items of a particular phonological
structure are "un-X" are often inherently circular: Semitic roots are
tri-consonantal, so any PS root with more or fewer consonants has to be
explained away ("weak" consonants that are elided everywhere, frozen
causative prefixes, fossilized compounds, borrowings).

Fifth, new words that are formed within descendant languages using
natural metaphors can lead to the appearance of a word being far older
than it is, or can be. For example, Hebrew uses the word for pomegranate
for hand grenades also (a very reasonable metaphorical extension). If
Arabic were to do likewise (I don't know whether it does or not, but,
hey, it's my example!), that could lead to an inference that speakers of
*Central Semitic (the latest common ancestor of Hebrew and Arabic), some
4,000 years ago, had hand-grenades in their arsenals. Similarly, if new
words are formed using transparent compounding out of inherited
elements, that too can lead to the illusion of lexical inheritance. Here
I don't have to make up an example, as we have the classic example of
pseudo-Proto-Algonquian "whiskey" which Bloomfield first discussed.

So, something that sounds as if it should be simple in fact isn't.
Within the past few weeks, Don Ringe has had a series of guest posts on
Language Log that address this sort of question in much greater detail.
Don has been thinking about this more and more recently than I have.


--
 =======================================================================
Alice Faber                                       faber at haskins.yale.edu
Haskins Laboratories                            tel: (203) 865-6163 x258
New Haven, CT 06511 USA                               fax (203) 865-8963

------------------------------------------------------------
The American Dialect Society - http://www.americandialect.org