________________
200
/
Jijñäsä
skilfully taken into consideration and thought all the causes of virtues and faults (and ) having discarded all (the other ) sons of the king (as not coming upto her standard )."14 The popularity which this motif acquired may be gauged by the fact that only three decades later, Mahārāja Mātrvishnu, a mere feudatory chief of Gupta emperor Budhagupta, is found describing himself as the one "who, by the will of (the god) Vidhātņi, was approached (in marriage choice ) by the goddess of sovereignty, as if by a maiden choosing (him) of her own accord (to be her husband)."15
There are a large number of ancient Indian records which do not contain any date or are dated in an era the identification of which is not beyond doubt. The probable dates of such records are usually determined with the help of their palaeographical peculiarities and other indications provided by their contents. In this connection it is important to note that palaeographical features can, at the most, suggest the general period of a record, and not its absolute date. Further, they cannot become the sole basis of fixing its date. Actually, the chronology of the evolution of a script itself depends upon those records the dates of which we determine by means other than their palaeographical features. For instance, Fleet and other competent epigraphists placed the records of the Vākātaka king Pravarasena II in ca. 700 A.D. and opined that there is nothing in the palaeography of his grants to controvert such a conclusion. But now we definitely know that Pravarasena II could not have flourished later than the second quarter of the fifth century A.D. Thus, a modification in the probable dates of those records which are regarded as either contemporary to or earlier or later than the Vākāțaka grants has become necessary. Of course, now our knowledge of the evolution of the Gupta script is far more advanced than it was in the days of Fleet and it is possible to suggest a more accurate date of a record on the basis of its script alone; but, even now, it is highly risky to fix the date of any king on the evidence of one or two letters of his records. It should not be forgotten that even in the same record, evidently written or engraved by the same person, the shapes of the same letters may vary considerably. For example, in the Deoriya stone image inscription in the word sarvasatvānam the first sa is engraved in western style of the Gupta Brähmi, while the second sa belongs to the eastern group. The Prayāga prasasti of Samudragupta possesses sa of eastern style in all places except in Kaushthalapuraka (1.20) in which the western shape of the letter is used. Consequently, epigraphists usually do not see eye to eye on the question of the dates of such undated records. The Nachne-ki-Talāi and Ganj inscriptions of Pịthvīşeņa, for instance, are placed by some competent epigraphists in the fourth century A.D., 17 and by other equally competent authroities in the fifth century A.D. 18 The legends on coins provide even a more slippery ground for palaeographists. For instance, some coins provide very uncommon shapes of letters and medial signs; some times we find both the forms of letter ma current in that period even on one coin. The palaeographical argument, therefore, cannot and should not be regarded as the sole basis of the date of an epigraph or coin; it should be studied in the context of other lines of evidence.19
While evaluating epigraphic evidence one should give due consideration to the difference between positive, circumstantial, corroborative and explanatory types of evidences, something to which, unfortunately, sufficient attention is not paid by quite a few scholars. It has not been fully realized that in the field of epigraphy and historical research a positive evidence or argument is comparatively always the best and should form the sheet-anchor of a suggestion. Circumstantial evidence is usually quite reliable but it needs utmost caution and restraint on the part of researcher. Corroborative evidence becomes worthy of consideration only when it is tagged to positive or circumstantial evidence. But
