Pramāṇa · Search · About · API

Indian Palaeography c14: 6. Brāhmī Writing Styles Between the First and the Fourth Centuries A.D.

Indian Palaeography c14

Language
English

Contents

6. Brāhmī Writing Styles Between the First and the Fourth Centuries A.D.

c14.2

BRĀHMĪ was marked by a definite 'pen' style in the previous period, with the main body of the letters equal in length and the verticals drawn with tapering thickness. The medial vowels and the long verticals (e.g. in letters such as *a*, *ka*, *ra* and *la*) received no particular attention, except that the medial *i* sometimes had a tall flourish and *ra* had its lower tip curved. This 'pen' style was much elaborated in this period. The elaboration is the result of three notable tendencies in writing: (i) a predilection for cursive forms dictated by the swift flow of the hand in writing, e.g. in the forms of *na*, *ma*, *la*, *sa* and *ha* in the Kosam region (Pl. VIII *a*, *b*); (ii) a desire for an ornate style, as is seen especially in the Bijayagadh inscriptions (Pl. IX *a*, *b*); and (iii) fondness for symmetry and proportion noticeable in the inscriptions from Jogeyyapeta and Nagarjunikonda (Pl. IX. 12). While the first tendency deformed the letters and hence paved the way for new shapes, the second and the third added ornamental curves, triangles and rectangles, and thus created different styles. The different ways of writing resulted in the gradual introduction of newer forms. The writer, by his personal mannerisms, prepared the way for a particular style of writing, which led to the growth of a definite school in due course. But for the present the style was fluid. Its ornamental additions were essentially superfluous, and there was no fundamental difference in the main body of the letters. Brāhmī maintained its unity throughout the subcontinent. This unity was the result, not of any political integration as in the time of Aśoka, but of those socio-religious forces which were let loose after the invasion of the Kushāṇas and which gathered round different nuclei in different regions by the end of this period to develop into local cultures. From the point of view of palaeography this led to regional scripts, the beginning of which is traceable from the fifth century A.D. onwards, with slight variations of date in some regions. These grew out of the various styles of writing that prevailed in this period. It is to the definition of these styles that the present chapter is devoted.

c14.3

*New approach* The materials have in the past been studied from the point of view of the ruling dynasties, and various styles of writing have been labelled Kushāṇa, Western Kshatrapa, Śātavāhana, or Gupta. This approach is understandable, as the materials available for study have been mainly the inscriptions of kings belonging to these dynasties. But with the varied inscriptions at our disposal this approach is no longer necessary, as it limits our vision and leaves many important varieties unexplained. What is essential is to discover the different styles of writing as revealed in the inscriptions, to locate them in space and time, and to follow their movements. On this basis the inscriptions of a given region will be classified and analysed, and the materials obtained will be compared with those of the other regions in order to discover interconnexion.

c14.4

Recently an attempt has been made by Dr. Van Lohuizen De Leeuw¹ to study the development of the ‘Kushāṇa’ script at Mathura on the basis of a few letters. She takes the forms of the vowel *u*, the medial *u* in *ku*, the subscript *ya*, and the consonant *ya*, and tries to arrange the inscriptions in chronological order, as the base of *u* becomes more and more crooked, or as the length of the subscript *ya* increases, or as the middle vertical of *ya* becomes inclined and its left vertical has or has not a loop. This study of a gradual development is not a new method. It is similar to the old idea of the gradual reduction of the verticals from Aśokan Brāhmī onwards—a conception which was based on misunderstanding and hence was rejected in the last chapter. Similarly, Dr. Lohuizen’s assumption fails miserably when applied to a wider horizon and tested by comparative materials from the other regions. This wider perspective demands a new approach to the problem.

c14.5

¹ *The ‘Scythian’ Period*, Leiden, 1949.

c14.6

*Classification* It is not possible to speak of a uniform Kushāṇa style as applicable throughout their empire during their rule. The inscriptions from Mathura, as analysed on Pl. VIII *a, b*, reveal a variety of styles, and when these are compared with a group of Kanishka inscriptions from Kosam, Sahet-Mahet and Sarnath, the difference is clearly marked. The Kanishka inscription of the year 14 (Pl. VIII. 4) and the Huvishka inscription of the year 33 (Pl. VIII. 6) have features of their own and can be explained only when they are placed side by side with the inscriptions from the Kosam region. The ‘early group’¹ of the inscriptions from Mathura has a greater affinity with the early inscriptions of the Western Kshatrapas than with the Kushāṇa inscriptions from Mathura. The style is further changed in the records of the later Western Kshatrapas, when it received new stimulus from Mathura. This style of the Western Kshatrapas is characteristic of Gujarat and Western Malwa. In Eastern Malwa another ornamental style developed, which reached its highest evolution in the Bijayagadh inscriptions. The two stages of the Western Kshatrapa writing are also marked on the Śātavāhana records of Western Deccan, where a new ornamental Deccani style takes shape. It is this style which influences the characters of the south. On the one hand, its development is traceable through the Banavasi and Maḷavalli inscriptions to those of the Kadambas, and, on the other, through the inscriptions of Amaravati it evolves into the highly ornate style of Nagarjunikonda and Jogeyyapeta. In Madhya Pradesh, which was a backward region, odd inscriptions, showing influences from both the Gangetic Valley and the Deccan, have been found. Lastly, the early Gupta inscriptions are reducible into regions, as they continued the earlier styles in the fourth century A.D.

c14.7

¹ For the list see Bibliography at the end of the chapter, p. 104.

c14.8

*Main features*

c14.9

There was a fundamental break from the earlier practice as regards the emphasis laid on the different limbs of the letters. The new changes resolve themselves into four main groups:

c14.10

(1) The first relates to the shaping of the head-mark (Fig. 8. I–III). In the inscriptions of the Mathura Kshatrapas the verticals by their tapering thickness received a new shape, with the head growing in prominence (Fig. 8. Ia). This feature continued for some time in the Kushāṇa inscriptions, but survived longer in the inscriptions of the Western Kshatrapas and the Śātavāhanas. Very soon it degenerated into a short head-mark, from which the verticals depend (Fig. 8. I*b*). It helped to keep the letters in line and was also useful in maintaining their equal breadth. This head-mark was referred to by earlier palaeographers as the ‘serif’. It is present in almost all the inscriptions of this period, especially in North India. But very soon it assumes newer shapes. The evolution proceeds on two different lines, as shown in Fig. 8. I, II. The simple

c14.11

[Figure: EVOLUTION OF THE HEAD-MARK (I) (a) (b) (c) (d) (e) (II) (a) (b) (c) (d) (III) MOHENJODARO POTSHERD BHA DA (IV) KUSHĀṆA MEDIAL VOWELS KĀ KI KĪ KU PŪ GṚI VE VAI BO PAU NAṀ (V) EARLY GUPTA SHĀ HKA HPA (VI) WESTERN KSHATRAPA NAU (VII) OTHER NEW SIGNS DDHAM ṄGA ṄGHA FIG. 8]

c14.12

head-mark shown by a single line, as we see in the Mathura inscriptions, in due course is thickened (Fig. 8. I*c*) by the habit of repeating the lines two or three times. This thick head-mark is the characteristic of the Eastern Malwa style. This tendency further results in turning the thick line into a square head-mark (Fig. 8. I*d*), as is seen in the Malwa group of the early Gupta inscriptions. In these cases the square is solid, but later on we meet with a hollow square head-mark (Fig. 8. I*e*), in the inscriptions of the Vākāṭakas and others. This type was called ‘box-headed’ by the old palaeographers.¹ In the second variety the head-mark develops a notch in the middle (Fig. 8. II*b*), giving place to an angular line which ultimately results in a triangular head-mark (Fig. 8. II*c*), as is seen in the inscriptions from Bijayagadh and Nagarjunikonda. This is achieved by a three-way turning of the pen instead of repeating on the same line for thickness. This is easily understood from an inscribed pottery fragment discovered in the (Kushāṇa) stūpa at Mohenjodaro² (Fig. 8. III), which makes the process clear. Here we have a hollow triangular head-mark. Other good examples are the Shorkot inscription of the year 83³ and the Majhgawan plates of Mahārāja Hastin.⁴ This style was called 'nail-headed' by the old palaeographers.⁵

c14.13

¹ J. F. Fleet, *C.I.I.*, iii. 19.

c14.14

² J. Marshall, *Mohenjodaro and the Indus Civilisation*, iii, pl. XC.

c14.15

³ *Epigraphia Indica*, xvi. 15.

c14.16

⁴ Fleet, op. cit., pl. XIV.

c14.17

⁵ Ibid., p. 19.

c14.18

The terminology adopted in this book is given below:

c14.19

(*a*) Line head-mark (Fig. 8. I*b*). (*b*) Block head-mark (Fig. 8. I*c*). (*c*) Notched head-mark (Fig. 8. II*b*). (*d*) Solid square head-mark (Fig. 8. I*d*). (*e*) Hollow square head-mark (Fig. 8. I*e*). (*f*) Solid triangular head-mark (Fig. 8. II*c*). (*g*) Hollow triangular head-mark (Fig. 8. II*d*).

c14.20

(2) The second change is noticeable in the style of the medial vowels. The new system is shown in Fig. 8. IV. The medial *ā* is now a slanting vertical. The medial *i* and *ī* assume rounded forms of several ornamental varieties (see also Fig. 10. IV). The medial *u* bends its base in a greater or lesser degree. The second stroke of the medial *ū* is added at an incline to the left of the first stroke. The medial *ṛi* also has a curved form. The strokes of the medial *e*, *ai*, *o*, and *au* bend upward. These changes are generally adopted in this period in North India, but the older system survives in the Deccan inscriptions.

c14.21

There are other changes affecting local varieties. In the Mathura inscriptions of the Kushāṇa period, and as an influence in a few others, the *anusvāra* is shown as a short line in place of the dot. In the inscriptions of the later Western Kshatrapas the medial *au* has its double stroke on the right joined in a curve (Fig. 8. VI). In the early Gupta inscriptions the medial *ā* is shown by a long horizontal line with an additional tick downwards (Fig. 8. V), a practice that became common in the subsequent period.

c14.22

(3) The third change is the result of the influence of cursive hands. In these cases the letters assume new forms. They are shown in Fig. 9. The Deccani *ta*, which had a hooked stroke attached to

c14.23

[Figure: EVOLUTION OF CURSIVE FORMS TA DA ṆA DA NA MA SA HA LA SYA FIG. 9]

c14.24

the right of the vertical, first bends its vertical, and later as the hook is drawn in one and the same action, a loop develops at the junction of the vertical and the hook. The original round-backed *da*, which had its mouth open to the left, first has its upper and lower appendages inclined, and then, after the head-mark becomes the general fashion, the lower appendage vanishes. The cursive form again receives a rounded back, but with the mouth open to the right. The letter *na* first bends its base, and, finally, a loop develops at the junction of the vertical and the base. The original stepped *da* first slants its middle bar, then curves its angles with the elongation of the lower stroke, and finally develops a round-backed form in the Deccan. In the letter *ṇa* the upper stroke plays a predominant role, though the base also bends as in the case of *na*, and rarely develops a loop (not shown in the figure). The upper stroke bifurcates on either side of the vertical and gradually tears asunder the vertical, leading to the shape of 𑀡. The two curves separate farther and stand on the original base. This is the open-mouthed *ṇa*. The triangular base of the letter *ma* is changed. First it flattens out and then the triangle is placed on one side. In one case the base of the triangle merges with the longer arm, producing a tailed *ma*. In the other case the apex of the triangle opens and we get an open-mouthed *ma*. In the case of *sa* the left hook becomes a part of the left stroke, and when they are drawn in a single action with the remaining limbs the looped *sa* is produced. Both the letters *la* and *ha* play upon their originally insignificant tick, which gradually becomes rounded and ultimately finishes off the original bar to produce the hooked *la* and *ha*. Another significant change is seen in the subscript *ya*, which in the Mathura inscriptions develops a looped variety until finally this loop absorbs the base and the remaining verticals, resulting in the hooked variety of the subscript *ya*.

c14.25

(4) The fourth change is the result of fondness for flourishes. There are several ways in which this is expressed. They are shown in Fig. 10. IV–VI. In the first place the medial vowels develop curls, as shown in Fig. 10. IV. In the second place the long verticals bend either at the lower end or at the upper (Fig. 10. V). In the third place some of the letters become notched in the middle, as shown in Fig. 10. VI. Finally, the upper and lower flourishes balance on either side of the main letters in the centre, producing a calligraphic effect, as in the case of the inscriptions from Nagarjunikonda and Jogeyyapeta.

c14.26

*New signs*

c14.27

In this period for the first time we meet with the initial *ṛi*, expressed by a simple cross in the Andhau inscription of Rudradāman. The *m halanta* (vowelless) is met in the word *siddham* in some cases, and is expressed by placing a short *ma* at the lower righthand side of the compound *ddha* (Fig. 8. VII). The consonant *ṅa* is met with in combination with *ga* and *gha*, and is expressed by a sign similar to the Roman capital letter ‘E’ minus the middle bar (Fig. 8. VII). We also meet with the use of *jihvāmūlīya* (the *visarga* in combination with *ka*) and *upadhmānīya* (the *visarga* in combination with *pa*). The first is expressed by a double-axe sign placed vertically on the letter *ka*, and the second by a double loop placed horizontally over *pa* (Fig. 8. V). A new form of long *ī* with a dot on either side of a central vertical is also met with. This form is earlier seen in the cave inscriptions¹ of the south. The use of the letter *ḷa* in the Allahabad pillar inscription of Samudra Gupta is not due to influence from the south, as Fleet² supposed. It is well established in the inscriptions at Mathura and Sanchi. In the Allahabad pillar inscription of Samudra Gupta we also find the verses numbered.

c14.28

¹ See above, p. 73.

c14.29

² Op. cit., p. 4.

The Indus Zone

c14.s1.1

The Kharoshṭhī script was generally used in this region, and Brāhmī writing is found only exceptionally.³ Some inscribed fragmentary potsherds were discovered in the excavation of the Kushāṇa stūpa at Mohenjodaro.⁴ They are all written with ink. One piece has letters with the hollow triangular head-mark already described. The letters on the other sherds have only a line head-mark The *anusvāra* is expressed by a short line and the letter *da* has its mouth open to the right. The letter *ha* shows a loop in place of the tick and *dha* has an almost triangular form with the apex downward. The end verticals of the letter *ya* are curled and *bha* is of the broad type with its horizontal notched. The base of *ma* has both oval and triangular forms. It is difficult to speak of a local style on the basis of these fragments. They show obvious connexion with the Western Malwa style. They can be dated in the second century A.D.

c14.s1.2

³ See below, p. 251.

c14.s1.3

⁴ Marshall, op. cit., pl. XC.

c14.s1.4

Of much later date are the inscribed pottery fragments from Ṭor Ḍheri in the Loralai district of North Baluchistan.⁵ These are also written with ink. Here the medial vowels *i* and *ī* have a curly flourish, but medial *ā* optionally has the later form used in the Gupta records. The triangular base of *ma* has an open mouth. On palaeographical grounds the inscriptions should be dated in the close of the fourth century a.d.

c14.s1.5

⁵ A. Stein, ‘An Archaeological Tour in Waziristan and Northern Baluchistan’, in *Mem. A.S.I.*, no. 37, p. 93, pls. XVII and XIX.

Mathura Styles

c14.s2.1

The first series of the inscriptions is included in the ‘Early Group’ (Pl. VIII. 1). In these cases the verticals show the tapering thickness, familiar to us from the inscriptions of the Mathura Śaka Kshatrapas. The medial vowels generally follow the old style, except that medial *i* does not show the tall flourish, but the *anusvāra* is marked by a short line. The development is seen in the form of the initial *a* (Fig. 10. 1*a*). Here prominence is given to the upper left curve while the lower one has an additional tick. The triangular bases of *kha*, *ma* and *va* receive greater attention. *Na* and *ṇa* have optionally bent bases. But *da* retains the older form. Only rarely is the new form with the mouth open to the right seen. *Bha* is of the broad type (Fig. 10. 11*b*) with a notch in the middle. Similarly *ya* has a wider base. *Ra* has its lower tip curved. The subscript *ya* maintains its tripartite form. On the whole the style of writing is pre-Kushāṇa but post-Mathura Kshatrapa. It seems that the end of the Kshatrapa rule of the House of Rañjuvula and Śoḍāsa was followed by disturbances, during which these inscriptions were produced.

c14.s2.2

With the coming of the Kanishka group of the Kushāṇas to Mathura the style of writing changed rapidly (Pl. VIII). This change appears to be the result of the introduction of the ‘pen’ style in the earlier period. The general popularity of writing is evidenced by the large number of private records, which show attempts at reproducing the accepted forms in various ways. A common feature in all these inscriptions is the use of the line head-mark in place of the tapering verticals—a practice which is quite understandable in popular writing. This fact accounts for the poor quality of engraving, in which respect these private records contrast strikingly with inscriptions on the statues discovered at Mat near Mathura. Nevertheless, there is a certain unity of form in all the inscriptions from Mathura, except in two records¹—one the Kanishka inscription of the year 14 (Pl. VIII. 4) and the other the Huvishka inscription of the year 33 (Pl. VIII. 6). The specific letters which distinguish these two from the others are the tailed variety of *ma* and looped variety of *sa* in both, and hooked *la* and *ha* in the first. As these forms became a popular feature in the Gupta records, the date of this inscription of Kanishka has always been doubted. Dr. Lohuizen, on the basis of this doubt, propounded the theory that the inscription omitted the first digit and that it should properly be dated to the 114th year of the Kanishka era. She makes no mention of the inscription of Huvishka. On the other hand, she has included a number of other inscriptions in her re-dated group. But none of these, except one¹ which has a looped *sa*, has the special forms of the Kanishka inscription. Obviously these two inscriptions of Kanishka and Huvishka are intruders in this region. As I shall show, their affinity lies elsewhere, with the Kosam group. To accept this is to reject the whole classification of Lohuizen's thesis, because the remaining discrepancies pointed out by her can be explained on the ground of different hands being employed in the different inscriptions.

c14.s2.3

¹ A newly discovered inscription of Kanishka, dated in the 4th regnal year, uses the hooked variety of *ha*. *Indian Archaeology, A Review*, 1956–7, p. 39.

c14.s2.4

¹ *Epigraphia Indica*, ii. 209, no. 37.

c14.s2.5

Leaving aside these two inscriptions, we find minor variations in the others. Among the vowels the forms of *a* and *u* vary much. *A* shows developments in two distinct ways, as shown in Fig. 10. I. In both the varieties the upper left stroke becomes important, but while in (*a*) this upper stroke slants downwards to meet the vertical and the lower stroke maintains its curved form, in (*b*) the upper stroke makes an angle before meeting with the vertical and tries to assimilate the lower stroke, which sometimes appears below the line. Henceforth the variety (*a*) will be called the curved type and variety (*b*) the angular type. The letter *u* rarely slants its base upwards but more often bends its tip downwards. The bend is more marked in the medial *u*. The short *i* has three short lines in place of the dots, while the form of *e* is almost standardized with the apex of its triangle on the left. *O* has its middle vertical inclined. *Ka* has its middle horizontal variously curved, and in *Kha* the triangular base is now prominent. *Ga* has both angular and rounded forms, while *Gha* has its verticals straightened. *Cha* shows variations in the inverted beaked type, sometimes with graceful bends. *Chha* is of the double-looped variety, but the older oval variety is seen optionally. *Ja* maintains the form of the Roman capital letter ‘E’ with its vertical sometimes curved inwards, and rarely outwards, while the lowest arm occasionally bends its tip downwards. The older forms of *jha*, *ṭa*, *ṭha*, *tha* and *ba* continue. The stepped *ḍa* has its angles rounded. *Ḍha* has a more or less squat form. *Ṇa* appears in all the varieties except the open-mouthed one. *Ta* has its lower appendage rounded, but the Deccani *ta* (with its appendage to the right of the vertical) also appears. The round-backed *da* with mouth open to the right now replaces the older type. *Na* has its base variously bent. *Pa* occurs only in the angular variety, with the left arm some-

c14.s2.6

FIG. 10

c14.s2.7

times curved. *Bha* exhibits two distinct lines of development, as shown in Fig. 10. II. In type (*a*) the middle bar slants upwards and the right arm is elongated, both making an angle at the apex. This will be hereafter called the angular *bha*. In type (*b*) the middle bar is elongated and finally notched, with two curved arms at either end. This will be called notched *bha*. *Ma* has its triangular base emphasized by lengthening its lower side. The base of *ya* is flattened and takes various forms, as shown in Fig. 10. III. This base bends gracefully to give differential effect to the three uprights, which receive special attention. The left arm develops its line head-mark into a curl and finally into a loop, and the middle arm inclines to the left, while the right presses downwards to balance the loop in the left. *Ra* has either a straight vertical or a bend at its lower tip. The long vertical of *la* is normally straight or gently curved outwards; only rarely it bends inwards. *Ḷa* continues the older form—the sign of *ṭa* with an additional hook below. *Va* generally has a triangular form, but occasionally a rounded form is also seen. In both cases the upper stem has vanished. *Śa* has mostly the rounded form, with its right arm optionally longer. The inner stroke makes three varieties: in some it is inclined, in others it is horizontal, either halfway or full length. The three arms of *sha* are generally straight, but occasionally curved. Its inner stroke again is of three varieties as in the case of *śa*. The form of *sa* has definite angles, with the left hook usually drawn in one sweep with the left arm. *Ha* has either straight or curved arms, but greater variation is seen in the additional tick. The conjuncts are formed in the usual fashion, but the most distinctive conjunct of the Mathura style is the hooked variety of the subscript *ya*. The tripartite form (i.e. with three uprights on a horizontal base) as a second element in a compound is rarely seen at Mathura.

Sanchi Style

c14.s3.1

Allied to the Mathura style is the writing observed in the Kushāṇa inscriptions at Sanchi. Only three definite inscriptions have so far been noted, all engraved on images. No. 52¹ refers to the year 28 during the reign of Vāsishka and no. 53 to the year 22 and King Vaskushāṇa. The evidence suggests that the main structural history of the Sanchi monuments was over before the coming of the Kushāṇas. The letters show the line head-mark. The vowel signs adopt the Mathura style. The most remarkable is the medial *u* in *ku*, which is excessively bent. *Chha* maintains the older oval form, but *da* has a rounded back with its mouth opening to the right. The form of *ṇa* is of the italic [Figure:]-type. *Bha* is of the angular variety and the subscript *ya* is hooked.

c14.s3.2

¹ These numbers refer to Marshall and others, *Monuments of Sanchi*, iii, pl. cxxxviii, nos. 52, 53, 54.

c14.s3.3

Inscription no. 56 on the same plate of Marshall has been dated by N. G. Majumdar in the fourth century a.d., probably because of the appearance of the tailed *ma* and open-mouthed *ṇa*. But these appear in the Kosam region much earlier. The style of writing is not very different from that of the other inscriptions, except that the head-mark in this case is a solid triangle. This inscription may be dated in the beginning of the third century A.D.

c14.s3.4

*Eastern styles*

c14.s3.5

Two different styles were employed in the inscriptions of the eastern region. The first is illustrated in its early stages by the Kanishka inscriptions from Kosam, Sahet-Mahet and Sarnath, and in its later ornamental development by the Sarnath inscription of Aśvaghosha, dated in the year 40 (Pl. VIII. 10), the Allahabad Municipal Museum inscription of the year 23, and the Pahladpur inscription. This style is occasionally met with in the seal inscriptions at Bhita, Basarh, and Sahet-Mahet as late as the fifth century A.D., and it persists longer in the Gupta coinage as an optional style. In this eastern style the line head-mark does not appear at all, but the verticals faintly recall the tapering style of the Mathura Kshatrapas. In the Kanishka group of inscriptions the medial vowel signs follow the Mathura style, but there are other differences. The distinctive letters are *a*, *da*, *bha*, *ya* and *ra*. The older *a*, with its two slanting strokes on the left meeting the vertical at the same point, persists here side by side with the curved variety of the Kushāṇa *a*. Similarly, the angular *da* of the Mathura Kshatrapa type is used along with the round-backed *da*. *Bha* is of the notched type. *Ya* never has the loop at the left arm, nor does it develop into a hooked form in combination with other letters. The lower tip of *ra* is always bent.

c14.s3.6

In the other group of the inscriptions the change is more marked. The Allahabad Municipal Museum inscription is absolutely in the Mathura Kshatrapa or pre-Kushāṇa style as regards the head-formation, the application of the medial vowels, and some of the letters such as *ka*, *kha*, *ga*, *gha*, *cha*, *chha*, *ja*, &c. The most remarkable is the older form of the round *da* with its mouth open to the left. But other letters such as *na*, *ma*, *śa*, *bha* and *ha* are also different. The Sarnath inscription of Aśvaghosha and the Pahladpur inscription agree in the use of the older fashion of the medial vowels. They also retain the older type of *a* with its two left strokes bifurcated. The subscript *ya* keeps the tripartite form. In these cases we find the survival of the older style in Eastern India, though showing some influence from the Kushāṇa inscriptions of the eastern region. To this same group belongs the Sahet-Mahet Buddhist stone inscription,¹ in which we find the round-backed *da* and the hooked *ha* of the Kauśāmbī region, but no looped *sa* is seen.²

c14.s3.7

¹ *An. Rep. A.S.I.*, 1908–9, pp. 133–5.

c14.s3.8

² There is probably a mistake in the published estampage.

Kauśāmbī Style

c14.s4.1

The inscriptions of the Magha rulers of Kauśāmbī display another style of writing. They are dated between the years 52 and 139 of an unspecified era. They preserve the tapering style of the eastern region and in the majority of the cases they follow the older system of marking the medial vowels, though occasionally the Mathura style of vowel marking intrudes. But they make a complete departure in preferring the looped *na*, the open-mouthed *ṇa*, the tailed *ma*, the open-mouthed *ma*, the hooked *la*, the looped *sa*, and the hooked *ha*. A definite Kushāṇa influence is seen in the use of the hooked variety of the subscript *ya*, while the looped form of *ya* is seen when used singly. These cursive forms of the letters are familiar to us from the Allahabad pillar inscription of Samudra Gupta, and later, in the fifth century a.d., these forms became current in the Gangetic Valley. On the face of this evidence there has been a controversy as regards the era used in these inscriptions. But a comparison makes the distinction clear. By their tapering style they are related to the eastern style of the Kushāṇa period, while the Gupta inscriptions of the eastern region all have the line head-mark—a tendency which became current in this region much later—and are an ornamental development of this Kosam style of writing. This cursive style is distinctive of the Magha inscriptions in this region. Though some of these cursive forms occur exceptionally at Mathura in the Kanishka inscription of the year 14 and the Huvishka record of the year 33, they must be taken as intruding elements from this region. Therefore there should be no hesitation in dating these inscriptions in the Kanishka era. The use of this era in no way implies the subordination of the rulers of Kauśāmbī to the Kushāṇas, but merely shows the continuation of a system of dating that was made popular by Kanishka in this region. In fact the presence of these rulers and the total absence of Kushāṇa inscriptions in the eastern region after Kanishka support the suggestion made by Dr. N. P. Chakravarti³ that this area slipped from the Kushāṇa control.

c14.s4.2

³ *Epigraphia Indica*, xxxi. 175.

The Backward Area of Madhya Pradesh

c14.s5.1

The earliest group of inscriptions (Pl. VIII. 12) found in this region comes from Bandhogadh in the former Rewa State, engraved on the walls of artificial caves. The inscriptions are dated from years 51-90 of an unspecified era, and they speak of three generations of kings. Their palaeography reveals a mixed style ranging from crude engraving, for example in inscription no. XI,¹ to well-formed letters such as those of inscription no. IX. The head-formation is also irregular. Some inscriptions have the line head-mark, some the semblance of the tapering style, and in a few the development of the solid triangular head is seen. This mixed tendency is also apparent in the treatment of the medial vowels. Generally the eastern style is preferred, but the Mathura style is also known in medial vowels such as *e* and *o* while the Nanaghat style is found in the case of *mā*, *me* and *mo*. In the formation of the letters a similar mixture is noticed. *A* preserves the old form, with the left strokes making two curves or a short straight line, and joined to the right vertical by a small bar. The angle of *u* is rounded, as is the triangle of *kha* in some cases. *Ga* has both angular and rounded forms. The left arm of *gha* is optionally curved. *Cha* broadens its loop and does not exist in the beaked variety, and *chha* optionally copies the Deccani variety, which has its stem over the left loop. Similarly, the looped variety of the Deccani *ta* is also optionally used here. *Da* retains the older cursive form along with the Kushāṇa type of the round-backed *da*. *Dha* is of two varieties—the bow type and the triangular form. The base of *na* is bent. *Pa*, *pha* and *ba* sometimes have their left arms notched. *Bha* is usually of the angular variety. *Ya* optionally has the loop at the left arm. *Ra* sometimes has its lower tip bent. *La* has three varieties: the cursive type of the hooked *la*, that with a straight vertical, and another with its vertical curved inwards. Similarly, *ha* has the hooked variety as well as the ordinary form. The subscript *ya* maintains its tripartite form. The evidence suggests that, though Bandhogadh is not far from Kauśāmbī, its main contacts were with the Deccan on the one hand, and Mathura or Sanchi on the other. The dates of the inscriptions should be assigned to the Kanishka era.

c14.s5.2

¹ The numbers refer to Dr. Chakravarti's publication in *Epigraphia Indica*, xxxi.

c14.s5.3

Another poorly engraved inscription comes from Gunji (Pl. IX. 8) in the former Śakti State in Chhatisgarh, and is dated in the 5th year of a local ruler called Kumāravaradatta. There is no definite style in the writing, and the writer mixes different forms. He shows acquaintance with the line head-mark, but does not give it regularly. He does not even follow the rule of the equalization of the vertical. Forms such as *ka*, *ta*, *pa*, *bha*, *ma*, *ya*, *ra*, *la*, *sa* and *ha* recall those which were current in the last period. But there are other forms, such as the triangular *dha* with one arm curved, the triangular form of *va*, the optional use of the triangular base for *ma*, the triangular *kha*, and the double-looped *chha* (optional), that point to a connexion with Deccani forms of towards the end of the second century A.D. Only the cursive *da* shows some resemblance to this form in the inscriptions of Ushavadāta in Western Deccan; otherwise there seems to be no point in comparing this inscription with those of the time of Nahapāna. The inscription may be placed in the beginning of the third century A.D.

c14.s5.4

An inscription on a wooden pillar is found at Kirari near Chandarpur in Chhatisgarh. Here the letters are more developed. They are characterized by the solid triangular head-mark, and the medial vowels follow the Deccani style and show ornamental curves in the case of *i*. *Chha* is double-looped, and *ṇa* and *na* have curved bases. The verticals of *ka* and *a* are gently curved. *Kha* and *va* have triangular bases. *Ya* is of the tripartite form. The inscription may be dated to the middle of the third century A.D.

c14.s5.5

*Eastern Malwa*

c14.s5.6

In the last chapter it has been shown that Eastern Malwa came under the influence of the Mathura Kshatrapa style of writing. The last main stage of construction at Sanchi took place during this period. Subsequently it came under Kushāṇa domination,¹ during the rule of Kanishka and Vasishka. Their inscriptions have been dealt with earlier. After this period Eastern Malwa became a bone of contention between the Western Kshatrapas and the Śātavāhanas, with probably occasional independence. During this period a distinctive style of writing developed in this region, as evidenced by several *yūpa* inscriptions (Pl. IX. 6) dated between 282 and 335 of the Kṛita era (A.D. 225–78) and a Śaka inscription from Kanakhera, not far from Sanchi, also of this century. All these inscriptions show some influence from the Deccan, especially in the case of the last. The first definite break from the Kushāṇa style of writing, known from a few inscriptions at Sanchi, is seen in the marking of the medial vowels. The slanting strokes of *ā*, *e* and *o* do not occur, but all these strokes are horizontal. This change is due to southern influence, as is suggested by the medial vowel *o* applied to *ma* in its middle—a tradition which originated from the Girnar version of the Aśokan edicts and is maintained in the Nanaghat inscriptions and later Sātavāhana records. The Śaka inscription further introduces the round-backed *ḍa* (as seen in the conjunct *ṇḍa*), the looped *na*, and the looped *ṇa*, all from the Deccan. But there are other features peculiar to this region. The head-formation of the verticals is noteworthy. It is a thick line, verging on block formation, sometimes appearing as triangular. The other important feature is the development of ornamental flourishes, a feature which also marks the Deccan records of the later half of the second century A.D. The forms which connect this style with other North Indian inscriptions are the prominence given to the triangular base in *kha*, the beaked type of *cha*, the curved appendage of *ta*, the looped *ya*, *śa* with its right arm elongated, and the hooked variety of the subscript *ya*. Thus Eastern Malwa at this time was the meeting ground of the Deccani and the North Indian styles of writing.

c14.s5.7

¹ See above, p. 88.

North-Eastern Rajasthan

c14.s6.1

At Bijayagadh (Pl. IX. 7), in the former Bharatpur State, the most highly ornamental letters have been found in two inscriptions, one of the Yaudheyas and the other of a local king Vishṇuvardhana, dated in the (Vikrama) year 428 (A.D. 372–3). All the letters are characterized by a solid triangular head-mark. Ornamental flourishes occur in all the medial vowels, and graceful curves and bends are found in the verticals and horizontals of the letters. Influences from other regions can be seen in the looped varieties of *na* and *ṇa*, which may be ultimately traceable to the Deccan by way of Eastern Malwa, but the Kauśāmbī source cannot be overlooked, as it is definitely present in the use of the open-mouthed *ṇa* in the inscription of the Yaudheyas. There is not much difference in time between the two inscriptions.

Gujarat and Western Deccan

c14.s7.1

After the Aśokan edicts the next group of the inscriptions of this region belongs to the time of the early rulers of the Sātavāhana dynasty. The standard style of the Western Deccan is revealed in the Nanaghat inscriptions, which, as has been shown in the last chapter, cannot be dated earlier than the first century A.D. Next in the series is the Nasik inscription of Haku Siri¹ (Pl. IX. 1). This follows the Nanaghat style, and there should not be much difference in date between it and the Nanaghat inscriptions. No headmark is seen in the inscription and no particular attention has been paid to the equalization of the verticals, though, generally speaking, uniform length is maintained. The medial vowel-marks follow the same old style. Letters such as *ga*, *gha*, *ṭa*, *ṭha*, *ṇa*, *ta*, *tha*, *da*, *na*, *pa*, *ya*, *la*, *sa* and *ha* are in the same tradition. But definite change is seen in the curving of the verticals in *a*, *ka* and *ra*. *Ḍa* has its angles rounded. *Cha* has developed a beaked form, while triangularity is unmistakable at the base of *ma* and *va*. On the basis of these features it can hardly be doubted that this inscription belongs to about the middle of the first century A.D.

c14.s7.2

¹ *Epigraphia Indica*, viii. no. 19.

c14.s7.3

The next series of inscriptions belongs to the period when the Śakas and the Yavanas had already become influential in this region. From the numerous private records we select the Karle inscriptions published by M. S. Vats² and Nasik inscription no. 18.³ No head-mark is seen in the letters, but a faint trace of the tapering style can be observed in some of them. The angularity is well marked and some of the long verticals are curved. The connexion with the older style is seen in the forms of *a*, *da* with its mouth open to the left, and the crescentic form of *ya*, and in a disregard for the equalization of the verticals. But the Nasik inscription establishes the connexion between this style and that of the inscriptions of Nahapāna. These inscriptions may be dated in the last quarter of the first century A.D. and the beginning of the second century A.D. With the coming of the foreigners in this region there is a marked improvement in the style of the writing. These inscriptions have not been shown in the plates as it is hard to distinguish them from those referred to in the following paragraph.

c14.s7.4

² Ibid., xviii. 325–9.

c14.s7.5

³ Ibid., viii. pl. V.

c14.s7.6

The next group of inscriptions of the Western Deccan belongs to the reigns of Nahapāna and Gautamīputra Śātakarṇi (Pl. IX. 2), which partly overlap the earlier series and partly come soon after them. Here again no head-mark is found in the letters. Tapering is faintly noticeable in a few inscriptions of Nahapāna from Nasik. The Karle inscriptions of Nahapāna and Gautamīputra Śātakarṇī show a surprising and remarkable difference in style from those of Nasik, but they are all, no doubt, contemporary. Hence the only conclusion that we can derive is that the older style of writing continued side by side with the new style in Western Deccan right down to the first quarter of the second century a.d. The older style is clearly seen in the persistence of the older system of marking the medial vowels. The change is specially noticeable in the ornamental signs for medial *i* and *ī* and the medial *ā* in *jā*. The two different systems are especially clear in the forms of *a*, *ka*, *kha*, *ga*, *gha*, *ḍa*, *ta*, *pa*, *bha*, *ya*, *la*, *sa* and *ha*. In the older style no particular care is taken over the length of the verticals. But in the new style the verticals are equalized and angularity is clearly marked. The letter *da*, with its mouth open to the left, retains the older form, though it is in some cases cursively drawn. The changes do not indicate borrowing from the Kushāṇa style of Mathura and Sanchi. The optional form of *kha* still gives prominence to the upper hook, while its triangular appendage is insignificant. The stem of the double-looped *chha* is moved on to the left. *Ḍa* is here round-backed. *Bha* is of the broad variety, though the notch does not yet appear. *Ya* does not show a loop or curl, and in the conjuncts maintains its tripartite form. The form of *sa* is still rounded. These fundamental differences clearly show that the writing style of the Western Deccan at this time cannot be derived from the Kushāṇa style of Mathura and Sanchi. On the other hand, they preserve the older features of the Nanaghat style mixed up with the forms derived from the Śaka Kshatrapas of Mathura. It is possible to see some connexion between the Nahapāna inscriptions from Nasik and the ‘early group’ of the inscriptions from Mathura,¹ which may be dated between the Mathura Kshatrapa style and the Kushāṇa style. Could it be that these Kshatrapas were survivors from the Mathura Kshatrapas in the Deccan? On such an assumption we could understand the presence of the Yavanas and the Śakas in the second half of the first century a.d. in the Deccan, where they make a solid contribution in bringing a new style of writing from the north.

c14.s7.7

¹ See above, p. 85.

c14.s7.8

Another definite change is seen in the inscriptions of the Chashṭana group of the Western Kshatrapas (Pl. IX. 5). All their records so far known are from north Gujarat and Kathiawad. The new style that these inscriptions introduce did not develop from the older Western Deccani style, but, as the detailed description will show, borrowed much from Eastern Malwa and from the Kauśāmbī region. It seems that some new cultural elements must have entered Gujarat during the rule of the Chashṭana group of the Western Kshatrapas. The changes are traceable in three distinct features. (i) For the first time in this region special attention is paid to the head-mark. The Junagadh inscription of Rudradāman shows the line head-mark, while a few others have a solid triangular head-mark. The tapering style of the earlier Deccani inscriptions is not found here at all in this period. (ii) Some new shapes of letters are introduced. These remained confined to this region and hardly reached the Deccan. Among them may be mentioned the form of *kha* which gives greater prominence to its triangular base, as opposed to the Deccani *kha* where the upper hook is important. *Ya* with its left arm curled is seen only here. The tailed *ma* and the hooked *ha* are used optionally in the Gadha (Jasadan) inscription of Rudrasena. The hooked type of the subscript *ya* is also employed optionally in this region. There are other forms of the letters which penetrate even to the Deccan. The most important are the forms of *ta* with its rounded appendage and the round-backed *da* with its mouth open to the right. (iii) The third remarkable change is seen in the ornamental forms of the letters. The curves and notches in the verticals and horizontals and the flourishes in the medial vowels are noteworthy. These flourishes seem to be a further development from the earlier practice seen in the Deccani inscriptions. This characteristic is shared equally by the later Śātavāhana inscriptions and it may be that this Gujarat style influenced the ornamental forms of Eastern Malwa. Other Deccani influences are traceable in the use of the medial vowels, the round-backed *da*, the simple *na* and *ṇa* with bent base, and the medial *o* optionally added to the middle of *ma*.

c14.s7.9

In the Western Deccan the change imperceptibly set in from the time of Vāsisthīputra Śrī Puḷumāvi (Pl. IX. 3) who was a contemporary of the early rulers of the Chashṭana group. It was from his time onwards that records of the Śātavāhanas are found in the Eastern Deccan and as far south as the Bellary and Shimoga districts. But the Śātavāhanas seem to have lost for good their territories in Malwa, Northern Gujarat, and Kathiawad to the Western Kshatrapas, who were the real masters of the overseas trade, as is evidenced by their silver currency. The only exception was Gautamīputra Yajña Śrī Śātakarṇi, who managed to lay hands on the region temporarily and to issue some silver currency. The Śātavāhanas from this time onwards became predominantly a Deccan power, and the style in their inscriptions influenced the growth of the southern characters. In the inscriptions of the time of Puḷumāvi a mixed style is seen—the older Deccani style continued side by side with the new influences which possibly came from Northern Gujarat and Kathiawad. In Amaravati and other places of the Eastern Deccan the older Deccani style is seen at this time. The ornamental style developed alongside it, and by the time of Yajña Śrī Śātakarṇi this became the predominant way of writing in the Deccan. In the new style the line head-mark is occasionally seen, the curves and the notches characterize the verticals and horizontals, and ornamental flourishes are added to the medial vowels. Two special forms are the looped *ta* and the triangular *dha* with its left side curved. The subscript *ya* retains its tripartite form. Among the older forms that persisted are *kha* with its prominent upper hook, *chha* with its stem to the left of its double loop, the round-backed *ḍa*, the Deccani *ta*, the cursive angular *da* with its mouth opening to the left, the medial vowels attached to the middle of *ma*, the tripartite *ya* showing no curl or loop at the left arm, and *la* with its long vertical curved outwards. The new forms included the angular *a* with an additional hook attached to the angle of the upper left stroke, *o* in the shape of reversed 's', the notched type of the beaked *cha*, the round-backed *da* with its mouth open to the right, and *la* with its vertical curved inwards. In all these cases the older system of applying medial vowels continued, the only change being that the vowel signs were now more ornamental. In fact in the Deccan, in the period from the time of Haku Siri to that of the Nagarjunikonda inscriptions, ornamental forms in the shape of curves began to appear. They assumed standard forms in the time of Puḷumāvi and Yajña Śrī Śātakarṇi.

Eastern Deccan

c14.s8.1

A number of Buddhist sites, such as Amaravati, Nagarjunikonda, Jogeyyapeta, Ghantasala, and Salihundam, have been explored in this region. The earliest inscriptions have been found at Amaravati. As has been shown in the last chapter, the writing here is a regional adaptation of the Nanaghat style, while the Bhattiprolu casket inscriptions are a rude version of the same style, with some mistaken forms and the evolution of a new letter *ḻa* to meet the needs of the local dialect. On the basis of the Nanaghat style these inscriptions have been dated in the first half of the first century A.D. This style continued until the time of Puḷumāvi, when new ornamental writing gradually penetrated into this region. The early inscriptions of Chanda¹ fall in this pre-Puḷumāvi period. To the same period belongs an unpublished inscription on a coping stone, recording the name of Rājakumārī Saṁmalī, the style of which is similar to no. 12 of Chanda. The second group of the inscriptions of Chanda is characterized by the optional head-mark, the curving of the verticals, and the extensive use of some of the forms from Western Deccan such as *ta*, *bha*, *ḍa*, and the crescentic *ya*—the older forms obviously affected by the new ornamental style. To the same group belongs the inscription published by Mr. D. Barrett.² These inscriptions should be dated to the time of Puḷumāvi. The inscriptions from Salihundam fall in this same group. The remaining inscriptions from Amaravati can again be divided into two groups: the first (Pl. IX. 11) belongs to the time of the later Śātavāhanas from Yajña Śrī Śātakarṇi onwards; the second shows a style identical with that of the inscriptions of the Ikshvāku rulers of this region, who must be dated towards the close of the third century A.D. There are only minor differences between the ornamental writing of Eastern and Western Deccan during the rule of the later Śātavāhanas. These include the optional form of *i*, which has three small crescents in place of dots, the rare use of *kha* with a prominent triangular base, and the optional use of the oval *va* earlier seen in the Bhattiprolu casket inscriptions. The Nagarjunikonda style is a further development from that of the later Śātavāhanas. It is characterized by the following features: (i) the head-mark is a solid triangle; (ii) the writing shows definite draftsmanship in so far as the main body of the letters is reduced in size to a definite proportion and aligned in the middle with flourishes going up and down; (iii) the verticals are exceptionally elongated; and (iv) the curves are exaggerated. The new forms of the letters seen here are the looped *na*, the looped *ṇa*, and optionally the hooked *ha*. The other change is seen in the medial *o* as applied to *to*, *tho*, and *no*. Here we have two separate strokes attached to the top of the letters in such a manner that one diverges from the other. This practice, together with the use of the solid triangular heads, looped *na*, and hooked *ha*, suggests that the writers of these inscriptions received some influence from the Middle Ganges Valley. However, it must be noted that all the inscriptions of the Ikshvākus do not show the same artistic excellence. There are some¹ which are in a very poor hand. Here the long verticals in medial *i* and *ī* are not seen, and even the loop in *ṇa*, *na*, and *ta* is omitted. This second group is a degenerate form of the main Ikshvāku style.

First 68 of 126 passages. Read on

Linked texts (18)