Re: Celtic and Afro-Asiatic?

Joerg wrote:
> I have recently read an article by Theo Vennemann in which he
> proposes that the peculiarities of the Insular Celtic languages
> are due to a "Semitidic" (i.e., Afro-Asiatic) substratum, and
> I don't really know what to think about it.
>
> Six years ago (March 7, 1999), Sally Caves posted the following list
> of common features of Celtic and Semitic to this list:
I think it is generally agreed that many of these similarities
either fall out from typological generalizations common to all
VSO languages, or from languages generally.  This does not mean
there is no chance for an influence in either direction (more
likely Semitic => Celtic, via Berber and extinct Celtic languages
in Spain), but that such influence is likely to be only tentative.

> 1) Conjugated prepositions (prep. + pronominal object in a single word.
> 2) Word order: VSO, N-Modifier, Prepositions
Both of these are common to VSO languages in general.  E.g., Nahuatl, a
VSO language clearly not influenced by a "Semitidic" substratum, has
prepositions that agree in person and number with their arguments.
[2] also clearly is the result of general patterns of headedness.

> 3) Relative clause linker: invariant particle, not relative pronoun.
You mean, like English "that"? What about Italian "che"? What about
all those isolating languages of East Asia, most of which I seem
to recall use a form unrelated to the question word for their
relative clauses?

> 4) Relative clause technique (oblique): copying, not gapping,
>     i.e., "the bed, I slept in it," meaning "the bed that I slept in."
Many unrelated languages have this.  See Comrie's book on language
universals.

> 5) Special form of the verb peculiar to relative clauses.
Um, countless languages have this.  E.g., Meskwaki has the conjunct
order, which does the same thing.  (Although there it's a little
more complicated:  by a strange kind of grammaticalization of that
fact, all traditional tales are in the conjunct order for the most
part, since there is at least the implication that someone said that...)

> 6) Polypersonal verb (subject and object both marked).
Since when do most Semitic languages have polypersonal verbs, or
Celtic languages for that matter?!?  To be sure, Akkadian and other
Semitic languages have object *clitics* that may be attached to the
verb.  But these are not agreement markers in the normal sense.

> 7) Infixing/suffixing alternation: Object marker is infixed to
>    the verb if there is a preverb, suffixed otherwise.
Lots of languages have such alternations.  Tagalog is only the most
famous example.  Tagalog is also IIRC VSO.

> 8) Definite article in genitive embeddings may occur
>   only on on the embedded noun: "house the-man" ="the man's
>   house."
Note that in English this is also the case:  we may not (unlike
Greek or German) say *"the the man's house"; by a pragmatic implicature
that has been grammaticalized, possessors are definite and therefore
what they possess is too.

> 9) Nonconcord of verb with full-NP subject: verb can fail
>     to agree with the subject, depending on word order.
But this happens in ancient Greek and modern English!

> 10) Verbal Noun (Vn: object in genitive), not Infinitive
>       (object in same case as with finite verb).
Um, Georgian is exactly the same way.  In his book on polysynthesis,
Baker claims that all truly polysynthetic languages are characterized
by this.  He's wrong in the absolute, but there does seem to be this
tendency with polysynthetic languages.

> 11) Predicative particle: in copular or nominal sentences,
>       the predicate is marked with a particle homophonous
>       to a "local" preposition: "He (is) in a farmer"="he is a farmer."
> 12) Prepositional periphrastic: BE + Prep + VN, e.g.,
>       "He is at singing"  [TEONAHT'S  "she is with singing"]
You mean, just like Old English, Middle English, and many modern
English dialects?  (I suppose Vennemann could counter that that's
because of the Celtic and Semitic substrate in Britain, but then
that begs the question.)

> 13) DO periphrastic: DO + VN, e.g. "He does singing."
Lots and lots of languages have light verb constructions like this,
e.g., Hindi.

> 14) Notional adverbial clause expressed as "and" + finite
>       clause
> 15) Nonfinite forms usable instead of finite main-clause verb
You mean, just like ancient Greek, or Meskwaki?

> 16) Word-initial change, expressing a variety of syntactic
>        functions
Actually, Meskwaki (like most Algonquian languages) has a kind of
ablaut just like this, called "initial change".  It's frequently
used in relative clause constructions, but in a number of other
unrelated things as well.  (Interestingly, Meskwaki also tends
to VSO order, though I don't see a clear connection between those
facts.)

I don't know much about the Semitic initial change, but how is
it like Celtic lenition?

> 17) Idiomatic use of kin terms in genitive constructions, e.g.
>       "son of sending" = messenger; "son of land" = "wolf"
Of all of the above, the only ones that stand out to me as odd are
(11), (14) and (17).  But in each there's no obvious reason why
they have to have arisen by contact.  When you get to know a language
well, you find that it has all sorts of crazy idioms, some of them
entirely contradicting the overall patterns of the language (like
"Butter bei die Fische" in German).  For example, I help edit my
friend's papers for publication, and it took him a while to learn
that in English, one cannot use a finite form with a topicalized
adverb "as", as in "as noted in paragraph 3...".  I find that many
nonnative English speakers try to say something like "as it was
noted in paragraph 3".   This is the kind of construction that
one must simply learn; it does not fit into the larger schema of
the language very well.

Anyways, I hope this post will show that Vennemann's arguments are
probably ultimately misleading, and certainly do not constitute
proof.  Given an ounce of skepticism, it's really remarkable that
this story has held on as long as it has.

Joerg later wrote:
> A problem with the assumption of an Afro-Asiatic substratum
> in Insular Celtic is that there is no trace of Afro-Asiatic
> anywhere in between (Iberian peninsula, France).  Instead we
> have Basque along the way, which has nothing to do with AA,
> neither genealogically nor typologically.  But that doesn't
> mean that an Afro-Asiatic language in Britain is impossible.
Actually, it's not inconceivable.  The Phoenicians had trade
interests in Britain, as Britain was one of the better sources
of tin in the ancient world.  So they could have founded emporia
along the same lines as those the Greeks founded throughout the
Mediterranean world.  But AFAIK there's no archaeological evidence
for such an emporion, and so any link is little more than a
remote possibility.

Patrick Littel wrote:
[snip a lot]
> Oh, and if I look around, I can probably find you a book claiming
> Semito-Mayan contacts as well.  Might be a fun read!
I agree with most of your comments.  If you're looking for such a book,
do a search for Ignatius Donnelly. He was a retired lawyer in the 19th
century who had all sorts of crackpot ideas, and spent his free time
researching possible evidence for Atlantis.  I remember reading (back
when I was younger, and had WAY too much freetime) that he made some
linguistic connections between the name of Noah and the Flood survivor
of various Native American cultures.

 =========================================================================
Thomas Wier	       "I find it useful to meet my subjects personally,
Dept. of Linguistics    because our secret police don't get it right
University of Chicago   half the time." -- octogenarian Sheikh Zayed of
1010 E. 59th Street     Abu Dhabi, to a French reporter.
Chicago, IL 60637

Reply