The Sanskrit Inscription of Śaṅkara and Its Interpretation in the National History of Indonesia

It is a well-known fact that Sanskrit had a relatively shorter and less prolific lifespan in the epigraphy of Indonesia, particularly in the Javanese epigraphic record, than in that of other Southeast Asian regions. All the more precious, therefore, are the rare oppor-tunities to add a Sanskrit inscription to the historical record of Java and learn more of how the Sanskrit language was deployed on the island to represent events recorded for posterity. In this article, I offer my edition and interpretation of the inscription referred to in Indonesian publications as Prasasti Sankhara ( sic , with kh ), that is, the Inscription of Śaṅkara ; debunk the entirely unfounded interpretation which it has received in successive reprints and editions of the greatly influential Sejarah nasional Indonesia (National history of Indonesia); and show the real historical interest of this inscription.


Introduction
J.G. de Casparis, who died in 2002, was the last great specialist of Indonesian epigraphy who was competent in Sanskrit, but he stopped publishing inscriptions early on in his career. If I am not mistaken, not a single previously unpublished Sanskrit inscription from Java has been published by a trained Sanskrit scholar in nearly 70 years, since an article by De Casparis dating to 1961.1 The work of this scholar has been greatly influential. This is true especially for two large volumes, dating respectively to 1950 and 1956, which contain editions and translations of numerous inscriptions in Old Malay, Old Javanese, and Sanskrit, along with elaborate discussions of the historical problems on which the epigraphical documents in question throw light. De Casparis followed in the footsteps of several earlier scholars by focusing his attention on the dynastic history of Central Java in the second half of the first millennium: the questions of the geographic origin of the Śailendra dynasty; of royal patronage of Hinduism versus Buddhism; and of the possible role of foreign, that is, non-Javanese, political power in these matters. In order to weave a consistent historical narrative, he often resorted to rather extreme speculations to tie rare facts together, and has been justly criticized for this by his French contemporary, Louis-Charles Damais, who devoted to De Casparis's books a book-length review of 225 pages in befeo volume 54 (1968), published two years after his death. In the apologetic opening footnote of this review (p. 295, n. 1), we read:2 Actually, I believe that there is already an excess of theories that have been put forward particularly regarding the ancient history of Java, and I wanted above all to warn against the acceptance of solutions that may be seductive but that seem to me insufficiently supported by the documents, or even plain conjecture. I would prefer, in other words, to return to the scrupulous study of original documents while avoiding as much as possible to confuse more or less hypothetical deductions-which one is of course entitled to make, on the condition that it happens with measurewith facts that are proven beyond doubt or are sufficiently corroborated.
The same questions and speculative reconstructions have not ceased to occupy historians of ancient Indonesia of the generation after De Casparis, and we will see that the interpretation of the inscription of Śaṅkara has suffered from a biased, nationalistic framework of interpretation and from a deficient competence in Sanskrit on the part of the responsible historian-epigraphist. Both De Casparis and Damais were acknowledged as teachers by the late Boechari (d. 1991), who in turn taught the current generation of senior epigraphists in Indonesia, and has during his career prepared numerous 'provi-Bijdragen tot de taal-, land-en volkenkunde 177 (2021) 1-26 sional transcriptions' of newly discovered inscriptions, most of which were never formally published during his lifetime but have been collected and published in the volume Melacak sejarah kuno Indonesia lewat prasasti / Tracing ancient Indonesian history through inscriptions (2012). One of the inscriptions that were included there is the inscription of Śaṅkara. 3 Boechari, definitely a very serious and admirable scholar, despite the weighty criticism that I will bring forward in this article, was responsible for many of the epigraphic contributions to volume ii, entitled Zaman kuno, of the Sejarah nasional Indonesia (sni), the National history of Indonesia, a major collective publication which was launched in the 1970s and of which new editions and reprints have appeared continuously since the first edition in 1975, most recently in an edisi pemutakhiran, or updated edition, that appeared in 2008 and has been reprinted several times since.4 Without quoting more than one verse-quarter of its original text, but referring to his own transcription, Boechari in sni vol. ii uses data from the inscription to support far-reaching reconstructions of Indonesian dynastic and religious history, on the basis of the imagined conversion from Śaivism to Buddhism of its protagonist, named Śaṅkara, who is presumed to have ruled as king in the eighth century ce. His dating of the inscription on palaeographic grounds, and his reading into it of the 'conversion' of a 'king' , allow Boechari to tie together the patronage activities of monarchs whom he would otherwise have felt constrained to attribute to two different dynasties, one of which would ostensibly have to be 'foreign' , according to a narrative that was dominant in Boechari's time and hasn't entirely lost its appeal to some scholars even today.5 To set the stage, I need to quote what are presumably Boechari's own words at length, while highlighting some elements that demand special attention:6 3 Because it was never published during his lifetime, the inscription is missing in the overview, by now rather outdated, of Sanskrit inscriptions from Indonesia and Malaysia that was offered by De Casparis 1991. 4 On Indonesian national and nationalist historiography, see Reid 1979. I cite sni vol. ii from the 4th edition (Bambang Sumadio 1990) and the updated edition (Bambang Sumadio and Endang Sri Hardiati 2008). 5 By way of example, let me cite Van Naerssen 1947, Bosch 1952, and Jordaan and Colless 2009 For further references, see the section 'A critical review of the single-dynasty and two-dynasty theories' in Long 2014 (pp. 79-84). Mark Long, an amateur scholar whose work is laudable but also shows many weaknesses, ignores the short paper on the same topic included as Chapter 11, 'Satu atau dua dinasti di kerajaan Matarām Kuno' , in Boechari 2012. Boechari's thinking was strongly influenced by the reaction published by Poerbatjaraka (for instance, in his 1958 article) to the aforementioned articles by Van Naerssen, Bosch, and others. 6 I am translating from sni vol. ii (1990), pp. 91-3, including notes 20 and 21. I do so with From the inscription of Sojomerto it is clear that Dapunta Selendra was a follower of Śaivism. When and for what reason the kings of the Śailendra dynasty began to follow Buddhism may perhaps be known from an inscription owned by Adam Malik, which we will provisionally refer to by the name Inscription of Śaṅkara. This inscription is in Sanskrit, but regrettably only its final part has been found. It seems that the inscription was engraved on two stones, but the first stone which contained the beginning of the text is not available. We therefore do not know when the inscription was issued-if it contained any date at all. In view of the unevenness of the back of the inscription, and the presence of parts that stick out, it seems that this inscription used to be placed in some building.
The remaining part contains the information that at a certain time the father of King Śaṅkara fell ill, and for eight days he suffered badly because of a burning fever. Finally, he died without his priest teacher being able to cure him. For this reason, King Śaṅkara then felt afraid of his teacher, whom he considered wrong, and he consequently abandoned his devotion to Śaṅkara (the god Śiva). The concluding part of the inscription indeed suggests that King Śaṅkara later became a follower of Buddhism, for it is stated among other things that he gave a favour to the bhikṣusaṅgha.
[…] Actually this interpretation is still rather doubtful. The matter rests on the interpretation of the sentence that runs so yan tyaktānyabhaktir jagadaśivaharāc chaṅkarāc chaṅkarākhyaḥ, above all the ablative case jagadaśivaharāc chaṅkarāc. Can this sentence be translated by 'him, who is called Śaṅkara, who abandoned his devotion to others, (among others) from Śiva, who removes the world's disorder'? Indeed this sounds quite odd, but keeping in mind that the final portion of the inscription suggests that the King Śaṅkara became a follower of Buddhism by making a donation to the bhikṣusaṅgha, it is very possible that this oddness results from the limited comprehension of the writer of the inscription, who very probably was a native scholar, with regard to Sanskrit grammar.
[…] What also attracts attention is the information that King Śaṅkara refers to his teacher as 'a bad teacher' (anr̥ taguru). silent adjustment of the punctuation and spelling of some words, including the fact that Boechari consistently writes Śangkhara instead of Śangkara (my Śaṅkara), and the insertion of a few quotation marks. The corresponding pages of the (only very superficially modified) edisi pemutakhiran (2008) are 117-8.
Bijdragen tot de taal-, land-en volkenkunde 177 (2021) 1-26 If this interpretation is correct then we find here an epigraphical source that offers information about the conversion of religion from Śaivism to Buddhism, and the king who converted was King Śaṅkara, whose name has not hitherto been found in the sources known to us.
This inscription being incomplete, the year to which it dates is unknown. However, from a palaeographic perspective it may be estimated that it dates from the middle of the eighth century ce. It thus very probably constitutes epigraphical evidence for the theory of Poerbatjaraka based on information in the Carita Parahyaṅan. In other words, it is very likely Poerbatjaraka who was right with regard to the pedigree of the Śailendra dynasty, namely that they were native Indonesians, and that there was only one dynasty, the Śailendra dynasty, whose members initially followed Śaivism, but from the government of Rakai Panaṅkaran onwards became followers of Mahāyāna Buddhism, later to convert again to Śaivism after the government of Rakai Pikatan.
We need to mention here the idea that the complete title of Rakai Panaṅkaran was very probably Rakai Panaṅkaran Dyah Śaṅkara Śrī Saṅgrāmadhanañjaya.7 Summarizing the historically important points, Boechari believed that Śaṅkara was a 'native Indonesian'; that he was a king of the Śailendra dynasty, none other than the Rakai Panaṅkaran famous from the inscriptions of Kalasan (778 ce), Mantyasih i (907 ce), and Wanua Tengah iii (908 ce);8 that he was the son of King Sañjaya; that he was initially a devotee of Śiva who later converted to Buddhism; and that he did not write the inscription himself but that whoever did had only a mediocre command of Sanskrit. Against the background of these ideas, let us turn now to the inscription itself.

Provenance, Preservation History, and Physical Aspects of the Stone
The original provenance of the inscribed stone is unknown. Boechari first refers to the inscription as a recent discovery in two closely related papers published in 1975,9 and mentions as 'very likely' one collector's report that it would have 7 This idea has been elaborated in a short paper by Bambang Budi Utomo (1989). been unearthed in Central Java near Sragen, to the east of Solo.10 I do not know why Boechari considered this specific provenance to be so plausible, but at least it seems safe to assume that the provenance lay somewhere in Central Java. After its discovery, the stone was acquired by Adam Malik, an important political figure under the Sukarno and Soeharto presidencies. After his death in 1984, his widow, Nelly Malik, took the initiative of founding a museum, but around the year 2000, the heirs were no longer able to keep this private museum going and several pieces were sold off so that the collection disintegrated.11 10 My translation of sni ii (1990), p. 90 (= 2008:116), n. 16: 'According to information from Mr. Adam Malik, the seller of the inscription said that the stone originated in West Java. But there is information from a collector in Sala who told Boechari that the stone originates from near Sragen, to the eastern side of Sala, and that at the site in question remains of brick constructions are still found. That collector is still keeping a head of a Buddha statue of beautiful manufacture, made of terracotta. Keeping in mind the tendency of dealers in archaeological artefacts never to mention the provenance of the pieces they sell, and keeping in mind the contents of the inscriptions, it is very likely that it is the information from the collector in Sala which is correct.' 11 I rely here on information from an item published on 6 September 2008 on the online news portal liputan6.com (see 'Museum Adam Malik tinggal kenangan' , https://www By the time I started studying the inscription, in 2009, nobody in Jakarta was able to tell me where the stone was kept. In that phase of my research, I had to rely on a black-and-white photograph published in sni ii (1990, Fig. 1). The exact whereabouts of the inscribed stone in question are still unknown to me, but I do know that the stone resurfaced in 2019 in a private collection in Indonesia. The new owner had the kindness to send archaeologist Véronique Degroot a recent photo that I include here as Fig. 2. Based on this photo, and on the observations of Véronique Degroot, I can affirm that the preparation of this stone and the engraving of the text reveal a workmanship unrivalled in Indonesian epigraphy.
Although, as stated by Boechari, this single stone does not preserve the whole of the original inscription, the surface of the one stone that we do have has been exquisitely polished and stunningly well preserved, even to the point that there are traces of possibly original red paint on and around the inscribed surface. The stone is not large, measuring perhaps c. 50 × 70cm. I agree with Boechari's suggestion that it was meant to be placed in a recess of the wall, such as a niche or a space between two pilasters (Fig. 3). The upper face of the stone is quite smooth. It is possible that it had been smoothed for the joint with the upper stone to be seamless. This upper stone, bearing the first part of the inscription, would have been attached to the wall with horizontal tenons as well.

Edition and Translation of the Text
My edition has been established on the basis of the aforementioned photographs. In footnotes, I report variant readings found in Boechari's edition (B) that I have myself prepared for publication since his death (Boechari 2012:473-5) on the basis of two frequently varying typescripts (Figs. 4 and 5). In my edition, I hyphenate the Sanskrit text and insert apostrophes (') to mark elision of vowel a-due to sandhi, although no avagraha sign is actually written in any instance. For these and other aspects of the transliteration system (notably the .liputan6.com/news/read/164858/museum-adam-malik-tinggal-kenangan, accessed 7-12-2020); an item that was published on detikNews.com on 26 November 2008 but is no longer available now (defunct url: https://news.detik.com/berita/d-1043374/museum -adam-malik-nasibmu-kini); and a blogpost by Djulianto Susantio ('Museum Adam Malik tak tersisa' , https://hurahura.wordpress.com/2013/01/08/museum-adam-malik-tak -tersisa/, accessed 7-12-2020).
Bijdragen tot de taal-, land-en volkenkunde 177 (2021)  Undated typescript by Boechari: Prasasti batu koleksi Bapak Adam Malik: Transkripsi sementara menurut baitnja. The spelling -nja suggests it is datable to the early 1970s at the latest use of capital letters for akṣara vowels), which I apply also to Boechari's readings in my footnotes, see Balogh and Griffiths (2020). My translation is on many points different from that of Boechari, but I will discuss only the historically important differences in the next section.

Historical Commentary
Regarding the structure of the text, it is clear that we have preserved only the final portion of an originally longer inscription. We can be certain that the stanza here numbered v was indeed the inscription's final stanza, both because of its contents (a wish for long life that is reminiscent of the Hampran and Balekambang inscriptions)16 and because of its metrical form (the Puṣpitāgrā metre that is used is a typical metre for concluding stanzas in Sanskrit poetics).17 Any possible dating of the inscription would have been contained in its opening portion and is now lost to us. We are therefore dependent on palaeographic arguments to establish its date. I will return to this issue below. The king mentioned in the final stanza remains unnamed in the available part of the inscription. I do not think the epithet arāti-nihan 'destroyer of enemies' needs to be connected with the complex of epithets that are similar in meaning but not in sound (sarvāri-mada-vimathana, vairi-vara-vīra-mardana, vīra-vairi-mathana) found in inscriptions explicitly related with the Śailendra dynasty (Bosch 1952:116)-arāti-nihan can easily be considered as a stereotypical royal epithet, although I cannot exclude that the similarity is more than coincidental.
Contrary to Boechari's assumptions, there is absolutely no evidence that the author of the text was any other than the protagonist Śaṅkara,18 nor that 16 Hampran,  (827 Śaka), one from the village Kilipan and another from the village Sampū Alas, our text contains no date, toponym, or other information that would allow us to confirm any particular identification, but only the first would be compatible with the date bracket that I will propose for the inscription below. At least it can be observed that none of these Śaṅkaras was king, and that the name Śaṅkara-spelt Śaṅkhara in some of these inscriptions but not in ours-was quite common on Java in the ninth century ce.
I also disagree with Boechari's arguments that are based on the assumption that the author of the text had only limited competence in Sanskrit. My impression is that the author of the text, who identifies himself as Śaṅkara in stanza iii, was quite an accomplished Sanskrit poet. Only one real error is found in the text: the misspelling of voiced da for unvoiced ta in pāda c of stanza v, which may have been induced by the pattern jagadapagada-. The spelling of ḍ for ḍh (lines 6 and 15) reflects a regional spelling standard and cannot be deemed a mistake.
In stanza i, the phrase dadataḥ ṣaṭ-suvarṇaṁ vyayārtham· contains what may be the earliest attestations in Javanese epigraphy of the terms suvarṇa and the term vyaya to denote, respectively, a particular weight in gold and expenditures related to pious foundations. Although either one of these terms by itself 19 I have nowhere seen this made explicit, but I assume that Boechari felt that the idea 'that the complete title of Rakai Panaṅkaran was very probably Rakai Panaṅkaran Dyah Śaṅkara Śrī Saṅgrāmadhanañjaya' was supported by the possibility of explaining the Old Javanese word panaṅkaran as derived with paN-..-an circumfix from a base śaṅkara. Personally, I find it much more likely that the underlying base is the word recorded with spelling sǝṅkǝr in Zoetmulder 1982, as spellings with ǝ and a were in almost free variation in early specimens of Old Javanese. A line in the kakavin Bhomāntaka (38.19d) shows the close association of the idea of a 'limit' with the power of a king: aṅhiṅ sira tiki makadaṇḍanīti sira sǝṅkǝr amuput i jagat 'only he administers the law; he is the outermost limit of the world' (the translation 'But only he who has chastisement as his policy is the spoke that reaches to the outer limits of the world' proposed by Teeuw and Robson 2005:263 seems less satisfactory). The same association transpires from a phrase in the unpublished Pūrvādhigama (cited by Zoetmulder 1982 under the entry rājanīti): vruha ri səṅkərniṅ śāsananira '(the king) should know the limits of his commands' .
Bijdragen tot de taal-, land-en volkenkunde 177 (2021) 1-26 is a perfectly normal Sanskrit word, their co-occurrence reveals that the particularly Javanese mix of Sanskrit metrological and administrative terminology was already in place by the time of the inscription's composition. The oldest dated co-occurrence that I am aware of is found in the ninth-century Old Javanese inscription Kwak i (801 Śaka In stanza ii, I find the sequence śr̥ ṇvan vacanam iti somewhat peculiar, but I have assumed that it can perhaps stand as equivalent to śr̥ ṇvan etad vacanam 'hearing those words' . Not to speak of Boechari's obviously impossible interpretation of the sequence tyaktānyabhaktir jagadaśivaharāc chaṁkarāc in stanza iii, which he himself admitted to be odd and whose oddness he explained away by accusing the poet of incompetence, the late scholar was clearly not aware of the possibility of interpreting an-r̥ ta-guru-bhayas in this stanza quite differently, namely by taking guru not as 'teacher' but in its literal sense of 'heavy' , and joining it with bhayas rather than with an-r̥ ta. It is this solution that yields my translation 'greatly fearing untruthfulness' , which makes much better sense in the context. In stanza iv, I am unable to detect specifically Śaiva or specifically Buddhist tenets. The emphasis on the salvific power of Gnosis is common to both religions. The main problem of interpretation, however, and one that is related to this question about stanza iv, lies in stanza v. What is clear, is that there is no question of a donation to the bhikṣusaṅgha, let alone of any conversion to Buddhism. The problem, as I see it, is whether any connection with Buddhism may be assumed at all. With the words dharma and bhikṣusaṅgha, we would at first sight seem to be placed squarely in a Buddhist context. But this interpretation, which would have as corollary that a Śaiva poet Śaṅkara composed a final stanza in support of Buddhist institutions, is not necessarily correct. The use of the word bhikṣu, to begin with, is not the monopoly of the Buddhists. It generally denotes a beg-Bijdragen tot de taal-, land-en volkenkunde 177 (2021) 1-26 gar in some of the most fundamentally 'Hindu' texts in Sanskrit.21 It is also used in manuals (paddhati) for sectarian Śaiva practice to refer to initiates belonging to a set-up under an ācārya, but in one context only, namely when they speak about going out to beg for food (bhikṣā). The author Somaśambhu, for instance, details various possible arrangements for food in a Śaiva monastic context, and the topic is complicated. But it is clear that begging is highly rated, since it seems that the Śaiva initiate raises up the level of everyone's Dharma by receiving alms (dharmonninīṣayā bhikṣāṁ bhikṣur abhyādadīta tām).22 The word possibly gained a more general usage in Śaiva circles in Indonesia, where it occurs in a few texts in the Old Javanese language, some of them specifically Śaiva.23 It is noteworthy, in this connection, that one of the most common and earliest attested Javanese terms to designate a (Śaiva) mendicant is viku, which is generally believed to be a loanword based on a Middle Indo-Aryan form of bhikṣu.24 Our devout Śaiva poet was therefore quite possibly aware of the Śaiva meaning of the term. He may have been playing consciously on its Buddhist associations by placing it in the direct vicinity of the words dharma and a verb-form cinotu that is derived from the same root as caitya, and moreover inserting it in a compound with saṅgha, but the latter word too is not the exclusive property of Buddhist Sanskrit, nor is the word dharma or the root ci. However, it is also possible that, despite Śaṅkara's personal devotion to Śiva, this final stanza is to be read at a more general level of public affairs, and that the community of Buddhist monks is actually intended, or perhaps that all mendicants are intended without regard for their religious affiliation. This would then help to explain the use of the term agrya-sukhī which, as far as I am aware, has no Śaiva but does potentially have a Buddhist connotation.25 The word kulapati, on the other hand, is certainly not a Buddhist term but is found in explicit association with 'Hindu' (Vaiṣṇava and Śaiva) temples and monastic institutions elsewhere. See, for instance, the inscription of Muṇḍeśvarī Hill in Bihar, India, re-edited and discussed by Neuss (2003), or the following inscription from a Vaiṣṇava context in Nepal, dating to the 660s ce, where we read:26 kaś cit prasāda-śāsana-paṭṭako 'bhūt sa ca pūrva-rāja-vibhramato naṣṭo 'dhunāsmābhiś cirantana-vyavasthānupālana-jātādaraiḥ sa eva prasādaś cira-sthitaye śilā-paṭṭakābhilekhyena prasādī-kr̥ to 'tra ca maryādā kulapatinā deyā kārtika-śuklaikādasyām mārga-saṁskārārthan taṇḍūla-mānikā 4 24 See item 5 in Hoogervorst 2017. 25 Cf. the prayer sarve sattvāḥ sarve prāṇāḥ sarve bhūtāś ca kevalāḥ | sarve vaḥ sukhinaḥ santu sarve santu nirāmayāḥ |, discussed by De Casparis (1956:338) on the basis of a Javanese attestation. 26 Regmi 1983, Vol. 1, no. 122, lines 7-11. I am grateful to Diwakar Acharya for helping me with the translation and informing me about the date of this inscription. In an email of 11 May 2020, he informs me: 'This inscription is now situated near a Śaiva temple but internal evidence tells that it concerns management and use of the property donated to the temple of Lokapālasvāmin, obviously a Vaiṣṇava deity. For, it mentions that Kārtika śukla dvādaśī (the next day after Viṣṇu rises from his sleep; the day of pārāyaṇa of the vaiṣṇava cāturmāsya) is the day of installation of the deity and all other festivals mentioned are on ekādaśī or dvādaśī; it also mentions varāhayātrā.' Bijdragen tot de taal-, land-en volkenkunde 177 (2021) 1-26 'There was a charter of a (royal) grant (prasāda) but that got lost due to the negligence (vibhrama) of the previous kings, and now by us, being attentive to maintenance and protection of ancient arrangements, that very grant has been gracefully bestowed through an inscription on a slab of stone for the sake of its long endurance. And in this connection, there is the rule: on the eleventh (tithi) of the waxing (fortnight) of Kārtika, 4 mānikās of rice should be given for the sake of repair of the (processional) path by the person in charge of the temple (kulapati).' For a Śaiva context, I may refer to the Cambodian inscription K. 180 (dated 869 Śaka / 948 ce), stanza xxiv (ed. Coedès 1913:20): sveṣāṁ māheśvarāṇāṁ yaḥ kulānāṁ patir āśrame māheśvarāśramābhikhye rājñāṁ kulapatir mmataḥ 'The one who was the chief of his own groups in the cloister of the Māheśvaras (that is, Śaivas) called Māheśvarāśrama was considered the abbot (kulapati) of the kings.' The term kulapati is not found, as far as I am aware, in other Sanskrit sources from Indonesia, but it has been borrowed into Old Javanese. There do not seem to be any occurrences in Old Javanese literary works-and in epigraphic contexts too, from Java and Bali, the term is not very common-but there is a noticeable spike in inscriptions of tenth-century East Java. The meaning is not evident from the contexts that I have seen, but the following three passages from the lists of witnesses in two unpublished inscriptions issued by King Sindok suggest that the word had the same range of meanings as illustrated by the Sanskrit inscriptions cited above:27 In no Old Javanese context that I have seen is kulapati used in connection with a recognizably Buddhist institution. Returning now to our inscription, it seems possible, perhaps even likely, that Śaṅkara, who refers to himself in the third person in stanza iii, does so again in stanza v and that the kulapati is none other than himself. Whether or not this interpretation is correct, the above evidence has made clear that the kulapati intended in stanza v was probably an abbot of a Śaiva monastic institution or manager of a Śaiva temple.

The Problem of Dating
Let us now turn to the question of dating. Boechari dated the stone to around the same period as the Kanjuruhan inscription from near Malang in East Java, which dates to 682 Śaka or 760 ce, but felt it might be a bit earlier.28 This means that the inscription would have to be placed somewhere close in time to the inscriptions of Canggal (653 Śaka, 732 ce) and Balekambang, both of which use the stiff/squarish script style that is usually called 'late Pallava' ,29 and the Hampran inscription (672 Śaka, 750 ce), which, like Kanjuruhan, uses the more cursive script style usually called 'Kawi' . It is this last script style that we also see deployed in the present inscription, but with a workmanship, glyph-size, and elegance that far surpasses the Hampran (Fig. 6) and Kanjuruhan (Fig. 7) specimens-a difference that complicates palaeographic comparison. 28 Boechari 1975b28 Boechari :26-7, n. 9 (= 2012: 'Our supposition that it might originate from the eighth century a.d. is based on palaeographical considerations. The script shows close resemblance to that of the stone inscription of Kāñjuruhan dated 760 a.d., and actually seems even more archaic.' 29 I am now less confident about my proposal (Griffiths 2012:477) to date the Balekambang inscription to the second half of the seventh century. It is really impossible to be sure, so a broad date bracket 650-750 ce is perhaps advisable.
Bijdragen tot de taal-, land-en volkenkunde 177 (2021)  It was no doubt Boechari who, in a parenthetic remark in sni ii (1990:124 [2008:156]), mentioned the similarity of shape of the virāma signs in the Kanjuruhan and Śaṅkara inscriptions and assigned this shared feature to the middle of the eighth century. But in fact the inscription shows two different shapes: contrast its shape in vyayārtham· in line 1 with that seen in kavitvam· in line 15, dharmam· in line 16, and jīvyāt· at the very end. And in his review of De Casparis's work, Damais (1968:504) had already pointed out that this pointer is not useful for narrowing down any specific dating between the middle of the eighth and the middle of the ninth century ce.
I do not believe that the problem can be solved exclusively on the basis of palaeographic arguments and would like to draw attention to the relevance of text-internal elements. The first is the very elusive possibility, mentioned above, that arāti-nihan in stanza v is to be connected with epithets of similar meaning in the Śailendra inscriptions which date to the last quarter of the eighth century ce. The second is the equally elusive possibility of identification of our protagonist Śaṅkara with one of the Śaṅkaras from the inscriptions, with better chronological and geographical grounding, listed above. Even the earliest of these (Kayumwungan, 746 Śaka / 824 ce) would situate us three quarters of a century later than Boechari's estimate.30 The third is the fact that we now have three Sanskrit inscriptions from Central Java-Hampran, Balekambang, and Śaṅkara-which share the characteristic of expressing a wish for long life in their final stanzas.31 Of these three, only the Hampran inscription bears a date, which Damais (1970:43) thought to be equivalent to 750 ce. Personally, I am inclined to attach chronological significance to this last element, and provisionally to retain Boechari's dating estimate, although it is impossible to exclude that a poet several decades or even a century later would have chosen to express the same theme, while the exceptionally monumental character of the script used for our inscription renders palaeographic comparison with any of the other early Kawi-script inscriptions from Java methodologically problematic, so that a later date cannot be excluded on palaeographic grounds either. Until new data become available that will make it possible to narrow it down, it is necessary to allow a wide date bracket, say 730-830 ce. 30 I do not have access to a usable reproduction of the Kayumwungan stele to compare its script with that of our inscription. From Sundberg's comparison (2008:116-9) of the script of the Kayumwungan stela with that of the short inscriptions on the reliefs of the hidden base of Borobudur, I gain the impression of a script style somewhat different from that which we are trying to date in the Śaṅkara inscription. But a difference in style does not necessarily mean a difference of date. 31 See n. 16 above.

Conclusions
The interpretation of the inscription in Sejarah nasional Indonesia reveals a number of tendencies that reflect the comparatively unsophisticated level of historical analysis that was the norm in the 1970s and 1980s and that is, alas, still the norm today in studies of the early history of Indonesia. These are: (1) an unwillingness to admit that the pieces of the historical puzzle that have been passed down to the present are more likely to be disconnected from each other than to be able to be woven together into a coherent historical narrative; (2) a belief that all important inscriptions were issued by monarchs; and (3) the assumption that royal patronage of religious institutions was to the benefit of either Śaiva or Buddhist establishments and that simultaneous patronage of more than one religion was exceptional.32 His framework of historical analysis being determined by such misconceptions, and insufficiently controlled by his ability to understand what the Sanskrit text actually says, Boechari could present as 'epigraphical evidence' a series of ideas that constitute an almost complete misrepresentation of the data. The larger-scale narrative built up on such a basis, seen in the second paragraph of the following quotation (sni ii, 1984:109 [= 2008:138]), is therefore very misleading: From the above analysis, we can imagine that the king of Matarām Saṅ Ratu Sañjaya had built back up the kingdom Ho-ling after King Sanna had fallen in battle due to a hostile attack, and the centre of his kingdom has been destroyed. In the year 717 ce, Sañjaya was inaugurated as king of Mǝḍaṅ, which was possibly situated at Poh Pitu. In the year 732 ce he established a sanctuary for the worship of the liṅga on top of the Gunung Wukir, as a symbol of the fact that he had once again subjugated the minor kings around him, who had formerly acknowledged the suzerainty of King Sanna. But at a certain moment he fell ill and died, in tremendous suffering during eight days, because he wanted to comply with what his teacher had said. His son, called Śaṅkara, or possibly in complete form Rakai Panaṅka-32 With regard to point 1, see the seminal article of Henige (1975)  ran Dyah Śaṅkara Śrī Śaṅgrāmadhananjaya, because he was in fear of the teacher who was wrong, then abandoned Śaivism, becoming a follower of Mahāyāna Buddhism, and moved the centre of his kingdom eastwards, possibly around Sragen at the eastern side of Bengawan Solo, or to the area of Purwodadi/Grobogan.
My new reading, translation, and interpretation mean that the inscription loses any pertinence to dynastic history, or to the existence of Buddhist vestiges in the Sragen area.33 The protagonist of this inscription was called Śaṅkara and not Śaṅkhara,34 and his lifetime cannot confidently be assigned to the middle of the eighth century. This figure was not a king but a Śaiva poet with a very sound command of Sanskrit, who possibly served as abbot of a community of Śaiva mendicants (bhikṣusaṅgha), or as temple manager-depending on what kulapati means. The inscription records how he fulfils a promise made to his father to carry out the pious foundation of a temple (prāsāda), presumably to house a liṅga as representation of Śiva, a private foundation but one that appears to have been perceived as somehow supportive of royal power. It thus constitutes one of the very few non-royal Sanskrit inscriptions of Java. And it offers us a glimpse of the religious motivations and the financial aspects of such a foundation. Finally, we gain with this inscription a precious piece of evidence for a refined local practice of Sanskrit epigraphy that seems to be on par with the contemporary tradition in Cambodia.35 33 On these Buddhist vestiges, see Boechari 2012:200-1. 34 See the editorial note * drafted by me on p. 474 of Boechari 2012: 'Melihat bacaan di sini, dengan jagadaśivaharāc chaṃkarāc chaṃkarākhyaḥ sebagai hasil sandhi untuk jagadaśivaharāt śaṃkarāt śaṃkarākhyaḥ, agaklah mengherankan bahwa secara konsisten di seluruh kesempatan Boechari merujuk pada prasasti ini beliau salah menulis nama tokoh ini dengan ejaan Śaṅkhara. Kesalahan tak sengaja ini kami perbaiki di segenap kasus di dalam buku ini.' See also my n. 6 above. 35 Pollock (2006:125-32) has given a useful comparative overview of the history of the use of the Sanskrit language in Cambodia and on Java. However, I believe he was quite mistaken about Cambodia and not entirely correct regarding Java in noting the following about the early Sanskrit inscriptions of these cultures (2006:130): 'As in the case of Khmer inscriptions, these Javanese texts are all royal records; inscriptional practice seems not to have extended outside the court-one significant difference from the Sanskrit cultural order in South Asia, where nonroyal records abound.' As other examples of non-royal (or at least not explicitly royal) Sanskrit inscriptions from early Java, I cite the Tuk Mas inscription, as well as the Hampran and Balekambang inscriptions that have already been mentioned above. For instances of non-royal Sanskrit inscriptions from Cambodia, I refer to Goodall 2015 and 2019.