Pramāṇa · Search · About · API

Indian Epigraphy: A Guide to the Study of Inscriptions in Sanskrit, Prakrit, and the Other Indo-Aryan Languages c10.s3: 5.4 Dating of Inscriptions

Indian Epigraphy: A Guide to the Study of Inscriptions in Sanskrit, Prakrit, and the Other Indo-Aryan Languages c10.s3

Language
English

Contents

5.4 Dating of Inscriptions

5.4.1 Undated or inadequately dated inscriptions; paleographic dating and problems thereof

c10.s3.t1.1

K. V. Ramesh, "The [F*]utility and (F)utility of Palaeography in Dating Undated Inscriptions," *Studies in Indian Epigraphy* [=JESI] 3, 156–62; RIE 51–62 (ch. V: "Scriptal Ramifications and Palaeographical Dating of Undated Inscriptions," a revised and expanded version of the preceding).

c10.s3.t1.2

In the cases—all too frequent, especially among earlier inscriptions—where a record has no explicit date, or only a date whose absolute value cannot be definitely ascertained, epigraphists must resort to estimating the date on the basis of paleographic analysis by comparing the script of the undated record with specimens of similar scripts from other dated or datable inscriptions. Because this method is inherently imprecise, paleographic dating should in general be treated as a last resort. Other dating techniques, such as historical analysis based on references to persons or events of known date, or chronological analysis based on linguistic or archaeological data, may supersede or be used in conjunction with paleographic estimates. But the fact remains that, since so many inscriptions are undated or inadequately dated, paleographic dating, imperfect as it may be, must be employed in a great many cases.

c10.s3.t1.3

The inherent imprecision of this method alluded to earlier is due mainly to the variability in writing styles and forms at any given place and time in history. Though it is true that in the long run scripts, like languages, follow more or less consistent courses of development, this does not mean that at any given time and place one and only one form of a script is in use. On the contrary, there will always be a variety of script forms for different contexts and purposes in simultaneous use in any particular place; and the stage and pace of development will by no means always be the same in different, even adjoining regions. Individual and idiosyncratic variations must also be kept in mind; for example, in the past, as in modern times, an older person tended to write in a more old-fashioned style than a younger contemporary.²¹ But despite these and other limiting factors, it has often been the practice to assign relatively precise dates to undated inscriptions on the basis of simple paleographic comparisons, which are sometimes chosen arbitrarily and evaluated in a superficial manner.

c10.s3.t1.4

²¹ Cf. Woodhead, *The Study of Greek Inscriptions*, 29 and 62–3. The limitations of paleographic dating are by no means peculiar to India but rather constitute a general principle of paleographic studies.

c10.s3.t1.5

Given the relatively undeveloped state of comparative paleography in India, as well as the inherent limitations of the method, it will not be surprising that the results obtained are often unsatisfactory. Any number of instances could be cited where widely varying dates have been proposed by reputable scholars for the same inscription. For instance, the undated Võ-cạnh inscription, the earliest Sanskrit inscription from Southeast Asia (see 4.3.7.8), has been dated on paleographic grounds to anywhere from the third to the fifth century A.D.²² The situation is hardly any better in the case of inscriptions with dates in regnal years of unknown rulers or in unspecified eras. The year 30 of an unspecified era in the Muṇḍeśvarī inscription of the time of Udayasena, for example, was attributed by R. D. Banerji (EI 9, 285 and 290) to the Harṣa era, equivalent to A.D. 635, while N. G. Majumdar (IA 49, 25) assigned it to the Gupta era and dated the inscription to A.D. 348–49.²³ Ramesh (RIE 55–7) cites several other cases where dates for a given inscription arrived at by paleographic estimate vary by two centuries or more. The problem is particularly acute in connection with the dating of the earliest inscriptions (other than those of Aśoka), where, due to the paucity and brevity of the materials, all undated, it is all but impossible to establish a convincing relative, let alone absolute, sequence for paleographic dating. Moreover, given the inherent flexibility of paleographic dating, there is often a temptation (not always resisted) to manipulate it in order to confirm a preconceived notion of an inscription's date or to prove some historical argument (cf. RIE 55).

c10.s3.t1.6

²² See 4.3.7.8 and references given there in n. 117; see esp. Kamaleshwar Bhattacharya, *Artibus Asiae* 24, 1961, 219–24.

c10.s3.t1.7

²³ See also more recently G. S. Gai in IEBHA 129–31, who would place the inscription in the early to middle sixth century.

c10.s3.t1.8

The dating of inscriptions by paleographic estimate alone is thus an approximate method at best, even as carried out by reputed experts in the field. In recent years a few scholars, notably Ramesh and A. H. Dani in his *Indian Palaeography*, have attempted to refine our knowledge of the regional, local, and stylistic subvarieties of the various Indian scripts. Ramesh has emphasized the importance of distinguishing "settled and stylized 'urban' writing" from " 'rural' or rustic paleography" (RIE 58–9; cf. JESI 3, 157–61). Inscriptions in the latter style, according to Ramesh, tend to appear to be later in date than contemporary records in more formal writing.²⁴ He therefore concludes (RIE 62; cf. JESI 3, 161–2) that in estimating the date of an inscription on paleographic grounds one must take into consideration first of all the nature of the writing, whether "urban" or "rural," and make suitable adjustments; and second, one must also keep in mind considerations of "dynastic, regional or religious influence" on the relative speed or conservatism in paleographic developments. Finally, he declares, "Where palaeography is the lone consideration in suggesting a date for an early inscription which contains no other sort of supplementary internal evidence, the suggested date must always be taken to represent the date so suggested + or –100 years" (ibid.) The examples cited above of widely differing dates for particular inscriptions would seem to confirm this rule, since the estimates tend to vary over a range of roughly two hundred years.

c10.s3.t1.9

²⁴ Chronological illusions created by differences between formal and informal styles are also by no means peculiar to India; see Woodhead, *The Study of Greek Inscriptions*, 64.

c10.s3.t1.10

Dani, too, has emphasized the limitations of paleographic dating, warning against the common practice of relying on one or a few "test letters" for paleographic dating, since "we must allow a margin for the transition period in which the new and old forms were used simultaneously" (DIP, 132). Paleographic dating "can never supply infallible dates" (ix); "this is too much to expect from palaeography" (11).

c10.s3.t1.11

Thus while it is true that in certain cases, where a relatively large and cohesive body of securely dated material is available for comparison, a somewhat higher level of precision may be attainable, as a general rule of thumb it would be wise to adopt Ramesh's principle of plus or minus one hundred years for the range of accuracy of paleographic dating. In the past many epigraphists have paleographically dated inscriptions to a range as short as a quarter of a century, but such precise claims should not be uncritically accepted.

5.4.2 Dated inscriptions

c10.s3.t2.1

Inscriptions may be dated in any number of ways, ranging from a simple year date (regnal or era) to detailed specifications of the year, month, *tithi* (lunar day), weekday, and/or other calendrical and astronomical data. From the point of view of epigraphic methodology, it is convenient to discuss the systems of year dating and those of more specific dates separately.

c10.s3.t2.2

5.4.2.1 Year dates

c10.s3.t2.3

*5.4.2.1.1 Regnal years* The practice of dating records in the regnal years²⁵ of a king, as opposed to years of a continuous era, was widespread in early centuries throughout India, beginning with the inscriptions of Aśoka. In the early centuries of the Christian era several dynasties such as the Sātavāhanas, Ikṣvākus, and Vākāṭakas continued to record only regnal years in their inscriptions, and in the medieval period the Pālas and some other eastern dynasties and the Pallavas, Coḷas, and Pāṇḍyas in the south still used regnal years exclusively or primarily.

c10.s3.t2.4

²⁵ The precise calendrical significance of a "regnal year" is not always completely clear. That is to say, we do not always know on exactly what date or occasion (accession to the throne, formal coronation, etc.) the regnal year began, and how this regnal year was correlated with the current calendrical year. In some cases, this seems to have produced calendrical complications; see, e.g., the cases noted in "Regnal Years" (IA 39, 1910, 217–24), and below in connection with the Cālukya-Vikrama era (5.5.1.13).

c10.s3.t2.5

A typical simple formulation of a regnal year date would contain the ruler's name and titles in the genitive case with a word for 'year' in the locative, such as *saṃvatsare* (often abbreviated *saṃvat* or *saṃ*) or less commonly *varṣe* (or their Prakrit equivalents), plus the date in words, numerical figures, or both; for example, *ūdākasa dasama-savachare*, "in the tenth year of Ūdāka," in the Pabhosā cave inscription (SI I.95–7).²⁶ More elaborate formulae, particularly in Sanskrit verse inscriptions, may involve other constructions such as the locative absolute, as in the Eraṇ boar inscription of Toramāṇa (SI I.420–2): *varṣe prathame pṛthivīm pṛthukīrttau pṛthudyutau / mahārājādhirāja-śrī-Toramāṇe praśāsati* . . . ("While the Great King of Kings, the Illustrious Toramāṇa of vast fame and glory was ruling the Earth, in (his) first year . . ."). Further details of the date such as month, *tithi*, weekday, and so on, may or may not be included with the regnal year (see 5.4.2.2).²⁷

c10.s3.t2.6

²⁶ Unusual types of early regnal dates include the Bhārhut pillar inscription (CII 2.2, 11–12) dated only "during the reign of the Śuṅgas" (*suganaṃ raje*) without a specific year, and the Hāthīgumphā inscription (SI I.213–21) which chronicles the king's reign year by year (*padhame vase* . . . *dutiye ca vase* . . . *tatiye puna vase* . . . , etc.).

c10.s3.t2.7

²⁷ Some medieval and later records from Orissa recorded regnal years counted according to the *aṅka* system in which the first year and all years whose unit figure is 6 or 0, except for 10, are skipped over (i.e., the years 1, 6, 16, 20, 26, etc.); see IA 19, 1890, 255–6, and Monmohan Chakravarti, JASB 72, part 1, 1903, 100. Thus, e.g., a period enumerated as twenty-five regnal years according to the *aṅka* system corresponds to only twenty-one normal years (EI 28, 1949–50, 242). There were evidently several variations of this system, which is sometimes said to be based on a sixty-year cycle; see further G. Ramadas, JBORS 17, 1931, 179–80; R. Subrahmanyam, *Inscriptions of the Sūryavaṃśī Gajapatis of Orissa*, xxxi–ii; and A. K. Bhattacharyya, *A Corpus of Dedicatory Inscriptions from Temples of West Bengal*, Appendix B, 190–1.

c10.s3.t2.8

Unless, as is occasionally the case, such regnal dates are accompanied by a second date in an established era, they do not by themselves provide an absolute date for the record. Such regnally dated inscriptions must therefore be subjected to the same techniques of dating by paleographic or other methods as are applied to undated inscriptions. It is often possible, however, to establish approximate or occasionally even exact values for such dates by means of historical synchronizations for the rulers concerned. Thus in the case of the regnally dated Aśokan inscriptions, reasonably accurate absolute dates have been established by the synchronization of Aśoka's regnal years with the contemporary Greek kings to whom he refers in his rock edict XIII (see 4.3.1.1 and 7.1.1.1). Among the regnally dated inscriptions of various dynasties of the early Christian era and later, we often have a fair number of records of successive rulers on the basis of which a dynastic succession list can be drawn up. The relative chronology of the dynasty can then often be correlated with known absolute dates; thus the regnal dates of the Vākāṭakas can be established with fair precision on the basis of their relations with the Guptas, and the Pālas can be approximately dated by their interactions with the Gurjara-Pratīhāras and Rāṣṭrakūṭas. (In each of these cases, of course, the latter-mentioned dynasty dated its records in known historical eras.) Even where such historical synchronizations are not available, paleographic dating may be applied as a supplementary technique, and when we have a sequence of regnally dated records of a particular dynasty this method is considerably more reliable than when applied to a single record or a few isolated records.

c10.s3.t2.9

In this way, regnally dated inscriptions can often be ascribed absolute dates with relative precision, at least as compared to undated inscriptions. The chronological reconstruction of a dynasty from regnally dated records is of course still only approximative, as the period of reign for each king can be given only as a minimum, based on the highest regnal date preserved for him in known inscriptions; the actual time of his rule, assumed to be as great as, or more likely greater than, this figure, can only be estimated (cf. 7.1.1.1).

c10.s3.t2.10

*5.4.2.1.2 Years of an era.* Beginning in the first century B.C. some inscriptions were dated in continuous eras.²⁸ This system gradually became more prevalent and was the standard practice in most regions by the early centuries of the medieval era. These continuous or historical eras evidently arose as continuations or extensions of a regnal reckoning of a given king by his successors, who instead of beginning a new reckoning commencing from their own accessions continued counting years from the previous king's dates (SIE 242–3). This development is explicitly indicated in the Mathurā pillar inscription (SI I.277–9) dated in the fifth regnal year of Candragupta II (*. . . śrīcandraguptasya vija(*ya)-rājya-saṁvatsa(*re) paṁcame 5*) and in the sixty-first year of the "continued reckoning" (*kālānuvarttamāna-saṁvatsare ekaṣaṣṭhe* [*sic*] *60 1*), that is, of the Gupta era (5.5.1.6).²⁹ Such an extended regnal-year system would then become a dynastic reckoning, and this in turn would often continue in use, sometimes with a change of name, long after the founding dynasty had fallen.³⁰

c10.s3.t2.11

²⁸ Systems of continuous historical reckoning seem not to have been an indigenous Indian notion, as the earliest genuine historical eras were founded by foreign rulers. But once established, such systems were initiated by many later dynasties of Indian origin (SIE 235–50).

c10.s3.t2.12

²⁹ Cf. also a similar dating formula in a newly discovered inscription of [Gupta] year 121 from Mathurā (Govindnagar): *saṁvatsaraśate ekaviṁśottaraśate* [*sic*] *100 20 1 kālānuvarttamānasaṁvatsare . . .* (EI 40, 1973–74, 20–2).

c10.s3.t2.13

³⁰ See the various examples discussed in section 5.5, particularly the Vikrama and Śaka eras (5.5.1.3 and 5.5.1.4).

c10.s3.t2.14

The name of the era is often but by no means always explicitly noted. In the early dates of a given era, the name was often

c10.s3.t2.15

5.4.2.2 Specific dates

c10.s3.t2.16

Beginning in the last two centuries B.C., the practice of recording dates by the day as well as year gradually came into use. The earliest examples, mostly in inscriptions of Indo-Greek and Indo-Scythian kings, usually give a (regnal) year, month, and day of the month.³⁴ The Shinkoṭ reliquary inscriptions of the time of Menander and Vijayamitra (SI I.105, l. 2), for example, contain the date *vaṣaye paṁcamaye 4 1 veśrakhasa masasa divasa paṁcaviśraye*, "In the fifth (5) [regnal] year [of Vijayamitra], twenty-fifth day of the month Vaiśākha." Several inscriptions from this period use Macedonian rather than Indian months; the Sui Vihār inscription of the eleventh year of Kaniṣka, for example (SI I.139, l. 1), is dated on the twenty-eighth day of *daïsiṁka masa*, that is, of Daisios.³⁵

c10.s3.t2.17

³⁴ Occasional examples of this dating system are found as late as the Gupta period; see David Pingree, "A Note on the Calendars Used in Early Indian Inscriptions," *JAOS* 102, 1982, 355–9 (esp. 357).

c10.s3.t2.18

³⁵ See Pingree, op. cit., 355–6, for further details and examples.

c10.s3.t2.19

Some inscriptions of the first to fifth centuries A.D., mostly Śaka and Kuṣāṇa inscriptions in Brāhmī from the Mathurā region, are dated in year (regnal or era), season (*grīṣma*, *varṣā*, or *hemanta*), number of the month in the season, and day of the month. For instance, the Mathurā Jaina image inscription of the time of Huviṣka (SI I.155–6) is dated *savasare 40 4 . . . gṛsyamasa* [= *grīṣmamāsa*] *3 divasa 2*, "In the year 44, month 3 of summer, day 2." Pingree describes this system as "an amalgamation of the Indian tradition of three seasons (*ṛtus*) . . . with the northwestern tradition of counting twenty-nine or thirty days in each month" (op. cit., 357).³⁶

c10.s3.t2.20

³⁶ Cf. Sircar, SI I.121 n. 3, characterizing this system as "a compromise between the Greek . . . and Indian . . . systems of dating," and SIE 246.

c10.s3.t2.21

Inscriptions of approximately the same period from the Deccan, the south, and the northeast are usually dated by year, season, number of the lunar fortnight (*pakṣa*) [rather than month] within the season, and day of the fortnight. For example, the Kārle cave inscription of the Sātavāhana king Vāsiṣṭhīputra Puḷumāvi (SI I.210) is dated *savachare catuvise 20 4 hemaṁtāna pakhe tatiye 3 divase bitiye 2*, "In the twenty-fourth (24) [regnal] year, the third (3) fortnight of winter, the second (2) day." Dates of this type are common among inscriptions of the Sātavāhanas and Ikṣvākus in the Deccan as well as among northern kingdoms such as Maghas of Kauśāmbī and the kings of Bāndhogarh (Pingree, op. cit., 358).

c10.s3.t2.22

The classical dating system, with the year, month, fortnight of the lunar³⁷ month (*pakṣa*), and lunar day (*tithi*)³⁸ of the fortnight,³⁹ first appears sporadically in inscriptions of the early centuries A.D., mainly those of the Western Kṣatrapas. For example, the Guṇḍā inscription (SI I.181–2) is dated *varṣe triyuttaraśate* [*sic*] *100 3 vaiśākhaśuddhe paṁcama . . . tithau*, "in the year one hundred and three (103), in the bright (fortnight) of Vaiśākha, on the fifth *tithi*." This system gradually spread, first through the Deccan and central India, and eventually throughout the subcontinent; by the Gupta period it appears to have become more or less standard.⁴⁰ The Udayagiri inscription of the time of Candragupta II (SI I.279), for instance, is dated *saṃvatsare 80 2 āṣāḍhamāsaśukle*[read *-ai*]*kadaśyām*, "In the [Gupta] year 82, on the eleventh [*tithi*] of the bright [fortnight] of the month of Āṣāḍha."

c10.s3.t2.23

³⁷ In later centuries in certain parts of India, mainly Bengal and the far south, the solar months (Meṣa, Vṛṣabha, Mithuna, etc.) were sometimes cited instead of lunar months; see SIE 223 and A. Venkatasubbiah, *Some Śaka Dates in Inscriptions*, 66–9.

c10.s3.t2.24

³⁸ On the precise definition of *tithi*, see the "Technical note on *tithis*" (5.4.3.3).

c10.s3.t2.25

³⁹ Occasionally the day of the full month, undivided into fortnights, is given, either in place of or in addition to the *tithi* of the fortnight; see EI 8, 1905–6, 289.

c10.s3.t2.26

⁴⁰ Details in Pingree, op. cit., 359.

c10.s3.t2.27

To this basic information additional data of varying types and amounts may be added. The *nakṣatra* occupied by the moon is often noted in later inscriptions but also occurs as early as the second century in the Gunḍā inscription noted previously (*rohiṇī*[sic]*-nakṣatra-muhūrtte*). The weekday (*vāra* or *vāsara*) is first clearly noted⁴¹ in the Eraṇ inscription of the time of Budhagupta, [Gupta] 165 = A.D. 484 (SI I.335, l. 3) in the phrase *suraguror ddivase*, "on Thursday." In subsequent centuries the citation of the weekday becomes more or less standard. (This practice is important in that it permits the verification of dates in later inscriptions; see 5.4.3.2.)

c10.s3.t2.28

⁴¹ But see also SIE 219 n. 2.

c10.s3.t2.29

Other types of calendrical and astronomical details sometimes added to dates of inscriptions of the Gupta period and later, especially in connection with dates in the Śaka era,⁴² include the *yoga* (period in which the combined longitudinal motion of the sun and moon equals one *nakṣatra* or 13° 20'), *karaṇa* (half-*tithi*), and *lagna* (zodiacal sign rising).⁴³ Sometimes still further data such as the season (*ṛtu*), *rāśi* (sun's zodiacal sign), half-year (*ayana*, i.e., *dakṣiṇāyana* or *uttarāyaṇa /udagayana*), and time of day in *ghaṭikās* ("hours" of forty-eight minutes) are provided. Special astronomical circumstances such as the sun's entry into zodiacal signs (*saṃkrānti/saṃkramaṇa*), especially the *viṣuva(t)-saṃkrānti* or entries into the equinoctial signs and the *dakṣiṇāyana*- and *uttarāyaṇa-saṃkrānti* or entries into the solstitial signs, may also be mentioned. Particularly common are references to solar or lunar eclipses (*grahaṇa*, *uparāga*), since these were considered especially auspicious times for the granting of gifts.⁴⁴ Citations of eclipses are useful for verifying dates, since their occurrences can be readily checked in various standard reference sources such as Th. Ritter von Oppolzer's *Canon der Finsternisse*.⁴⁵

c10.s3.t2.30

⁴² Kielhorn, IA 20, 1891, 412 = KS I.610.

c10.s3.t2.31

⁴³ See Fleet, IA 18, 1889, 162 n. 1.

c10.s3.t2.32

⁴⁴ F. Kielhorn, "Die Sonnen- und Mondfinsternisse in den Daten Indischer Inschriften," *Nachrichten von der Königlichen Gesellschaft der Wissenschaften zu Göttingen*, Phil.-Hist. Klasse, 1896, 59–75 = KS II.826–42. See also Venkatasubbiah, *Some Śaka Dates*, 19–27.

c10.s3.t2.33

⁴⁵ Wien: Kaiserliche Akademie der Wissenschaften, 1887. English translation, *Canon of Eclipses*, by Owen Gingerich (New York: Dover, 1962).

c10.s3.t2.34

A typical example of a complex date from a late inscription is that of the Ajmer slab inscription containing the *Harakeli-nāṭaka* (IA 20, 212), reading *saṃvat 1210 mārga-sudi 5 āditya-dine śravaṇa-nakṣatre makarasthe candre harṣaṇa-yoge bālava-karaṇe*, "In the [Vikrama] year 1210, the bright [fortnight of] Mārga[śīrsa], [*tithi*] 5, Sunday, in the *nakṣatra* Śravaṇa, with the moon in Makara, the *yoga* Harṣaṇa, and *karaṇa* Bālava."⁴⁶

c10.s3.t2.35

⁴⁶ For an even more elaborate date, including specific times of day, see the Burhānpūr mosque inscription (Appendix, no. 14), ll. 5–6. In the case of such precise dates, the day and time cited in the inscription is presumably that of the ceremonial performance of the gift, foundation, etc., commemorated in the inscription, rather than that of the actual incision of the inscription itself.

c10.s3.t2.36

Various abbreviations were often employed in the notation of dates. The seasons (in the older dating systems) were often abbreviated as *he* or *hema*, *va* or *vā*, and *gri* or *gī* for *hemanta*, *varṣā*, and *grīṣma* (or their Prakrit equivalents), respectively. The days of the bright (*śukla*/*śuddha*/*sita*) and dark (*kṛṣṇa*/*bahula*/*asita*) fortnights are commonly abbreviated as *śudi* or *sudi* (= *śukla-divase* etc.) and *badi*/*vadi* (= *bahula-divase* etc.) or *vadya*,⁴⁷ respectively. (For further examples of abbreviations, see BPLM 160 n. 5 and SIE 220.)

c10.s3.t2.37

In many inscriptions a formulaic phrase such as *etasyāṃ pūrvāyām*, “At this aforementioned [time (*velāyām*), or lunar day (*tithau*)],” *asyāṃ saṃvatsara-māsa-divasa-purvāyām*, “On this aforementioned year, month, and day,” or the like, would be added after the date by way of introducing the body of the document.⁴⁸

5.4.3 Conversion and verification of inscriptional dates

c10.s3.t3.1

H of an Indian year can be determined only if the nature of the year (current or expired) is known and the month is provided. Further complications arise in certain cases, for instance, in connection with eras such as the Kalacuri-Cedi or Lakṣmaṇasena (5.5.1.5 and 5.5.1.15), whose precise epochs are uncertain or variable.

c10.s3.t3.2

5.4.3.2 Conversion and verification of specific dates

c10.s3.t3.3

The determination of correspondences for specific dates involves complex calculations for which several methods are available, for example, those given by Jacobi in EI 1, 406–8 and by S. B. Dikshit in CII 3, intro., Appendix II, 145–59.⁵² But as a matter of convenience, the usual practice is to consult the tables provided in Swamikannu Pillai's *Indian Ephemeris*, providing equivalents in the European calendar for dates in the more common Indian eras which are adequate for general or preliminary purposes. But due to the complexities of Indian calendrical reckoning, they should not be treated as conclusive.⁵³

c10.s3.t3.4

⁵² These systems yield the corresponding "English" or "old-style" date, i.e., the date according to the Julian calendar which was current in England until September 2, 1752.

c10.s3.t3.5

⁵³ Mainly because Pillai's dates are based on mean beginnings of *tithis*, whereas the dates recorded in inscriptions were based on the real local beginnings of *tithis* (see the "Technical note on *tithis*," 5.4.3.3).

c10.s3.t3.6

It is therefore always desirable to verify an inscriptional date whenever sufficient data (at a minimum, year, month, *tithi*, and weekday) are provided; that is to say, the data should be checked and if necessary calculated to confirm (or deny) that the given weekday (or eclipse, *saṃkrānti*, etc.) did in fact fall on the specified date, thereby confirming (or denying) the tentative European equivalent for the date in question. The verification of a weekday and/or of other astronomical and calendrical specifications such as eclipses, *saṃkrānti*, *nakṣatra*, *yoga*, cyclical year of Jupiter (see 5.5.2.2 and 5.5.2.3), and so on involves complex calculations requiring specialized knowledge of astronomical systems, which are facilitated by the systems and tables provided by Jacobi (EI 1, 408–37) and the other sources listed at the beginning of section 5.4.3.⁵⁴ Such calculations, however, involve several variables and complicating factors which must be taken into consideration, and which are not always explicitly stated in the document itself. These include the era, which is often not specified; the nature of the years, whether expired or current, also frequently unspecified;⁵⁵ whether the year began in Caitra (*caitrādi*, the "northern" year) or Kārttika (*kārttikādi*, the "southern" year), or occasionally in some other month such as Āṣāḍha or Śrāvaṇa; and whether the months are counted as ending with the full moon (*pūrṇimānta*, the "northern" system) or the new moon (*amānta*, the "southern" system),⁵⁶ or whether they are solar rather than lunar months.⁵⁷ When any or all of this information is unknown or uncertain, multiple calculations must be made for each possibility before a final verdict is passed (Jacobi, 406).

c10.s3.t3.7

⁵⁴ For examples of the full procedure for calculation of the Western equivalents of Indian dates, see Jacobi, 408–12.

c10.s3.t3.8

⁵⁵ Or sometimes incorrectly specified; see Kielhorn, IA 25, 1896, 266 = *KS* II.727.

c10.s3.t3.9

⁵⁶ See Jacobi, EI 1, 406.

c10.s3.t3.10

⁵⁷ See n. 37. According to Venkatasubbiah (*Some Śaka Dates in Inscriptions*, 66–7), the names of the lunar months are sometimes applied to what are really solar months, and vice versa; such deviations can be determined only by calculation.

c10.s3.t3.11

*siddhānta*, *Ārya-siddhānta*, etc.) was used to calculate the *tithi*, due to their differing solar models and parameters for the movements of the heavenly bodies; the moment of the expiry of a *tithi*, which is critical to calculations of dates, may vary by as much as an hour or even more according to different *pakṣas* (Jacobi, EI I, 403). It is therefore necessary to check results by the different systems in any case where the result is in doubt, for example, where the weekday calculated by the general system falls one day before or after the one specified in the document, or where the end of the astronomical *tithi* falls within three hours of sunrise.⁶⁰ Moreover, in calculating the date for a given inscription adjustments must always be made for the conversion of mean Laṅkā time as given in the tables to true local time according to the latitude and longitude of the place of issue, insofar as this is known (Jacobi, 412; cf. n. 53). Here, too, the variations between mean time and local time may affect the moment of the end of the *tithi* and thus produce a different civil date.

c10.s3.t3.12

⁶⁰ Kielhorn, IA 19, 1890, 21 (= KS I.514) n. 1. Kielhorn (ibid., 354-7 = I.556-9) lists several cases in which recalculation according to different astronomical systems enabled him to verify a date which might otherwise have appeared incorrect, and thus to authenticate a record which might otherwise have wrongly been suspected to be spurious.

c10.s3.t3.13

By way of example of these and the other complications involved in the verification of inscriptional dates, we may cite the Cāndpur boar inscription (ASIR 10, 97; Kielhorn, IA 19, 354 = KS I.556), which is dated *sa[ṃ*]vat *1207 jyeṣṭha-vadi 11 ravau*, that is, Sunday, the eleventh day of the dark fortnight of Jyeṣṭha, [Vikrama] 1207. On the basis of the *Sūrya-siddhānta*,⁶¹ Kielhorn calculated the following possible equivalents of this date:

c10.s3.t3.14

⁶¹ From Jacobi's tables in IA 17, 1888, 145-81. These were superseded by his tables in EI 1, which were based on European systems rather than on any one Indian *pakṣa* (441).

c10.s3.t3.15

1. If the [Vikrama] year 1207 is taken as a northern (*caitrādi*) current year and the month as *pūrṇimānta*: Thursday, May 5, 1149 A.D.
2. Northern current year, *amānta* month: Friday, June 3, 1149
3. Northern expired year, *pūrṇimānta*: Tuesday, April 25, 1150
4. Northern expired year, *amānta*: Wednesday, May 24, 1150
5. Southern (*kārttikādi*) expired year, *pūrṇimānta*: Monday, May 14, 1151
6. Southern expired year, *amānta*: Tuesday, June 12, 1151

c10.s3.t3.16

None of these dates gives a satisfactory equivalent for the weekday of the inscription's date, namely, Sunday. However, when the fifth alternative (southern expired year, *pūrṇimānta*) is recalculated according to the *Brahma-siddhānta*, the civil day numbered by the eleventh *tithi* of the dark fortnight of Jyeṣṭha is the day *before* that arrived at by the *Sūrya-siddhānta*, namely, Sunday, May 13, 1151, and this is presumably the correct equivalent of the date of the inscription.

c10.s3.t3.17

Even when a date cannot be verified by the usual methods, further circumstances must be taken into consideration before dismissing it as "irregular." Kielhorn, for example,⁶² notes numerous instances in which the date of an inscription cannot be verified for the year given, but does work out for the preceding or following year, or sometimes for the preceding or following year but one. Such cases are so common that they probably do not indicate forged or erroneous dates but rather some kind of consistent calendrical irregularity or dislocation, perhaps resulting from confusion between expired and current reckonings (cf. Sewell, *Indian Calendar*, 40 n. 2). Such cases, and related problems of seemingly irregular dates, have been treated at length by Venkatasubbiah in *Some Śaka Dates in Inscriptions* (esp. 31–48 and 95–6), who suspects that they are connected with various complications involving the calculation by different systems of the year of the sixty-year cycle of Jupiter (2–19; 96). Venkatasubbiah shows that a given Indian date may theoretically have as many as twenty-eight possible Western equivalents (72), all of which must be duly considered and tested before a date is condemned as “irregular,” and he is highly critical of arbitrary emendations by scholars of dates which “seem to them to be irregular” (viii), warning that “one has to be very circumspect in pronouncing any given date to be irregular” (93).

c10.s3.t3.18

⁶² IA 19, 1890, 364-9; 24, 1895, 4-5 = KS I.566-71; II.682-3.

Linked texts (10)