Pramāṇa · Search · About · API

Indian Palaeography c12: 4. Aśokan Brāhmī

Indian Palaeography c12

Language
English

Contents

4. Aśokan Brāhmī

c12.2

THE Indian alphabetic¹ system, as developed fully by the Sanskrit grammarians, is not at present known from actual specimens of writing till a very late period. This system, so far as it was needed by the popular language of the inscriptions, found expression at the same time in two scripts, the Brāhmī and the Kharoshṭhī. These were used in the inscriptions issued by a king who bore the title of Devānaṁpiya Piyadasi, belonged to Magadha (South Bihar), had his capital at Pāṭaliputra (modern Patna), and whose name, as given in two of the records, was Aśoka. He has long been identified with the Mauryan king of the same name, reigning between *c.* 272 and 231 B.C. The country over which he ruled was called Jambudvīpa, its limits being defined by the places where the inscriptions have been found as well as by the names of the bordering kingdoms of the Choḍas, Pāḍas, Satiyaputas, and Keralaputas in the south and those of the Yona (Greek) kings in the north-west. It covered almost the whole of the subcontinent of India and Pakistan north of the Kaveri Valley. Throughout this vast empire the inscriptions, called (*dhamma-*) *lipi*, *libi*, or *dipi*, were written in these two scripts—the Kharoshṭhī used in the Indus Valley zone, and the Brāhmī in the remaining parts. The Brāhmī script of these inscriptions is termed here ‘Aśokan Brāhmī’, as it is well defined in age, language, content, and common character.

c12.3

¹ The term ‘alphabetic’ has been used here in a popular sense, as the Indian scripts are generally so called.

Material of Writing

c12.s1.1

The inscriptions are engraved on stone pillars (*silāthaṁbha*) and stone slabs (*silāphalaka*). The pillars are monolithic, made of Chunar sandstone, chiselled to a round shape, rubbed, and highly polished. The places where the pillars are found lie in the Ganges Valley, except Rummindei and Nigali Sagar, which are in the Nepalese Terai but not far from the river Gandak—a tributary of the Ganges—and Sanchi in Central India, on the main route from Delhi and Kanauj to Gujarat and the south. Probably these pillars were transported by river, and short distances overland were covered by bullock-carts, as was done later in the time of Sulṭān Fīroz Shāh Tughlaq, who transported some of the pillars. The huge expenditure on transport suggests that the stone-cutter’s art was not then widespread, and, combined with this fact, the difficulties in carrying the pillars to distant places by road explain very well their absence outside the Gangetic region.

c12.s1.2

The term *silāphalaka* (stone slab), used by Aśoka, must refer to the blocks of rock on which the edicts are engraved, as only one stone slab (the Calcutta Bairat slab) has so far been discovered. The rocks are of different types—granite, quartzite, trap, and others, whichever was available on the spot, but the outstanding block in the locality was always chosen for the inscriptions. No attempt was made to prepare the surface, except in the case of the pillars, where engraving was done after polishing. In the case of rocks a naturally flat surface was selected. As a result the pillar inscriptions are better written, while the rock inscriptions are comparatively rough.

Technique of Engraving

c12.s2.1

Rock engravings of about this period and of slightly later date¹ have been discovered in India at several places. Here the figures are drawn by scratching. But the technique of the engraving of the Aśokan inscriptions was far in advance of these. We find the technical terms *lipikara* or *dipikara*, derived from the original Achaemenian *dipi* and used with the verb *likhite* or *lekhite*, from the root *likh*, ‘to draw’ or ‘to write’. In the Shahbazgarhi R.E.² the verb used is *nipistaṁ* or *nipesitaṁ*, formed from the root *ni-pis*, equivalent to the modern Persian *navishtan*, also meaning ‘to write’. Hence the word *lipikara* or *dipikara* is taken to mean ‘writer’. Bühler translates it as ‘clerk’. These terms from the Aśokan inscriptions do not explain the complete process involved in engraving on stone. But it is significant that the emphasis was laid on ‘writing’, because the writer’s part was of primary importance.³

c12.s2.2

¹ D. H. Gordon, *Prehistoric Background of Indian Culture*, Bombay, 1958, pp. 98–118.

c12.s2.3

² For a complete list of the Aśokan inscriptions, see Bibliography, pp. 48–49.

c12.s2.4

³ Compare Aśoka’s remark in the 14th R.E.

c12.s2.5

Another term for engraving found in literature is *chhindati*, from the root *chhid*, meaning 'to cut'. These two verbs together describe the whole process of engraving. There were two stages: the first was the writing of the inscription on the stone by a *lipikara*, and the second was the actual work of cutting the letters on the stone. The former was done by a literate man, but the latter was the work of a stone-cutter who was probably illiterate, judging from the way he blindly followed the long wavy lines of the writer. The letters were cut by chiselling along the lines drawn by the writer, but the chiseller did not always follow the same process of forming the letters as was done by the writer. The writer, of course, was responsible for the shape and the style of the letters, but the chiseller made slight variations; e.g. while cutting the letter *ya* he might begin with [Figure:] or [Figure:], and then add the second hook as [Figure:] or [Figure:]. These variations do not make any fundamental difference in the actual form of the letters. The very technique of engraving implies such variations. But it is important to note that the engraver's part was secondary in the Aśokan inscriptions. It was the writer who dictated the style of writing.

c12.s2.6

It was on the writer's method of producing the letters on a given surface that their actual shape depended. Leaving aside individual mannerisms, which are also a factor to be considered, we can detect the final influence of the tool of writing on the shapes of the letters. Individual mannerisms, if common to a large group of writers in a given region, determine the regional school, but the introduction of new tools results in a new style of writing, with consequent simplification or elaboration of shapes. After all, writing is an art, and its principles are not different from those governing other arts. In the Aśokan inscriptions the writer used the simplest of tools—a longish piece of chalk, charcoal, or haematite—so that the lines drawn were uniform in width. His hand is clearly discernible in the long wavy lines at Girnar and Jaugada. This device cannot result in the formation of the so-called 'serifs' on the tops of verticals. Bühler was mistaken in detecting these marks in Aśokan Brāhmī. Their formation is due to the introduction of a new tool in a subsequent period (see Chapter 5). The size of letters again is not a fundamental factor, for it depends on the available space. More important is the ratio between the verticals and the horizontals; double is usual, but longer forms, as in the Mysore inscriptions (i.e. Brahmagiri, Siddhapur, and Jatinga Ramesvara), and shorter (as in Kosam P.E.), are also to be found.

c12.s2.7

*Method of reproduction*

c12.s2.8

It is almost impossible to reproduce on paper all the peculiarities of the writer and the engraver. The photographs and the tracings of the inscriptions hitherto published do not tell us as much as we can learn from the original stone. Moreover, one cannot be certain whether one has given all the variations of a particular letter unless one traces all its occurrences in the inscriptions. Herein lies a difference between handwriting and mechanical reproduction. The most important point is to find the manner in which the letters were actually formed; once we discover this, the shapes emerge of themselves. In Pls. III and IV, illustrating Aśokan Brāhmī, no attempt has been made to trace the letters. They have been copied from the published photographs after determining the various methods of their formation, and this process has given the important varieties. It has also helped to correct many mistakes made by earlier palaeographers. The letters have not been drawn to scale, as this was not possible from the published photographs, and hence no conclusion should be derived from the drawings as regards their size and proportion. The rare types are shown in brackets.

c12.s2.9

*Arrangement of plates*

c12.s2.10

There is at least fifteen to twenty years' difference between the writings of the earliest and the latest records, but no gradual improvement in the engraving of the letters is to be noticed, except that the pillar inscriptions are better written than the rock inscriptions. Hence the letters cannot be arranged on the basis of chronology. It would again be wrong to separate rock inscriptions from pillar inscriptions, as there is no difference in the actual method of writing. The regional groupings—northern and southern (according to the theory of Bühler and, following him, others), eastern and western—do not work out, as exactly the same methods are traceable in different regions. The various inscriptions show individual characteristics, and I have found them hard to reduce to any one or two systems. Even on one particular rock different hands are traceable. As a result it was considered desirable to copy in separate rows as many hands as possible. These rows of alphabets have then been classified in two plates (III *a*, *b*) showing letters from the pillar and minor pillar inscriptions and the Kalasi R.E., all of which come from north of the Vindhyas. The pillar inscriptions are on a prepared surface, but they hardly differ in style of writing from the rock edicts, which are on a natural rock surface. Pl. IV *a*, *b* depicts the writing from the remaining rock edicts and minor rock inscriptions, all of which come from south of the Vindhyas. The plates also include the Bairat slab, the Gujarra and Sasaram M.R.I., and the Barabar cave inscriptions. Except for the last these inscriptions are engraved on natural rock surfaces. The few North Indian inscriptions are given here for the sake of comparison with the inscriptions of the south. But, as the plates show, there is hardly any recognizable basis on which to differentiate between northern and southern styles.

c12.s2.11

The inscriptions suggest that knowledge of writing travelled along the imperial routes, and that the imperial scribes were the carriers of Aśokan Brāhmī to the distant parts of the empire. Otherwise it is difficult to explain how the same type of the script is to be found in the far-flung areas of north, south, east, and west. Since in the north-west Aśoka used Kharoshṭhī, Aramaic, and Greek¹ scripts, it is very unlikely that he would not have used regional scripts in other parts of his empire, if such had existed. Aśokan Brāhmī does not tend to show regional variations, but the Aśokan inscriptions mark the process of cultural infiltration into the distant parts of this subcontinent.

c12.s2.12

¹ 'Une bilingue Gréco-Araméenne D'Aśoka', by Daniel Schlumberger in *Journal Asiatique*, 1958, pp. 1–48.

Style of Writing

c12.s3.1

The writing can be said to follow the ink style, as is clear from the use of dots in some letters, e.g. *kha*, *ja*, and *ḍa*, and from the free movement of the hand resulting in round forms. It is seldom that the scribe attempts to achieve angularity, and in most cases where he does so, e.g. *u*, *e*, *o*, *ka*, *ga*, *jha*, *ña*, *ḍa*, *ṇa*, *ta*, *na*, *ba*, and *bha*, he is governed by the original form of the letters rather than by his own inclination. The optional examples are *a*, *ā*, *gha*, *cha*, *ja*, *ṭa*, *ḍha*, *da*, and *śa*. There is no abruptness in writing, the only exceptions being in the minor rock inscriptions of Mysore State, where in the formation of *a*, *ja*, *ma*, *yi*, and *hi* the hand stops and begins from another direction to complete the letters. But this defect might be due to the engraver, and not to the writer. Each letter is distinctly formed, and no cursive writing in the way of continuous drawn-out lines is noticeable. This last tendency is a characteristic of most Indian scripts, though in a very few cases, especially in much later mercantile documents, this rule is broken. The cursive hand, if at all traceable, is detected in the medial vowels and the conjuncts. To this general rule the letters *cha*, *ja*, and *va* are exceptions, as their formation is much varied. In the Kauśāmbi Queen's Edict (hereafter called Kauśāmbi M.P.I.) the letters *da* and *ḍa*, and in the Kalasi R.E. in one place the letter *ḍa*, are very cursively drawn.

The Initial Vowels in Aśokan Brāhmī

c12.s4.1

Only six initial vowels were used in Aśokan Brāhmī. Of these, three are basic: *a*, *i*, and *u*; while three others, *ā*, *e*, and *o*, are secondary forms.

c12.s4.2

*The letters a and ā.* Bühler distinguishes 'eight most important' forms merely from the external shape. From the point of view of technical formation they fall under three broad heads:

c12.s4.3

(i) In this series the vertical is drawn first and the two curves to the left of the vertical subsequently. The varieties originate as the two curves come closer to each other and finally their one arm merges with the other (Fig. 4. I). This series is found in most of the columns, but is rare at Girnar.

c12.s4.4

(ii) In this series the curves on the left are replaced by straight lines which make an acute angle. They are drawn first and then the vertical is made. Sometimes the vertical is bent to touch the apex of the angle, or a short bar is drawn to join the apex with the vertical (Fig. 4. II). The first two examples are seen in the Mysore inscriptions (Pl. IVa), the third in the Girnar R.E., Sopara R.E., and Erragudi, Rajula Mandagiri, and Gujarra M.R.I.s. The last can be seen in the Dhauli and Jaugad R.E.s, and Kosam P.E., Girnar R.E., and Mysore inscriptions.

c12.s4.5

(iii) In the third series the vertical is drawn first and then one of the left strokes is brought to the middle of the vertical either straight or cursively, and from this meeting-point another stroke goes out either straight or cursively (Fig. 4. III). These forms are distributed fairly evenly in most of the columns, but are rare at Girnar.

c12.s4.6

The long *ā* is generally formed by adding a horizontal bar to the right of the vertical where the left strokes meet. But in the Girnar R.E., Gujarra M.R.I., and Erragudi and Rajula Mandagiri M.R.I.s the horizontal bar is added to the top or near the top of the vertical.

c12.s4.7

FIG. 4

c12.s4.8

*The letters i and e.* The letter *i* consists of three dots placed trianglewise, and when these dots are joined by lines they form *e*. In most cases the apex is on the right; sometimes it is on the left, as in the Girnar R.Es.; sometimes it is at the top, as at Maski, but very rarely at the bottom, as in the Sopara R.E. and Sarnath M.P.I. *E* is never seen with its apex downward. The horseshoe form of *e*, which Bühler finds in the Kalasi R.E., is unjustifiable. Three triangular points are the main thing in the formation of *i* and *e*, and they cannot result in a horseshoe form. At best the lines joining these points might be slightly curved, and very rarely this is noticeable. It is not shown in the plates illustrated.

c12.s4.9

*The letters u and o.* *U* is formed with two straight lines meeting generally at a right angle, and very rarely at an acute angle, the angle opening on the right. The vertical line is usually longer than the horizontal, suggesting that the former was drawn first. *O* is formed by adding a horizontal bar to the top left of *u*, but the whole letter is made in one action starting from the top left end, and consequently in *o* the middle vertical line is inclined (Fig. 4. IV). In the Dhauli and Jaugad R.E.s the letter *o* is formed in the reverse fashion, as shown in the same figure.

The Consonants in Aśokan Brāhmī

c12.s5.1

Thirty-three consonants, including the rare North Indian *ḍa*, were used in Aśokan Brāhmī. Their formations are described below:

c12.s5.2

(i) The letter *ka* consists of a simple cross. Occasionally the vertical line is lengthened, as in the Mysore inscriptions, Erragudi R.E., and Rummindei M.P.I. But in these cases, except at Erragudi, all the letters have longer verticals. At Siddhapur alone the horizontal line is inclined on the left. Bühler's 'dagger-shaped *ka*' is not traceable. That was the result of a new technique (see Chapter 5).

c12.s5.3

(ii) The letter *kha* resembles a fish-hook upside down, with a dot or a circle at the end of the longer arm, but the primary form is the latter, as shown in the last chapter (Fig. 4. V). This dot or circle is very rarely missing (Delhi-Topra P.E., Pl. III*a*). The long arm is bent in some cases, thus making a cursive form (Delhi Topra P.E. VII). In the Mysore inscriptions the dot is thickened, recalling the circle. *Kha* with a circle is confined to North Indian inscriptions but is found in the Dhauli and Jaugad S.E., while *kha* with a dot is common to all the inscriptions except the Kalasi R.E. *Kha* with a triangle in place of a circle, noticed by Bühler at Kalasi, is hardly observable. The triangle is really a cursive circle, as shown in Pl. III*a* (Kalasi R.E.).

c12.s5.4

(iii) The letter *ga* is formed by two straight lines meeting at an acute angle which opens downwards. Sometimes the apex of the angle is rounded off as in the Barabar cave inscriptions (Pl. IV*a*).

c12.s5.5

(iv) The letter *gha* is formed from the rounded form of *ga* turned upside down, but its left arm is extended upwards and a vertical stroke is added in the middle (Fig. 4. VI). This round form is very common, but in the Kalasi R.E. (XI–XIV) and rarely in the Dhauli and Jaugad R.E.s (Pl. IV*a*) it has a flattened bottom. The S.E. has no angular form as noted wrongly by Bühler.

c12.s5.6

(v) The letter *cha* has three main varieties: the first is formed by drawing a semi-circle to the bottom left of a vertical line. In this case the vertical line was probably drawn later (Pl. III*a*, Delhi Topra P.E. VII). In the second variety the vertical line is drawn first and then the semi-circle is made. As a result it is always less than the arc of a semi-circle. This second variety, when written cursively, produces the angular *cha*. It was probably formed in a single action beginning from the vertical (Fig. 4. VII). The second variety is most common, while the third is an optional form.

c12.s5.7

(vi) The letter *chha* is formed from *cha* by duplicating the semi-circle on the right, and thus completing the circle. Usually the circle was drawn first and then the vertical. In the reverse process, i.e. when the vertical was made first, the circle took the form of an oval; e.g. the Delhi-Topra P.E. (I–VI), and Sanchi M.P.I. Sometimes this resulted in two circles (Pl. III*a*, Kalasi R.E. I–X). The *chha* with a notched head-mark, ☿, shown in Bühler’s chart, from the Kalasi R.E., is a mistaken reading. This is a much later development (see Chapter 5).

c12.s5.8

(vii) The letter *ja* has two basic forms, the angular with three arms and the round with double curves; the latter is most common, and the first is optional in many inscriptions. The round form has several sub-varieties, depending on the formation of the middle arm (Fig. 4. VIII). Of these (*b*) is not the only type in the southern inscriptions, as is maintained by Bühler, though it is confined to the Girnar and Mysore inscriptions. (*c*) and (*d*) are present in the northern inscriptions along with others.

c12.s5.9

(viii and ix) The letter *jha* has only the angular form made by adding on the right an angular hook to the middle of the vertical so that the mouth opens upwards. Variations are seen in the shortening of the hook as in Pl. III*a* (Delhi-Topra P.E. VII), or in the angle between the hook and the vertical (Fig. 4. IX). *Ña* is this form of *jha* upside down with a stroke added to the top left.

c12.s5.10

(x) The letter *ṭa* has only one main form, a semi-circle open to the right, but more often it is not a full semi-circle (Erragudi R.E., Pl. IV*a*). Sometimes the angular form is cursively drawn by flattening the upper and lower arms (Pl. IV*a*, Gujarra and Rupnath M.R.I.; and Pl. III*a*, Delhi-Mirath, and Sarnath inscriptions). The process is given in Fig. 5. I.

c12.s5.11

(xi) The letter *ṭha* is a full circle.

c12.s5.12

(xii) *Ḍa* is formed by adding a vertical stroke to the top of *ṭa*, but the semi-circle of *ṭa* is straightened, and it looks like a step. It is drawn from the top downward in one action. It has only the angular

c12.s5.13

FIG. 5

c12.s5.14

variety. The round *ḍa* found in the Kauśāmbī M.P.I. (Pl. III*a*) and optionally in the Kalasi R.E. (XI–XIV) is an exception.

c12.s5.15

(xiii) *Ḍha* seems to have been originally formed by adding a loop on the right to the bottom of *ḍa*, as is seen at Siddhapur. But this angular variety is rarely seen (Delhi-Topra P.E. VII). The round form is very common. It is also drawn from the top downwards.

c12.s5.16

(xiv and xv) The letters *na* and *ṇa* should be taken together. *Na* is formed by drawing a perpendicular line on a horizontal base; while in the case of *ṇa* an additional horizontal line is given at the top of the vertical and at the same time both the horizontals are shortened in relation to the vertical line (Fig. 1). In some examples from the Mysore inscriptions the vertical is exceptionally prolonged.

c12.s5.17

(xvi) The letter *ta* is formed in several ways. The simplest is formed from an inclined straight line from the middle of which another short stroke is given at an angle either on the right or on the left as the case may be; e.g. 人 or 入. The curving of the additional stroke makes other varieties: 𑀢 or λ. A different formation is obtained by appending an angle just below the vertical: 人, and this last also makes a variation when the angle is cursively drawn into a curve: 𑀢. Two or three forms are found in the same inscription and hence no regional classification can be made.

c12.s5.18

(xvii) The letter *tha* has a dot in the centre of *ṭha*, i.e. a full circle (Fig. 1).

c12.s5.19

(xviii) The letter *da* has two main varieties, the angular and the round. The primary form is made up of a semi-circle, open on the left, with the addition of short verticals at either end; e.g. כ → 𑀤. In the other variety the semi-circle gives place to an angular form; e.g. 𑀤. The lower vertical is sometimes cursively drawn, suggesting that the letter was usually formed in a single action beginning from the top, e.g. 𑀤. The Kauśāmbi M.P.I. and Sarnath M.P.I. have very cursive forms.

c12.s5.20

(xix) The letter *dha* has two main varieties. The first resembles the Roman capital letter ‘D’, and the second is its exact reverse. The first is very common. Bühler is right in deriving it from *da*. In this case the verticals meet at the centre of the semi-circle instead of going farther away from it as in *da* (Fig. 1).

c12.s5.21

(xx) The letter *pa* resembles a fish-hook drawn to the right. Very rarely the angular form is seen (Pl. III*b*, Kalasi R.E. XI–XIV).

c12.s5.22

(xxi) *Pha* is formed by adding a loop to the hook of *Pa* (Fig. 1). Sometimes we find very cursive forms, when the vertical is inclined (Pl. IV*b*, Girnar R.E.; Pl. III*b*, Delhi-Topra P.E. VII).

c12.s5.23

(xxii) *Ba* makes two varieties, a square or a rectangle. The first is very common.

c12.s5.24

(xxiii) *Bha* is a sign similar to *jha* but turned upside down (Pl. IV*b*, Rupnath M.R.I.). Usually there is a notch between the vertical and the angular hook. But actually this letter is formed from *ba* (Fig. 5. II).

c12.s5.25

(xxiv) The letter *ma* has several varieties. The basic form is the bottom loop or circle with two additional strokes above. When the letter is drawn in a single action from the top stroke with one sweep, the result will be the *ma* of the Girnar R.E. (Fig. 5. III*a*). If in this action the line bends slightly at the beginning, the result will be something like Fig. 5. III*b*; if the line bends again at the end, we get another form (Fig. 5. III*c*)—a variety which is seen in many inscriptions. But the engraver has not followed the writer's principle. We can see him at Siddhapur (Pl. IV*b*) first drawing the circle and then supplying the upper strokes separately (Fig. 5. III*d*); in the Erragudi M.R.I. (Pl. IV*b*) the upper strokes are drawn straight at an incline.

c12.s5.26

(xxv) The letter *ya* has two basic forms, the crescentic and the segmented. The first is made by drawing a vertical in the middle of the arc of a circle; while the second is formed by dividing this arc into two halves. Both have angular varieties. The process is shown in Fig. 5. IV.

c12.s5.27

(xxvi) The letter *ra* is not found in all the inscriptions. It has two varieties: (*a*) the straight vertical, and (*b*) the serpentine vertical.

c12.s5.28

(xxvii and xxviii) Both the letters *la* and *ha* are formed in the same way, with the only difference that the one is the exact reverse of the other (Fig. 5. V). Cursive forms of *la* can be seen in the Kalasi R.E., and of *ha* in the Delhi-Topra P.E. VII. The tick sometimes makes a little hook (see Dhauli and Jaugad R.E., Pl. IV*b* for *la*, and Erragudi R.E., Pl. IV*b* for *ha*). At Sarnath this tick is formed by a twist as shown in Fig. 5. V*b*.

c12.s5.29

(xxix) The main variety of *va* is a circle appended to a vertical. But the circle becomes longish when the letter is drawn in a single action, and more often it appears like a loop on the right bottom of the vertical (Fig. 5. VI).

c12.s5.30

(xxx–xxxii) All the three sibilants are found in Aśokan Brāhmī, but the commonest is the dental *sa*, occurring in all the inscriptions. The other two are present in the Kalasi R.E. (XI–XIV), while palatal *śa* is found also at Kalasi (I–IX), at Maski, in the Mysore inscriptions and in a conjunct with *va* at Bairat. The cerebral *sha* has also been read by Hultzsch and others at Sarnath and in the Kauśāmbi M.P.I., but this reading is mistaken. As the forms of *sa* and *sha* do not differ much, there was often confusion on the parts of the writer, the engraver, and the reader. The same confusion prevails in the Mahasthan Brāhmī inscription.¹ All the three sibilants are found in Aśokan Kharoshṭhī, and the fact that the North-western Chapaḍa uses them in the Mysore inscriptions suggests that their occurrence is due to the hand of a person coming from the north-west.

c12.s5.31

¹ See below, p. 56.

c12.s5.32

The palatal *śa* resembles the form of *ga* with an additional stroke appended in its inner side (Fig. 5. VII). Both the angular and round forms are known. In the angular form the inner stroke is attached to the left arm almost parallel to the right arm, and in the round form the stroke falls from the apex.

c12.s5.33

The dental *sa* has an additional hook in the same direction attached to the long arm of the form of *pa* (Fig. 5. VIII). This hook is sometimes angular as in the Erragudi inscriptions (Pl. IV*b*), and is sometimes lengthened (Pl. III*b*, Kalasi R.E. I–IX); in some cases this hook starts just where the long arm of *pa* ends, and in rare cases it drops from the curve of *pa* (Pl. III*b*, Kalasi R.E. XI–XIV). In this last variety it is often confused with *sha*. In rare cases, e.g. at Girnar (probably owing to the engraver's mistake), the long arm is extended and the curve is added later, e.g. [Figure:] (Pl. IV*b*). At Sarnath the cursive form is also seen, in which case the long arm is bent inwards.

c12.s5.34

In cerebral *sha* the curve is doubled and the separate hook is dropped. The main point to note is that the lower curve in *sha* is the primary one and it never takes the form of a hook, while the upper one is secondary, answering to the hook. This distinction is clear when we take into consideration the addition of the medial vowel *u*, which is always added to the primary curve (Fig. 5. VIII).

c12.s5.35

(xxxiii) In the pillar edicts one more letter, *ḍa*, is observed. It is used optionally only with three words: *eḍaka*, *duḍi*, and *paṁnaḍasa*—the first two being the names of animals. It is formed by adding a dot at the bottom of the line of *ḍa*. This is the parent of the modern North Indian *ḍa*, in which case also a dot is placed below the form of *ḍa*.

The Formation of the Conjuncts

c12.s6.1

The conjuncts are formed by placing one consonant on the top of the other. Unlike the practice in modern Indian scripts, the first letter is fully formed and the second is cursively added. According to the later rule the first letter should be mutilated, so as to drop its vowel, as the Sanskrit grammar requires, but this is not the case in Aśokan Brāhmī.¹ In Aśokan Brāhmī the following conjuncts are found:

c12.s6.2

¹ Cf. Bühler's remark on p. 15 of his *Indian Palaeography*, where he is mistaken.

c12.s6.3

(i) Having the second element *ya*:

c12.s6.4

*kya*, *khya*, *gya*, *chya*, *tya*, *dhya*, *nya*, *bhya*, *mya*, *vya*, *sya*, and *hya*.

c12.s6.5

In all these cases *ya* comes later, except in the Girnar R.E. where *vya* is spelt like *yva*. G. H. Ojha takes this as a mistake, and, judging from the practice in all other inscriptions, I think Ojha is right.

c12.s6.6

(ii) Having the second element *va*:

c12.s6.7

*tva*, *dva*, *rva*, *śva*, *sva*, and *hva*.

c12.s6.8

In all these cases *va* comes later, and is added, sometimes with its vertical and at times without it, to the right, centre, or left.

c12.s6.9

(iii) Having the second element *ba*. There is only one word, *dvādasa*, which is spelt sometimes with *va* and sometimes with *ba* in the Girnar R.E. Does this mean that the writer of the Girnar R.E. hailed from Bengal?

c12.s6.10

(iv) Having the second element *ra*:

c12.s6.11

*kra*, *tra*, *dra*, *pra*, *vra*, and *sra*.

c12.s6.12

This conjunct is very ingeniously formed. As the letter *ra* makes a serpentine vertical, in conjunction with another letter it replaces the straight vertical of that letter. When this conjunct is drawn, the letter *ra* is formed first and the other consonant follows. *Ra* is fully drawn and the second letter is added in abbreviated form. Normally the first component of a conjunct is placed above the second, but this is an exception. It is for this reason that the same sign is read as *sra* and *rsa*, or *vra* amd *rva*. In the case of *dra* in Mysore inscriptions both the verticals are serpentine. On the evidence of this principle the reading of *dra* in Jaugad R.E. IV, line 3, by Hultzsch is not justified. Similarly, in the Bairat M.R.I., line 3, the reading *rve* is not correct. We do not find the serpentine *ra* in these cases. The idea of *repha* had not yet developed. For the same reason the reading of *dhra* in Girnar R.E. I, line 12, and edict XIII, line 9, and Kalasi R.E. I, line 4, is mistaken. A knotty problem arises from the *bra*, read twice in Girnar R.E. IV, lines 2 and 6, simply because the left vertical of the rectangular *ba* has a slight notch in the middle. This notched vertical cannot be taken to be *ra*, firstly because it is exceptional (and actually doubtful in the second example), and secondly (which is fundamental) because it does not conform to the principle given above.

c12.s6.13

(v) Having the second element *sa*:

c12.s6.14

*rsa*. See my remark in (iv) above.

c12.s6.15

(vi) Having the second element *ṭa*:

c12.s6.16

*ṣṭa* is found in the Girnar R.E. III and IV. But *ṭa* is written first and then *sa* is cryptically added below it. This may be an exceptional case like that of *ra* (but see below, (vii)).

c12.s6.17

(vii) Having the second element *ta*:

c12.s6.18

*sta*, *pta* (but read as *tpa*).

c12.s6.19

In both these cases *ta* is an adjunct of the first letter which is fully drawn. The reading of *pta* as *tpa* in the word *ātpa-pāsaṃḍa* is grammatically justified, but whether that was the actual pronunciation at the time is doubtful. If it was so, it has to be reckoned as an exception like *ṣṭa* in (vi). But more probably this was not the case, if we remember how the letters are reversed in popular language; e.g. *kāchū* for *chākū* in Punjābī. Hence even the *ṣṭa* of (vi) occurring at Girnar is a mistake, just as is *yva* at the same place.

c12.s6.20

(viii) Having the second element *pa*:

c12.s6.21

*spa* occurs at Girnar in the normal fashion.

c12.s6.22

(ix) Having the second element *ma*:

c12.s6.23

*sma* and *hma* are used according to the normal rule.

c12.s6.24

(x) Having the second element *ha*:

c12.s6.25

*mha*. In the word *baṁ-mh* (or *hm*) *a-na* two spellings are followed, the first at Girnar and the second at Kalasi (R.E. XI–XIV). The confusion in the spelling of this word survives in the modern Indian languages.

The Medial Vowels and the Anusvāra

c12.s7.1

Bühler (pp. 14–15) seems to suggest that the addition of the medial vowels to the consonants is based on the same principle as that of the conjuncts, and for his proof he quotes the examples of the medial signs *o* and *u*, especially Pl. III*a* (Kalasi R.E. I–X), where the full form of *u* is subjoined. But this principle is not borne out by other examples. In fact the signs for the medial vowels are based on the three basic initial vowels: *a*, *i*, and *u*, and from these basic forms other signs are developed, just as in the case of the initial vowels, obviously by the Sanskrit grammarians.

c12.s7.2

(i) The medial vowel *a*, according to the rule of Sanskrit grammar, is present in all the consonants, and hence there is no separate sign for it.

c12.s7.3

(ii) The medial vowel *i* is marked primarily by adding a vertical stroke on the top right of the consonant (e.g. *ji* and *ṭi*), but in order to distinguish it from the vertical of the consonant it was usual to draw a short horizontal bar between the two verticals. In the case of the letters *ba*, *ṭha*, and *tha* the vowel is added to their middle right.

c12.s7.4

(iii) The medial vowel *u* is marked by a vertical (e.g. *nu*), horizontal (e.g. *ku* and *dhu*—the Kalasi example is an exception), or slanting (e.g. optional cases of *chu* and *khu*) stroke at the bottom right of the consonant.

c12.s7.5

(iv, v, and vi) The long vowels (i.e. *dīrgha* in Sanskrit) are formed by an additional stroke in the same manner: two verticals on the top for *ī*, and two strokes downward (e.g. *sū* and *pū*) or horizontal (e.g. *dhū*) for long *ū*. The principle governing the long medial *ā* is the same as that of deriving the initial *ā* from the initial *a*, i.e. the addition of an horizontal stroke to the top right, but sometimes it is also placed in the middle right (e.g. *ṭhā*, *thā*, *bā*, and *mā* only in the Girnar R.E.).

c12.s7.6

(vii and viii) The medial vowels *e* and *o*, which, according to Sanskrit grammar, are the *guṇa* forms of *i* and *u*, are derived from them by the addition of a horizontal stroke to the top (or middle) left of the consonant, but this addition cancels the vertical stroke of *i* and transfers the horizontal stroke of *u* from bottom to the top right (Fig. 5. IX, X).

c12.s7.7

(ix and x) The vowels *ai* and *au* are again, according to Sanskrit grammar, the *vṛiddhi* of *e* and *o*, and hence as before an additional stroke to the left is given (Fig. 5. IX, X). Actually the medial vowel *au* does not occur in Aśokan Brāhmī, but this principle is observed in the later inscriptions.

c12.s7.8

Throughout the formation of the medial vowels the hand of the Sanskrit grammarian is clearly observable. The signs follow the grammar, and not vice versa.

c12.s7.9

In some examples cursive forms of the medial vowels can be seen. In these cases the angular shape of *i* becomes rounded (*pi* in the Erragudi R.E.); *u* in *tu* slants upward (the Dhauli and Jaugad R.E.); *o* in *go* (the Rummindei M.P.I.) becomes one long stroke going across the top of the consonant. In the Girnar R.E. the strokes for *mā*, *me*, and *mo* are given in the middle instead of at the top of the consonant.

c12.s7.10

The *anusvāra* consists of a single dot placed on the right top, sometimes in the middle, and very rarely at the bottom right of the vowel or the consonant. In some cases where the medial *i* is found (e.g. *kiṁ*, Pl. IV*a*; *tiṁ*, Pl. III*a*) the *anusvāra* is placed at the angle of this vowel.

c12.s7.11

In the Jaugad S.E. there are three *svastikas* with curved arms and three *ma*’s engraved in the margin, one of the latter having a possible dot in the centre of its circle. It could be read as *maṁ*, and taken to stand for *maṅgala*, meaning auspicious. In the Rajula Mandagiri M.R.I. trisceles occurs. As these symbols are definitely auspicious marks, the attribution of alphabetic meaning to one of them seems far fetched.

Punctuation in Aśokan Brāhmī

c12.s8.1

Punctuation is an exception rather than a general rule in Aśokan Brāhmī. In the pillar edicts especially and in others rarely we notice distinct spaces left between the words, but it should be noted that this was not so in all the cases. The idea of writing each word separately had developed but was not regularly followed. On the other hand, in the Kalasi R.E. (XI–XIV) the vertical line, i.e. the *daṇḍa*, in later scripts equivalent to the full-stop, appears generally along with the use of the cerebral *sha*. In this rock edict each edict has been separated from the other by placing a curve resembling the closing end of a parenthesis at the finish. In the Sahasram M.R.I. the *daṇḍa* regularly occurs at the end of the sentence., In the Maski M.R.I. it is again marked, but not at the proper place.

Conclusion

c12.s9.1

On the whole the inscriptions of Aśoka are well written. They were revised by the scribes after engraving. We notice the addition of missing letters here and there and also the erasing of mistaken repetitions. Aśokan Brāhmī shows two or three forms of the same letters, depending on the style of writing or engraving, but no one style can be attributed to any particular region. Sometimes two reversible varieties are noticed, e.g. D or Ɑ for *dha*; ƪ or ʃ for *o*; ⊦ or ⊣ for *no*; but, except for the second form of *o*, which is peculiar to the Dhauli and Jaugad R.E., no regional classification is possible.

c12.s9.2

BÜHLER, J. G., *Indian Palaeography*, Eng. tr. by J. F. Fleet, published as an appendix to the *Indian Antiquary*, xxxiii, 1904. HULTZSCH, E., *Corpus Inscriptionum Indicarum. I. Inscriptions of Aśoka*, Oxford, 1925. OJHA, G. H., *Bhāratīya Prāchīna Lipimālā* (*The Palaeography of India*), 2nd ed., Ajmer, 1918. UPASAK, C. S., *The History and Palaeography of Mauryan Brāhmī Script*, Nalanda, 1960.

List of Aśokan Inscriptions

c12.s10.1

A. *Rock Edicts* (abbreviated as R.E.)

c12.s10.2

(i) The Girnar R.E. in Junagarh, Bombay State.

c12.s10.3

(ii) The Kalasi R.E. in Dehra Dun district, U.P.
(iii) The Dhauli R.E. in Puri district, Orissa.
(iv) The Jaugad R.E. in Ganjam district, Andhra.
(v) The Bombay Sopara Fragment in Thana district, Bombay State. Hultzsch published only the fragment of R.E. VIII. R.E. IX is published in *Indian Archaeology, A Review*, 1956–7, Delhi, p. 73, pl. LXXXIXA.
(vi) Erragudi R.E. in Bellary district, Madras. Only R.E. IV and VIII are published in *An. Rep. A.S.I.*, 1928–9, pl. LXI.

c12.s10.4

The R.E. in Kharoshṭhī

c12.s10.5

(vii) The Shahbazgarhi R.E. in Peshawar district, West Pakistan.

c12.s10.6

(viii) The Mansehra R.E. in Hazara district, West Pakistan.

c12.s10.7

B. *The Pillar Edicts* (abbreviated as P.E.)

c12.s10.8

(i) The Delhi Topra pillar.

c12.s10.9

(ii) The Delhi Mirath pillar.
(iii) The Lauriya Araraj and Lauriya Nandangarh pillars in Champaran district, Bihar.
(iv) The Rampurva pillar in Champaran district, Bihar.
(v) The Allahabad Kosam pillar.

c12.s10.10

C. *The Minor Pillar Inscriptions* (abbreviated as M.P.I.)
(i) The Sanchi pillar in Bhopal, M.P.
(ii) The Sarnath pillar in Banaras district, U.P.
(iii) The Rummindei pillar in the Nepalese Terai.
(iv) The Nigali Sagar pillar in the Nepalese Terai.

c12.s10.11

The following two inscriptions engraved on the Allahabad-Kosam pillar are included in this category:

c12.s10.12

(v) The Queen’s P.E.
(vi) The Kauśāmbi P.E.

c12.s10.13

D. *The Minor Rock Inscriptions* (abbreviated as M.R.I.)
(i) The Rupnath rock inscription in Jabalpur district, M.P.
(ii) The Sahasram rock inscription in Shahabad district, Bihar.
(iii) The Bairat rock inscription in Jaipur, Rajasthan.
(iv) The Calcutta Bairat slab inscription or The Bhabhru Edict.
(v) The Maski rock inscription in Raichur district, Mysore.
(vi) Erragudi M.R.I. in Karnul district, Madras, published in *I.H.Q.*, 1933, pp. 114–20; and also ibid., 1931, pp. 737–40.
(vii) Rajula Mandagiri M.R.I. in Bellary district, Madras, published in *Epigraphia Indica*, xxxi, no. 28. (viii) Gujarra M.R.I., published in *Epigraphia Indica*, xxxi, no. 27. (ix–xi) The Brahmagiri, Siddhapur, and Jatinga Rameśvara M.R.I. in Chitaldrug district, Mysore.
(xii) The Barabar Hill cave inscriptions (three in number) in Gaya district, Bihar.

Linked texts (14)