The revise the paper of the Eurasian nomad history

THE

HORSE THE

WHEEL AND

LANGUAGE

HOW

BRONZE-AGE RIDERS

FROM THE

EURASIAN STEPPES SHAPED THE

MODERN WORLD

DAVID W. ANTHONY

Princeton Univenity Press

Princeton and Oxford

2-007­

"T I

2.:'2 Chapter 10 Domesticatioll ofthe Hone 223

horse bones in southeastern Europe, central Europe, the Caucasus, and

northern Kazakhstan.

THE ECONOMIC AND J\1ILlTARY EFFECTS

OF HORSEBACK RlDlNG

A person on foot can herd about two hundred sheep with a good herding dog. On horseback, with the same dog, that single person can herd about five hundred. 3s Riding greatly increased the efficiency and therefore the scale and productivity of herding in the Eurasian grasslands. More cattle and sheep could be owned and controlled by riders than by pedestrian herders, which permitted a greater accumulation of animal wealth. Larger

herds, of course, required larger pastures, and the desire for larger pastures would have caused a general renegotiation of tribal frontiers, a series of

boundary conflicts. Victory in tribal warfare depended largely on forging alliances and mobilizing larger forces than your enemy, and so intensified warfare stimulated efforts to build alliances through feasts and the redis­ tribution of wealth. Gifts were effective both in building alliances before conflicts and in sealing agreements after them. An increase in boundary conflicts would thus have encouraged more long-distance trade to acquire prestigious goods, as well as elaborate feasts and public ceremonies to forge alliances. This early phase of conflict, caused partly by herding on horseback, might be visible archaeologically in the horizon of polished stone mace­ heads and body decorations (copper, gold, boars-tusk, and shell ornaments) that spread across the western steppes with the earliest herding economies

about 5000-4200 BCE.36 Horses were valuable and easily stolen, and riding increased the effi­

ciency of stealing cattle. When American Indians in the North American Plains first began to ride, chronic horse-stealing raids soured relationships even between tribes that had been friendly. Riding also was an excellent way to retreat quickly; often the most dangerous part of tribal raiding on foot was the running retreat after a raid. Eneolithic war parties might have left their horses under guard and attacked on foot, as many American Indians did in the early decades of horse warfare in the Plains. But even if horses were used for nothing more than transportation to and from the raid, the rapidity and reach of mounted raiders would have changed raid­ ing tactics, status-seeking behaviors, alliance-building, displays of wealth, and settlement patterns. Thus riding cannot be cleanly separated from

warfareY

.Many experts have suggested that horses were not ndden in warfare until after about 1500-1000 BCE, but they failed to differentiate between mounted raiding, which probably is very old, and cavalry, which was in­ vented in the Iron Age after about 1000 BCE.'~ Eneolithic tribal herders probably rode horses in inter-clan raids before 4000 BCE, but they were not like the Buns sweeping out of the steppes 011 armies of shaggy horses. What is intriguing about the Huns and their more ancient cousins, the Scythians, was that they formed armies. During the Iron Age the Scyth­ ians, essentially tribal in most other aspects of their political organization, became organized in their military operations like the formal armies of urban states. TIlat required a change in ideology-how a warrior thought about himself, his role, and his responsibilities-as well as in the technol­ ogy of mounted warfare-how weapons were used from horseback. Prob­ ably the change in weapons came first.

Mounted archery probably was not yet very effective before the Iron Age, for three reasons. The bows reconstructed from their traces in steppe Bronze Age graves were more than 1 m long and up to 1.5 m, or almost five feet, in length, which would clearly have made them clumsy to use from horseback; the arrowheads were chipped from flint or made from

bone in widely varying sizes and weights, implying a nonstandardized, individualized array of arrow lengths and weights; and, finally, the bases of most arrowheads were made to fit into a hollow or split shaft, which weakened the arrow or required a separate hollow fore shaft for the attach­ ment of the point. The more powerful the bow, and the higher the impact on striking a target, the more likely the arrow was to split, if the shaft had already been split to secure the point. Stemmed and triangular flint points, common before the Iron Age, were made to be inserted into a separate foreshaft with a hollow socket made of reed or wood (for stemmed points), or were set into a split shaft (for triangular points). The long bows, irregu­ lar arrow sizes, and less-than-optimal attachments between points and arrows together reduced the military effectiveness of early mounted ar­ chery. Before the Iron Age mounted raiders could harass tribal war bands, disrupt harvests in farming villages, or steal cattle, but that is not the same as defeating a disciplined army. Tribal raiding by small groups of riders in eastern Europe did not pose a threat to walled cities in Mesopotamia, and so was ignored by the kings and generals of the Near East and the eastern Mediterranean. 39

The invention of the short, recurved, compound bow (the "cupid" bow) around 1000 BCE made it possible for riders to carry a powerful bow short enough to swing over the horse's rear. For the first time arrows could

:':'-1 Cbi/flcr IU

be fired behind the rider with pcnttratillg power. l1Jis manCllver, Lner known as the "Parthian shot," \\',lS immortalized as the iCOJlic inugc uf the steppe ardler. Cast broJlze soekered anowhe:lds of stambrd weights and sizes also appcarcd in the Early Iron Age. A socketed armwhead did Jlot require a split-shaft mount, so arrow, vvith socklted arrowheads did Ilot oplit despite the power of the bms; they also did not need a separate fore shaft, and so arrows could be simpler and mUfe streamlined. Reusable moulds were invented so that smiths could produce hundreds of sucketed arrowheads of standard weight and size. Archers now had a much wider field of fire-to the rear, the front, and the left-and could carry dozens of standardized arrows. An army of !11uunted archers could now fill the sky with arrows that struck with killing powcr. 41J

But organizing an army of mounted archers was not a simple matter. 'Ihe technical advances in bows, arrows, and casting were meaningless without a matching change in mentality, in the identity of the fighter, from a heroic single warrior to a nameless soldier. An ideological model of fighting appropriate tor a state had to be grafted onto the mentality of tribal horseback riders. Pre-Iron-Age warfare in the Eurasiall steppes, from what we can glean from sources like the Iliad and the Rig Vcda, prob­ ably emphasized personal glory and heroism. Tribal warfare generally was conducted by forces that never drilled as a unit, often could choose to ig­ nore their leaders, and valued personal bravery above following orders. ll

In contrast, the tactics and ideology of state warfare depended on large disciplined units of anonymous soldiers who obeyed a general. These tac­ tics, and the soldier mentality that went with them, were not applied to riders before 1000 BCE, partly because the short bows and standardized arrOWS that would make mounted archery truly threatening had not yet been invented. As mounted archers gained in firepower, someone on the edge of the civilized world began to organize them into armies. That seems to have occurred about 1000-900 BCE. Cavalry soon swept chari­ otry from the battlefield, and a new era in warfare began. But it would be grossly inappropriate to apply that later model of mounted warfare to the Eneolithic.

Riding began in the region identified as the Proto-Indo-European homeland. To understand how riding affected the spread ofIndo-European languages we have to pick up the thread of the archaeological narrative that ended in chapter 9.

CHAPTER ELEVEN

/ ~ The End of Old Europe and the Rise of the Steppe

By 4300-4200 BCE Old Europe was at its peak. 111.e Varna cemetery in eastern Bulgari"\ had the most ostentatious funerals in the world, richer than anything of~e same age in the Near East. Among the 281 graves at Varna, 61 (22%) coh.t:ained lllore than three thousand golden objects to­ gether weighing 6 kgR'U Ib). Two thousand of these were found in just tour graves (1, 4,36, and 3). Grave 43, an adult male, had golden beads, arm rings, and rings totaling ,516 grams (3.37lb), including a copper axe­ adze with a gold-sheathed ha dle. l Golden ornaments have also been found in tell settlements in the lo~er Danube valley, at Gumelnita, Vidra, , . and at Hotnitsa (a 31O-gm cache a golden ornaments). A few men in these communities played prominent s iaI roles as chiefs or clan leaders, symbolized by the public display of shini gold ornaments and cast cop­

per weapons.

Ihousands of settlements.. ;'ith broadly simila ~ceramics, houses, and fe­ male figurines were occupi~d between about 4500 nd 4100 BCE in east­ ern Bulgaria (Varna), th~/upland plains of Balkan race (KaranovoVI), the upper part of the Lo~er Danube valley in western B'u~aria and Roma­ nia (KriVOdol-salcu~,aI and the broad riverine plains of th~ower Danube valley (Gumelnita) ( gure 11.1). Beautifully painted ceramic essels, some almost 1 m tall an fired at temperatures of over 800

a

C, line the walls

of their two-si:0ri(1 houses. Conventions in ceramic design and ri aI.were shared over larg regions. The crafts of metallurgy, ceramics, and eve1,1 flint working beca e so refined that they must have required master crafrspe­ cialists who fere patronized and supported by chiefs. In spite of this, power was ~ot obviously centralized in anyone village. Perhaps, as John Chapman pbserved, it was a time when the restricted resources (gold, cop­ per, Spondylus shell) were not critical, and the critical resources (land, tim­ ber, labor, marriage partners) were not seriously restricted. This could have prevented anyone region or town from dominating others. 2

225

.;88 Notes to Chapter II Notes to C!)i1pter II .;89

( 33'''-I':-()~)\()rsc8 in the Ca\lcasu~ I relied 011 the text of J. (onfercnu.: paper by l\lezhlullli:J.n (1990). A fn\' hvrses might hJVC pa~sed ~bfQUl!:h tlte,C;LUcaslls tntu !lortht:rn Iran br:fore 30()O

BCE, indicated I~~'-:f'~btc h(l;:-'C teeth at the site or Q~lhn.:~tan, we."! u{ Tc:hcran (see Ma,hk"ur 2(~o.Jhnd~a po"ihle horse . t Codin Ttpe (see Cilbcrt lYY1)~ N" dehnite horse

rtm:!i.J.H:rha·\.T been identified in CJstcrn Iran, Cc:ntra the Indian subcontinent 1Il depos­

<rfS'~htcd earlier tban 2000 BeE. cL1illlS to the contrary nutwlth:,tal1 ~'or a review of this Jebate, see l\ltadow and Patel ] Y~7~

34~ Fur central European horses, see See Beneeke 1994; Bi;konYI 1~7~; and Pesh 1'180

n 15. Khazallo\, lYY4:32. 36. For war and the pn:stigc trade, ~ee Vchik 2002. 37. 1he American Indian analogy is descrilH.:d 10 Anthon)' 1986. 1hc most detailed analysis of the effects of horsehack riding and hon,e ket.:ping on Plains Indi"lIl cultures is Ewers 1955.

38. Onc argument against riding before) 500 BeE was that steppe horses were too small to ride.1his is not true. More than 70% of the horses at Dercivka and Botai stood 136-144 cm at the widlers, or about 13-14 hands high, and some were 15 hands high. They werc the .same size as Rom:m cavalry hor~es. Another argumtnt is that rope J.nd leather bits were inadequate for controlling horses in hattle. 111is is also not true, as the American Indian~ demonstrated. Our

SUN)' students at Cobleskill also had "no problem" controlling horses with rope bits~ 111e third

is that riders in the steppes rode sitting hack on the rump of the horse, a manner suited only to

riding donkeys, which did not exist in the steppes~ We have rebutted these doubts about Eneo­

lithic riding in Anthony, Brown, and George 2006~ For the arguments against Eneolithic rId­

ing, see Sherratt 1997a:217; Drews 2004:42-50; Renfrew 2002; and E~ Kuzmina 2003:213~

39. 1nl' remains of a bow found in Berezovka kurgan 3, grave 2, on the Volga, in a grave of Pokrovka type probably dated about 1900-1750 BCE, had bone pbtes reinforcing the shaft and

hone tips at the ends-a composite bow. 1be surviving pieces sug-gest a length of 1.4-1.5 m, almost five feet from tip to tip. See Shishlina 19YO; and Malov 2002~ For an overview of early

archery and bows, see Zutterman 2003.

40~ I am indebted to Dc. Muscarella for some of these ideas about arrow points. For a dis­

cussion of the initial appearance and usage of socketed bronze arrowheads, see Derin and Mus­

carella 2001. For a catalogue and discussion of the early Iron Age socketed arrowheads of the

Aral Sea region, see I tina and Yablonskii 1997. Socketed bronze spear points were made in the

steppes as early as 2000 BCE, and smaller socketed points hegan to appear occasionally in

steppe sites about the middle of the Late Bronze Age, around 1500 BCE, but their potential

was not immediately exploited. The ideal bows, arrows, and arrowheads for mounted archery

\ evo~ved slowly,

L..-:1. For tribal warfare, see Keeley 1996.

CHAPTER 11, THE END OF OLD EUROPE AND THE RISE OF THE STEPPE -~--~

L ~e gold at Varna, see Bailey 2000:203-2~ntaineand Jordanov 1988; and Eleure 1989,

2. Chapman 1 3~ For off-tell settlemen

1998~

4. The decreas

Perry and H 00; and Bond et a1. 2001. For the Piora tion in the Swiss Alps, see Zoller

1977 r indicators of cooling in about 4000 BCE in the Creenlan cores, see O'Brien et a1.

1 95. For climate change in Central Europe in the German oak tree rings, s

2002. For the Pontic steppes, see Kremenetski, Chichagova, and Sbishlina 1999~

5~ For the flooding and agricultural shifts, see Bailey et a1. 2002~ For overgrazing and so' erosion, see Dennell and Webley 1975~

6. Fur Ji!aV'l, see Comsa 1Y76. 7 TIle pollen (hange~ ;Hl: dL:~nibl'd ill .f\1arinova 2UU3

8. Cast UJppU ohjects bcgall l(1 ~Lppe~H !"cgld<lrly in western Hungary witll the Lasinja- Balatoll ,~ulturt at ah"uI 4000 BCE; sec B:iIlHy 19Y5; ab" Parzinger lYn~ ~

9. T[)d(}rova J<)95:9U; Chernykh lSi92:S2. 'HIe burning ofhollSLS mIght have ~)L'CIl an inten­ tiuridl ritual acl during the Enco!ithic; see Stn'<lnovil. 1997. But the final hrcs.that cOIlc;umt.:d tilt EncoJirhic towns of thc' lower Danuhe valley and tbe Balkans about 400p BeE wert fol­ lowed by rcgiun-widc ;lb:.llldonment ;llld abrupt culture change. Regjon-widt/abandonments of Luge settlements in the North American S"llthwest (1lO0-1400 CE) and il/Late ClaSSIC Maya

sItes (7UO-C)OO CE) in .f\1esoamerica were a%ociated with inkn<.,c warfa!~; set Camerun <'1I~d Tomb 1\1\13111" kmd of climate ,hift that struck the lower D'llluhe \'~Ile) ahout 41l1()~3800 BCE would not have made tell ~ettll'fncl1ts llninh<lbitahle. \Varfare yfH:rcf()n~ SCL'J11S a likely explanation. /.

10. For evidence (If ovcrgranng and soil l'fU:-..iOIl at the cnd of trt Karanovo VI period, see Dennell and \i\'ehky '1~75; for the dt')truction of Eneo\ithic YunaJ~ite, see .f\1erpert lSi95; and Nikolo"a 2()00~ , /

11 Todoro\'a 1~95. .... /

12, Sce Ell is 1984 for cc~arnic workshops, ~lllJ Popov lCJ?1 for flint wurkshops. I use the Russian spelling (Tripolye, Ti>masho\'ka) rather than the Ul#ainian (Tripil'ye, Tomashi\'ka),

because JOJ.llY site names such a\., Tripol\'e MC estahlished i,Jthc literature outside Ukraine in . \. / their Russian spelling. \

13~ On the demo graphics, see Dc\~achcv2003; and ~1'son 1 Y7Y. On the fligbt of Bolgrad- Aldcni refui':ces, see Sorokin lY8Y~ \

14. On Tnpolye Bl warfare gcnerall~ See Dergachev 003, 1998b; aud Chapman lY99~ On Drutsy 1, see RVlLdina and Engovatova 1~. For much fthe other information in this section ., C \

I have relied on the review article by Cherny"'h lY82~

15. The Cucuteni C designation refers only to :J. t pe of shell-tempered pottery. The Cueu­ teni chronology ends with Cucuteni El' Cucute:1i ware appeared first in sites dated to the

Cucutcni A/Tripolye Bl period and ultimately aminated ceramic assemblages, See Ellis 1 Y84:40-48.

16~ 111e source of the steppe influence on Cue teni pottery is usually identified as the early Sredni Stag culture, phase Ib, for Telegin; or th Skelya ulture, for Rassamakin~

17~ Shell-temper adds to the durability a impact r istance of vessels that are regularly

submitted to thermal shock through reheatin , and also inc ases the cooling effect of evapora­

tion, making a shell-tempered pot good for ooking or storin cool drinking V'later. Cucuteni C

ware and fine painted wares were found to ether both in pit-h ses and large two~storied sur­

[Jce houses. Contextual differences in t e distribution of Cueu ni C ware and fine ware in settlements have not been described. A some sites the appearance of Cucuteni C wares seems

abrupt: Polivanov Yar had traditional rag-tempered coarse wares i the Tripolye B2 occupa­

tion but switched to shell-tempered wares of different shapes and esigns in Tripolye Cl,

whereas the fine painted wares show d clear con:inuity between the tw hases. See Bronitsky

and Hamer 1986; Gimblltas 1977; nd Marinescu-Bilcu 1981.

18~ For the horse-bead maces ee Telegin et a1. 2001; Dergachev 1999; heorgiu 1994; and Covedariea and Kaiser 1996~ \

19. For the skull shapes, se Necraso\' 1985; and Marcsik 1971. Gracile 'lylediterranean"

Tripolye skulls have been foe d in ritual foundation deposits at Traian (Tripolye' B2).

20~ For Mirnoe, see Bu 0 and Stanko 1981.

21. For the eastern mi ration, see Kruts and Rizhkov 1985~

22~ 111e Iron Age ste eotype of nomadic cavalry seems to lie behind some of the writings of Merpert (1974, 1980) d Gimbutas (1977), who were enormously influential.

! 23. The "awkwar seat" hypothesis is based on Near Eastern images that show riders sitting

awkwardly on the rse's rump, a seat more suited to donkey riding. Donkeys have low withers

1

i• i

J