Language Log

« previous post |

The first part of this virtuoso study of the Afro-Eurasian archeolinguistics of the donkey and its concomitant terms in diverse languages across vast expanses of land from East and North Africa to the heartland of East Asia was described in "A Sino-Iranian tale of the donkey's Eurasian trail" (5/10/24).  This post summarizes the second part of the study, which appears here:

Samira Müller, Milad Abedi, Wolfgang Behr, and Patrick Wertmann, "Following the Donkey’s Trail (Part II): a Linguistic and Archaeological Study on the Introduction of Domestic Donkeys to China", International Journal of Eurasian Linguistics, 6 (2) (October 16, 2024), 294-358.

The first two paragraphs of the Abstract were reproduced in the Language Log post cited in the first paragraph above, so there is no need to repeat them here.  Here is the third paragraph of the Abstract, which appears at the head of the just published Part II of the article:

While Part I focussed on the question how and when donkeys might have been introduced to China, Part II retraces the origins of Iranian and Chinese terminologies concerning the donkey. Section 3 first provides a chronological overview of words attested in Persian literature and Iranian contact languages. The second half (subsections 3.5–3.6) proposes several possible explanations for the etymology of the word for donkey in Iranian languages. Section 4 discusses Chinese donkey related terminologies in five subsections, laying a special emphasis on the interrelation with words for its hybrid relatives. Finally, section 5 concludes this paper with a discussion considering all findings of both Parts.

The linguistic and archeological evidence adduced by the authors under the following headings is plentiful and detailed:

• 3 The Donkey’s Sombre Passage through Iranian Languages
  • 3.1 Linguistic Evidence for ‘Donkey’ in Old Iranian Languages (1000–400 BCE)
  • 3.2 Old Iranian Contact Languages
  • 3.3 Linguistic Evidence in Middle Iranian (400 BCE–900 CE)
  • 3.4 Middle Iranian Contact Languages and the Term for the ‘Donkey’
  • 3.5 The Borrowing Scenario
  • 3.6 Internal Innovations in Indo-Iranian? An Alternative ‘Colour Scenario’
  • 3.7 Interim Conclusion on the Iranian Data
• 4 How Donkey Terms Entered Chinese
  • 4.1 A Term for a New Companion
  • 4.2 A Door for the Donkey
  • 4.3 The “Donkey-Donkey”
  • 4.4 Mulling over Mules
  • 4.5 Interim Conclusion: Searching for the Best Name

Because we at Language Log have especially stressed the value of evidence for and from Tocharian, I wil quote three paragraphs focusing on that language:

Tocharian. At least three branches of Indo-European (Tocharian, Iranian, and Indic) used identical vocabulary to designate donkeys or donkey-like equids: Toch. B khare– ‘ass’ (hapax), YAv. xara-, and Skt. khara-. The Tocharian form stems from Buddhist Skt. khara– ‘donkey’ (Adams 2013: 263). Toch. B khare– ‘ass’ only appears once in an unclear context (Adams 2013: 263), making it another unsafe hapax. Toch. B also has another form for ‘donkey’ and ‘ass’, kercapo-, which is a cognate or borrowing from Sanskrit gardabhá– ‘donkey, ass’ < *gordebho– (Adams 2017: 1368). Adams has pointed out, however, that the borrowing must have occurred very early, i.e. before the merger of the non- high vowels in Indo-Iranian. A common inheritance from PIE seems therefore more likely (Adams 2017: 1368). If so, this word must initially have had another meaning than donkey, since it was used long before the long-eared domestic animal was introduced to Indo-Iranian populations.

Apart from the word khare– ‘donkey’ and the seemingly older form kercapo, there is Toch. B etwe ‘mule’. As Peyrot (2018) has convincingly shown, Toch. B etswe ‘mule’ is a loanword from Proto-Iranian *atswa– ‘horse’ (cf. Av. aspa– and Khot. aśśa-). Tocharian inherited its word for ‘horse’ directly from Proto-Indo-European *h₁eḱu(o)- (Beekes 2011: 38) as Toch. A yuk and Toch. B yakwe ‘horse’ show. Peyrot considers Toch. B etswe as a kind of horse or horse-like animal that was sufficiently different from the “normal” Tocharian horse. He believes that Toch. B etswe was introduced through an Iranian dialect possibly spoken in the north or the east of Tocharian speaking area (Peyrot 2018: 270–271).

As a preliminary conclusion regarding the case of Old Iranian, it is worth stressing the lack of direct and safely attested evidence for a word meaning ‘donkey’. The comparative study of the forms for ‘donkey’ in Sanskrit, Tocharian, and Old Iranian seems to support the idea that the suspected Avestan term, xara-, may indeed refer to a ‘donkey’ or a ‘donkey-like equid’. Nevertheless, what can be the possible reasons for the unsafe spelling ⟨x-r, x-a-r, x^v-a-r⟩ of this term in Avestan? The writing system of Avestan is famous for its highly precise alphabet of 56 vowels and consonants. Do the different spellings thus reflect the initial variance of a longer loanword adaptation process or was it caused by another, possibly internal, development? The situation in Sanskrit and Tocharian is not much clearer than in Old Iranian. We are confronted with either an unsafe hapax in Tocharian or a lack of attestation in Vedic texts. Already Lüders (1935: 85) highlighted the lack of attestation of Skt. khara– ‘donkey’ in Vedic texts and interpreted Skt. khara– as an Iranian loanword into Sanskrit. As far as we can see, the only attestation of Skt. khara– ‘donkey’ in Vedic texts belongs to the Paippalāda-Saṁhitā of the Atharvaveda. There, it is mentioned as kharo (Bhattacharya 2016: 1723), the Sandhi form of Skt. kharas ‘donkey.NOM.SG.M’.

The final section of Part II, "Concluding Discussion: Retracing the Donkey’s Trail", offers a good summary of the achievements of Part II, so I quote it in full here:

As discussed in Part I section 2, reported archaeological finds of donkey remains from China are very rare and their clear identification as either wild or domestic lacks a comprehensive analysis in most cases. However, recent studies on modern donkey bones and one reported archaeological find from the ninth century CE grave of Lady Cuī in Xī’ān both show that these donkeys descend from the African wild ass. Apart from the debated depiction of a possible donkey on a Western Hàn (25–220 CE) pictorial tomb brick from Zōu county in Shāndōng province, clear art historical evidence on the presence of donkeys in China proper only appears during the early Medieval period. In these depictions, donkeys mainly serve as riding and transport animals. In contrast, north of the Chinese core area, in the Ordos region of modern Inner Mongolia, several bronze ornaments in the shape of donkey-like equids with bent necks and long ears that are typically dated between the sixth and second centuries BCE have been found. Considering that this region has been home to the Mongolian wild ass (Equus hemionus hemionus), i.e. the nominate subspecies of the Asiatic wild ass (Equus hemionus), until today (cf. Yang Weikang 2007), chances are high that the depicted beasts on the Ordos bronzes were indigenous and hunted for their meat. Further variants of Asiatic wild asses are found across Eurasia populating the lower arid desert and steppe regions, and Tibetan wild asses (Equus kiang) inhabit the north-western Himalayan plateau. However, it seems that none of these elusive beasts have ever been domesticated. While some of them were interbred with donkeys at different points in time, due to their genetic difference⁵² their offspring was likely sterile.

When looking at the Chinese philological evidence, it appears that the domestic donkey started to gain recognition by the second half of the third century BCE. As it was still considered a rare animal which was gifted to the northern states of the Eastern Zhōu realm, it was primarily held in stables as a symbol of wealth. This probably started to change during the reign of the Qín dynasty after it unified the former Zhōu states and expanded its area of influence northwards. The donkey was presumably present and used in the Tarim Basin as beast of burden to transport goods between the oasis settlements since the first millennium BCE.⁵³ While the donkey might have been less suitable than the Bactrian camel to cross the desert regions, it was the perfect beast to transport heavy loads across the piedmont regions along the rims of the Tarim Basin where various semi-nomadic settlements thrived. Furthermore, it also facilitated the crossing of the surrounding mountain ranges, the Kūnlúnshān 崑崙山 in the south and the Tiānshān 天山 in the north. Apart from their role in transport, the advantages of donkeys in agriculture should not be forgotten (see e.g. Brodie 2008). Donkeys were ideal for everyday transport of water, supplies, and people to the fields, as they were easy to control, and played a critical role in harvest efficiency since they could also be used as draft animals⁵⁴ to pull heavy carts. Societies of the Tarim Basin originally were cattle breeders (see e.g. Yu Chong et al. 2022), but with the proceeding aridification it must have been exceedingly difficult to maintain the herds (cf. Hinsch 1988: 138). While cattle are strong and can serve as powerful draft animals, they need an ample amount of water and time to ruminate. In contrast, donkeys profit from a faster fibre consumption and the ability to retain water in their bodies for an extended period of time (De Santis et al. 2021: 4). The passage about the practice of the Kingdom of Wūchá to raise only donkeys but no cattle reflects this earlier change from cattle to donkey breeding which took over in the region around the Tarim Basin after the donkey had been introduced. As the use of donkeys spread through the Tarim Basin, people living in the vicinity of the Héxī Corridor, among them the Róng herdsmen providing the First Emperor of Qín with what might have been mules (see Part I section 2.3), were among the first to introduce the new equid to China. It is important to notice that by the third century BCE horses were still more common than donkeys in the Héxī region, thus hybrid breeders were probably more likely to let their rare donkey jack sire some horse mares, which would give birth to mules, than to endanger some precious donkey jennies by trying to breed hinnies.⁵⁵ The exceptional status of hinnies, possibly reflected in the ‘special donkey’ term jùxū 駏驉 (see section 4.3), might be connected with this initial rarity of the hybrid, whereas mules became a common sight in the Chinese heartland much sooner.

By the time the Xiōngnú confederation formed in response to the Qín expansion to the north (Di Cosmo 1994: 1116), donkeys and mules were bred alongside camels, especially in dry regions, and thus became known to Chinese historiographers as ‘special domestic animals of the Xiōngnú’, even though the peoples using the beasts did not necessarily identify as Xiōngnú themselves. Finally, when Hàn Emperor Wǔ attacked the Xiōngnú, he learnt about the effectiveness of donkeys, and when the southern Xiōngnú accepted their formal submission to China in 51 BCE, the sturdy beast of burden had definitely found its place among the domestic animals of the Central Plains. But how did donkeys arrive in the Tarim Basin? Considering that the donkey had been a popular animal in West Asia since its first domestication before the third millennium BCE, it is noteworthy that it only gained popularity on the eastern side of the Pamir Mountains by the 1st millennium BCE. While its introduction to the Tarim Basin might be connected to an increase of mobility between the uprising city states in the region, other factors could also have led to the rather late arrival of domesticated donkeys. Both the Assyrian and the Mesopotamian Empires made great use of donkey caravans for transporting goods through regions lacking suitable waterways (cf. Algaze 2008) and as depictions in Persepolis show, the Achaemenids were surely acquainted with the animal. But the question of when and how the animal gained popularity in the eastern areas of the Persian Empire is not easily answered. Iranian philological and linguistic evidence seems to reflect a rather diffuse origin of the name for the beast during the Old Iranian period. Only in Middle Iranian, sources of which are dated between the fourth century BCE and the tenth century CE, we find clear evidence for an actual animal called xar ‘donkey’. In southern Iran, donkeys became possibly known through contacts with Mesopotamia in the third millennium BCE, when they were used to connect regions which were difficult to access by boat (Clutton-Brock 1992: 65). Despite Potts (2011) being convinced that donkeys were well known, if not domesticated, on the Iranian Plateau already by the fourth millennium BCE, most evidence seems to support the idea that donkeys spread to northern Iran along with the expansion of the Old Assyrian Empire. It is important to notice that regions of northern Iran were used as breeding centres for donkeys with indigenous wild Asian asses (Goulder 2018: 84). According to the cuneiform manuscripts of Kültepe-Kanesh, a trade hub in modern Anatolia, donkeys were excessively used under the Old Assyrian reign during the Middle Bronze Age (ca. 2000–1750 BCE). An extensive caravan business called for good donkeys which were bred in the capital Aššur. A special breed of donkey is often mentioned in the texts: the so-called ‘black donkey’ (emāru[m]/imēru[m]) which is identified with the modern Damascus ass, a large, strong and dark-haired breed (Atici 2014: 244–245). Donkeys were often sold for good money at the caravan’s destination if the goods to be transported back to the capital required less donkey power (Barjamovic 2018: 152). In this way, donkeys might also have found their way to the Iranian plateau, where they possibly became known as the ‘black’ variety of asses in contrast to the local wild asses with which they could be interbred to produce “donkey-likes” (PIr. *khara-tara-), i.e. mules or hinnies (Kümmel 2017: 279). However, the PIr. word *kh-er-o- meaning ‘dark’ as suggested in section 3.6 “A new etymology for YAv. xɑrɑ-?” could also be a translation from Old Assyrian ṣa-lá-mu-um ‘black’, which was used to describe the prized black donkey (Zarins 1978: 4). Yet again, the fact that donkeys were brought to northern Iran in order to be crossbred with local wild asses and possibly also horses could have triggered an association with a mythical beast of fertility after which the strange equid was finally named (see section 3.1 “Commentary”).

It should be stressed that in Iran, like in China, the donkey was not the primary equid, but a “latecomer”. As such, it was constantly compared to the horse which played a central role in many societies east of the Zagros mountains. Upon its introduction, the nobility might have found pleasure in the animal as a ‘rare breed’, but there was no cultural background which would have given the donkey a ritual, cultural and social role comparable to the one it enjoyed in West Asia (cf. Greenfield et al. 2012). Thus, donkeys were destined to become animals of the common people and easily gained a bad reputation or even developed into negative symbols as soon as they became more available. This supposedly happened on the Iranian plateau⁵⁶ as well as later in China and possibly in many regions in between (cf. Shirasaka and Watanabe 2019). Various vernacular variants for the donkey’s name might well have led to multiple written forms in Iranian literature, thereby making its etymology extremely difficult to track.

At least since the reign of the Achaemenids (550–330 BCE), donkeys maintained a certain importance as beasts of burden.⁵⁷ As it can be seen in Achaemenid administrative documents, donkeys were used in Bactria, which means that they had been established as beasts of burden in Eastern Iran by the fourth century BCE (see section 3.2 “Achaemenid Aramaic”). It is possible that the fall of the Achaemenid Empire conditioned a stronger eastward orientation among some of the former satrapies, which could have facilitated the spread of domestic donkeys to the western end of the Tarim Basin.

The late borrowings seen in Tocharian could be in line with the comparatively late introduction of donkeys to the area, since people living on the northern rim of the basin relied on the Tarim River for east-west transport, while on the southern routes travel on foot, camel, or horseback was more common. The fact that the Toch. B word for ‘mule’ is the only donkey-related word which can be identified as a loanword from Old Iranian (*atswa– ‘horse’) beautifully shows that mules were probably first to arrive in the Tarim Basin, i.e. before their donkey parents, since they were more resistant against the adverse conditions on the passageways through the Pamir mountains (see section 3.2 “Tocharian”). On the southern rim of the Tarim Basin, however, the settlements were more dispersed and not connected by a direct river system. The piedmont region had to be traversed on foot or on camel back to travel between the villages nestling around the meltwater streams flowing from the Kūnlún mountain range. It was probably in this region that the donkey was an exceptionally welcome domestic beast, as it enabled the extension of agricultural activities as well as better interconnection – in a similar way as it had enabled the flourishing of the Late Uruk culture three millennia before.

We need an adjective that stresses the manifest, manifold virtues of the donkey.  If we had one I would apply it many times over to the authors of this outstanding collaborative article, who have worked so hard, so long, and so meticulously on it.  We owe them a great debt, as we do to the donkeys who have served humanity dutifully for more than seven millennia — but don't push this admirable (to my mind) creature too far!

Suggested readings

• "So many words for 'donkey'" (3/17/23)
• "'Little competent donkey'" (9/18/20)
• "Joshua Whatmough and the donkey" (9/1/09)
• "Flip Donkey Doodleplunk?" (2/22/18) —
• "Dialectology in 2020" (10/14/20) — "donkey's dick" (in the comments)
• "Donkeys in Cyberspace!" (6/9/08)
• "Wild Ass Homestay" (1/30/11)

October 19, 2024 @ 6:32 pm · Filed by Victor Mair under Language and animals, Language and archeology

Permalink