The Javanese rulers, however, did not only cover their island with caṇḍi śilā , “temples of stone”, but they also commissioned scores of poets and other literati to create what they called caṇḍi bhāṣa, “temples of language”.^[10] Building “language temples” turned out to be a wiser decision than covering the whole of Yawabhūmi with “stone temples”, and even more so than hoarding fabulous wealth as the rulers of Śrīwijaya did in Suwarṇabhūmi (“Land of Gold”, i.e. Sumatra).

The oldest known dated Old Javanese inscription is from AD 804, that is almost one and a half centuries after the Old Malay inscriptions of Śrīwijaya. But, unlike Śrīwijaya where no more Old Malay inscriptions were apparently issued, hundreds of Old Javanese inscriptions have been found in Java, covering a period of more than six centuries. Since there must have been some connection between the level of literacy and the issuing of charters in the vernacular language, the abundance of the Old Javanese inscriptions must be an indication of the extent of literacy in ninth-century Java. This is perhaps only to be expected. During the seventh century, Walaing (Chinese: Holing) was apparently already such a famous centre of learning that a Chinese scholar went there to translate Buddhist texts into Chinese under the guidance of a Javanese guru. The erection of the Buddhist Borobudur and the śivaite Prambanan temple complexes, probably during the first half of the ninth century (Soekmono 1979:457-472; de Casparis 1956:309-311), shows that the dynastic changes that occurred in central Java after Sañjaya’s accession to the throne did not diminish the importance of the region as a centre of learning. The thousands of bas-relief sculptures of Borobudur, which are based on various Buddhist texts (Krom 1927; Bernet Kempers 1980), and those of Prambanan based on a version of the Rāma saga (Stutterheim 1989), are clear testimony of continued vigorous study of literary texts.

It is not surprising, therefore, that the oldest “temple of language” that has come down to us, the Rāmāyaṇa kakawin, was also the product of the central Javanese period.^[11] Surprisingly, this poem is based neither on the best-known version of the Rāma sagas, namely Valmiki’s epic, nor on the version depicted in the bas-reliefs of Prambanan temple, but on the Rāvaṇavadha, “The slaying of Rāvaṇa”, a sixth or seventh century poem by an Indian poet named Bhaṭi. The choice of Bhaṭi’s poem, rather than Valmiki’s epic, to serve as the poet’s model in writing his poem, is remarkable. For the latter is not only the best-known version of all the Rāma sagas, but its language is also much easier than that of the former. Bhaṭi himself says of his work that this poem “can be understood only by a commentary; it suffices that it is a feast for the clever and that the stupid come to grief in it as a result of my love for learning” (Keith 1956:116). Whatever reasons prompted the choice, however, the completed result was without doubt a masterpiece, the gem of all that has been produced by the Old Javanese kawi, “poets”. To later generations it became the ādikakawin, that is the first and foremost among the Old Javanese poems (Hooykaas 1958). It is also testimony of the high level of scholarship that must have existed in central Java at that time. The poet’s ability to grapple with a text which posed so many problems clearly shows that his knowledge of Sanskrit must have been considerable and that he must have had complete mastery of his own language to have been able to render this difficult Sanskrit prototype satisfactorily. With the creation of both temples of stone and of language of the magnitude of the Borobudur and the Old Javanese Rāmāyaṇa, it is no wonder that Java of the ninth century continued to attract people from other countries. An inscription from Champa, for instance, tells us that a high official from what is today the central coast of Vietnam went on a pilgrimage to Java “to acquire the magical science”. “The pilgrim’s quest”, comments Wolters (1982:27), “may epitomise Java’s reputation for possessing esoteric knowledge”.

The obvious advantage of the “temple of language” over “temple of stone” is of course its transportability. Once completed, a book, or more likely its copies, could be transported to the remotest part of the country. Even in those days it was apparently not uncommon to purchase books and build up some kind of a library in one’s residence in the countryside. Prapañca tells us in the Nāgarakṛtāgama 29.2 that his friend, Kṛtayaṛa, who lived in a village some distance from the capital, used to occupy himself “with the appraisal of kīrti (valuable) books. Having been bought they were well taken care of, put into safe keeping” (Pigeaud 1960-63). In fact, one does not have to transport literature in its physical form to convey its message. Once its contents have been mastered, one can leave the book aside and relay its message orally to audiences anywhere. In this manner, throughout the long history of Sanskritization, literature became the most potent instrument in the dissemination of the Sanskritic culture. The Mahābhārata and the Rāmāyaṇa epics in particular played a crucial role in spreading this culture from the confined walls of the kraton to the countryside. As Srinivas (1989:61) observes in the context of Sanskritization in India, the epics “have not only transmitted to the people a knowledge of the great gods of Hinduism and certain basic theological ideas, but have also helped to spread common culture throughout the century. The epics, and the innumerable stories included in them, constitute the foundation of the literature in every Indian language. The fact that the institution of harikatha, or public reading of the epics and the Pūraṇas by trained masters of the art, was a popular pastime made it possible for Sanskritic Hinduism to reach even the illiterate masses.”^[12]

The earliest evidence of the penetration of the Sanskrit epics into rural areas is found in the Sangsang copper plate inscription issued in the name of King Balitung in AD 907 to confirm the granting of freehold to the monastery of Dalinan (Sarkar 1971-72:II, 85-98). To celebrate what must have been the most important event of the year, if not of decades, for the whole region, a variety of performances was given on that occasion. Apart from singing (mamidu), dancing (mangigal), a Kicaka dance, play-acting (mamirus) and clowning (mabañol), the inscription also mentions a wayang performance (mawayang) of the story of “Bhīmaya-kumara”, and a recital (macarita) of the “Bhīma-kumara” and “Rāmāyaṇa” stories. Both “Kicaka” and the “Bhīma-kumara” obviously refer to a certain episode from the Wirā ṭaparwa, the fourth book of the Mahābhārata (Zoetmulder 1974:208-209). Since there is no evidence that a Javanese translation or adaptation of the Mahābhārata already existed at that time, it seems likely that the narrator of the Bhīma-kumara episode recited a Sanskrit text, and then, as in a harikatha, explained it to his audience in Javanese. The mabasan in present-day Bali, in which people gather to listen to a recital of an Old Javanese poem and its interpretation in Balinese, no doubt originated from this kind of macarita.