Bronze, iron, gold, silver

« previous post | next post »

In our ongoing quest to link up linguistics with archeology, we have had numerous posts involving Iranian-speaking peoples spreading from west to east and bringing culture and language with them.  When I say "culture", I mean technological as well as spiritual, artistic, architectural, and other aspects, plus social customs and political organization.  Because the Iranian-speaking peoples were so active in spreading diverse manifestations of culture, I often refer to them as Kulturvermittlers par excellence.

Among the more prominent features of culture that Iranian-speaking peoples transmitted across Eurasia was metallurgy.  That includes all four of the main metals:  bronze, iron, gold, and silver.  The first two were mainly for weapons and implements, and where they went, they transformed military affairs, agriculture, and daily life.  The changes that bronze and iron brought about amounted to revolutions of civilization.  Gold and silver were primarily for ornament and embellishment, and the Iranian-speaking people created breathtakingly beautiful works of art out of these precious metals

What prompted today's post was coming across this fascinating database of Eurasian silver vessels from the 3rd- to 13th-century, containing many Sogdian and Sasanian vessels.  There is even a map of hoards, and one can click on each separate hoard in the menu.  I was particularly intrigued by the continuity of Iranian iconography, e.g., weaponry and costume, but above all the persistence of the cross potent (☩ [U+2629]), which I have elsewhere written about as a sign of the magi, that already worked its way into the oracle bone script around 1200 BC.

In previous posts, we have described how the Iranian-speaking peoples were among the first to ride in chariots and to mount on horses, so we can think of them as being highly mobile.  They were also responsible for the spread of key instruments and modes from the Middle East to Central Asia and thence to East Asia (more on that in forthcoming posts).  So we can call the Iranian-speaking peoples masters of metallurgy, mobility, and music, but much more as well.

In their peregrinations across Eurasia, the Iranian-speaking peoples often encountered ethnic groups who spoke different languages.  For example, during the medieval period, the Sogdians formed close alliances with Turkic-speaking peoples.  In earlier times, Scythians lived in association with Tocharians.  One thing we should keep an eye open for is evidence of exchanges between Germanic and Iranic peoples and between Finno-Ugric and Iranic peoples (documentation of such contacts does exist in specialist literature).  Naturally, it is also essential to glean specific words that passed from one language to another, but just knowing such words themselves does not demonstrate how they got where they ended up.  For that we need bones, bodies, burials, and artifacts — hard archeological evidence — and the symbolism they carry or convey.

Selected readings

[Thanks to Petya Andreeva]


  1. Chris Button said,

    January 29, 2021 @ 9:06 pm

    Some random remarks with lots of nice ə/a alternations:

    I tend to support the idea that 金 *kə̀m "gold" is attested in the oracle-bones. I do think it is related to 柑 *kám "orange"

    銅 *láŋʷ "copper" is related to 彤 *lə́ŋʷ "red, vermilion"

    鐵 *ɬə́c "iron" seems related to 錫 *ɬác "tin" (where *ɬ- → s-) and perhaps was loaned into Balto-Slavic:

  2. R. Fenwick said,

    January 30, 2021 @ 3:57 am

    Intriguingly, the single Indo-Iranian metallurgical term borrowed into North-West Caucasian arises not from Iranian, but from Indic, via a very early loan at the Proto-North-West Caucasian level. Ubykh daʃʷanə́ ~ dáʃʷanə "silver", Proto-Circassian *tːəʒənə́, and Proto-Abkhaz-Abaza *raʥənə́ all go back to Proto-North-West Caucasian (my reconstruction) *raɖʐynə́ "id.", placing it firmly alongside Sanskrit árjunam "silver" in particular rather than any cognate of Proto-Indo-Iranian *ʜr̥j́atám "id." (cp. Avestan ərəzatam, Old Persian ardatam, Sanskrit rajatám "id."). There are some other strikingly clear Indic Kulturwörter reflected in Ubykh loans; I'm hoping to treat these in a paper in the near future.

  3. Chris Button said,

    January 30, 2021 @ 7:16 am

    There is a suggestion that 銀 *ŋrə̀n "silver" is the source of Tocharian A nkiñc / Toch. B ñkante "silver".

    Blažek (2018) has a good discussion about the proposal and ultimately rejects it. He is, however, too accepting of the Tibeto-Burman cognates of 銀 since there are phonological issues there too. Sagart (1999) notes that 銀 appears late in Chinese, but then still wants to directly associate the Tibeto-Burman forms. Rybatzki's (1994) discussion of Mongolian in relation to 銀 adds some more fodder to the mystery.

  4. R. Fenwick said,

    January 31, 2021 @ 8:53 am

    The Chinese form is interesting indeed under the circumstances.

    Though I can't speak to the relatedness of the Sino-Tibetan terms, on the Tocharian side I think Blažek is also far too quick to discard Witczak’s (1990) hypothesis proposing an origin in PIE *h₂reǵn̥tóm via the single slight irregularity of an assimilation to *h₂neǵn̥tóm. His alternative suggestion of a link to Sogdian n’ktync requires the acceptance of three phonologically quite separate irregularities (including a quite similar irregular assimilation, as well as a deletion by haplology and vocalic metathesis) on the grounds that these have Tocharian parallels. But these deformations are required before the Sogdian word ever hits Proto-Tocharian, undermining the relevance of these purely internal (i.e. characteristic of either Tocharian A or B, but not attested in both, therefore impossible to ascribe to the period of Proto-Tocharian unity) parallels.

    Also, while true that *r…n → *n…n is not overtly attested in Tocharian, a very similar resonant assimilation of *r—*r…l → *l…l—is: cp. Toch B lakle ‘pain, suffering, sorrow’, most likely cognate with Homeric Greek λυγρός ‘miserable, unhappy’ (see Adams 2013. The Greek term could alternatively be a dissimilation, of course, but if these terms go back to PIE *lugRós (where *R is either *-l– or *-r-), the form we’re dealing with is a suffixally-stressed derivative of the zero-grade root (*leug-), making it probably a Caland adjective in *-– and suggesting that the Greek form is the archaic one.

    Ultimately, I'm with Witczak. The assimilation PIE *h₂reǵn̥tóm → Late Pre-Proto-Toch IE *h₂neǵn̥tóm might not have exact internal parallels, but it’s a common enough sporadic sound change cross-linguistically, and from what I can find the descent from *h₂neǵn̥tóm into Tocharian B is thenceforth essentially regular (Toch A nkiñc, instead of the regularly-expected *nkänt, is probably a back-modelling from the derived adjective nkäñci ‘of silver, silvery’, with subsequent i-umlaut of the vowel in the vicinity of palatals, common in both Toch A and B). Also, the Toch A derivative nkäñci is paralleled by Toch B ñkañce, both showing fully regular descent from an IE *-yo-derivative (presumably Late Pre-Proto-Toch IE *h₂neǵn̥tyo– → Proto-Toch *näkä́ntye-), which further undermines a Sogdian origin.

  5. Chris Button said,

    January 31, 2021 @ 10:47 am

    … might not have exact internal parallels, but it’s a common enough sporadic sound change cross-linguistically …

    I think that's the crux of the issue regarding the validity of loanword origins.

    When using the comparative method to reconstruct a proto-language, the evidence comes from the phonological basis of regular sounds laws. However, when loanwords are introduced, it becomes a question of surface phonetics and how those might be processed within the phonological system of the recipient language, which isn't going to match that of the donor.

    Unfortunately, phonetics doesn't tend to play much of a role in historical reconstructions. A classic example would be the suggestion that Old Chinese had nasal prefixes that caused the voicing of obstruent onsets. Far more likely is that the nasalization resulted from the incompatibility of voicing with obstruents, which led to the use of nasalization as an articulatory mechanism to preserve the voicing (a phonetic phenomenon well-attested cross-linguistically).

    There are still degrees of reasonableness, though. My ramblings over on the "Zoroaster/camel" thread ( about 駱 almost certainly stretch those boundaries too far (there's pretty much a way to make a linguistic argument around any sound change if you try hard enough). Nonetheless, the reason I threw those ideas out there on the off chance it would stimulate an idea from someone else is that there is sound evidence elsewhere that it must be a loan.

    Given the antiquity of 駱, its specific horse rather than camel sense, and its interchange with other first syllables, I suspect that it most likely was added to 駝, representing (uš)tra-, internally and is unrelated to the source:

  6. Chris Button said,

    January 31, 2021 @ 10:50 am

    However, when loanwords are introduced, it becomes a question of surface phonetics and how those might be processed within the phonological system of the recipient language, which isn't going to match that of the donor.

    It's probably worth adding that the loans don't tend to enter at the same time either, which just adds to the confusion.

RSS feed for comments on this post