National Institute of Informatics - Digital Silk Road Project
Digital Archive of Toyo Bunko Rare Books

showing the type of the S'aka or early Kuṣana period in India, and as being of older appearance
than that I remembered from the single tablet with Brāhmī text, N. xx. 1, excavated on my first
visit to the Niya Site.¹

M. Boyer's
decipher-
ment of
Brāhmī
record. When, early in 1917, I was able to turn my attention to this little relic of Indian writing and
presumably Indian language from the Han Wall, the analogy of the silk strip T. xv. a. i. 3, with its
Chinese inscription and its almost identical breadth, led me to hazard the conjecture that the Brāhmī
legend, too, might prove to contain some record descriptive of the roll of silk fabric from the edge
of which this strip had been cut off. But the only support I could see for it was in the Sans-
krit word paṭa (paṭṭa), 'piece [of fabric]', which appeared in Dr. Hoernle's tentative transcript
of the otherwise unintelligible legend as supplied in his Appendix F. Remembrance of the
most valuable help received from M. Boyer in respect of the Kharoṣṭhī inscriptions at Mirān
made me turn once more to this exceptionally qualified collaborator. His painstaking scrutiny
of the legend, aided by repeated examination in the original of the few characters not abso-
lutely clear in the reproduction, has been rewarded by gratifying results. M. Boyer, in letters
of March 13 and April 4, 1917, determined the reading as :

[ai]ṣṭasya paṭa giṣṭi saparīśa.

Mention of
'forty-six
spans'. Apart from the first word, in which the initial akṣara, owing to a hole in the silk, is incom-
plete and hence not quite certain, he was able to interpret the short record in a manner which appears
to me philologically very convincing, and which accords remarkably well with archaeological
considerations. Accepting paṭa in the obvious sense of 'piece [of cloth]', already recognized by
Dr. Hoernle, he takes saparīśa as a Prākrit equivalent of Skr. ṣaṭ+catvāriṃśat, 'forty-six'. The
fact that in the Prākrit of the Kharoṣṭhī tablets from the Niya Site catvāriṃśat 'forty' appears
as caparīśa, and that in Pāli the same decimal numeral is contracted from cattālīsaṃ into tālīsaṃ
when compounded with single numerals (e. g. in cuttālīsaṃ, 'forty-four'), makes this interpretation
of saparīśa phonetically quite acceptable. This reading of the word as a number necessarily suggests
that the preceding word giṣṭi may designate a measure. No such term is found in Sanskrit, but it is
just from such a form that we can most appropriately derive the word gitth, gith, meaning 'span',
which M. Boyer quotes from modern Panjābī,² and which, as Sir George Grierson has been good
enough to point out to me, is found also in Kashmiri gīth, with the confusion between cerebral and
dental typical in Dardic or 'Piśāca' languages.

Record
notes length
of silk roll. This interpretation of giṣṭi appears to me all the more convincing because, if the record on the
edge of the silk piece referred to the round roll of silk itself—it is always in this rolled form that
silk is carried in Chinese trade nowadays, just as the roll L.A. 1. 002 proves it for antiquity— there
was an obvious reason for its showing the length of the piece. The other details which the Chinese
inscription on T. xv. a. i. 3, Doc., No. 539, records, about its weight, price, etc., were not always essential,
especially for the foreign trader carrying his purchased goods to distant countries with different
measures, money, etc. The width of the silk was always visible to him and his purchasers without
opening the roll of silk. But the length he had certainly to note for his own convenience, if the
troublesome unrolling was to be avoided on every occasion. In short, while the Chinese inscription
is such as would naturally recommend itself to the producer or wholesale exporter of the fabric as a
guiding record, the Brāhmī note, in a strange script and language, was just a brief memorandum in-
tended by the trader from the West for his own guidance. Chinese inscriptions of similar length and