CHAPTER EIGHT

A New Beginning

Cognitive Numismatics I

In 1941, three pioneers perished at the height of their professions. Edward T. Newell was at the time one of the world’s greatest numismatists; Sir Arthur Evans was its most celebrated archaeologist, and Professor James Westfall Thompson was the prolific sitting president of the American Historical Association. What is most striking about their passing is the subsequent history of their respective fields over precisely the same period. Today, numismatics remains very much Newellian in its methods and interests. The ghost of Newell could feel quite at home reading a recent journal or book in his area of expertise, written likely as not by a knowledgeable collector having no formal academic training in the discipline—indeed, few still are the degrees actually granted in numismatics. The same cannot be said of Evans, whose brand of archaeology has long ceased to be practiced. In fact, the methods if not the interests of Evans would be something of a scandal in any excavation today, and the knighted hero could scarcely understand the terminologies and techniques of current archaeological research. Thompson’s brand of history has fallen from favor among those less enthralled by his own regard for kings, popes, saints, artists, and intellectuals. In the AHA Presidential Address that he never lived to deliver, Thompson would have said something anathema to the “new social historians” of the next generation: “the prime movers of human affairs…are Law and Government, Religion, Literature, and Art.”¹ This capitalized hierarchy leaves little room for the lowercase histories of the lower-class peoples now in vogue. Clearly, the past seventy years have seen the rise of a New Archaeology and a New History—but not yet a New Numismatics.

This is not to say that progress has stalled in the study of coinages; previous chapters have amply demonstrated the fruits of many labors. But in spite of hard-won advances, there is still a lot more that coins can tell us about places like Bactria. The key is for numismatics to become relevant to the kinds of questions being asked today. The new archaeologies and histories have pushed aside traditional preoccupations with ruling elites, pretty objects, descriptive catalogues, and eventful narratives. Instead, the focus has shifted to the study of ordinary people, everyday objects, explanatory models, systems analysis, and scientific hypothesis testing.² Hence we find a proliferation of subfields ranging from women’s history, ethnohistory, and cliometrics to processual, postprocessual, and logicist archaeology. Historians now investigate phenomena such as hybridity, urbanization, and colonialism as seriously as their predecessors tracked the reading habits of Thomas Jefferson; archaeologists study olive pits and pollen samples as thoroughly as their forebears did palaces and pyramids. Meanwhile, most numismatists remain fixated on the same questions that bothered Vaillant, Bayer, Pellerin, and Wolff centuries ago. For this reason, many experts in the history and archaeology of Hellenistic Bactria are finding numismatics of secondary importance in answering the dominant questions of the day.³ What can we know about the subhistoric peoples of Bactria, who lived below the threshold of our written sources? How were identities constructed and negotiated in a world too often reduced to a stark struggle of Greek versus non-Greek? How do we correctly combine different kinds of material evidence in order to understand the dynamics of the region? Why do we stress events rather than processes?

For nearly three hundred years, Bactrian numismatics has concentrated on the names, number, and order of kings; their kinship and marriage ties with each other; their personalities and appearance; their rivalries and their conquests. Archaeological and epigraphic discoveries have done little so far to alter this numismatic obsession with those who ruled rather than those who were their subjects. History and archaeology have become bottom-up in orientation, whereas numismatics remains top-down. We look at the coins and wonder too much about the kings and not enough about their kingdoms, as if this approach simply cannot be helped. As one numismatist has recently put it:⁴

The coins naturally put the focus on the kings who issued them rather than on broader knowledge of these societies, but this is a limitation we have to accept.

That is a false constraint. Perhaps it is time to try a different methodology in our search for the lost civilization of Bactria. Let us turn, therefore, to what might be called cognitive numismatics.

Taking a cue from that branch of the New Archaeology known variously as processual or cognitive, let us forget for a moment the kings of Bactria and concentrate instead on the nameless and faceless people around them.⁵ Who appears on the coins matters less than who made them (this chapter) and how they were used (next chapter), for these themes give us unique access to the thoughts and actions of at least part of the subhistoric population. Coins preserve patterns of behavior across a broad spectrum of society, and we begin with the mint and the mind-set of those working there. Coins, after all, are the products of labor and the patterns of thought that guide that labor. Cognitive numismatics, like its older cousin cognitive archaeology, asks how material objects may illuminate the mental processes that shaped them. How did anonymous ancient craftsmen, both individually and as a group, plan out a complex task or respond to new challenges in the workplace? These are key questions because, as is well known, “the amount of information which survives in our [written] sources on the subject of the minting of coins is meager in the extreme.”⁶

We must first of all treat coins as a process instead of simply as end products. An archaeological chaîne opératoire (“operational sequence”) makes explicit the various steps necessary to create an artifact, from acquisition of raw materials and manufacture to its use and eventual abandonment, or loss, or reuse.⁷ The Bactrians gained access to gold, silver, copper, tin, and cupronickel by means of mining, panning, war, and trade.⁸ The volume of surviving coinages suggests that considerable effort was expended in these endeavors, although it is not yet possible to quantify the workforces and military expeditions marshaled specifically to acquire these metals.⁹ Alluvial gold occurs in many rivers in and around Bactria, and lode deposits have been identified throughout the region. Silver and copper mines, some of them ancient, exist in the Badakhshan and Hindu Kush mountains. One of the world’s largest deposits of copper, worked since the second century B.C.E., lies sixteen miles (25 km) southeast of Kabul at Ainak. Local sources for cupronickel are also known. Tin, necessary to make bronze, was mined very early in Afghanistan.¹⁰

Our knowledge of mining operations elsewhere in the ancient world provides a grim picture of what it may have been like for the slaves, condemned criminals, and desperately poor who were the likely diggers in Bactria.¹¹ In dangerous underground shafts, the rock had to be heated by fires and then cracked with splashes of water. Large chunks of ore had then to be pounded into smaller and smaller pieces by relentless labor. High-temperature cupellation separated the gold and silver from base metals. It should not, of course, be assumed that the requisite metallurgical skills were limited to the Greeks who colonized the region, since metalworking was practiced in Central Asia long before and after the Hellenistic period. Following the large-scale exodus from Ai Khanoum, those who salvaged the site took pains to recover in a systematic way the gold and silver left there. Their ingots have been excavated alongside a balance and weights, and these workers marked one lump in a non-Greek language yet to be deciphered.¹²

Once the requisite raw materials had reached the mint, skilled artisans armed with a cognitive map of the desired mintage could transform the metal into money.¹³ Here is where we see through the eyes of some nonroyal Bactrians and discover their daily habits. Most ancient coins were made by hammering a piece of preweighed metal, called a flan or blank, between two carved dies.¹⁴ The metal required a complicated system of assaying in order to produce a uniform alloy, which was then cast, often en chapelet (using molds), to the appropriate weight standard.¹⁵ The flan would then be cleaned with an acid treatment before being hammered between the dies. One shortcut in this method was to use old coins as blanks, resulting in an overstrike. The frequent inability to obliterate altogether the designs of the host coin (the undertype) provides valuable chronological data. For example, the mints of Heliocles II overstruck the coins of numerous contemporary or earlier rulers, including Eucratides I, Agathocleia, Strato I, Antialcidas, Philoxenus, and Hermaeus.¹⁶

By numismatic convention, the lower die is called the obverse (or heads) and the upper die is named the reverse (tails). Workers had to engrave each of these two dies as a mirror image of the desired coin. Thus, to produce the dies for any set of coins, the die cutter took up his tools and carved everything backwards, including the inscription. This could be quite challenging and time-consuming, with commensurate pressures whenever production rates increased.¹⁷ Especially for tiny coins with minute details, the use of some sort of magnifier may have been necessary, such as a plano-convex lens; otherwise, the craftsmen had to be young, myopic, or both.¹⁸ Even so, eyestrain was certainly an occupational hazard.¹⁹ Needless to say, this method of manufacture meant that no two dies were exactly the same, and therefore coins minted using the same die (or dies) were identifiable as die-linked.

Only a few coin dies actually survive from the ancient world, but one of them happens to be from Bactria and, more precisely, is said to have been recovered at Ai Khanoum.²⁰ (See plates 6–8.) Extant dies are rare not so much because few were ever made; the coins in the Qunduz Hoard alone represent the production of 973 separate dies (425 obverses, 548 reverses).²¹ Surviving dies are uncommon because they had to be closely controlled while in use, and afterward they were generally destroyed or deposited in temples to keep them out of unscrupulous hands. If the extant Bactrian die came from Ai Khanoum, it may have been stored in the palace treasury or, more likely, consigned to one of the city’s temples. It was not being used at the time when the site was abandoned because its outdated purpose was to strike the reverses of tetradrachms issued by Demetrius I.

This die is about the size and shape of a small sewing spool (39 mm high; 37 and 35 mm diameters at the ends: 1.5 × 1.4 × 1.4 inches). It is made of heterogeneous bronze, containing 78.4–84.2 percent copper, 15.3–21.1 percent tin, and a trace of iron.²² It seems likely that its concave shape accommodated tongs or some other clamping device that held the die firmly against the flan as the hammer did its work. The coinage of Demetrius I was struck with an adjusted die axis of 12:00, meaning that the designs on both faces were aligned upright when the finished coin was spun on its vertical axis. To achieve this, the mint workers had to orient the two dies properly for each strike. The small notch on the hammered end of the die may have served this purpose, since it happens to be aligned with the bottom of the reverse composition.²³ Because reverse dies took so much of the force, they tended to wear out more quickly than the obverses, which were anchored securely into an anvil. This surviving example does not seem particularly worn or cracked on its engraved surface, although the hammered end has been badly damaged on the side opposite the notch, and this may explain why the specimen was withdrawn from service. Another possibility is that the die was intentionally broken and retired following the demise of Demetrius. No coins have yet been identified as products of this particular die. While tens of thousands of strikes are possible with a good die, this example may have left little or no trail in the numismatic record of Hellenistic Bactria.

If we could be certain of the die’s provenance, which of course we cannot, it might be possible thus to link a particular monogram to a specific place. While is not especially common on coins found at Ai Khanoum, it is noteworthy that this monogram appears on coinages struck by Demetrius I, Euthydemus II, Agathocles, Pantaleon, Antimachus I, and Eucratides I.²⁴ This means that disappeared at the same time that Ai Khanoum was abandoned, and it is the only major monogram to have done so.²⁵ Nevertheless, a Demetrius die employed elsewhere could have been sent to Ai Khanoum for decommissioning in its treasury or one of its temples. Also, a monogram used at Ai Khanoum need not end with the city’s fall, since a magistrate represented by, say, could have evacuated the area and then continued to use the same monogram at a different locale. Thus, we must stop short of attributing all Bactrian coins bearing to the mint at Ai Khanoum.

We can, however, see plainly the handiwork of the die’s inscriber, no matter where he happened to live and work. It must have taken some skill to fashion a serviceable die out of hardened metal. To give the standing nude figure of Hercules his lifelike pose, the artisan had to sink the hero’s impression into the metal. The deeper the cuts and drills into the die, the higher the relief on the resulting coins—the face of Hercules, for example, and the dots governing the shallower strokes for the monogram and lettering. This intaglio technique was the same as that for engravers of gemstones and signet rings—commissions the die cutters may also have performed.²⁶ The worker responsible for Demetrius’s die did a fine job of rendering Hercules in the act of crowning himself, with his club and lion pelt in hand.²⁷ This artwork accompanies well the inscription that frames it.

As part of his cognitive map, the worker had obviously to be mindful of etching backwards, since letters more than images could betray carelessness or error. Carving the Greek words retrograde could be tricky, especially for letters such as B, Σ, A, and P that had to be reversed. Fortunately for the engraver, twelve of the seventeen letters inscribed on this die were symmetrical on a vertical axis and therefore were engraved the same way whether forward or backward: and . The propensity for error was therefore relatively low on this task, as compared with carving backward something more demanding, such as . The difference is analogous to writing the common names TOM and STAN retrograde: the S and N require greater care to invert properly and are more likely to cause trouble.

Obviously, numismatists can study the engraving of thousands of dies without actually seeing each one, since the coins struck from the dies faithfully preserve their features. Even the wear and tear on dies, such as distinctive die-breaks, can be tracked. We have only to visualize the missing dies as mirror images of the extant coins, either by reversing photographs of them or by making impressions in clay. In this fashion, it is possible to recreate the cognitive maps of the Bactrian workers as they planned and executed their various tasks. For example, the chaîne opératoire for carving a die usually began with the image in the central field, followed afterward by any necessary text. This procedure is evident on many Bactrian coins where the lettering has been squeezed into remaining spaces after the type has been finished. Similarly, some letters have been cramped or displaced to accommodate the central image. (E.g., the lettering beneath the jugate busts in fig. 20.)²⁸ This is evident, too, on those issues where the legend has intentionally been interrupted by the artwork.²⁹

The inscriptions on Bactrian coins reveal a lot about the men chiseling them. When tasked to engrave the inscription, one option would have been for workers to ignore the finished product as a mental template and to proceed instead by copying an already reversed text. Doing so would not require as much thought about the actual words; mere duplication would not even oblige the engraver to understand what he was writing.³⁰ But the artisans indeed did know. It can be demonstrated that Bactrian die cutters were thinking in Greek as they toiled, because whenever words in the Greek coin inscriptions fall out of alignment, they generally do so at the end rather than the beginning.³¹ This is the tendency of engravers who were thinking out each word in Greek, engraving the text retrograde from its beginning as though spelling it out from right to left (producing a coin to be read left to right). On tetradrachms of Eucratides with the arched legend engravers sometimes used the long spears of the Dioscuri to delineate the beginning and end of these words. Often the text on the coin commences with evenly spaced lettering (…) but ends with a jumble of mashed letters at the end (…), because room has run out at the spear points.³² This happens if the artisan knows (or believes he knows) the language and starts each word at its real (not reversed) beginning. Simply put, our unseen workers were spelling Greek, not just replicating a string of meaningless symbols set before them. As an example, persons unfamiliar with the Cyrillic alphabet might copy backwards a Russian text by working as readily from the end of the text as its beginning. But given a text in familiar English, French, or German, the same person is more likely to work retrograde while starting at the text’s beginning.

Figure 20. Eucratides commemorative tetradrachm, showing his parents, Heliocles and Laodice. Note the displaced lettering beneath the jugate busts, indicating that the images were carved prior to the inscription.

Another proof that die cutters did not simply follow an already reversed text can be found on a coin of the Indo-Greek king Lysias.³³ Here the name of the king appears backwards on the coin, meaning that it was accidentally engraved forward as on the die. The workman thought and inscribed the name without remembering to invert it on the die. Some early Seleucid issues from Bactria show similar lapses.³⁴ It is significant, of course, not only that these errors were made in the first place but also that coins struck from faulty dies were not withheld from use. This compounds the nature of all stumbles, because botched coins in circulation represent several missteps within the mint.

Errors, in fact, can provide our best evidence for what was (or was not) on the minds of Bactria’s workers. On another bronze coin struck in the name of Lysias, for example, someone did catch an error and then, rather halfheartedly, tried to fix it.³⁵ The die cutter absentmindedly omitted the letters IK in the middle of the epithet . His solution was simply to insert these missing letters incongruously at the bottom of the die, just ahead of the king’s name. No supervisor apparently noticed or cared about this slipshod behavior. Such coin legends remain as material expressions of ancient states of mind. One is reminded of the Roman whose thoughts must have wandered as he engraved a child’s tombstone. The distracted workman looked up his model and etched what he saw: HIC IACET CORPVS PVERI NOMINANDI (“Here lies the body of the child name to be supplied”).³⁶

Cognitive numismatics tries, literally, to connect the dots in order to understand planning as well as execution in the engraving process. We can discover on many coins the telltale signs of preparatory draftsmanship, usually in the form of small guide dots sunk into the die. These could be drilled to help form individual letters and the monograms (as on the Demetrius die), or to align the inscription as a whole, or both. The guide holes for individual letters probably served to keep the text properly sized, since rounded letters engraved without the use of dots often appear too short: or .³⁷ While these dots are a common feature on Bactrian and other Hellenistic coins, the oversized monsters of Eucratides and Amyntas reveal them most clearly. They allow us to see that the mind of the man who created dies for the latter coins was fastidious and focused; the former was indecisive, careless, and yet also perhaps brilliant.³⁸

The die cutter of the Amyntas double decadrachms had firmly in mind a finished product that exhibited a semicircular arrangement of letters for the royal title and epithet.³⁹ To achieve this end, he first devised a tool to use as a master template for his dies, probably in the form of a simple multipointed punch that transferred to each new reverse die a set of two concentric circles of guide dots. This arrangement of dots appears on all the reverse dies for this series, and the dots extend into the exergue (lower quadrant) where they are not needed, since the king’s name was engraved there horizontally.⁴⁰ This means that there are guide dots even where none would be necessary, showing that indeed the master template was applied as step one in the engraving process (fig. 21). Next, within the inner circle of dots, the engraver neatly framed the coin type. Finally, he cut (backwards) the upper inscription within the arc of concentric circles, and the lower inscription straight across at the bottom. The strokes forming individual letters (except, of course, for round omicrons) were guided by an independent set of deeper dots. It is obvious here as on other coins that the spacing of the text accommodates the already engraved type (at the tip of Zeus’s scepter, for example) and not the other way around.

Figure 21. Reverse of a double decadrachm of Amyntas, shown as a mirror image in order to replicate the appearance of the die (dot pattern emphasized).

As we follow the process of the artisan who some years earlier made the dies for Eucratides’ famous twenty-stater coin, we observe a very different pattern of thought and behavior.⁴¹ This anonymous worker seems not to have begun his commission with an appropriate cognitive map, since he obviously had to erase part of his work and start again. Clear traces of this debacle can still be observed on the reverse. Unlike the engraver of Amyntas’s dies a half-century later, this die cutter did not use a visible template for the overall design. Nevertheless, the centering of the type is quite expertly done. The only oversight on this part of the die is that the engraver failed to continue the shaft of the background sarissa at the horse’s rump.⁴² When the worker finished the Dioscuri design, he began cutting the monogram n. Extraneous guide dots here suggest a slight repositioning of the symbol. The upper inscription was at first aligned retrograde straight across the top of the die.⁴³ The engraver did not anticipate that this text would not fit properly. When the problem finally became apparent, he chiseled out his first attempt and substituted a less crowded semicircular alignment that followed a precise arc for the letters B through M, after which the last six letters strayed outward. The king’s name was also cut slightly off center at the bottom of the die.

Obviously, this artisan did not bother to remove all traces of his fumbling, and he had no intention of starting fresh with a new die after already carving the Dioscuri on this one. His lack of foresight at one end of the chaîne opératoire, and his indifference at the other, give us pause. This was no innocuous local bronze issue, but rather the most ostentatious coin ever produced in antiquity. It must have been a noteworthy assignment to make this die, and it seems unlikely that sloppy work would go unnoticed—or even unpunished. What, then, was going on inside his workshop the day this man made this die? Was the engraver careless, carefree, overworked, or preoccupied? Was he an ancient free spirit given to trial and error even on the weightiest of royal commissions, or was he just in over his head?

Some scholars have theorized that this engraver was the very first to make a Eucratides die after that king ordered the title added to his coinage; therefore, the artisan did not anticipate the challenges of inserting these seven extra letters.⁴⁴ Even so, this means that the man gave no prior thought to this special task. He worked with an old cognitive map unhindered by grabbing a sketch pad or measuring tools. Now fresh numismatic evidence shows that this engraver was not the first or only worker to deal with the expanded text (see plate 9); tetradrachms with the longer inscription engraved horizontally have now come to light, albeit with different monograms ( and ).⁴⁵ Had the engraver of the twenty-stater die not seen already the crowding of letters on these circulating varieties? Did a better idea not occur to him until after he had carved himself into a corner? The innovation of the semicircular legend was a stroke of genius, but a belated one that tells us that some workers in Eucratides’ day did not plan ahead.

Another interesting example of a reworked die can be found in the Qunduz Hoard. Coins 47 and 48 with monogram , minted by Demetrius II, share obverse and reverse dies. These tetradrachms show a clear anomaly on the reverse, where the ends of Athena’s spear appear twice. Raoul Curiel and Gérard Fussman thought this to be the result of deliberate restriking after flipping the reverse die to correct the coins’ die axis from 6:00 to 12:00.⁴⁶ Colin Kraay suggested a simpler explanation.⁴⁷ Noting the difficulties surrounding the repositioning of the dies so precisely, Kraay argued that the extra spear points must have existed on the die itself. So, either an old die was partly ground down and reengraved, or some lapse occurred while cutting a die and the worker rotated it, erased most of it, and started fresh; one way or the other, the original spear points lingered as telltale signs of the process. This episode indicates a certain degree of stress in the mint, either from an initial engraving or striking error, or from the need to recycle quickly an old die. More will be said about the die-linked obverses of these same coins below.

Cognitive numismatics therefore tasks us to consider general working conditions as well as an individual worker’s competence. What was the rate of error on inscribed dies in Bactria as a whole, and what circumstances might explain those errors?⁴⁸ At times, engravers in Bactria so blundered their craft that they actually misspelled their king’s name, the sort of behind-the-scenes mistake that could hardly be made more public. From the area of Bactra, for example, we find an obol of Eucratides I with two letters missing from the king’s name: .⁴⁹ Is this mere carelessness, or the inability to recognize and rectify carelessness? Both possibilities seem significant for a basic understanding of life in Hellenistic Bactria. Periods of carelessness on a large scale betoken poor training, lax supervision, or stressful overwork, which may be signs of larger troubles beyond the mints. The inability to identify and correct mistakes raises the question of linguistic background and facility, which may be clues about the ethnicity of the workforce.⁵⁰After all, some coinages (of Heliocles II, e.g.) exhibit clear linguistic variations in the rendering of the king’s name in Kharoshthi (Heliyakreya, Heliyakresa, Heliyakrea). Furthermore, these variations seem to be regional, since they are specific to particular monograms.⁵¹ This may indicate local recruitment of die cutters. The practice noted earlier of engraving the text as though spelling out each word does not support the idea of completely Greekless die cutters who merely copied meaningless symbols, so occurrences of error in the Greek are indeed consequential. Do these deviations from expected norms mean that Hellenism occasionally weakened or broke down in Bactria?

The question of Bactria’s cultural matrix used to be argued on aesthetic grounds—for example, the artwork on these coins looks less Greek than on those; this royal portrait appears Sogdian, but that one is purely Hellenic; one side of the Eucratides gold giant (the reverse with the erasure) is the work of a talented Greek, but not the other. These historical judgments remain woefully subjective and qualitative. The notion that in Bactria good art is Greek and bad art is native invites bias and bigotry. The rightness or wrongness of a seemingly Greek sculpture simply cannot be measured with the same precision as that of a Greek text. Thus, unlike perceived lapses in art, errors in language can be quantified and lend themselves to objective analysis.

Using a controlled sample of dies that eliminates the selective bias of marketed, collectible coins or forged intrusions, reign-by-reign error rates may be ascertained for the Bactrian kingdom (table 1).⁵² The data encompass all mistakes, from simple one-stroke errors (e.g., failure to add the crossbar to letter A) to multiple and more complex blunders (e.g., missing or intrusive letters).

Overall, the Greek vocabulary engraved on Bactrian coins remained quite small and repetitive during this period (just two dozen words and names all together), yet the rate of error increased meaningfully during the reign of Euthydemus I and again before and after the fall of Ai Khanoum, which seems to have occurred during or soon after the reign of Eucratides I.⁵³ All the errors on the dies of Euthydemus I and Antimachus, however, are of the simple variety, whereas 55 percent of the flawed dies for Eucratides I and II have complex errors.⁵⁴ For Demetrius II, the percentage of error dies with complex problems rises to 75 percent, and for Heliocles it is 62 percent. This suggests that the mints of Bactria experienced both a rise in the rate, and in the magnitude, of error shortly before and after the abandonment of Ai Khanoum. These data are statistically relevant for the deterioration of work in the mints, for the probability of an increase in error this large occurring merely by chance is much less than 1 percent.⁵⁵

Table 2 uses as a control word in our sample the genitive-case form of the ubiquitous Greek title found on nearly every coin of every king. The engraving errors can be one-stroke oversights, or (especially in the reign of Heliocles) serious spelling lapses. The increase in both the rate and the severity of mistakes can readily be appreciated in these blundered dies from the workshops of Euthydemus I, Antimachus I, Eucratides I–II, Demetrius II, and Heliocles. What was going on?

TABLE 1

Incidence of Error on Bactrian Coin Dies, 250–130 B.C.E.

One might immediately assume that die cutters less familiar with Greek were responsible for these changes. This would be evidence of a culture mixte and a rapid influx of indigenous artisans employed at the mints.⁵⁶ But even so, it does not necessarily follow that good Greek could be written only by good Greeks. Sophytos at Kandahar, Atrosokes at Takht-i Sangin, Hyspasines at Delos, and Oxybazus, Oxyboakes, and others at Ai Khanoum were men with non-Greek names whose facility with the Hellenic language has been attested. (See chap. 6, above.) The latter who served in the treasury, of course, were supervised by personnel with Greek names. If we blame the rise of error on the linguistic incompetence of an increased non-Greek workforce, then there remains the problem of why their presumably Greek overseers did not catch and correct the lapses. Also, the notion that most ancient Greeks read everything aloud raises the question of whether their cousins in Bactria did so even with their currency. If so, this surely would have tested the mass of coinage through the mouthing of the inscriptions on them by Greeks all across the kingdom. Were they, too, less skilled in the language of their ancestral homeland?

 

TABLE 2

Incidence of Specific Engraving Errors in the Word , 250–130 B.C.E.

Control Word Error	King	Number
	Euthydemus I	1
	Demetrius II	1
	Heliocles	10
	Antimachus	2
	Eucratides I	1
	Heliocles	4
	Eucratides I	2
	Demetrius II	2
	Heliocles	7
	Heliocles	1
	Heliocles	5
	Heliocles	1
	Heliocles	1
	Heliocles	1
	Demetrius II	1
	Heliocles	1
	Heliocles	1

 

The chief excavator of Ai Khanoum does not think so. According to Paul Bernard, the Greeks in that city assiduously preserved their language as the veritable cement of their national identity: “They continued to speak and write Greek in an uncontaminated form right to the end, as evidenced by the inscriptions found during the excavations.”⁵⁷ Setting aside the lack of evidence for spoken language, it is true that at Ai Khanoum and elsewhere in Central Asia there is growing epigraphic documentation of the high standard of written, even self-consciously erudite, Greek. Clearchus set up the Delphic maxims with a verse preamble. Heliodotus honored his favorite kings. The non-Greek Sophytos boasted of his Hellenic education and put it on public display. But this is evidence from a certain stratum of the ancient population, one that had the means and inclination to set up expensive monuments at Ai Khanoum, Kuliab, and Kandahar. What about everyone else? The linguistic landscape of Central and South Asia was shaped by numerous factors, including intermarriage and bilingualism. Although we lack the same level of documentary evidence as found elsewhere in the Hellenistic world, particularly Egypt, it is significant that the treaties between the Seleucids and India included provisions for intermarriage.⁵⁸ The children of such unions might learn a mixture of languages, as had the Branchidae of Sogdia who became bilingual (“bilingues”) and gradually spoke a degenerated form of Greek (“paulatim a domestico externo sermone degeneres”).⁵⁹ The significance of local languages used alongside Greek has now been demonstrated by bilingual inscriptions on everything from coins and bowls to royal edicts.

Were die cutters and their supervisors as competent in Greek as were the ruling Hellenes and wealthier native peoples of the region? The coin inscriptions provide a negative answer, even though they actually appear on official state documents. Some people somewhere in Bactria’s cities, thinking in Greek, botched the Greek. Their handiwork does not exactly square with Homayun Sidky’s characterization of places like Ai Khanoum as “a purely Greek city” where “there can be no doubt that the inhabitants…were Greeks who retained their full ethnolinguistic identity” until the site was suddenly overrun and “all traces of Greek civilization [were] swept away.”⁶⁰ Sir William Tarn, who did not know of Ai Khanoum or the inscriptions, nevertheless reached the same conclusion based on his aesthetic judgments about the coins. He wrote of the nomad conquest of Bactria in the time of Heliocles:⁶¹

One section [of Greeks]—a very small one—is known to have been exterminated: the Greek coinage of Bactria remained fine to the end, and then the great Bactrian artists vanished from the world; no trace of their peculiar skill in portraiture ever occurs again, in India or anywhere else. The likeliest supposition is that one battle sufficed to destroy most of the Greek aristocracy.

Tarn’s old theory about the artist X and his school of gifted Greek die cutters (see chap. 4, above) draws a dramatic line: the end of X marks the spot where Hellenism succumbed to a holocaust. Those adhering to the narrative tradition in Bactrian studies tend to envision a sudden, stark, violent boundary between a Greek Bactria and then a barbarian Bactria. The change, whether explained archaeologically or numismatically, allegedly came about as a single tragic event. Cognitive numismatics, however, favors a different approach that uses quantitative coin analysis and focuses on processes. This allows us to understand that the Greek coinages and the texts on them were not fine to the end. A period of increasingly corrupted engraving, either tolerated or unnoticed, was under way in Bactria’s mints before the fall of Ai Khanoum and the eventual collapse of Hellenistic Bactria north of the Hindu Kush after the demise of Heliocles.

This may mean, of course, that the mints were being staffed in this period by workers whose Greek-language skills were marginal—either ethnic Hellenes who were losing their language or Central Asians not altogether fluent in Greek. But this trend must be examined in conjunction with a larger question: Why were their failings tolerated by those whose language skills allegedly remained high? The answer to this quandary may stem from production rather than personnel. A change in the ethnicity or education of the workers may be just part of a larger transformation in basic working conditions brought on by overtaxing demands. The first period of elevated incidence of error occurred during the reign of Euthydemus I, when simple one-stroke mistakes show up on the dies. This happened in the context of the disruptive Seleucid invasion of Bactria by Antiochus the Great. The capital, Bactra, in fact, endured a prolonged siege during which its dependency Sogdia seems to have gained its independence. These circumstances may have adversely impacted workers’ abilities or inclination to produce pristine dies at a time when coinage had to be manufactured quickly under wartime conditions. In other words, times of anxiety and urgency may be manifested in the mental lapses of die cutters and the inattentiveness of officials to those lapses. The escape of an obvious mint error into circulation from one of Euthydemus’s main mints exemplifies this problem.⁶² While striking a rare gold distater with monogram , a flan was set on the anvil die without first removing the previous strike. The hammer then drove down the reverse die, so that the fresh coin received the design of the punch and the same design, only incuse, from the last strike. This so-called brockage of a high-value coin apparently caught no one’s attention at the time.

The next, and far more serious, wave of error commenced before the abandonment of Ai Khanoum and grew worse there after. This runs counter to the notion that the city fell without any prior signs of trouble in Bactria.⁶³ Quantifiable mistakes on the coins betoken a buildup of significant strains throughout the Bactrian workforce, for the problem was widespread and not isolated to one or two mints or magistrates. The incidence and tolerance of error reflect a period of increased production demands, lax supervision, less training, and perhaps major internal distractions. Procedural as well as personnel changes in the mints may have abbreviated or eliminated a quality control step that allowed the manufacture of a less consistent product. One sign of such a change is not only the use of dies that were improperly engraved but also the continued employment of dies that had become badly worn or even cracked. The run of ten coins in the Qunduz Hoard that were among those struck from a progressively more defective obverse die suggests that quantity trumped quality.⁶⁴ Significantly, one reverse die used in this sequence had an obvious misspelling of the king’s name that slipped past the minters.⁶⁵ Another pair of coins in this run shared the reengraved reverse die discussed above, which itself indicated stress in the mint.⁶⁶ A further signal of new production obligations is the rapid increase in the number of monograms being used, whether they signify more mints or more magistrates, from about fifteen for Euthydemus I to over fifty for Eucratides I.⁶⁷

Chronologically, the coinage manufacturing system became strained under circumstances leading up to the nomadic incursions. Clearly, everything in the region was not fine right to the day (or, in Tarn’s scenario, the battle) that doomed the Greek rulers and their artisans. In fact, those invasions may have been the consequence rather than the cause of Bactria’s collapse. Whatever the immediate contingencies (civil war, plague, earthquake, environmental poisoning, or some other series of disasters) that prompted the Greeks to evacuate the city at Ai Khanoum, conditions may have already led the workers in the mints and their supervisors to abandon the highest standards of their craft. The nomadic invaders may simply have moved in to take advantage of some worsening situation, much as they had done in Sogdia when Greek fought Greek during the siege of Bactra. Conditions deteriorated further after the assassination of Eucratides as the stresses on the resources and peoples of Heliocles’ realm pushed the rate and severity of mint errors higher. The nomads have become a convenient scapegoat for the entire decline and fall of Hellenistic Bactria, but on this frontier—as later on the Roman—internal crises and systems collapse may be a better explanation.⁶⁸

What seems certain now is that maintaining a pure and uncontaminated form of Hellenism was not the paramount aim of the Bactrian population in those times of crisis. Care was still apparently taken in all other stages of the chaîne opératoire, ensuring that there was an adequate—indeed increased—supply of bullion properly assayed. Yet, flawed dies pounding out tens of thousands of coins each were tolerated. Was there no time to check one in five dies, or one in every few thousand coins? If not, the situation must have been dire. Perhaps, too, many of the coins were meant for those who had different expectations from the Greeks, such as foreign mercenaries. The mass-produced coinages, as opposed to a few monuments and inscriptions, prove that Bactrian society was in a state of flux and that normative behaviors were less guarded. That, at least, seems to be the lesson of cognitive numismatics as it examines the front end of the chaîne opératoire. Next, we turn to behaviors reflected beyond those mints.