The Payagyi urns and their dates

In 1911, excavations carried out near the Payagyi stupa at Sri Ksetra brought to light an assemblage of four burial urns, each one bearing a brief inscription in the Pyu language on its rim (Figures 1–4).Footnote ¹ By then, knowledge of the Pyu language was very limited, with the results of C. O. Blagden's pioneering study of the Pyu face of the quadrilingual Rajakumar (or Myazedi) inscription from Bagan just having come out.Footnote ² The director of the newly established Archaeological Survey of Burma, Charles Duroiselle, therefore turned to Blagden again for the decipherment and translation of the urn inscriptions.Footnote ³ Describing in great detail how he tried to find a meaning in what he read, Blagden proposed the following translations,Footnote ⁴ which have subsequently gained general acceptance:

Figure 1. Suryavikrama's relative urn inscription. Source: U Pe Maung Tin and G. H. Luce (comp.), Inscriptions of Burma (London, 1959), vol. 4, plates 354 a–d (Rangoon, 1956).

Figure 2. Suryavikrama urn inscription. Source: U Pe Maung Tin and G. H. Luce (comp.), Inscriptions of Burma (London, 1959), vol. 4, plates 354 a–d (Rangoon, 1956).

Figure 3. Harivikrama urn inscription. Source: U Pe Maung Tin and G. H. Luce (comp.), Inscriptions of Burma (London, 1959), vol. 4, plates 354 a–d (Rangoon, 1956).

Figure 4. Sihavikrama urn inscription. Source: U Pe Maung Tin and G. H. Luce (comp.), Inscriptions of Burma (London, 1959), vol. 4, plates 354 a–d (Rangoon, 1956).

This supposed dynasty of three successive rulers of a dynasty ending with –vikrama resulting from Blagden's readings remained in place until 1993, when another urn was discovered near the Payahtaung stupa at Hmawza (Sri Ksetra). The lengthy inscription consists of 1,127 characters written in five lines around the body of the urn.Footnote ⁵ In its text, which has been deciphered but not fully translated, ten names occur, of which six are introduced with ‘Sri’ and end with ‘ba’—the two suffixes that usually denote royal names. In the order of their appearance, these are:

• Sri Devamitra

• Sri Harivikrama

• Sri Si(ṁ)havikrama

• Sri Suriyavikrama

• Sri Bri(ṁ)thuvikrama

• Sri Adityavikrama

First of all, it should be noted that a third piece of evidence—the dated headless Buddha, which mentions that Harivikrama was the younger brother of Jayacandravarman—is of no relevance for this question, as it is the only reference for the name Jayacandravarman.Footnote ⁶ This leaves three names to be considered, which agree with the names found on the four urns mentioned: Harivikrama, Simhavikrama, and Suryavikrama. Of them, the succession from Harivikrama to Simhavikrama poses no problem; however, the listing of Suryavikrama after, instead of before, the two names does.

In general, we would assume that epigraphy ‘developed’ from relatively short annotations to longer pieces of greater literary quality and more precise information. The sequence from Suryavikrama's relative or wife (stating death year only) and Suryavikrama (died in year 50 at the age of 64) to Harivikrama and Siṁhavikrama (full information on year of death and age) and finally to Adityavikrama's long inscription would fit such a development. However, the sequence of names in the Payahtaung urn seems to make this development impossible, as it put the short inscriptions of Suryavikrama after those of Hari- and Siṁhavikrama. On the other hand, any rearrangement of the dates on the four urns does not seem to work either, especially in the way that U San Win has done. He revised Blagden's readings by swapping the meaning of the dates assigned to each name, so his reading rendered a new order of kings that was seemingly consistent with the dates of the Payagyi urns:Footnote ⁷

At first glance, U San Win's revision seems to solve the problem of the conflicting serial order of rulers found on the urns but, at the same time, his reading poses new questions, which are connected to the nature of the records and the way the urns are related to each other. Firstly, the Payahtaung urn appears to suggest a dynastic succession of kings—either fathers and sons or, as occasionally found in Myanmar history, fathers, younger brothers, and their children.Footnote ⁸ However, the dates proposed by U San Win do not easily allow a sensible relationship to be established for the three rulers mentioned in both records. Harivikrama, the first person in U San Win's list, died in the year 41, aged 57. This would set his year of birth at (41–57) –16.Footnote ⁹ The second in the list, Sihavikrama, died only three years later, but at the age of 80. He would then have been born in the year (44–80) –36, which is 20 years before his alleged predecessor. Finally, Suryavikrama, being 50 years old in the year 64, would have been born in the year 14.

According to this calculation, Sihavikrama was not only the oldest (or first-born) of the three, but also so much senior to the next in line that he should have been the uncle rather than the elder brother of Harivikrama. In either case, he should have preceded instead of succeeded Harivikrama in the serial order of the Payahtaung urn. The short reign of three years would confirm that he indeed became king at an old age (he would have been 77 at the time of succession). The third in line, Suryavikrama, could then have been the son of Harivikrama, who would have been 30 years of age at the time his son was born. This seems an acceptable relationship.

U Tun Aung Chain has addressed this question in a different manner.Footnote ¹⁰ Principally accepting U San Win's revised order of dates,Footnote ¹¹ he focussed on identifying and interpreting the relationship term ṭlo, which he understood as ‘brother-in-law’. This interpretation resulted in a succession going sideways, from Devamitra to the brother-in-law Harivikrama and then on to (brother) Harivikrama and his brother-in-law Simhavikrama, before finally the filial generation would succeed in the person of Suryavikrama. Though such an order of succession cannot be ruled out on arithmetic grounds, it seems highly unlikely that two successive rulers could have come to the throne as outsiders in relation to the family in power—all the more so as the suffix -vikrama in all of their names seems to indicate shared ancestry.

Before concluding this section, two more aspects of the urn inscriptions need to be considered. The first concerns the nature of the sources that we are examining, which are inscriptions incised on burial urns in this case. The immediate implication of this is that they would have served to record the time of death of the persons whose ashes the urns contain and, moreover, we would expect the dates to become more comprehensive and detailed the longer such a burial cult prevails.

In other words, an urn inscribed with one date only should be older than an urn with two dates, and an urn stating the year only should be older than one on which we find the year, month, and day. The second aspect refers to the keeping of personal records such as the date of a person's birth or the accession of a ruler. This would require a much more sophisticated ‘chancellery office’ in operation than the mere computation of the years by way of a linear calendar, which the Pyus presumably did. And, again, dates with few details can be expected to be older than those that are more comprehensive and, if only years are stated, then they should refer to the linear calendar rather than to individuals. U San Win's list, which has Suryavikrama (only one set of dates) succeeding Siha- and Harivikrama (two full sets of dates) and Suryavikrama's queen (year only) at the very end of a sequence of rulers (irrespective of how they are personally related to each other), violates this logic at two points.

However, with the lack of other evidence on which to decide this question, there is little we can do but to ignore the Hpayahtaung urn for the time being and focus on the four urns with single inscriptions alone, as they are the ones that contain the dates that are the subject of this article. This is what will be done in the following section. However, before considering the suggestions made by scholars to identify the eras used by the Pyus, it is necessary to remind ourselves that the readings of the dates themselves are by no means certain. This section therefore builds on the assumption that the Pyus indeed computed time by way of a continuous calendar, which commences at a given point in time. This in itself distinguishes them from many of their contemporary neighbours, with whom they were in contact but who much more frequently gave dating in regnal years than in a current era. We will have to come back to this at the end of the article.

Re-examining the eras

Blagden, who was the first to decipher and translate the Payagyi urn inscriptions, identified the era in question as the Myanmar/Cula–Sakaraj era of 638 CE, though he added that he made this assumption ‘for the sake of convenience’.Footnote ¹² Despite this cautioning, his assumption was as widely accepted by later scholars as his translations were, by Harvey, Ray, Luce, U Aung Thaw, and Aung-Thwin, for example.Footnote ¹³ The only challenge to this interpretation was raised by the head of the Archaeology Department at the time, Taw Sein Ko, who suggested that the choice of eras was between the Vikrama era of 58 BCE and the Saka era of 78 CE. He preferred the latter, but when he communicated the suggestion to Blagden, he rejected it on palaeographic grounds.Footnote ¹⁴

A major challenge to this accepted theory came in the mid-1960s when Myanmar scholars began to discuss alternative assumptions. One of them was U Than Tun, who, in 1969, proposed that the calendar used by the Pyus was that of the Indian Guptas, commencing in 319/320 CE. U Than Tun based his revision on the palaeography of the Pyu alphabet and its parallels to the Gupta script.Footnote ¹⁵ However, when taking the emerging synchronisms between the Pyu cities of Beikthano and Sri Ksetra into account, U Than Tun later changed his mind and proposed the Saka era of 78 CE instead of the Gupta era.Footnote ¹⁶ In doing so, he possibly followed U Aung Thaw, the excavator of Beikthano, who had suggested this identification of the era at a conference held at Yangon in 1966.Footnote ¹⁷

This line of argument was also followed by U Htin Aung—the long-time professor of English at Yangon University, who had suggested that era in his rather vitriolic attack on Luce and Luce's alleged distortion of Myanmar history.Footnote ¹⁸ As the title of his treatise indicates, U Htin Aung's justification for the reinterpretation of the era comes from the chronicles which state that the Pyus used the Saka era before replacing it with the Myanmar era. This is an obvious circular conclusion, as the suggestion of an earlier use of the Saka era only works on the premise that the urns were made and incised before 638 CE.

What U Htin Aung reminds us of, however, is that indigenous Myanmar traditions must not be ignored. The main embodiments of this tradition—the Myanmar chronicles—only sporadically predate the eighteenth century when the first important work was written and will therefore have to be read with caution, especially in relation to their statements on the early history of Myanmar.Footnote ¹⁹ There is, however, one central point on which the chronicles are unequivocal: the computation of time in Myanmar has always been achieved by way of a continuous, linear calendar. Only the starting point of a calendar could change if a ruler, his advisers, and astrologers saw the need for such a measure.

According to the chronicles, the oldest existing era, called the Ajjana or Anjana era, began in the year 68 before Gautama Buddha was born and ended with his parinibban 80 years later, having reached the year 148 at this point. This event marked the start of the Buddhist era (of 544 BCE), which replaced the Ajjana era.Footnote ²⁰ The Buddhist era in turn existed until the year 624, when King Thumodarit started a new computation by eliminating 620 years. The king's reform thus established (or rather introduced) the Saka era of 78 CE in Myanmar.Footnote ²¹ Finally, the Pyu King Poppa Saw Rahan is credited with a repetition of this exercise in Saka 562 (or 640 CE), at which point he established the current Myanmar era.Footnote ²² As said, the existing chronicles, which are seen to embody Myanmar traditions and customs, come from a relatively late point in time, but the evidence that they provide can be corroborated by way of sources going much further back in time. In the inscriptions from the Bagan period (eleventh to thirteenth centuries CE), all four calendrical systems mentioned in the chronicles are practically employed for dating in at least one instance in Myanmar history.

Beginning with the era very likely used least often, the Saka era of 78 CE was employed by King Alaungsithu in his long Pali inscription from the Shwegu-gyi temple, composed in 1131 CE.Footnote ²³ The actual date is stated at the end and in Sanskrit, which seems to indicate that the author was an Indian—either a brahmin or a Buddhist monk (as the rest of the inscription is composed in Pali verse). In their edition of the inscription, Luce and U Pe Maung Tin underscored that this was a ‘unique instance for the use of this calendar’.Footnote ²⁴ This seems correct for the Bagan period but, in the early fourteenth century, an inscription from Shwebo in Upper Myanmar gives the year 1262 of an unspecified era. If Saka, this would have been 1340 CE—a date perfectly in line with the two dates on the obverse of the record.Footnote ²⁵ Regardless of whether this second example is correct or not, it is obvious that the Saka era in Myanmar was used only in exceptional cases. This sets the country apart from the rest of Southeast Asia, notably the Angkorian empire, where the Saka era was widely used. Having said this, one should, however, not overlook that the very term ‘Saka era’ (sakaraj, or thekkayit in Myanmar pronunciation) became synonymous with ‘calendar’ in the Myanmar language (as much as in other languages of mainland Southeast Asia).

As could be expected, the Ajjana era is a similarly rare occurrence. The only examples of its use during the Bagan period come from two short painted inscriptions on the ceilings of temples 845 (Kutha) and 1471 (Theinmazi) at Bagan. Both consist of four horoscopes referring to important events in the Buddha's life: conception (Maya's dream), birth, enlightenment, and parinibban.Footnote ²⁶ In the case of the Kutha temple, the readings of the four dates are interspersed with a second inscription describing the ‘Buddhist paradise’ Uttarakuru (Figure 5). Apart from these two instances, which are directly related to the four ‘great’ events in the Buddha's life, the Ajjana era was apparently not used for practical purposes such as the recording of a donation.

Figure 5. Kutha ink inscription (left part). Source: P. Pichard, Inventory of the Monuments at Pagan (Paris, 1994–1995), vol. 4, no. 845.

The Buddhist era, in contrast, was commonly used throughout the Bagan period, from the time of King Kyanzittha and his son Rajakumar up to the end of the Bagan period and beyond.Footnote ²⁷ This widespread use of this era is again specific to Myanmar and makes the place an exception among Buddhist countries. The era probably originates from ancient Sri Lanka, where monks had attempted to determine the year of the Buddha's parinibban, expecting the future Buddha Metteyya to appear 1,000 years after the demise of his predecessor. The first chapters of the Sinhalese chronicle Dipavamsa, which may date from around the fourth century CE, are mainly concerned with calculating the years from the Buddha's parinibban by way of royal successions and monastic lineages running parallel to the dynasties.Footnote ²⁸ In ancient Sri Lanka, however, the era was not used again for dating either by Sri Lankan historiographers or by the monks or scribes of inscriptions for several centuries. The next instance of a date in the Buddhist era occurs in the twelfth century, when King Parakkama Bahu I unified the island's monastic order and subjected it to a set of rules to abide by.Footnote ²⁹ This was well after this era had come into use in Myanmar during the reign of King Kyanzittha and, given the well-established relations between the two countries by that time, it seems more likely that the calendar had been reintroduced to Sri Lanka by monks from Myanmar. Yet, while the calendar fell into disuse again in Sri Lanka soon after, it continued to be employed in Myanmar with great frequency. It was only in the fifteenth century that it was revived in an emerging Buddhist ecumene and spread to other places of the Southeast Asian mainland.Footnote ³⁰ From then on, it gained wide acceptance all over the Buddhist world.

It is not clear why King Kyanzittha or his son Rajakumar used the Buddhist era to date their inscriptions, as, by the late eleventh century, the Myanmar era was not only known, but also used widely. As far as we can see, it was used for the first time in the year 440 (1078 CE), closely followed by a royal donation made in the year 455 (1093).Footnote ³¹ Since then, the Myanmar era has been in use for dating practically anything written or printed in Myanmar that needed to have a date on it: inscriptions, manuscripts, printed books, newspapers, etc. At Bagan, the Myanmar and Buddhist eras were frequently used side by side, but there is not a single inscription from the period to have used regnal years or any other possible dating system on its own or instead of the Myanmar era.Footnote ³²

Pyu urns, Myanmar calendars, and Indian influence

The evidence presented so far has shown two things that any attempt to date the Pyu urns will have to take into consideration. Firstly, the Myanmar people preferred continuous and linear computation of time, using calendars that had a clearly defined starting point and ran independently of political or dynastic developments such as royal successions or transfers of the capital city. This did not exclude the possibility of having a new era established when circumstances required it; and this also included the possibility of using more than one such continuous reckoning of time side by side.Footnote ³³ Depending on the nature of the Pyu–Myanmar cultural continuum, this attitude towards the computation of time may have prevailed already during the Pyu period. This problem will be addressed below.

Secondly, the written evidence of the Bagan period demonstrates that the calendars in use at the time consisted of the Ajjana, Saka, Buddhist, and Myanmar eras. Of these four, two seem to have been imported from somewhere outside Myanmar—the Buddhist era, which probably came from Sri Lanka, and the Saka era, which could have come from either India or the Southeast Asian mainland.Footnote ³⁴ Crucially for discussion of the dates on the Pyu urns, the Gupta era was not among the eras in use and, as we will see next, there is no evidence that the era was known or used anywhere outside India (and, more precisely, a rather limited area of north India) at all. More importantly, the circumstance of its creation and spread effectively preclude its use by the people of Sri Ksetra.

Regarding the Gupta era, if we ignore two spurious inscriptions dated to the years 5 and 9 of that era, the earliest reliable examples for its use are Candragupta's II Mathura pillar inscription (year 61) and the Udayagiri rock inscription (year 82). Two more epigraphs stemming from the reign of this king bear the dates 88 and 93. Candragupta is generally believed to have ruled during the second half of the fourth century CE; the era he used (possibly, he had created it as well) referred to the founder of the Gupta Dynasty, Samudragupta I, hence its commencement in 319/320 CE.Footnote ³⁵ As stated previously, the first documented use of the Gupta era in India dates to the year 61 (380/381 CE) and, while the era may have advanced as far Rakhaing (Arakan), which borders on Bengal,Footnote ³⁶ the ‘late’ use alone throws doubt upon the idea that the calendar could have been employed for the dates on the urns from central Myanmar: not only would its spread have required direct links and exchanges between north India and central Myanmar for swift transmission; it would also imply that the two urns of Harivikrama and Suryavikrama predate the earliest use of the era in India. The different readings of Blagden and U San Win make no difference for this matter. As this is hard to imagine, and as the Gupta era is never mentioned in any record from Myanmar throughout, we can effectively rule out the possibility of the Gupta era being used by the Pyus.

Interestingly, the conclusion is virtually the same if the possibility of the use of the Saka era is examined with similar rigour. The Saka era was first used at around the end of the first century CE in north-west India, within the realm of the Kushana rulers.Footnote ³⁷ The year in which it commenced, 78 CE, probably refers to the accession of Kanishka I.Footnote ³⁸ Unlike the Gupta era, the Saka era was used right from the time of this king for practical purposes, such as the dating of Buddha images or other dedicatory inscriptions. However, it was not until the fifth century CE that the era was more widely employed in both India and Southeast Asia.Footnote ³⁹ The oldest Sanskrit inscription from Southeast Asia dated in the Saka era originates from Cambodia and bears the date 534 or 612 CE;Footnote ⁴⁰ a little later, the era also came into use in the Srivijayan realm, in Java and Champa. The use of the Saka era on the Pyu urns would thus again have predated by several centuries any other inscription from Southeast Asia using the era and, in addition, it would have required direct contacts between the Pyu heartland of central Myanmar and north-west India. That apart, the use of the Saka era in Myanmar seems to be less problematic than use of the Gupta era, as the dates on the urns do not predate its use in India and the Saka era was in fact used during the Bagan period, even if only rarely. Still, the problem of direct relations for providing a swift transfer of a new calendrical system appears to form a considerable obstacle to the possibility of the Saka era being used on the Payagyi urns.

Before ruling out the possibility that the Gupta and Saka eras (or in fact any other calendar of Indian origin such as the Kalachuri–Chedi or the Ganga eras)Footnote ⁴¹ could have been employed for the dates found on the Pyu urns, two more related questions will have to be considered: Were the people of central Myanmar susceptible to Indian culture generally, and if so, to what degree? Space will not permit the whole debate about the ‘Indianisation’ or ‘Hinduisation’ of Southeast Asia to be presented in any detail here, but the position of Myanmar in this process will have to be briefly assessed. It seems important to note that Myanmar, despite her geographical proximity to India, with some difficult overland routes (across the Bengal–Assam–Manipur borders) and a much easier accessible sea route linking the two regions, was probably less influenced by Indian culture than any other region of Southeast Asia. This is not to deny that the people inhabiting Myanmar during the first millennium CE had come into contact with Buddhism relatively early (though probably not through Ashoka's legates Sona and Uttara, who allegedly went to Suvannabhumi),Footnote ⁴² but surely before the third to fourth centuries CE. At that time, the first Buddhist monuments were built at Beikthano, imagery from Amaravati was transferred to Sri Ksetra, and monks there possessed seals written in a late Brahmi script.Footnote ⁴³ A little later, probably in the first half of the fifth century, extracts from the Pali canon were copied onto gold foils in the shape of palm leaves.Footnote ⁴⁴

While Indian Buddhism, predominantly originating from the regions along India's east coast, is well attested in the Pyu settlements of central Myanmar from around the second to third centuries CE, Hindu sculpture, Brahmins, and Sanskrit are less common in Myanmar. There are a few sculptures of Vishnu and other Hindu deities,Footnote ⁴⁵ but Sanskrit inscriptions are few and far between.Footnote ⁴⁶ The largest number and best-known examples probably come from Arakan—not accidentally the area closest to India—where King Anantacandra left behind a long, four-sided inscription that displays many features of contemporary eighth-century Sanskrit epigraphy from India.Footnote ⁴⁷ Sri Ksetra has yielded a Buddha image with an inscription in Sanskrit and PyuFootnote ⁴⁸ and a few votive tablets with Sanskrit texts as well as a mixed Tamil and Sanskrit inscription composed by a member of a South Indian trading guild residing at the city have come down from the Bagan period.Footnote ⁴⁹ Lastly, there is a huge but heavily obliterated Sanskrit inscription that was recovered from the moat of Bagan in 1990. Still unpublished, it is now kept in a brick shed opposite Bagan's Tharaba gate.

This overview, brief as it is, may suffice to illustrate that Indian culture entered Myanmar mainly in the form of Buddhist culture, whereas Brahmanic cults and Sanskrit were rather marginal. The predominance of South and Southeast Indian culture, as manifested from the mid-first millennium CE in the form of scripts and sculpture, moreover suggests that the routes of transmission led across the Bay of Bengal rather than through the jungle tracks of north-east India and northern Myanmar, though this question certainly requires further exploration. Brahmins and their expertise—including literature and poetry, ritual, and astronomy incorporating the computation of time—appear to have been in small demand.