Re: etymology and semantic domains

From: Don Wilkins (
Date: Thu Apr 23 1998 - 16:38:16 EDT

Rolf and Clayton have provided excellent insight into this sticky issue. I
would like to respond to Rolf's comments below.

At 04:00 PM 4/23/98 +0200, he wrote:
> For a functional equivalence translation the "kernel" (which is found by a
>semantic analysis rather than a grammatical analysis) is the fundamental
>translation unit. I have for a long time studied the question: Is it
>possible to create a linguistic model (without seeking recourse in the
>etymological fallacy) where the *single word* is the fundamental
>translation unit? I believe the answer is yes, but I repeat that it only
>works for the particular target group already mentioned. Because this model
>only works in the case of one target group it is no rival for functional
>equivalence tranlation, rather a useful complement. However, one
>interesting fact with my model is that there is little need on the part of
>the translator, either for etymological or for Semantic domain word
>studies, because of the role of the reader. If I should create a slogan in
>relation to this very simple model, I would say: "More power to the

First a question: is your *single word* linguistic model still based on the
semantic domain concept? At first I thought it was, but then your comment on
the translator's approach confused me. For that matter, what do you see the
translator using as his/her sources and methodology for this model?
Having asked that, let me suggest that in reality the translation process
runs far too wide a spectrum to be categorized as kernel vs. single word,
semantic domain vs. etymology, etc. It is most probably a delicate
combination of all methods and sources that have some claim to legitimacy.
E.g., I am as quick as anyone to condemn root-fallacy errors, yet Clayton
has already pointed out that there are cases in which the use of etymology
is valid. Also, in the real marketplace, readers have the rightful
expectation that a *translation* is not a commentary, as it can quickly
become if the translators try too hard to make the meaning of the text
"clear". When we say that we are trying to determine how the first-cent.
reader in this or that precise geographical location understood a word or
phrase, I heartily concur with the goal, but fear that the ultimate answer
is "ignoramus". How do we really figure that out, with any acceptable degree
of certainty? Maybe an anecdote will illustrate my concern. The best-read
professor I ever had during my grad work often criticized our feeble efforts
in Greek prose composition by pointing out that we were using inappropriate
(often poetic) diction. One of his complaints will forever ring in my
memory: "I can't find this word in Thucidydes!!" One day I thought I had
him, because I had checked and double-checked my Greek vocabulary in the
unabridged LSJ, to confirm that none was poetic or from the wrong period.
When I informed him of this fact, he pointed out that, as scholars, we ought
not to trust anything so limited as LSJ, and that the only way to know for
sure how a word was being used was to read the literature. Stunned, we asked
him how he became so well-read, and he said only that he had been privileged
to read Greek at least 8 hours every day for 10 straight years (doesn't
everyone? ;-)).
The point of mentioning this anecdote is not to discourage beginning or
intermediate Greekers! There probably are not many people on the planet who
have achieved this level of acquaintance with the Greek (I wouldn't be
surprised if a few of them were list contributors, myself excluded). Rather,
I am wondering who of the experts in Greek semantics have even begun to
approach such a degree of familiarity with the language that they could
become legitimate critics of the major standard lexicons, as this professor
did. Not that these experts do not have valid opinions, or that the lexicons
are infallible sources. Indeed, the best and thickest lexicons are such not
because they offer a myriad of possibilities of meanings for words, but
because they offer a good minimal number of references that the reader can
check for him- or herself. A researcher should read all these references
within their wider contexts, sorting them out by time periods and other
relevant factors. I would hope that the experts in semantics have done at
least that much. Perhaps they have done that and more, including global
computer searches, etc. (I really don't know). In that case, my hat would
be off to them (if I wore a hat), and I probably ought to take their works
as the best primary sources of meanings, while nonetheless checking their

Don Wilkins

This archive was generated by hypermail 2.1.4 : Sat Apr 20 2002 - 15:39:35 EDT