In another venue (Sundberg 2009: 310, n. 45), I have written of my inability to confirm de Casparis' (1956: 312) published reading of the word vālaputra on the Indonesian National Museum's inscription D28, commonly known as ‘Śivagṛha’, the ad 856 funereal stele of King Pikatan. Because his reading stands at the basis of a nexus de Casparis' claims about the history of the Ratu Boko plateau and his interpretation of an end to Buddhist Śailendra rule in Java, it is worth some energy to provide a minute analysis for those interested specialists who read Kawi.

My notes on my original observations on are no longer available to me, but I can offer an analysis de novo to refute de Casparis' reading thanks to an image of the pertinent section of the stone which was kindly furnished by Andrea Acri. The reading of the first six akṣara s of line 10 (marked by the red circle on the photograph) per de Casparis (1956: 312) is ‘hī vā la pu tra//’. In order to register my comments and objections— which are by no means as skillful or nuanced as I would have wished because I have not practised reading the script in over a decade—to most of de Casparis' transliteration, I have imposed Roman letters on to the photograph and will take up the pertinent features of the stone in turn.

1. (A) The concentric arcs to the right of the ‘A’ must be the elements that de Casparis reports as a ha, a character which in Kawi assumes the serpentine form ഗ. The proper form is obviously lacking on the stone.

2. (B) De Casparis reads a long vowel ī. While it is not out of the question that the open semicircle does indeed represent a superscript i rather than a subscripted ga from the line above (nu- merous examples of such an open form exist on the ad 792 inscription of Manjuśrīgṛha), there is no indication of a vowel-lengthening element. (For reference to how the i-vowel is lengthened in Manjuśrīgṛha, see the twenty-first akṣara of line 7).