Dear Luisma,

I would like to take this opportunity to demolish the claim made by Larry Trask and other Vasconists that seme, the Basque word for "son", reflects an earlier *sen-be.

The basis for this claim is a Latin text which says that Rufus, son of Sembexo, was betrothed to Prusilla, daughter of Troccus.

Now, bear in mind that this text is written in Latin, and that Rufus, the son, has a Latin name, whereas his father has an Iberian name. The writer could well have been a Latin monoglot.

It is only possible for Trask and his followers to analyse the name, Sembexo, in the way they do, because they simply do not know how Iberian personal names are usually constructed. And this is because they deny that Basque and Iberian are related. (They recognise the Iberian spoken in Aquitaine as being related to Basque and call it "Aquitainian", but refuse to accept that the Iberian spoken south of the Pyrenees was also closely related to Basque, even though it is almost identical to "Aquitainian".)

Sembexo is fairly obviously a compound comprising *sen "name" and *beso "arm".

*sen is not a common Iberian personal name element, but it is present in the name of the deity Senaicus (*sen-aiki "the exalted one of the name"), recorded on a Roman era inscription at Braga, Portugal.

Basque i-zen "name" contains an additional anterior i-, which is absent in Iberian and is also absent elsewhere in Dene-Caucasian: Chechen tse "name", Na Dene (Tlingit) sa/sen "to name".

*beso arm is also uncommon as an Iberian personal name element, but it is attested in the name of the Turdetanian regulus, Besadin "arm intellect".

The letter "x", in names recorded by the Romans in Aquitaine, probably represents the Basque "ts" and "tz", which is absent from Iberian dialects south of the Pyrenees.

To native Iberian speakers in Aquitaine, the name would have been *senbetso. The Romans could not tolerate the -nb- cluster, so they substituted -m- for -n-.

If Trask is right, and Sembexo is really to be analysed as *sembe-xo, then what does the "xo" part mean? And if seme is really a development of *sen-be, then what is *sen, and what is *be?

In fact, contrary to what Trask and his followers tell us, there is good evidence that seme existed in Iberian in exactly the same form as it now does in Basque.

There is the deity, Semnocosus (recorded in Roman sources), which appears to be a Roman rendering of *seme-no koso "I am the son" (a compound form typical of personal names), plus the common epithet *koso, which may be cognate with Basque gozo "sweet".

And there is the deity, Sisemius (recorded on a Roman era inscription at Moron de la Frontera, Andalusia). This would appear to be *sei-seme "six sons".

And there is also the French village, Semeac, Hautes-Pyrenees, which may have been *seme-acum, the Roman estate of a native Aquitainian called Seme.

Trask's motive for analysing seme as *sen-be is of course to provide evidence for his claim that intervocalic -m- in native Basque words always reflects and earlier -nb-.

While occasionally an intervocalic -m- reflects an earlier -b- (note, not -nb-) in stems (as opposed to compounds) (eg, Basque (h)emen/(h)eben, and Iberian eban, both "here"), normally this is not so.

Indeed, in some cases, the intervocalic -m- is also present in cognates outside Basque. For instance, Basque -kume "woman" (as in ema-kume) and Burushaski quma "concubine", Proto-Dagestanian *qama "woman", Proto-Yeniseian qam- "woman", etc.

And seme itself has at least one cognate outside Basque: Sumerian isim/isimu "shoot, sprout, offspring, descendant".

Intervocalic -nb- is very rare in native Basque stems, but it is certainly present in (z)enbor "tree trunk". (The dictionaries say that the initial "z" is present only in French dialects, but note Zenborain, just east of Pamplona.)

We know this because the -nb- in (z)enbor reflects a Proto-World intervocalic -mb-.

Let me give you some cognates to illustrate the point:-

Sumerian henbur "red shoots, stalks"

Proto-Yeniseian *tempal "root" (Kott tempul, Arin tembirgang/tenbir, both "root").

Indo-European: Scottish Gaelic tom "thicket, bush" (Proto-Celtic *tombos); English timber

Afro-Asiatic: Arabic tamr "date palm"

Uralic: Finnish tammi "oak tree"

Altaic: Kazakh, Kyrgyz and Tatar tamir "root"

Korean tombul "thicket"

Japanese teiboku "bush"

Amerind: Macro-Tucanoan temba "firewood"; Penutian tumaj "stick, wood"

If Trask is right - ie, intervocalic -m- in native Basque words really does reflect an earlier -nb- - then we would expect (z)enbor to be (z)emor. But it is not (z)emor, it is (z)enbor! Of all the numerous variants listed by De Kerexeta (and there are more than 20), not one exhibits intervocalic -m-.

Trask is quick to decry those linguists, archaeologists, historians, etc, who make elementary errors, yet in his analysis of the Iberian personal name, Sembexo, and Basque intervocalic -m-, the good professor has made two huge ones himself. Don't expect contrition.

With best wishes,

 

Angus J Huck