Pramāṇa · Search · About · API

Indian Palaeography c15.s4: A III. Mathura and the North-Western Region

Indian Palaeography c15.s4

Language
English

Contents

A III. Mathura and the North-Western Region

c15.s4.1

The palaeographical chart (Pl. XII) is prepared from the writing on (*a*) stone or rock (nos. 1, 2, 6, 7, and 9), (*b*) the iron pillar (no. 3), (*c*) copper plates (nos. 4, 5, and 8), (*d*) palm-leaf (nos. 10 and 12), and (*e*) birch-bark (no. 11). The technique of writing on each was different. On stone and rock the letters were in the last stage engraved with the help of a chisel. The engraver probably incised the lines drawn beforehand by the scribe. The plates could not be prepared in this fashion because of the nature of the material. The engraving on the plates must have been executed by the metal worker. The same technique was probably employed in the case of the iron pillar. For palm-leaves and birch-bark a broad or edged pen was used by the scribe, who wrote in ink. Here the proper style of the writer is visible. The copper plates¹ are at best faithful engravings of the original letters, but the metal worker may have contributed his own element to the writing irrespective of the graving tool used by him. Similarly, the engraver on stone or rock may have missed the details natural to the pen style. Hence there are bound to be differences in the styles of writing dealt with in this section. But there is one common element in all these writings, the original copy of the scribe, which was the determining factor in giving particular shapes to the letters. The engraver or the metal worker would not create new forms of the letters, but would only distort the given forms to suit his graving tool. It is therefore hardly necessary to assume a 'hieratic' style for the inscriptions and a 'literary' one for the manuscripts, unless we have definite evidence for a particular set of inscriptions. The writers had a common style and in general they must have followed the practice of the day, except for their individual mannerisms. Therefore I do not accept the opinion of Bühler, given in his 'Remarks on the Horiuzi Palm-Leaf MS.'² and Hoernle in connexion with the Bower MS.,³ that 'the masons who incised the inscriptions, or the writers who wrote the originals from which the masons copied,

c15.s4.2

¹ The technique of heating the copper plate for writing is adduced from the following verse occurring in the Nidhanpur copper plate:

c15.s4.3

² Max Müller, op. cit., pp. 90-91.

c15.s4.4

³ *Indian Antiquary*, xxi. 29-45.

c15.s4.5

Śāsana-dāhād-arvāg-abhinava-likhitāni bhinna-rūpāṇi
Tebhyo aksharāṇi tasmānn-aitāni kūṭāni.

c15.s4.6

'Because, after the burning of the plates, these newly written letters are of different form, therefore they are not forged.' tried to make the characters archaic'.¹ I believe that the manuscripts and the inscriptions follow the current styles of writing, though between them there may be stylistic differences due to the technical reasons already mentioned or to the writers' particular mannerisms. These differences can be detected, and it is hardly necessary to give priority to the manuscripts over the inscriptions in date.²

c15.s4.7

¹ *Indian Antiquary*, xxi. 30.

c15.s4.8

² See below, p. 151.

c15.s4.9

On stylistic and chronological grounds the writing is divided into the following groups:

c15.s4.10

(*a*) *The Mathura school of the Gupta period* (Pl. XII, 1, 2, 4, and 6)

c15.s4.11

No. 1 is taken from the Mathura inscriptions of Chandra Gupta II. The style has been fully described in the last chapter.³ No. 2 is derived from the Mathura inscriptions of Kumāra Gupta, dated in the (Gupta) years 113 and 135. It is in the same style as no. 1, but a few new forms are seen. The head-mark is a solid triangle and in some cases the foot-mark is also observable. No. 4 is taken from the Indor plate of Skanda Gupta, dated in the (Gupta) year 146. It represents the copper-plate style of Mathura. Here the head-mark is a simple line and no foot-mark is seen. No. 6 is derived from the Kura stone inscription of Toramāṇa. It shows the developed forms of the Gupta period. The head-mark is a small triangle but no foot-mark is seen.

c15.s4.12

³ See above, p. 104.

c15.s4.13

The initial vowels *i*, *e*, and *o* are of the same forms as the Middle Ganges Valley, but *a* has two regional varieties—one, seen in the Kumāra Gupta inscriptions and ultimately traceable from the Kushāṇa form of this region, was termed in the last chapter the 'angular' type.⁴ The Gupta form shows a development in curving the upper end of the left limb outwards, placing a head-mark over the right vertical, and thickening the left bottom angle into a triangular knot. This form will be henceforward called the angular form with a knot. The other variety of *a* is seen in the Indor plate and the Kura inscription. Here the left limb makes a double curve, outwards at the upper end and inwards at the lower. This is the regional variation of the Gupta form influenced by the first variety. The vowel *ū* in the Kura inscription has a fully developed curved base. The medial vowels follow both the Kushāṇa and the eastern system, as is seen in the forms of *ā* and *e*, but the medial *o* and *au* are given always in the Kushāṇa style. The right hook for the medial *ā*, as seen in Bengal and Nepal, is not found here at all. The medial *i* has its curve to the left, and in some cases is extended to the bottom of the letters, and the medial *ī* has its curve turned to the right. The medial *ṛi* follows the North Indian fashion of a hook turned to the right.

c15.s4.14

⁴ See above, p. 86.

c15.s4.15

*Ka* has its cross-bar bent. *Kha* has a rounded top, but *ga* optionally has a flat top. *Gha* maintains the straight base, and *ṅa* in conjuncts merges the top line with the head-mark. *Cha* is of the pointed beaked type, but the point gradually moves upwards. *Chha* maintains the double looped type and *ja* has three right arms branching from the vertical. *Ña*, as seen in the conjunct *ñcha*, is not found in the cursive form. *Ṭa* emphasizes the flat top, and in conjuncts such as *ṭṭa* it keeps its correct form, just as *ja* does in *jja*, and no twisted form is seen as in the eastern region. *Ḍa* has a short hook appended from the tip of the curved back. *Ṇa* has three varieties—the old Kushāṇa type resembling an italic *X*, the open-mouthed *ṇa* of the Kauśāmbī style but without the loop at the left, and a development from the first type. In this last variety the lower curves are joined together to form a thick line. This is the cursive form of the Kushāṇa *ṇa* and must be distinguished from the cursive form of the open-mouthed *ṇa*, which developed later in the Middle Ganges Valley. The two arms of *ta* meet, sometimes at an angle and sometimes in a curve. *Tha* is an oval with a mid-line. In conjuncts such as *sthi* the oval lies flat. *Da* is of the double-curved form, and *dha* narrows its lower end. *Na* keeps up the regional form with a curved base. *Pa*, *pha*, *ba*, and *bha* all have the angular forms. Only in the Kura inscription *pa* develops an acute angle, and in the Indor plate *ba* has its left side curved. The open-mouthed *ma* is seen in two varieties—the tailed one is more usual, but another with a half circle in place of the tail is found in the Kumāra Gupta inscriptions. *Ya* is of the tripartite form with an inner loop at the left arm. *Ra* is a straight vertical. Both *la* and *ha* have firm bases, two verticals and a hook. In the Kura inscription the base of *ha* is rounded, and optionally also has a notch. *Ḷa* maintains the form of Kushāṇa times. *Va* is generally triangular but in a few cases the angles are rounded. *Śa* has both the rounded and the flat top, and has a triangular foot-mark at the left limb, as in *ga*. In the Kura inscription the foot-mark is replaced by a curve turned to the left. *Sa* is of the hooked type, but in the Kura inscription the loop is seen in place of the left hook and in another variety the mouth of this triangular loop opens out, a type probably derived from Central India.¹ Other varying features of this inscription probably came from the same source. On the whole these inscriptions keep up the regional differences from the Middle Ganges Valley.

c15.s4.16

¹ See below, pp. 162 and 164.

c15.s4.17

(*b*) Pl. XII. 3 is taken from the Mehrauli iron pillar inscription of Chandra. The execution of the letters on the pillar is superb. Each letter has been boldly brought out, and the writing, if it truly portrays the scribe's hand, is inclined to over-emphasize the angularity. The head-mark as well as the verticals are thick-lined. If these special features are borne in mind, it is not difficult to place the inscription in time and area. No initial vowels are known from this record. The medial vowels follow only the Kushāṇa system, as they easily lend themselves to ornate forms, noticeable particularly in the up-curled form of the medial *i*, and in the rectangular head added to the long vertical of the medial *ī*. The right limbs of *ga*, *ta*, *bha*, and *śa* are longer than the left, as in all the inscriptions of the Guptas. Similarly the left limbs of *kha*, *ga*, and *śa* have triangular foot-marks, and the tops of these letters are flat. Identical with the Middle Ganges Valley forms are the open-mouthed *ṇa*, with a loop at the left, the looped *na*, the tailed *ma*, the hooked *la*, the looped *sa*, and the hooked *ha*. The presence of so many forms from the Kauśāmbī style definitely proves that the writer of this inscription came from the Middle Ganges Valley. The forms of the other letters do not militate against this view. The cross-bar of *ka* is horizontal with ticks at either end. *Cha* has an ornate form with its beak drawn up to the line of the head-mark. *Ja* has three arms attached to the vertical. Angularity is marked even in the double curve of *Da*. *Dha* narrows its lower end, a characteristic of fifth-century writing. *Bha* is of the angular type. *Ya* has lost its third left arm but the loop is traceable in the thickening left at this point. This type of *ya* is seen in the Gadhwa inscription of Kumāra Gupta. In the conjuncts *ṅa* merges its upper arm with the head-mark. *Ña* has the cursive form, *sha* has its mid-line slanting downwards, *tha* is horizontal, and *ha* and *la* in combination have a common vertical. There is no doubt that the inscription was written in the early fifth century A.D. by a writer from the Middle Ganges Valley.

c15.s4.18

(*c*) Pl. XII. 5 is taken from the inscription from Shorkot in the Punjab, dated in the (Gupta) year 83. The place where it was found is identified with the city of Śibinagara. The inscription is engraved on a metal bowl. The writing shows an individualistic style in adding looped ends to the bottom of the long verticals, as in *ka*, *ra*, *śa* and medial *u*. Another special feature is seen in the head-mark, which is a hollow triangle. If these individualistic mannerisms are set aside, the forms of the letters fit the style of the early fifth century A.D. Here the medial *ā* is a horizontal line, but the medial *i* curls its upper end, as in the Mehrauli inscription, where the sign is modified to fit the angular style. The medial *o* follows the Kushāṇa style. The cross bar of *ka* is angular. In conjuncts such as *kla* and *kshu*, *ka* resembles an *x* with a top mark because of the combination of this angle and the head-mark. *Gha* has a slightly inclined but straight base. *Cha* is of the beaked type. The arms of *ta* meet in a curve. *Tha*, as seen in the conjunct *tthā*, is oval with the mid-line curved, a form also known from Central India. *Da* is of the double curved type. *Ḍha* is of the bulged form. The curved base of *na* is angular. The angularity is also marked in the form of *pa*, both of whose uprights have hollow triangular heads. The left side of *ba* is curved. *Bha* and *śa* have their right limbs longer than their left. *Ma* is of the open-mouthed type with a curve at the left angle, as is seen in the Kumāra Gupta inscriptions.¹ *Sa* is of the hooked type. The hollow triangular head-mark and the habit of giving curved tips suggest that the writer may have come from Central India. The form of *tha* lends support to this conclusion.

c15.s4.19

¹ See above, p. 143.

c15.s4.20

(*d*) Pl. XII. 7 is taken from the Tusam rock inscription. The letters are poorly engraved on the rock. No attempt is made to copy the ornamental forms. It illustrates a simplified style, but definitely shows the new tendency of the post-Gupta period. It may be dated in the first half of the sixth century A.D. The head-mark is generally a solid triangle. The only initial vowel found here is *a*, which is of the curved form. Another new form is seen in one variety of *ya*, in which the left vertical is merged with the loop, which is drawn along with the middle vertical—the form thus produced approaching the bipartite type. *Ra* also shows a new tendency by having an upward slant at the foot. Other letters recall the Kushāṇa style. The most important are *kha* with its prominent triangular base and the curved base type of *na*. Influence from Kauśāmbī is seen only in the adoption of the open-mouthed *ṇa*. *Bha* keeps up the angular form. The open-mouthed *ma* has a curve at the left angle. *Sa* is of the hooked type. We also find here *ḷa*, whose double curves are equalized. We do not find any cursive forms in the conjuncts. The medial vowels follow both the Kushāṇa and the eastern systems.

c15.s4.21

(*e*) *Kangra style*

c15.s4.22

Pl. XII. 8 is taken from the Nirmand inscription of Samudrasena. It shows the Kangra style of the early seventh century A.D. The inscription is dated in the year 6, but it need not be ascribed to the Harsha era. The writing is in the simplified style, but influence from the *kuṭila* forms is clearly seen in the shapes of some of the letters. On the other hand, there are several typical forms that indicate a regional version of the *kuṭila* writing in Kangra. Certain individualistic features may be noticed first: the feet of the left limbs of *bha* and *śa* and the hook of *sa* have curved strokes added to them. Of the initial vowels *ā* is of the curved type, *u* has a curved base, and *ai* has a flat top with an additional slant to the top left. But *i* preserves its three-dot form in one case and in the other example shows two dots above and a tailed one below. The medial vowels are simplified versions of those seen in the Middle Ganges Valley. Of the consonants *kha*, *ga*, *cha*, *ja*, *ṭa*, *ḍa*, *na* (optionally), *pa*, *pha*, *bha*, *ya* (optionally), *la*, *śa*, *sha*, *sa*, and *ha* preserve the older regional forms. The bent right verticals are observable in *ṇa*, *tha*, *dha*, *ma* and *ya*. The new influences can be seen in the use of the looped *ka*, the bulged form of *tha* and *dha*, the double curved *da* with a tail, the head-form of *na*, the acute angled *ma* and *ya*, *ra* with a triangular foot-mark, and the rounded *va*. In the conjuncts again no cursive forms are seen.

c15.s4.23

(*f*) *Chamba style*

c15.s4.24

Pl. XII. 9 is taken from five stone inscriptions from Chamba. The writing is in different hands. The Panali Nala inscriptions show a style with the head-mark which is usually a hollow triangle, but the Gum inscription has solid triangular heads. Again the Panali Nala inscriptions are crudely engraved, but the Gum inscription is well executed. All of them clearly show the influence of *kuṭila* writing and forms identical with those seen in the Middle Ganges Valley at this time. Regional differences are no longer found. The inscriptions may be dated in the seventh century A.D. The only initial vowel found here is *a*, which is of the curved type with two arms of the curve projecting beyond the left vertical. The medial *i* and *ī* have their curves extended down to the bottom of the letters. *Ka* not only has the loop on the left, but also has an additional bent vertical on the right. *Ga* has the triangular foot-mark at the left limb. *Chha* is of the double-looped type. *Ja* merges its upper arm with the head-mark and bends the two lower ones downwards. *Ṇa* is of the open-mouthed type with its base sloping to the right. *Ta* is of the *kuṭila* form. *Da* has an upward slant at the tip of the curve. *Dha* is of the bulged form, and *na* is of the head type, but also shows a regional variety of the looped *na*. Both *pa* and *ma* have acute angles on the right. *Ya* is of the bipartite type. *Ra* clearly shows the foot-mark. *La* slants its base upwards with its hook curled in. *Va* when triangular has its base slanting, and when rounded bends its right vertical. *Śa* shows an eastern form, with its top curled and joined by a bar to the right vertical. *Sa* is looped and has another variety with its mouth open.

(g) The Manuscript Style of the Gupta Period

c15.s4.t1.1

Pl. XII. 10 is taken from the palm-leaf manuscript *Kalpanāmaṇḍitikā*. The style shows some individualistic features which must be noted first. As is natural with the broad pen, the head-mark is throughout given in a thick line, and similarly the interplay of thick and thin strokes is clearly noticeable. At the lower ends of the verticals, as in *a*, *i*, *ka*, *ga*, *ra*, *śa* and the medial *u*, the writer leaves behind a flourish or slant that must be associated with his special manner of lifting the pen. If we ignore these peculiarities, the forms agree with the style current in the early fifth century A.D. in the Mathura region. The alphabets show a complete development of the Mathura Kushāṇa writing. The only influence from the Kauśāmbī style is seen in the form of the open-mouthed *ma*, but this also is of the Mathura type with a curve at the left angle, as is seen in the Kumāra Gupta inscriptions.¹ The striking difference from the Kushāṇa style is marked in the use of the evolved angular *a* with a knot,¹ *kha* with its reduced triangular base and extended upper hook, the lengthening of the right limbs of *ga*, *ta*, *bha* and *śa*, the round-backed *ḍa* with a curved leg, the cursive form of the Kushāṇa *ṇa*, the oval *tha* with a mid-line, the angular *bha* with its middle bar broad and notched, the tripartite *ya* with its left arm merged in the loop, and *ra* with the foot-mark. The deciding factor in dating this manuscript is better illustrated by the medial *i* with its curve going to the left and the medial *ī* with its curve turned to the right—a practice not seen even in the Allahabad pillar inscription of Samudra Gupta. It is only from the fifth century onwards that this style is known. The initial *i* preserves its three-dot form, as is also seen in the Indor and Nirmand plates. The initial *u* and *ū* lower the tips of their base lines. The initial *e* and *ai* have flat tops, and *o* has its central line inclined, as is seen in the Mathura inscription of Chandra Gupta II. The medial vowels follow the pattern of the fifth century, the only notable exception being the medial *u* and *ū*, which do not have the exaggerated curve. On the whole the forms are quite in keeping with those known from the inscriptions of the fifth century a.d.

c15.s4.t1.2

¹ See above, p. 143.

c15.s4.t1.3

¹ See above, p. 142.

c15.s4.t1.4

*(h) The Bower MS.* (Pl. XII. 11)

c15.s4.t1.5

The palaeography of this manuscript has been discussed in detail by Rudolf Hoernle in the pages of the *Indian Antiquary*² and later in his edition of the manuscript. His method of study is different from the system followed here. He singles out one letter, such as *ya*, and bases his conclusions about the dating of the manuscript on the forms of it appearing in the text. But I do not place much reliance on the evidence of single letters because forms were not uniformly adopted even in the same region. I believe first of all in assessing the general tenor of the writing, then in analysing the technical processes involved in the production of the letters, and finally in observing the different forms resulting from these sources. My dating depends on the evidence of all these three factors. Hoernle's date of the manuscript is not borne out by this test. Before a date is assigned, I shall describe the palaeography in our terminology. Hoernle divided the manuscript into seven parts and observed rightly: 'all the seven parts of the Bower Manuscript are written in an essentially identical script.' He also realized that different hands were used to write the different parts of the manuscript, and these he divided into three groups: (i) parts I–III, (ii) part IV, and (iii) parts V–VII. Within this last group part VI shows a better hand than the rest. The writing material is birch-bark. A broad or edged pen was used, producing thick and thin strokes. No attempt was made to copy an ornamental style. Throughout the head-mark is distinctly given. It is a firm line, often with a notch in the middle. The foot-mark is also clearly visible at the base of the verticals. The marks in the letters *ka* and *ra* in part IV suggests that the writer lifted his pen with a back stroke and hence resulted in an upward slant at the foot. This tendency has been noticed in the manuscript described in (*g*) above. The letters *pa*, *pha*, *sha* and *ma* have acute angles on the right, and their right verticals are slightly bent. In some parts of the manuscript both *pa* and *ra* develop tails. This general similarity suggests that, though the hands are different, the date of the whole manuscript falls within one definite palaeographical period. From the point of view of palaeography it may be divided into two sections: (A) parts I–III, and (B) parts IV–VII. Of these part IV is in a very slovenly hand. Its letter forms show some cursive tendency but in general its essential features are the same as observable in other parts of (B). These two sections derive their letter forms from two different sources: (A) from what we have called the Rajasthani or Malwa style¹ and (B) from the Mathura style, but there are many borrowings from one to the other. Therefore it would be wrong to limit one section to only one source. Here an attempt will be made to trace the different traditions as they appear in the two forms.

c15.s4.t1.6

² *Indian Antiquary*, xxi. 29–45.

c15.s4.t1.7

¹ See below, p. 154.

c15.s4.t1.8

The initial vowels of section (A) are exclusively derived from Rajasthan. The vowels of section (B) also appear to have come from the same source, but other traditions are also observable. The most important is the vowel *a* in this second section, which shows a development of the type which I have called the angular form with knot.² This knot assumes the form of a solid triangle, and in parts V and VI the triangle opens its mouth. In this section the initial *i* consists of three dots, two above and one below.³ This form is seen in the Mathura region as well as in Malwa in the Mandasor inscription of Yaśodharman.⁴ This was the standard system of the three-dot *i*, probably evolved in Rajasthan. It is from this that the tailed *i* of the later times developed. The initial *e*, except in part IV which follows the other tradition, also differs from the form appearing in section (A). Here an upper stem projects from the triangular form and the left end of the triangle is slightly curved. In section (A) the upper stem disappears, but it leaves its trace by emphasizing the right arm of the triangle, which is generally bent. This *e* with a stem is again derived from Rajasthan. In section (A) the initial *a* is of the curved type with its two arms projecting beyond the left vertical, and the initial *i* has two dots above and a tailed one below. Both these forms first appeared in Rajasthan and Malwa, and later in the sixth century a.d. they were adopted elsewhere, probably with the migration of scribes after the Hūṇa invasion. No doubt the tail in the lower dot of *i* is the result of a general tendency in cursive writing. The initial *ī* maintains the Gupta form—a dot on either side of a vertical. The initial *u* curves its base downwards, and in part IV it is extended beyond the head-mark. The initial *ū* is formed by adding another curve going upwards in section (B) and downwards in section (A). The initial *ṛi* is a development from the cross form, as has been shown already.¹ *Ai* is formed by adding a horizontal stroke to the top left of *e*. *O* is a cursive form with its base curved down, and *au* has an additional upward curve attached to the form of *o*. In section (B) the Kushāṇa style of top strokes for the medials *e*, *ai*, *o*, and *au* is adopted, while the medial *i* has its curve extended downwards around the bottom of the letters. But in section (A) the eastern styles of the medial *e*, *ai*, *o*, and *au* are generally seen, while the medial *i* is not extended so far as to cover the bottom of the letters. But there is another type of the medial *i* used in section (B), the circle type which must have been derived from Rajasthan.² The consonants that show forms fundamentally different from the Gupta³ forms are *ka*, *bha*, *ya*, *śa*, and *sa*. In both the sections we have the looped type of *ka* used side by side with the cross-bar type, as is seen in inscriptions of the sixth century a.d. In both the sections the triangular foot-mark at the left limb of *bha* grows in prominence and the right limb curves downwards, and in section (B) it appears to have an open mouth in some cases. *Ya* has three varieties—the tripartite form with a loop, which is preserved in Malwa till the sixth century a.d., the tripartite form with an outward curl at the left arm, which appears in the sixth century a.d. in Bengal, Nepal, Bihar, and Rajasthan, and the bipartite *ya*, which also appears in the sixth century a.d. Here the right limb of *ya* shows the *kuṭila* form. Two forms of *śa* are seen. One has a curl at the upper end of the vertical, joined by a slanting bar to the right vertical, a form which developed in the sixth century a.d. This is seen in section (A). In the other section we have a flat-topped variety with its left limb ending in an upward curl, again a late tendency. The dental *sa* either has a triangular loop or opens the mouth of its triangle, both features appearing in the post-Gupta period.

c15.s4.t1.9

² See above, p. 142.

c15.s4.t1.10

³ Hoernle's drawing of one dot above and two below in part IV, as given in his table 1, is a mistake.

c15.s4.t1.11

⁴ See below, p. 159.

c15.s4.t1.12

¹ See above, p. 121.

c15.s4.t1.13

² See below, p. 157.

c15.s4.t1.14

³ We restrict the term 'Gupta period' to only two centuries, the fourth and fifth, and hence do not include all the inscriptions given in Fleet's *Corpus*, as was done by Hoernle.

c15.s4.t1.15

*Kha* does not have the foot-mark at the left limb, as is also the case in the inscriptions of Yaśodharman of Malwa.¹ *Gha* generally has its base divided into a straight right half and a curved left half. *Cha* is of the beaked type, and in section (B) the beak is exaggerated. *Ja* has its two lower limbs bent, and in parts III and IV the uppermost limb merges with the head-mark. *Jha* has an acute angle on the right. *Ṭa* has a flat top. *Ṇa* is of the open-mouthed variety, but its corners tend to become small closed loops. *Tha* and *dha* are of the bulged form. In section (B) *tha* is horizontal. *Da* is of the double curved type. *Na* is still of the Mathura type with a curved base. The right verticals of *pa*, *pha*, and *ma* are bent. *Pha* generally has an outer curl in place of the inner loop. *Ba* has its left arm bent. *Ra* has an upward slant or a tail at the foot. Both *la* and *ha* have a firm base, two verticals, and a hook. *Va* is triangular but its base is slanting. *Sha* has the form of *pa*, with an added horizontal line in the middle.

c15.s4.t1.16

¹ See below, p. 160.

c15.s4.t1.17

These features point to a strong connexion with the writing style of the sixth century a.d. The manuscript may be dated in the first half of the sixth century a.d. Its script is related rather to the Rajasthani style than to that of Mathura.

(i) Horiuzi Palm-leaf MS.

c15.s4.t2.1

Max Müller notes the following tradition regarding the ownership of the manuscript: 'We have good evidence showing that these leaves were brought to Japan in a.d. 609, and that they came from China. It is further probable that in China they belonged to the monk Yashi, who died in a.d. 577, and before him to Bodhidharma, who emigrated from India to China in a.d. 520’. Relying on this tradition, Bühler adds: ‘Leaving all probabilities aside it is certain that this MS., which evidently has been written by an Indian scribe, cannot date later than the first half of the 6th century a.d.’¹ Later he² remarks: ‘If we had no historical information regarding the age of the Horiuzi palm leaves, every palaeographist, I believe, would draw from the above facts the inference that they belonged to the beginning of the eighth century a.d.’ Finally, he draws the conclusion: ‘They clearly illustrate the truth of the maxim that the inscriptions are not safe guides for the investigation of the history of the Indian alphabet, but that in the development of the letters they lag behind the literary documents.’³ This maxim is probably true when two allied scripts are used—one for the inscriptions and the other for literary purposes or business documents. Examples can be cited of the simultaneous use of Hieratic and Demotic in ancient Egypt, or of Kūfic and Naskh in the Muslim world during the early centuries of the Hijra. Here the two scripts show development in their own ways. But when the same script is employed for inscriptions and literary documents, the difference will be merely stylistic unless a particular archaic form is deliberately preserved; e.g. in India and Pakistan even today only Naskh script is used for the letters of the Holy Qur’ān, though for other purposes Nasta‘liq script is used. This same Nasta‘liq turns into the Shikastā style in ordinary letter writing. But all three styles have been in use in the subcontinent at least from Mughal times. In the case of Indian inscriptions there is no reason to believe that the archaic form was deliberately preserved. On the other hand, the epigraphic writing shows changes that are traceable to the literary style because the new forms seen in the letters are the natural result of the technique of the pen rather than that of the engraver. In consequence of technical improvement new forms were being created, no doubt first in the literary style, but when these forms came into general use they were certainly adopted even in the inscriptions. An exceptional literary document may show one or two new forms, but in the general tenor of writing they must conform to the current system. It is on these grounds that the two manuscripts (*g*) and (*h*) pp. 147–51 have been dated in the same manner as the inscriptions. If we rely on the historical tradition about the Horiuzi MS. and antedate it by two centuries, there is no reason why the above two manuscripts should not also be antedated by a similar period. This will lead us to accept a date for the complicated forms of the Bower MS. even before the appearance of the simple style in the Gupta inscriptions of the fourth century a.d. This is hardly possible. Hence I prefer to rely on the evidence of general palaeography for the date of the Horiuzi MS. In this choice support is derived from the fact that so much importance was attached not only to the text but also to the forms of the letters that as late as the seventeenth century a.d. Ziogon attempted to reproduce an identical copy. It may be that the original copy was made by an Indian scribe at the time referred to by Max Müller, but the present copy could not have been written earlier than about a.d. 700 at the earliest.

c15.s4.t2.2

¹ Max Müller, op. cit., p. 64.

c15.s4.t2.3

² Ibid., p. 90.

c15.s4.t2.4

³ Ibid., p. 93.

c15.s4.t2.5

Bühler¹ speaks of five 'general principles visible in the formation of the letters'. The first refers to the Indian habit of writing each letter separately, the second to the triangular head-mark, and the third again to this head-mark, which at this time does not cover the entire breadth of the letter as it does in the later Nāgarī characters. These are the features shared by all the northern styles dealt with in this chapter. The fourth refers to the flattening of the tops of the letters *kha*, *ga*, and *śa*. This practice is known from the fifth century a.d. onwards. The last principle is described as 'the development of right hand verticals, projecting beyond the body of the letters'. In my terminology it can be described as the development of the tails—a feature which develops out of *kuṭila* writing and is seen nowhere in India before the close of the seventh century a.d. The best examples are the Nalanda stone inscription of Yaśovarman² and the proto-Nāgarī inscriptions of the Rāshṭrakūṭas.³ That the Horiuzi MS. also falls in the same chronological sequence is borne out by a detailed analysis of the letters, which are seen not only in the text but also given separately as alphabets in the last line of the second leaf.

c15.s4.t2.6

¹ Max Müller, *op. cit.*, p. 68.

c15.s4.t2.7

² See above, p. 129.

c15.s4.t2.8

³ See below, p. 164.

c15.s4.t2.9

The initial *a* is of the curved type with the curve projecting its two arms beyond the left vertical, a feature of the later period. The long *ā* is formed optionally by adding a curve to the right vertical, as had been the practice since Gupta times, but a new form with a vertical stroke on the right of short *a* is seen as in the Rāshṭrakūṭa inscriptions of the eighth century A.D. The initial *i* has two dots above and a tailed one below. The initial *ī* shows a new form, displacing the one which was in use since the Kushāṇa period. It shows four dots arranged in the form of a rhombus, the lowest one having a tail curved to the left. This form is also seen in Nepalese manuscripts from the tenth century A.D. onwards.¹ The other initial vowels are of the type seen in this period. The medial vowels have all become standardized, as is also seen in the Nalanda stone inscription of Yaśovarman. *Halanta* (a vowelless consonant) is marked by a short stroke attached to the bottom right of the letter. *Ka* is of the looped type. Flat tops are seen in *kha*, *ga*, *ṭa*, *dha*, and *śa*. The development of tails is observable in the right limbs of *ka*, *kha*, *gha*, *tha*, *da*, *dha*, *na*, *pa*, *ma*, *ya*, *la*, *va*, *sha* and *sa*. Remarkable late forms are *ṇa* with an angular right outer arm while the left still falls in a curve, *tha* with an outer loop at the top left, *pha* with an outer curl on the top right, the rounded form of *va* used both for *va* and *ba*, *bha* with a curved right limb and the left showing a notch; this last is an obvious simplification from the triangular foot-mark, though *sa* maintains the triangular loop at the bottom left with the only change that the loop is now filled up. Finally, *ma* has now merged its left tail in the twisted loop, as is also seen in the late inscriptions. On the whole the style is much simplified. The angularity is well marked and the bent verticals have mostly been straightened, leaving a trace in the tails—all these due to the hard pen that was used for writing, as has been rightly observed by Bühler.

c15.s4.t2.10

¹ Bendall's *Catalogue*.

Linked texts (4)