14. THE ORIGIN OF THE HEADLESS SCRIPT
(DBU MED) IN TIBET
Sam VAN SCHAIK (The British Library)
INTRODUCTION1
Writing is an important aspect of Tibetan culture, which has placed a high
value on the mastery of calligraphic skills. Tibetan writing comes in a great
variety of styles, which are often speci2c to particular social functions, but
there is a fundamental distinction between two scripts. On the one hand
there is a script with horizontal lines along the tops of many letters, like the
serifs of the Latin script, known as dbu can, or ‘headed’, and second there
is a script without these lines, known as dbu med, or ‘headless’. In the
latter there are numerous di,erent styles including a simple style for
teaching children, ornamental styles for o<cial edicts, and a very cursive
style for handwriting.2
My purpose in this paper is to look for the origin of Tibet’s dbu med
script. Theories on the origin of dbu med fall into two camps. The 2rst is
that it was invented, along with the dbu can script, based on models from
di,erent Indian alphabets. The second is that the dbu med script evolved
over time as the dbu can script was written quickly.3 That is to say, it is a
classic cursive script, according to the de2nition of ‘cursive’ in the current
Oxford English Dictionary:
1
I would like to thank the Leverhulme Trust for providing the funds for the research that
led to this publication, and Imre Galambos, Kazushi Iwao, Dan Martin and Burkhard
Quessel for their comments.
2
Styles of dbu med include the ḥbru tsha, the dpe tshugs (‘book form’), and the khyug
yig (‘running script’), and variations on these known as tshugs riṅ (‘long form’), tshugs thuṅ
(‘short form’) and tshugs chuṅ (‘small form’). These styles are explained in numerous
Tibetan calligraphy manuals, such as Bkras Lhun dgon (2003) and Śes rab ñi ma (n.d.) The
classic study of Tibetan writing, from which most later accounts are drawn, is the White
Beryl of Saṅs rgyas rgya mtsho (1653–1705). For a brief account, in English, of modern
Tibetan writing styles and their functions in an o<cial context, see French (1995: 155–158).
3
One of the 2rst people to put this theory in writing was the maverick Tibetan scholar
Dge ḥdun chos ḥphel in a newspaper article and his posthumously published White Annals.
See Dge ḥdun chos ḥphel (1994: Volume III: 269–271) and the reproduction of the
newspaper article in (Narkyid 1983).
412
SAM VAN SCHAIK
Written with a running hand, so that the characters are rapidly formed without
raising the pen, and in consequence have their angles rounded, and separate
strokes joined, and at length become slanted. 4
It is the purpose of this paper to demonstrate that analysis of the earliest
sources of Tibetan writing according to the principles developed in
European paleography makes it quite clear that dbu med was originally a
cursive script that developed out of dbu can. We can show that this early
dbu med was in use throughout the Tibetan empire, and that it was mainly
taught to o<cial scribes. We will also show how a variety of calligraphic
dbu med styles developed after the fall of the Tibetan empire in the midninth century. In these calligraphic forms of dbu med we begin to see the
forerunners of the dbu med styles that are known today.
ORIGINS: THE SOURCES OF THE TIBETAN SCRIPT
According to the Tibetan historical tradition, Tibetan writing was invented
when the emperor sent a minister, Thon mi Saṁbhota, to India to 2nd a
model on which to base the Tibetan alphabet. With gold given to him by
the emperor, he was able to procure the services an Indian scholar, a
Brahmin called Li byin.5 According to one of the earliest versions of this
story, from The Pillar Testament, Thon mi studied twenty di,erent scripts
with the Brahmin.6 Having learned these scripts, Thon mi returned to Tibet,
and formulated a Tibetan alphabet from the Indian scripts, having found
almost all of the letter forms he needed for a Tibetan alphabet in them.
4
From OED Online (http: //dictionary.oed.com).
Some Tibetan historians reconstructed from this Tibetan name as the Sanskrit name
Lipikara, which is actually a genuine Indic term for a scribe dating back to the Aśokan
period. While The Pillar Testament gives the Brāhmīn’s name as Li byin ti ka, other
versions shorten this to Li byin. A more recent attempt to Sanskritize the name makes it
*Kaṁśadatta (Sørensen 1994: 168, note 463). Some sources replace Li byin with a di,erent
teacher, called Lha rig paḥi seṅ ge (Skt. *Devavidyāsiṁha). The earliest appearance of this
alternative teacher, as far as I am aware of, is the Ma ṇi bkaḥ ḥbum (102a.4).
Hypothetically, this could be the personal name of the teacher, whereas Lipikara (if that is
indeed the name behind Li byin) is a profession, rather than a personal name.
6
The Pillar Testament (A): 105.10–106.5: thon mi sam bho ṭas rgya gar lho phyogs su
phyin nam bram ze li byin ti ka bya baḥi yig mkhan cig daṅ mjal nas/ bram ze de la khyed
kyis ṅa la yi ge slob daṅ ces źus nas gser deḥi phyed phul bas/ bram ze na re/ ṅas yi geḥi
lugs mi ḥdra ba ñi śu tham pa śes pas/ bod phrug khyod yi geḥi lugs gaṅ la slob zer te bram
zes bod phrug khrid nas rgya ḥtshoḥi ḥgram na rdo riṅs cig la yi geḥi lugs mi ḥdra ba ñi śu
tham pa bkra lam me ba bris brkos yod pa de bstan pas/
5
ORIGIN OF TIBETAN DBU MED
413
According to The Testament of Ba, the Brahmin accompanied Thon mi
back to Tibet and helped him formulate the Tibetan alphabet.
These early versions of the Thon mi story do not specify any one Indic
script as the basis for the Tibetan alphabet; nor do they distinguish between
the dbu can and dbu med scripts. However, many later versions of the story
state that Thon mi used two Indic scripts, Lañtsa and Vartula, with the
former being the basis of dbu can and the latter the basis of dbu med. The
Ma ṇi bkaḥ ḥbum seems to be the 2rst place this statement appears. 7 The
Lañtsa and Vartula scripts certainly were known in Tibet, but not until long
after the imperial period. They are scripts in the Siddhamātṛkā family
which were adopted by Tibetans no earlier than the eleventh century,
probably from Nepal, as calligraphic alphabets for rendering Sanskrit titles
on the title pages of Buddhist scriptures.
Since Lañtsa and Vartula were the Indic scripts that Tibetans were most
familiar with by the time of the compilation of the Ma ṇi bkaḥ ḥbum, it is
not surprising that they came to be taken as the ancestors of the Tibetan
script. However, the Indic writing style known as Gupta script, as seen in
Indic inscriptions from the 2fth or sixth century, is a very much better
model for Tibetan writing. With a few exceptions, every Tibetan letter
traditionally said to have been derived from Indian scripts can be traced to
the so-called Late Gupta style found in the inscriptions of North India
throughout the sixth century and in Nepal into the early seventh century. 8
The question of the geographic origin of the Tibetan script excited considerable interest, and disagreement, among Indologists and Tibetologists in
the 2rst half of the twentieth century. Some argued for a Central Asian
7
See Ma ṇi bkaḥ ḥbum (102a.5). The second version of the tale in the Ma ṇi bkaḥ ḥbum
replaces Lañtsa with Nagarī (186b.5).
8
The Gupta style was well de2ned by Bühler (1904: 65–71). For Bühler the Gupta style
was more or less identical with what he termed the northern alphabets during the fourth and
2fth centuries. Bühler employed the chronological classi2cation of Gupta —>
Siddhamātṛkā —> Nāgāri used here. Bühler identi2es Siddhamātṛkā inscriptions as early as
the sixth century and Nāgāri as early as the seventh, but their appearance as coherent styles
should be dated to the seventh century for Siddhamātṛkā and ninth at the earliest for Nāgāri.
Lore Sander (1968) subdivides the Gupta alphabet into Gupta A (third–fourth century),
Gupta B (fourth–2fth century) and Late Gupta (6th century); it is the latter that forms the
basis for most Tibetan letter forms. A. H. Dani (1963) identi2ed many regional sub-classes
of Bühler’s northern alphabets. His regional classi2cations, which include Nepal and the
Northwest, are very useful, though sometimes the dividing lines among the regional styles
are not as clear as they might be. For a summary of these developments see Salomon (1998:
38–40). The most useful single volume of late Gupta inscriptions is Fleet (1888).
414
SAM VAN SCHAIK
Figure 14.1: Detail from the Barābar Hill Cave inscription
(Fleet 1888, pl.xxxB).
source, others for Kashmir, others for Northern India and Nepal. 9 In my
own recent studies, I have observed that the epigraphical sources from
Northern India and Nepal provide the closest models for the Tibetan letter
forms, especially when compared with the earliest examples of Tibetan
epigraphic writing.10 This paleographical observation is reinforced when we
take into account Tibetan contact with North Indian and Nepalese cultures
in the early seventh century. A route to India via Nepal was used by
Chinese envoys and pilgrims, and during the early seventh century Tibetans
too were involved in Chinese diplomatic and military expeditions to
Northern India.11 It is likely that the Nepalese king Narendradeva and his
court were resident in exile at the Tibetan court in the 630s, and the
Nepalese architectural features seen in early Tibetan temples are evidence
of the cultural interaction between Tibet and Nepal during this period.12
9
For example, A. H. Francke (1911) argued for a Khotanese source of the Tibetan script.
This was strongly contested by Berthold Laufer (1918). F. W. Thomas (1951) was inclined
to favour Nepal as the most likely source. Shōju Inaba (1954) argued that a single
inscription from Gopālpur, near the current Indian-Nepalese border o,ered the best model
for the Tibetan script.
10
In particular, the inscriptions from Central Asia, Kashmir and Pakistan lack the looped
na and ma, and the distinctive form of tha found in the Tibetan alphabet. Furthermore, the
‘acute-angled script’ that is seen in the North Indian inscriptions of the late sixth and early
seventh century is not found with any regularity in the inscriptions from Central Asia,
Kashmir and Pakistan. For a detailed discussion see van Schaik (2011).
11
A route from Tibet through Nepal to India in the seventh century is indicated by
several accounts of Chinese envoys and monks travelling to India via Tibet in the years 6434 and 648. The existence of one such route was con2rmed by an inscription discovered in
1990 in Kyirong, near the border between Tibet and Nepal. It was written during the
Chinese envoy 王玄策 Wáng Xuáncè’s mission to India in 658 (see Sen 2001: 25–26).
12
This historical connection between Narendradeva and Tibet may be indicated the
statement in the Old Tibetan Annals (Pelliot tibétain 1288, l.10) that a Na ri ba ba was
returned to the throne in Nepal in 641. The Chinese Old Tang Annals (chapter 21) state that
the king was a vassal of Tibet in the year 647. The evidence for this reading of history is
ORIGIN OF TIBETAN DBU MED
415
What we appear to have in the earliest Tibetan inscriptions is a style
based on the simple and elegant Gupta letters of the 2fth and sixth century,
with some alterations based on an early precurser of the Siddhamātṛkā style
that is sometimes known as the ‘acute-angled script’. However, the Tibetan
alphabet shows no trace of inbuence from the fully-developed Siddhamātṛkā script of the latter half of the seventh century. So the formulation of
the Tibetan script would appear to be placed between the 2rst appearance
of the ‘acute-angled script’ in the mid-sixth century, and the evolution of
various letter forms that changed the script into the form known as
Siddhamātṛkā by the mid-seventh century.13 This is a surprisingly narrow
span of time, and perhaps not so surprisingly, it accords exactly with traditional Tibetan account that the Tibetan script was invented in the reign of
Sroṅ brtsan sgam po (629–c.649).14
Of course, we should compare these Indic sources with the earliest
known examples of Tibetan writing. These are the pillar inscriptions from
Central Tibet and the manuscripts dating from the Tibetan occupation of
Central Asia. The pillar inscriptions provide us with the earliest dated
source of Tibetan writing: the Źol pillar in Lhasa, dated to the 760s, approximately a century after the 2rst appearance of writing in Tibet. Analysis of
the writing on this and other pillars shows that it accords well with the
Indic sources, and is closer to them in some respects that later forms of
dbu can. Thus, despite the hundred-year gap between the 2rst recorded
instance of Tibetan writing and the earliest surviving examples of such
writing, a close comparison of sixth and early seventh century inscrip tions
from Northern India and Nepal with the Tibetan pillar inscriptions leaves
little room for doubt that Indian inscriptions such as these were the main
source for the Tibetan dbu can script.
discussed in Vitali (1990: 71–72).
13
Many Siddhamātṛkā letter forms, including the ma, ya and sa, are very di,erent from
the Gupta forms, and could not have served as a model for the corresponding Tibetan letters.
14
The 2rst instance of writing mentioned in the Old Tibetan Annals is dated to the year
655, a record of the results of a census of the previous year. This strongly indicates that a
script was formulated some years before this event, so that at least the beginning of the
process must have occurred during the reign of Khri Sroṅ brtsan. In the citation below the
opening curl is marked with @ and the reverse gi gu sign is marked with a capital I.
Pelliot tibétain 1288, ll.26–29: @/: /stagI lo la bab ste/ bstan pho mer khe naḥ bźugs
shIṅ/ blon che stoṅ rtsen gyis/ moṅ pu sral ḥdzoṅ duḥ bsduste/ rgod g.yuṅ dbye źing/ mkho
sham chen pho bgyi baḥi rtsis mgo bgyI bar lo gźig/ @/: /yos buḥI lo la bab steḥ/ /btsan po
mer khe na bźugs shiṅ/ blon che stoṅ rtsan gyIs/ /ḥgor tir/ bkaḥ/ grims gyI yi ge brIs phar lo
gchig/
416
SAM VAN SCHAIK
On the other hand I have found no Indic manuscripts or inscriptions that
bear close comparison with early Tibetan dbu med, and certainly nothing
that could be identi2ed as a source for the dbu med script. It might be
thought that the rounded letter forms found in South Indian inscriptions are
a likely source. Paleographers of Indic writing usually explain these
rounded letters as derivations from manuscript writing, showing the e,ects
of writing with a stylus—this ‘cursivization process’ will be discussed later
in relation to the Tibetan script. More recently however, such changes have
been interpreted as calligraphic elaborations, similar to those seen in
Northern Indian scripts of the same period (Salomon 1998: 39).
In the 2fth to seventh century inscriptions of the Kadambas and
Cālukyas, in a script usually called Grantha, there are rounded letter forms,
but they are not a convincing source for the Tibetan dbu med (Burnell
1968: 33–40 and Plates II and III). Many letter forms in these inscriptions
di,er radically from Tibetan forms—whether dbu can or dbu med.15
The very rounded style known as Vaṭṭeḷuttu, ‘rounded writing,’ an early
form of the Tamil alphabet, may initially seem to be another possible
source for dbu med. However, once again we 2nd far too many letter forms
—including ta, na, ma and ra—that di,er radically from any Tibetan
writing style. The rounded pa, ba, ya and la might appear somewhat similar
to dbu med forms, but even these letters di,er in their proportions from any
dbu med writing. We may compare, by contrast, the very close match
between the proportions of the letters found in North Indian inscriptions of
the sixth and early seventh centuries and the early dbu can inscriptions in
Tibet (See Burnell 1968: 47–52 and Plate XVII).
Therefore, the theory that dbu med developed out of dbu can—rather
than being invented at the same time based on a di,erent Indic script—is
supported by a lack of exemplars in the Indic manuscript and epigraphic
sources. But is there any positive evidence for dbu med as a development
out of dbu can? Recently, some attempts have been made to show how
dbu med might have developed out of dbu can.16 However, these have been
hampered by two problems: 2rst, they have lacked a methodology such as
those that have been developed in the 2eld of Latin paleography; and
15
These di,erences include: na: which does not appear in a looped form; pa and ba:
which have a characteristic ‘bulge’ at the lower left; ma: which is a closed loop; ra: in which
the lower stroke is a leftward curl which later forms a closed loop; la: in which the right leg
2rst curls above the letter and later forms a closed loop; śa: a very di,erent letter form.
16
In particular, see Narkyid (1983).
ORIGIN OF TIBETAN DBU MED
417
Figure 14.2: Example of the Vartteḷuṭṭu script (Burnell 1968: pl.XXXIIa).
second, they have been based on contemporary forms of the letters, rather
than the earliest known forms.
METHODOLOGY: THE PALEOGRAPHIC APPROACH
In our analysis of the development of the Tibetan script, we would be remiss to ignore the work that has already been done by paleographers in
other areas. The 2eld of Roman paleography, for example, can provide us
with models and examples of how to proceed. The paleography of Roman
writing is of particular interest here because the earliest examples of Roman
writing include both the large, angular writing style known as Roman
Capitals, and the smaller, more curved letters known as Roman Cursive.
Studies of Roman paleography have shown how Roman Capitals developed over time into Roman Cursive, and how Old Roman Cursive, an early
form of the cursive script, developed into New Roman Cursive. Central to
this analysis has been the concept of ductus, that is, the number, direction
and sequence of the strokes which the scribe uses to write a letter. 17
Breaking up the letter forms into strokes, and assessing the ease or di< culty of writing each stroke is central to the analysis of script develop ment,
and will be very helpful to us in examining early Tibetan writing.
Let us stay with Roman paleography for a moment longer. In the 1950s
Giorgio Cencetti examined the early examples of Roman Capitals and
Cursive. He noted that both styles appeared as early as the fourth century
BC, already evolved and formed (Cencetti 1956: 63–66). Then he set out to
17
Jean Mallon de2ned this use of the term ‘ductus’ in his 1951 work Paleogra2e
Romaine. Mallon was particularly concerned to show how the ductus of individual letter
forms was preserved in the transition from Old Roman Cursive to New Roman Cursive.
418
SAM VAN SCHAIK
show how the cursive forms could have evolved from the capitals, asking
the reader to consider what would happen to the capital letters if they were
written frequently, repeatedly, and at speed. Cencetti estimated that this
‘cursivization’ process resulting in the two kinds of writing happened
within the period 500–375BC, a similar span of time to the period elapsing
between the probable creation of the Tibetan script and our earliest examples of dbu can and dbu med.
Among the speci2c changes caused by ‘cursivization’ in Roman writing
identi2ed by Cencetti the most important are the changes in the direction of
strokes, and the combining of multiple strokes into a single movement—
known as a ligature. Another inbuential paleographer of the 1950s,
Bernhard Bischo,, identi2ed these e,ects of cursive writing, and added
several others to the list:
The principle consequences that follow from this kind of writing with more
rapid, lighter strokes could be described as follows: because they are
simpli2ed, the bourishes drop out. As a result of this more rapid writing, the
script is, when space allows, elongated by end-strokes up or down. Individual
strokes that are contiguous to one another or that can be brought together in
relative positions are joined. Angles are rounded o,, and di<cult curves are
smoothed out.18
Such descriptions assume that the same ductus—that is, the order and
direction of strokes—can often be observed in the cursive forms of letters.
However, as recent paleographers have shown, a more radical development
sometimes occurs in which the ductus itself changes. Here the ductus of the
original letter form mutates into the new ductus of the cursive form, which
is then written with a ductus that is more comfortable and/or quicker to
execute, but retains su<cient resemblance to the form of the letter as it is
commonly recognised at that moment in history. 19
Somewhat closer to Tibet, the derivation of a cursive script from noncursive styles has also been noted in Chinese writing. Though there are
many named styles writing Chinese, the cursive style has been shown to
derive from earlier styles through certain basic principles, much like those
identi2ed by paleographers of Latin writing.
Brieby speaking, the term Chinese Cursive Script is applied to Lìshū, Kăishū,
and Jyăndź [Jiǎnzì] when the strokes are executed with rapidity. In general
18
Bischo, (1990: 52). I have omitted his examples from Latin letters in this passage.
This process is described in Gilissen (1973). Delorez, in his study of Gothic cursive,
states that “the simpli2ed ductus is the essential feature of cursive script” (Delorez 2003:
126).
19
ORIGIN OF TIBETAN DBU MED
419
there are two ways through which the characters in Cursive Script are formed:
(1) the characters are written with an attempt to lift the tip of the writing
instrument from the paper as few times as possible; (2) the characters are
written in a simpli2ed form with certain strokes omitted or a complex element
being substituted by a simpler one. (Wang 1958: xxv)
In more recent, comparative studies, paleographers have surveyed a range
of world scripts, and noted the presence of similar principles of
‘cursivization’ where there have been similarities in writing implements,
the writing surface, and of course the universal similarity of the human
body. Thus, Albertine Gaur, in her wide-ranging study of the history of
writing, notes the principles we have discussed above across very di,erent
and geographically separated writing groups; for example, of the Aramaic
script she writes:
The outward appearance of the oldest Aramaic letter-signs di,ered little at
2rst from those of the Phoenician script, but gradually special characteristics
began to emerge: the tops of certain letters such as b, d, and r (originally
closed) became open; a tendency to reduce the numbers of separate strokes in
certain letters appeared; and 2nally angles became more rounded, and
ligatures were introduced—in other words, the whole script became slightly
more cursive. (Gaur 1987: 92.)
As we will see, all of the features mentioned here can also be observed in
early dbu med.20 One of the most thorough attempts to de2ne the principles
that inform the development of writing is found in the work of Peter van
Sommers. The graphetic principles de2ned by van Sommers are based on
empirical studies as well as the analysis of historical scripts, and address
issues like “how and where the hand approaches the writing surface, the
manner in which the hand and arm work as a stroke is made, how writers
and drawers anchor one stroke to another, and so on.”21
20
When Gaur writes, “the tops of certain letters...became open,” she is observing a
similar process to the omission of the ‘heads’ of many letters in dbu med. The tops of the
Phoenician letters in question are generally horizontal strokes that were not necessary for the
recognition of their forms.
21
van Sommers (1991: 4). Eight principles are illustrated by van Sommers. To summarize, the principles (as applied to right-handed writers, writing from right to left) are
preferences for: (i) drawing lines in the directions of two, 2ve and seven o'clock; (ii)
anchoring lines to a 2xed point; (iii) keeping close control by minizing the stroke area; (iv)
starting at the top left; (v) drawing circles anticlockwise; (vi) progressing from one stroke to
an adjacent one; (vii) completing similar strokes together; (viii) keeping paper contact.
420
SAM VAN SCHAIK
Though these graphetic principles are intended to apply to all writing,
and not cursive writing alone, some of them clearly a,ect the cursivization
process, since the greater the ease, the greater the speed of writing. In
particular, van Sommers sees the graphetic principle of maintaining paper
contact as the fundamental principle operating in the development of
cursive writing. He writes:
Let us turn to another constraint: it makes sense for writers and drawers using
instruments that leave scratches or residues (inscribing in wax, using pens,
pencils or brushes, for example) to maintain contact with the writing surface
as they move from stroke to stroke. This tendency is accentuated, according
to my studies, as the 2gure size gets smaller and writing speed gets faster...
Once a writer has started to move continuously around angles, there will be a
tendency for the angles to be changed into curves. (van Sommers 1991: 13.)
Drawing from the above studies, we can identify these main principles
governing the development of cursive scripts:
(i) Non-essential strokes—especially the heads of letters—are dropped.22
(ii) Adjacent strokes are joined in ligatures.23
(iii) Angles become curves.24
(iv) Stroke direction follows the line of easiest articulation.25
(v) End-strokes are lengthened and may curl in the direction of writing.26
22
This principle is seen in Bischo,’s account of the e,ects of cursive writing quoted
above, in which he refers to these non-essential elements, such as the 2nials in capital
letters, as “bourishes.” It is also comparable to the opening of the tops of certain letters in
the Aramaic script as described in the above quotation from Albertine Gaur.
23
Ligatures between di,erent letters are a common feature of cursive writing, but are not
found in early dbu med (nor indeed many of later dbu med styles either). This may be due
in part to the need in written Tibetan to lift the pen frequently to write the syllable-dividing
tsheg—either a dot or a short downward stroke.
24
A point that is implicit in the studies quote above, but not stated outright, is that a
sharp angle in writing requires the pen to come to a complete stop—however momentary
that stop may be—whereas in a curve the pen is always in constant motion.
25
As shown by van Sommers, and mentioned above, these are the directions of two, 2ve
and seven o'clock (for right handed writers).
26
Though the lengthening of such strokes is mentioned by Bischo, in the passage quoted
above, the curl in the direction of the next letter is not mentioned in the studies of writing I
have consulted. However, it is clearly one e,ect of fast writing, where the hand has begun
to travel towards the next letter before the pen has entirely left the writing surface. This curl
could be seen as an incomplete ligature towards the next letter.
ORIGIN OF TIBETAN DBU MED
421
The next step is to see whether these principles can be shown to apply to
the development of dbu med in Tibetan writing. Before we do that,
however, we must identify the appropriate sources for the study of early
dbu med.
SOURCES: THE EARLIEST EXAMPLES OF DBU MED WRITING
The manuscripts recovered from Central Asia by Western explorers in the
late 19th and early 20th centuries include the earliest examples of Tibetan
writing on paper—many of them dating to the period of the Tibetan occupation of this area in the 2rst half of the ninth century. While the epigraphic
material mentioned earlier is relatively limited in extent, there are thousands of these Tibetan manuscripts, and they contain a vast array of writing
styles. It must be remembered that the terms dbu can and dbu med are
themselves a later classi2cation, and are not found in any of the pillar inscriptions or Dūnhuáng manuscripts.
In gathering paleographical evidence for early dbu med, my 2rst step
was to consider only those manuscripts that could be dated to the period of
the Tibetan empire. Many of the Tibetan manuscripts from Central Asia
derive from the ‘library cave’ at Dūnhuáng, and these manuscripts may
date from as late as the beginning of the eleventh century, when the cave
was closed. On the other hand, those manuscripts that can be shown to date
from the Tibetan occupation of Dūnhuáng have a terminus ad quem of 848,
when the town was taken back by the Chinese.
I will suggest here a preliminary de2nition of 2ve basic groups of handwriting seen in the Dūnhuáng manuscripts from this period. These are very
broad characterizations, yet, as we shall see, they accord well with parti cular types and functions of the texts themselves.
(i) A style that emulates the proportions of the epigraphic writing from
Central Tibet, mainly found in historical, legal and divinatory texts that
may have been circulated from Central Tibet. We may classify this style as
dbu can, and since the proportions of these letters tend to be more square
than other styles, I will refer to it as ‘square style dbu can’.
(ii) A style that maintains the ductus of the square style, but is adapted for
faster writing. It is characterized by longer lines at the end of strokes, par ticularly noticeable in the vertical descenders of certain letters and the
horizonal strokes of the vowel signs. As there is little change to the ductus
422
SAM VAN SCHAIK
of the letters, we may classify this style as dbu can as well.27 It is best represented in the mass-produced copies of sūtras and other Buddhist texts,
and thus appears to be the style that was taught to the scribes (many of
whom were Chinese) recruited to produce these sūtras. I will refer to it as
‘sūtra style dbu can’.
(iii) A hastily written style found in brief military communications from the
Tibetan forts in Central Asia, and certain o<cial documents from Dūnhuáng.28 It is often referred to as “cursive dbu can” by F. W. Thomas and
others. There is much variation among these documents, and the style comprises mainly inconsistent alterations to the basic forms of the letters,
apparently deriving from the quick writing of one of the above taught
styles. Although I am reluctant to classify these hands as a ‘style’ per se I
will refer to them as ‘military style dbu can’.
(iv) A somewhat cursive style found in several o<cial manuscripts from
Dūnhuáng (e.g. Pelliot tibétain 999), the Bde khams area (e.g. Pelliot tibétain 1089, pictured here), and indeed two Dūnhuáng manuscripts from
Central Tibet (IOL Tib J 1459 and Pelliot tibétain 1085). The heads of the
letters are retained, but the ductus may be changed to facilitate writing
quickly. This style maintains consistent letter forms, and appears to be a
taught handwriting style. I will refer to it as ‘o<cial style dbu can’.
(v) A truly cursive style characterized by rounded lines (avoiding pen-lifts)
and the lack of ‘heads’ in many letter-forms. This style, mainly found in of2cial documents, I will refer to as ‘early dbu med’. As we will see below, it
appears to be a distinctive taught script.
In all of the above styles the basic forms of the letters found in the
epigraphic sources are altered to some extent. This is inevitable when the
medium and writing tools—pen and paper rather than stone and chisel—
di,er so much. Thus, even the documents that closely mimic the style of
the pillar inscriptions contain forms of some letters altered for ease of
writing, in accordance with the graphetic principles mentioned in the pre27
The triangular ba is a feature that distinguishes this style from the ‘square’ style.
Examples of the style can be seen throughout the wooden slips and paper fragments
collected by Aurel Stein from the Tibetan fort sites of Mīrān and Mazār Tāgh, catalogued
under the British Library pressmarks IOL Tib N (for the wooden slips) and Or. 15000 (for
the paper manuscripts).
28
ORIGIN OF TIBETAN DBU MED
423
vious section.29 In the sūtra and o<cial style dbu can, we see some changes
in the ductus that facilitate quick writing, and I think that this is one good
reason to suspect that they may be taught styles (the other good reason
being the many documents written in a similar style).
With the 2fth style, which I am calling ‘early dbu med’, there is a crucial
di,erence from all of the others in the consistent omission of the heads of
the letters. There are over thirty manuscripts written in this style (for a list
see the Appendix). This group of manuscripts is also surprisingly consistent
in its subject matter: generally o<cial issues of a local nature. All of the
dateable manuscripts in this group are from the Tibetan imperial period,
usually from the last decades of the occupation of Dūnhuáng—that is, the
830s and 840s. The texts are generally of local interest; most of the manuscripts were written in Dūnhuáng itself, though some originated elsewhere
in Bde khams, the large administrative district that included Tibet’s territories in Eastern Central Asia: two are from the o<ce of the bde blon, the
minister governing Bde khams, and one from a government o<ce in
Tsoṅ ka.30 We also have some evidence that this early dbu med style was
not limited to Tibet’s Central Asian territories. A letter from the palace
(pho braṅ) of Ḥon caṅ do, one of the headquarters of the Central Tibetan
government, shows both the o<cial dbu can and early dbu med writing,
evidently from the hand of a single scribe.31 Though the right side of the
manuscript is missing, and hence there is no o<cial seal, the oblique lines
that mark the blank part of the document and are unlikely to be found in a
copy, are present. Thus, it is likely that this is the original letter, originating
from Central Tibet.32
29
The Old Tibetan Annals (version 1) found in IOL Tib J 750 and Pelliot tibétain 1288,
is perhaps the manuscript most closely based on the epigraphic style. Nevertheless, there are
di,erences from the pillar inscriptions, including a triangular—rather than square—head on
the letter ga. This change can be explained by one of the van Sommers' graphetic principles:
that there is a preference for anchoring lines to a 2xed point (see van Sommers 1991).
30
The extent of the province of the bde blon is unknown. It seems to have been created
to include Tibet’s northeastern territories, and then expanded with further conquests. Hugh
Richardson (1990) addressed the question in some detail. He pointed out that several
Dūnhuáng documents (including the dbu med manuscripts IOL Tib J 1126 and Pelliot
tibétain 1111) mention a place called Źa, where the assembly of the bde blon was held, but
the location of Źa is still unknown.
31
The other manuscript that is thought to originate from Central Tibet is the letter from
the Lhan kar palace (Pelliot tibétain 1085). The writing style in this letter is similar to the
o<cial dbu can in the Dūnhuáng manuscripts.
32
This is further indicated by preliminary 2bre analysis carried out by Agnieszka
424
SAM VAN SCHAIK
Figure 14.3: IOL Tib J 1359(B): Register of scribes.
Reproduced by kind permission of The British Library.
Figure 14.4: IOL Tib J 1126: Letter from the o<ce of the bde blon.
Reproduced by kind permission of The British Library.
Even further away from Dūnhuáng, at the western limits of the Tibetan
empire, there are a number of scratched inscriptions on stone, sgra<to left
by the Tibetan occupiers of the region. The greatest concentration of these
are found along the banks of the Indus River at Alchi, northwest of Leh in
Helman-Wazny. IOL Tib J 1459 is composed of Thymelaeaceae (Daphne) 2bres, the most
common 2bre used to make Tibetan paper. Among the manuscripts analysed so far, none of
the manuscripts locally produced in Dūnhuáng during the Tibetan period is composed of
these 2bres.
ORIGIN OF TIBETAN DBU MED
425
Figure 14.5: IOL Tib J 1459: Letter from the palace ( pho braṅ) of
Ḥon caṅ do. Reproduced by kind permission of The British Library.
modern Ladakh. Here there is a ruined fort and bridge that were once of
some strategic signi2cance to the Tibetan empire. Many of the inscriptions
are Buddhist dedications written next to crude pictures of stūpas, giving the
name of the person who dedicated the ‘stūpa’. Some of the inscribed names
include o<cial titles like blon and stoṅ dpon, and so appear to have been
written by o<cials and higher-ranking miltary personnel.33
The Ladakh area, known in the Tibetan records as Mar(d) yul was under
the control of the Tibetans by 719, when an o<cial census was held there,
and it was used as the base for the conquest of neighbouring Bru źa.
Various references to Bru źa in the inscriptions and other sources indicate
that the region remained under Tibetan control for most of the eighth
century, and possibly well into the ninth. Phillip Denwood, who photographed and translated many of these inscriptions, estimates that they date
from the period between the Tibetan occupation of the area in the mideighth century to the collapse of the empire in the mid-ninth century.
Tsuguhito Takeuchi, on the other hand, has suggested that they may date
from after the fall of the empire, when a local Tibetan kingdom was established in the area.34 Both Denwood and Takeuchi agree that the orthography
33
Most of the names are followed by the instrumental particle and the verb bris,
“written/inscribed by...” The pictures of the stūpas appear to have been considered religious
o,erings. Another such rock-inscribed stūpa discovered by Stein in the Darkōt pass
(between Yarkand and Kashgar) has the message: rmeḥor ḥirni dor kyi yon. The last syllable
was read by A. H. Francke as om, but the reading of yon, which is quite clear from Stein’s
photograph, provides the much better sense that the inscribed stūpa is the religious gift ( yon)
of the person named here. See Stein (1928: I.45 and II.1050–1051 Appendix L).
34
For Denwood’s view, see Denwood (2007: 50, 52), and Snellgrove and Skorupski
426
SAM VAN SCHAIK
Figure 14.6: Sgra<to on rock at Saspol, Ladakh.
Photograph courtesy of B.R. Mani, Archeological Survey of India.
of these sgra<ti matches that of the Dūnhuáng documents from the Tibetan
imperial period. I would agree with this assessment, and note in addition
that both dbu can and dbu med styles are represented in these sgra<ti. The
dbu med style is seen in Denwood’s inscriptions 2 and 7, and is also found
in other recent photographs from the same site.35 The sgra<ti are too brief
to allow us to compare them in detail with the Dūnhuáng dbu med documents, but we can note the appearance of na and ma without heads, and
more particularly, the completely rounded ba, the u-shaped pa and the
three-stroke sa—all letter-forms found in the Dūnhuáng dbu med documents. Despite the inconclusive nature of these sgra<ti, they o,er further
evidence that this dbu med style may have been taught throughout the
Tibetan imperial area.36 We now turn to the analysis of the speci2c features
of the early dbu med style in the Dūnhuáng documents, and its relationship
to early dbu can writing.
(1980: 163). For Takeuchi’s see Takeuchi (forthcoming (b)).
35
I was able to see these photographs thanks to Dr. B. R. Mani, Director of the
Archeological Survey of India.
36
The sgra<ti that appears to share the characteristics of the Dūnhuáng dbu med manuscripts is as follows: (i) Denwood’s Plate 84(B): stoṅ pon rtsa; (ii) Denwood’s Plate 84(F):
stoṅ pon khrom; (iii) Archeological Survery of India photograph (Saspol bridge): smar dbaṅ
po bdaṅ bzaṅ.
ORIGIN OF TIBETAN DBU MED
427
ANALYSIS: THE LETTER KA
The traditional method of paleography is to analyse and compare the forms
of key letters, and we will follow that method here, using the letter ka to
show the variety of styles and the speci2c features of the dbu med style. I
will begin with the pillar inscriptions because the development of the letter
forms can be much better understood when these are taken as the original
model.
In the pillar inscriptions, the letter ka is composed of four lines: (i) a
horizontal ‘head’ (mgo), and then from left to right, three vertical lines descending from the head, (ii) a stroke angled or turning to the left known in
later Tibetan calligraphy as the ‘tooth’ ( mche ba), (iii) a straight line known
as the ‘central arm’ ( dbus lag) and a slightly longer straight line forming
the right side of the letter, known as the ‘leg’ ( rkaṅ ba).
Epigraphic writing
Figure 14.7: Źol pillar, 767(?).
In the earliest example of the letter ka here, from the Źol pillar, the three
vertical lines are almost the same length. The ‘leg’ is only very slightly
longer than the other two strokes, and is exactly the same length as the
‘head’. This gives the letter a very square appearance, an appearance that is
characteristic of all the letter forms on this pillar.37 Both the ‘tooth’ and the
‘central arm’ are angled toward the left.
37
It should be noted however that because the Źol pillar is very high, letter images taken
from photos may be subject to foreshortening.
428
SAM VAN SCHAIK
Figure 14.8: Lhasa Treaty Pillar, 822.
By the time of the Lhasa Treaty Pillar, some half a century later, the leg is
much longer than the other vertical strokes, and longer than the head as
well. This may be an e,ect of non-epigraphic writing on an epigraphic
inscription, both the length and the angle of the ‘leg’ being features of a
stroke written with pen and ink.
(i) Square style
Figure 14.9: Old Tibetan Annals
(IOL Tib J 750), early to mid ninth c.
The Old Tibetan Annals manuscript presents us with a style of dbu can that
closely follows the epigraphic style, yet we see certain divergences from
the styles of the pillar inscriptions, due to the e,ects of writing with pen
and ink on paper.
ORIGIN OF TIBETAN DBU MED
429
(ii) Sūtra style
Figure 14.10: Aparimitāyurnāma sūtra (IOL Tib J 310.1210),
early to mid ninth c.
This style shows even more clearly the e,ects of pen-and-ink writing on
the proportions of the letters. The ‘leg’ becomes even longer, and both the
‘tooth’ and ‘central arm’ now point to 7 o’clock (one of the three easiest
directions for writing according to van Sommers’ graphetic principles).
Small ticks appear at the end of some strokes as the hand moves on to the
next stroke before the pen lifts away. It seems that the ductus of the letter
may have been altered here so that the ‘central arm’ and the ‘leg’ are com pleted in a single stroke, with two more strokes for the ‘head’ and ‘tooth’,
reducing the number of strokes from four to three.
(iii) Military style
Figure 14.11: Military communique (Or. 8212/1852),
Mazār Tāgh, early to mid ninth c.
The rather awkward style of this military communique shows the same
basic letter form somewhat distorted by the swift movement of the writer’s
hand. However, the ductus of the letter seems to be unchanged.
430
SAM VAN SCHAIK
(iv) O<cial style (dbu can)
Figure 14.12: O<cial despatch from the
bde blon (Pelliot tibétain 1089), mid ninth c.
Here it is even more clear than in the sūtra style that while the head of the
letter has been retained, the ductus has been changed to facilitate writing
quickly. The number of strokes taken to write the letter has been reduced to
three.
(v) Early dbu med
Figure 14.13: O<cial despatch from the bde blon
(IOL Tib J 1126), mid ninth c.
This form of the letter shows further changes in the ductus, reducing the
number of strokes to two. It is clear that, as above, the ‘central arm’ and the
‘leg’ are completed in a single stroke, with the ‘tooth’ completed along
with a now vestigal ‘head’. Since there is no true ‘head’—no separate
vertical stroke—we can identify this as an dbu med letter form. Other accidental features of fast writing are also observable, such as the tick on the
end of the ‘leg’ seen in this example.
431
ORIGIN OF TIBETAN DBU MED
epigraphic
square
sūtra
military
o<cial
dbu med
Figure 14.14: The letter ‘ba’, taken from the same sources.
The general development of the letter ka seen in these manuscripts can be
described according to the ductus. The simple manuscript dbu can style
represented in the square style above requires four individual strokes. The
sūtra style and o<cial style appear to allow a form in which two adjacent
strokes are combined into one, allowing for a faster execution of the letter
with three strokes rather than four. In the early dbu med manuscripts, two
more adjacent strokes are combined, resulting in a ka that can be completed
in just two strokes. Other letters show a similar development. The letter ba
is a good example. Here we can see how the epigraphic letters are closest
to the Indic inscriptions that were their models, and how these letter forms
were then modi2ed in the development of written styles.
These analyses show us that all of the manuscript styles can be analysed
to some extent as transformations of the epigraphic style along the lines of
the basic graphetic principles looked at earlier. Such transformations are
least evident in the styles that closely follow the epigraphic style, and are
most fully realized in our early dbu med documents. There is something of
a sliding scale here: the ‘sūtra’ and ‘o<cial’ dbu can styles can be called
dbu can in that they preserve the heads of the letters, yet in terms of ductus
they are sometimes closer to the dbu med style (the letters ka and ba both
being a good examples of this). We may still feel con2dent in setting the
dbu med style apart in that (a) it represents the fullest development of the
cursivization process and (b) it is the only style that consistently omits the
‘heads’ of letters. Its general features may be listed as follows:
(i) Pen-lifts and stops are avoided whenever possible—resulting in more
curved lines and fewer sharp angles.
(ii) Features of letters not necessary for recognition are dropped—in
particular, the heads of many letters.
(iii) Straight descenders (like the śad) curve away to the left or right, or in
an ‘S’ shape.
(iv) End-strokes (like the gi gu and na ro vowel signs) are lengthened.
432
SAM VAN SCHAIK
CONTEXT: THE SCRIBES
Is it possible that this early dbu med was taught as a distinct style, rather
than being an accidental e,ect of scribes writing quickly? I believe there
are good reasons to think that it was. The 2rst reason is the consistency of
the ductus of this style across the early dbu med manuscripts, which are
from di,erent locations and in the handwritings of di,erent scribes. The
second reason is that the early dbu med manuscripts have a remarkable
consistency of subject-matter. As the list in the Appendix shows, they are
all o<cial or semi-o<cial documents. The early dbu med group of manuscripts is composed mainly of o<cial registers of land and people, contracts
for sales and loans, and letters to or from local o<cials. Many of these
manuscripts contain seals, either the square o<cial seals, the small round
personal seals or the so-called ‘2nger seals’.38 Those without such seals
should probably be considered to be copies or drafts.
Apart from these, the only other examples of early dbu med are the
various notes written by the editors of the large-scale project to produce
multiple copies of the Prajñāpāramitā sūtra, which was done at the command of the Tibetan emperor and overseen by the local government at
Dūnhuáng. These editors may have been of minor o<cial rank, unlike the
scribes, who were generally ordinary householders or monks.39 Though the
scribes of the Prajñāpāramitā sūtras sometimes signed their names in a
cursive style, in all of the examples I have seen, it is not the true early dbu
med, but rather a cursive form of the ‘sūtra-style’ dbu can, retaining the
ductus, and often the heads of the dbu can forms. These hands seem more
comparable to the military documents, in which the dbu can letters are
altered in an ad hoc fashion as an e,ect of the scribe’s writing quickly.40 It
seems reasonable to conjecture that the ordinary people conscripted to copy
these sūtras need not have been taught more than the single dbu can style
required for the task. The editors, on the other hand, may have held an
o<cial position and been trained in clerical writing styles.
38
On the Tibetan seals see, Takeuchi (1995: 107–115).
The status and regulation of the scribes in Dūnhuáng is demonstrated by IOL Tib J
1359. See the translation and discussion in Takeuchi (1994).
40
Compare for example Pelliot tibétain 1005, a list of missing pages from sūtras, written
by an editor, with Pelliot tibétain 1127, a scribe’s scrap paper ( glegs mtshas) showing
examples of his or her cursive writing.
39
ORIGIN OF TIBETAN DBU MED
433
Figure 14.15: IOL Tib N 398: Wooden slip from Mīrān with Tibetan alphabet,
early ninth c. Reproduced by kind permission of The British Library.
This brings us to another issue. In later Tibetan culture, as dbu can became primarily the script for printed books, dbu med became the script
used for handwritten documents, and for teaching the alphabet to children.
This, it seems, was not the case during the imperial period. At this time,
some centuries before the advent of woodblock printing for Tibetan, some
form of the dbu can script was used for the majority of handwritten documents, as we have seen, whereas the early dbu med script appears in
relatively few documents from this period, within a speci2c genre of
o<cial and semi-o<cial manuscripts.
That the dbu can script was the basic writing style for learning the
Tibetan alphabet is borne out by the military documents in which the
writers are clearly writing at speed, yet do not have the early dbu med
forms at their disposal; instead they rather haphazardly alter the dbu can
forms. Further support for this supposition is found in the numerous writing
excercises among the Dūnhuáng manuscripts, in which the Tibetan alphabet, or a standard line of text, is written one or more times. Invariably,
these are in the dbu can script.
In functional terms we might compare the early dbu med style with the
cursive chancery styles of Europe. It was taught to a speci2c class of
scribes, and used for a speci2c class of documents. The social context of
the early dbu med documents is the o<cial milieu of Dūnhuáng and the
surrounding area of Tibetan administration. Since the dbu med script derives from the principles of ease in writing, it lends itself to fast writing.
The usefulness of this in an o<cial scribe can easily be imagined. Taking
down letters and other documents from dictation, for example, would
require a quick hand, and we do indeed have some evidence that dictation
was used in the writing of o<cial documents in Bde khams.41
41
We 2nd the verb spad (or ḥpad) in the colophons of Pelliot tibétain 1071, 1089 and
1113. This is the action done by the 2rst person in the colophon, while the second person is
described as writing ( ḥbris) the document. This has been translated by Yamaguchi (1980: 43
n.105), Scherrer-Schaub (2007: 303 n.159) and Iwao (forthcoming) as “dictated.” As Iwao
points out, Pelliot tibétain 1071v contains the colophon “Dictated by Ji Rom ga. Written by
Mo Ḥgom mye cha.” ( ji rom gas koṅ gyis ḥpad / mo ḥgoṁ mye cha gyIs bris). Here it is
clear that the scribe is distinct from the person composing the document.
434
SAM VAN SCHAIK
It is likely that the early dbu med script was one of the writing styles
imported into the area of Bde khams when the Tibetan administration was
established there. Our documents must be the work either of Tibetans
stationed in the area, or local Chinese (and other non-Tibetans) trained in
this particular writing style in order to write local o<cial documents. We
might speculate that the larger administrative o<ces, like that of the bde
blon, had dedicated clerical scribes, such as were found in later Tibetan in stitutions, while at less important centres like Dūnhuáng, o<cials may have
had to draft their own documents.
These scribes may well have been trained in other styles as well. They
certainly at least had the command of an dbu can style. This is indicated by
the manuscripts written in the style I have called ‘o<cial dbu can’. These
manuscripts are functionally very similar to those written in early dbu med.
For example, of two despatches from the o<ce of the bde blon, one is
written in o<cial dbu can (Pelliot tibétain 1089), the other in early dbu
med (IOL Tib J 1129). We also 2nd a number of manuscripts that seem to
be written by a single scribe who alternates between an dbu can and an dbu
med style. We saw one example of this in the letter from Ḥon caṅ do (IOL
Tib J 1459) pictured above. There are other examples as well. In an edict
from the minister at Loṅ cu, forbidding abductions of Chinese women and
children in Dūnhuáng by Tibetan o<cials (Pelliot tibétain 1083), we 2nd
just the 2nal line written in dbu med. In another letter (this time probably a
copy of a letter) found in IOL Tib J 856, the scribe shifts from an dbu can
style halfway through the letter to an dbu med style. In each case, since the
dbu med writing occurs towards the end of the letter, it appears that the
scribe switched (perhaps not intentionally) to this script as he or she attempted to 2nish the copy more quickly towards the end.
So it seems that the early dbu med style, under the Tibetan empire, was
mainly used for drawing up o<cial documents—registers, contracts and letters—of a local nature. We can say with some con2dence that this was the
case in Dūnhuáng and the administrative area of Bde khams in general. We
have some evidence that the early dbu med style was also used for this
purpose in Central Tibet, perhaps also in the far western territories of the
Tibetan empire; more epigraphic and manuscript evidence from these areas
might help us to determine whether the early dbu med style was in fact
taught and written throughout the Tibetan empire by the 2rst half of the
ninth century. But such conclusions must wait.
ORIGIN OF TIBETAN DBU MED
435
FURTHER DEVELOPMENT: CALLIGRAPHIC ELABORATION
The next step in the study of the development of dbu med is to trace its
further development, but this is a formidable task. The style that I refer to
in this paper as ‘early dbu med’ occurs in manuscripts dating to the Tibetan
imperial period, most of them from a rather short span of time between the
820s and 840s. As we have seen, most of these manuscripts have a similar
social function, and were probably written by a speci2c class of scribes.
Unsurprisingly then, we 2nd the dbu med style in these manuscripts quite
consistent in its basic features.
By contrast, the Dūnhuáng dbu med manuscripts dating from the postTibetan period, that is, from the late ninth century and tenth centuries,
testify to the emergence of a number of divergent dbu med styles. This
accords with certain traditional accounts of the development of dbu med. It
is said that there were two basic styles, the “tradition of Li” (li lugs) and
the “tradition of Ldan” (ldan lugs).42 The Li tradition is thought to have
died out while the Ldan tradition survived, but unfortunately we have no
examples that might help us to attach these two traditions to speci2c styles
of dbu med. In any case, a recent version of this traditional account states
“Later, there was no universally accepted script because the master scribes
[each] adopted their individual style of writing” (Ribur Ngawang Gyatso
1984: 29.) This situation is said to have continued until a standardization
attributed to the prince of Gyantse, Rab brtan Kun bzaṅ ḥphags pa (13891442).43 His work is said to form the basis of the models described by Saṅs
rgyas rgya mtsho (1653–1705) in his White Beryl, which became a basic
textbook for Tibetan calligraphy.
The post-imperial Tibetan Dūnhuáng manuscripts would date to well
before this trend toward standardization. The description of a situation in
which each scribal master develops his own style does seem to 2t the tenthcentury Dūnhuáng manuscripts rather well. The main di<culty in discussing the development of dbu med after the fall of the Tibetan empire is
trying to establish whether a group of manuscripts in one style of dbu med
represents an established style, or just the pecularities of a particular
scribe’s handwriting. I have argued elsewhere that we can sometimes
identify a group of manuscripts as being in the hand of one particular scribe
42
An inbuential well-known Tibetan account of these traditions can be found in The
White Beryl: 17–34.
43
This prince is better known as the sponsor of the Them spaṅs ma edition of the bkaḥ
ḥgyur (see Ribur Ngawang Gyatso 1984: 29-30 and Harrison 1996).
436
SAM VAN SCHAIK
Figure 14.16: Letter, 960s. IOL Tib J 754(a), letter 3.
Reproduced by kind permission of The British Library.
(cf. Dalton, Davis and van Schaik 2007). In this context, however, I would
like to make some fairly general observations about the dbu med manuscripts from the post-Imperial period.
We have a number of letters written in dbu med in the post-Tibetan
period (Takeuchi 1990). In these letters we can detect a di,erence, a
development in the dbu med style. A good example in Pelliot tibétain 1129:
the calligraphic elements that appear in this later dbu med letter include a
distinction between light and heavy lines, longer bourished strokes for the
vowels, and a return to sharper angles in some letters, such as da and ra.
Since these developments often increase the number of strokes required, it
seems that the desire for calligraphic e,ect had some impact on the further
development of some dbu med styles, along with the need to write at speed.
If we look at the ductus of individual letter forms, we can see that some
letters now require more pen strokes than the early dbu med forms. This is
the case with ka, which is often written in later dbu med with three strokes
ORIGIN OF TIBETAN DBU MED
437
Figure 14.17: IOL Tib J 321 showing the later dbu med ‘ka’ with the
early dbu med ‘ka’ for comparison (IOL Tib J 1126).
Reproduced by kind permission of The British Library.
Figure 14.18: IOL Tib J 321 showing the later dbu med ‘ga’ with the early
dbu med ‘ga’ for comparison (IOL Tib J 1126).
Reproduced by kind permission of The British Library.
rather than the two of early dbu med.44 Another letter form that is very
common in later dbu med manuscripts is the open-topped ga, although the
early form of ga with a looped head also continues to appear, sometimes in
the same manuscript as the open-topped form.
In one special case we can also see a development toward the letterwriting style of later, and contemporary Tibet known as khyug yig
(‘running script’). This is one of the letters of passage contained in the
manuscript IOL Tib J 754, probably from Tsoṅ kha but written by a
Tibetan called Smar khams Rin chen rdo rje. Though this is by no means
equivalent to the fully-formed khyug yig, his handwriting displays a buidity
and the very long bourishes for vowel signs that characterize the khyug
style.45
Another important change in the post-imperial period is the content of
the manuscripts written in dbu med. In this later period, many of them are
44
These manuscripts are too numerous to list here in full, but here is a sample: IOL Tib J
321, 331, 594, 647, Pelliot tibétain 149, 322, 626
45
On the dating and social context of this manuscript, see van Schaik and Galambos
(2011). For models of khyug and tshugs thuṅ scripts, see Bkra lhun dgon (2003).
438
SAM VAN SCHAIK
Figure 14.19: Letter, 960s. IOL Tib J 754(a), letter 2.
Reproduced by kind permission of The British Library.
Figure 14.20: Examples of dbu med in Buddhist manuscripts from the post-imperial period,
clockwise from top left: IOL Tib J 321, 552, 437, 1.
Reproduced by kind permission of The British Library.
Buddhist texts, rather than secular documents. Thus, along with the increasingly calligraphic nature of the dbu med script, there seems to have been a
change in its status, allowing it to represent Buddhist texts, even the
scriptures containing the Buddha’s word. This change of status is accompanied by a change in the character of the writing. The calligraphic nature of
the script is particularly evident in the Buddhist texts, some of which are
written in a very formalized dbu med. Here, some features deriving from
the action of quick writing that are found inconsistently in early dbu med
ORIGIN OF TIBETAN DBU MED
439
(like the ‘s’ shaped bend and right turning ticks at descenders) are formalized in the script itself. In these Buddhist dbu med manuscripts, the need to
write quickly seems to be less signi2cant than the need to create an
attractive script.
It is in these manuscripts that we can also begin to see the forerunners of
di,erent classi2cations of dbu med in the later tradition. The style seen in
manuscripts like IOL Tib J 321 above, that almost seems a hybrid between
dbu can and dbu med, is comparable with the style called ḥbru tsha, a form
of which is still current in Bhutan. We have a few examples of a style that
looks like a forerunner of the ‘book form’ ( dpe tshugs) or ‘book script’
(dpe yig) that became the most popular script for Buddhist manuscripts (as
against printed texts, which were in the dbu can script) in Tibet in later
centuries.46
In general then, we can say that the early dbu med appears to have been
developed principly to provide letters that can be written as quickly as
possible while retaining su<cient di,erences between each letter to allow
them to be legible. After the end of the Tibetan empire, a trend towards increasing the calligraphic features in dbu med tended to obscure the original
impetus for its development. This movement between cursive and calligraphic styles is not unique to Tibetan writing, and has been observed in
the development of European scripts as well. The paleographer Albert
Derolez describes this very process in the evolution of the ‘documentary
script’ in Europe, and concludes:
Seen in this way, the history of script might be described as an alternation of
increasing cursivity, on the one hand, and consolidation and calligraphy, on
the other.47
I have attempted here merely a brief sketch of the development of dbu med
after the Tibetan imperial period, based on the Dūnhuáng manuscripts.
Further study is needed to trace the evolution of di,erent dbu med styles,
ideally creating a genealogy of writing styles that bridges the gap between
46
Takeuchi (forthcoming(a)) identi2es IOL Tib J 358 as similar to dpe yig. IOL Tib J 82
seems even more similar to the ‘book form’ in its alternation of heavy vertical and light
horizontal strokes.
47
Derolez (2003: 5). Elsewhere Derolez attempts a detailed description of “the various
ways of introducing greater formality in an informal cursive script” (Derolez 2003: 128–
130). These include (i) a reduction in the number of ligatures, (ii) a move back to a more
complicated ductus (i.e. more strokes), (iii) a move toward shading (the calligraphic
alternation of wide and narrow strokes) where a broad-nibbed pen is used, and (iv) an
increasing angularity.
440
SAM VAN SCHAIK
Figure 14.21: IOL Tib M 50: Manuscript from Khara Khoto, 12th–13th century?
Reproduced by kind permission of The British Library.
the Dūnhuáng manuscripts the later manuscript material from Tibet proper.
In this, Tibetan manuscript collections from the eleventh century onwards,
such as the Khara Khoto and Tabo collections will play an important part.48
The manuscripts from the Tangut city of Khara Khoto, probably dating
from the twelfth century, have barely been studied, but they can certainly
help us to trace the development of dbu med further. These manuscripts
show that the developments seen at Dūnhuáng continued along a similar
route. The calligraphic features of the later Dūnhuáng dbu med styles are
further formalized, so that heavy and light pen strokes, and the angle and
curvature of letter forms becomes very consistent, and the style very attractive. Examples among these, each showing a di,erent style of dbu med,
include IOL Tib M 50, 54 and 55.
CONCLUSIONS
To summarize the conclusions reached here: the origins of dbu med seem
to be in the o<cial bureaucracy of the Tibetan Empire. The script de veloped over a century or so from the original letter forms of the Tibetan
script, the angular dbu can. When scribes wrote o<cial documents quickly,
48
For some examples of the Tabo manuscripts see the plates in Scherrer-Schaub and
Steinkellner (1999).
ORIGIN OF TIBETAN DBU MED
441
the principles of ease and speed that can be observed in scripts all over the
world changed the forms of the letters they wrote. Individual strokes were
joined in ligatures, angles became curves, and strokes that were not required for the recognition of letters were dropped. This latter e,ect meant
that the ‘heads’ of the letters were no longer part of this script, leading to
its later being called ‘headless’ ( dbu med).
By the early ninth century, if not earlier, this cursive and headless style
was quite distinct from the angular letters with heads. As the scribe had to
employ a quite di,erent method to write most of these dbu med letters, the
script must have been taught independently of the dbu can styles. The
scribes who were taught this dbu med style were those who wrote o<cial
correspondence. Whether they were o<cials themselves, or professional
scribes working in an o<cial capacity, they had at their command two
distinct writing styles: not only the basic dbu can script, but also the dbu
med script for situations where quick writing was required.
After the fall of the Tibetan empire in the mid-ninth century, there was a
profusion of di,erent forms of dbu med writing. The large-scale government of the imperial period had allowed the same style of writing to be
taught to o<cial scribes everywhere. With the fragmentation of the empire,
this regularization was no longer possible. Di,erent calligraphic forms of
dbu med developed. The function of the script became more varied as well,
and was now to be seen in religious manuscripts as well as secular
documents. One aspect of this development was that characteristics of the
dbu med letter forms that in the imperial period had developed out of the
principles of ease became encoded into the new calligraphic styles as
ornamental features.
For many years, scholars working with the Dūnhuáng manuscripts have
noticed di,erences in writing styles and speculated that these may be a way
to date the manuscripts.49 While this may never be possible to the accuracy
that we would like, I have tried to demonstrate that is is possible to identify
benchmarks that will help us to distinguish manuscripts written during the
Tibetan imperial period from those written in the later ninth and tenth
centuries. The detailed analysis of speci2c letter forms is the key to moving
from a connoisseur’s personal sense of di,erent writing styles to a
de2nition that is explicable to all. I hope that the analysis contained here,
preliminary and incomplete though it is, has made a convincing case for the
development of dbu med out of an early form of dbu can in the Tibetan
49
See Scherrer-Schaub and Bonami (2000) and Takeuchi (forthcoming(a)) for important
preliminary suggestions toward this end.
442
SAM VAN SCHAIK
imperial period, and also shown how we might continue to trace the
development of dbu med styles after the fall of the empire.
APPENDIX: MANUSCRIPTS IN THE EARLY DBU MED STYLE
Manuscripts marked with an asterisk may be dated with some con2dence to
the Tibetan imperial period (usually to the last decades of that period: the
820s to 840s). Manuscripts marked with an obelisk begin in dbu can. The
OTC numbers listed after some manuscripts refer to the numbers in
Takeuchi (1995).
Letters
Or.8210/S.2228: letter/petition concerning the estate (lha ris) of an
unnamed temple*
IOL Tib J 856(A&B): letter copy to the monks of Shāzhōu asking for
protection†
IOL Tib J 897: letter
IOL Tib J 1126: letter from the bde blon concerning a shortfall of grain
(with square seal)*
Pelliot tibétain 1077: a series of petitions*†
Pelliot tibétain 1080: letter to Daṅ za Źaṅ ceḥu
Pelliot tibétain 1083: letter from the Blon chen po at Loṅ cu (with square
seal)*†
Pelliot tibétain 1200, 1201,1202: three letters to Hongbian*
Pelliot tibétain 1217: letter from the Źaṅ blon chen po at Tsoṅ ka (with
square seal)*
Contracts and other local legal matters
Or.8212/194(a): contract for hiring a man (OTC40)
IOL Tib J 1374: contract for the sale of a woman (OTC7)
IOL Tib J 914, 1379: receipts for the repayment of wheat (OTC34,35)
Pelliot tibétain 1084: document regarding a dispute over the sale of cattle
Pelliot tibétain 1087: promissory letter to attend an assembly (for
judgment)*
Pelliot tibétain 1088: scribal practice including two contracts (OTC3)*
Pelliot tibétain 1094: contract for the sale of an ox (OTC1)*
Pelliot tibétain 1096: judicial document on the case of a lost horse
ORIGIN OF TIBETAN DBU MED
443
Pelliot tibétain 1104, 1203: ledgers for the loan of grain by a temple
(OTC27,28)*
Kozlov 4, Pelliot tibétain 1297/1, IOL Tib J 1018, Pelliot tibétain 1088/2,
Pellliot tibétain 1115, Or.8210/S.7133: contracts for the loan of grain
(OTC18–22, 25)*
O<cial registers
IOL Tib J 508, 1404 & 1486: land register*
IOL Tib J 839: o<cial register*
IOL Tib J 1359: rules for scribes and register of names*†
Pelliot tibétain 1000,1001,1002,Or.8210/S.10828: register of texts
memorized(?) by monks and nuns
Pelliot tibétain 1111: o<cial accounts of two granaries in Dūnhuáng*
Editing of o<cially sponsored sūtras
Pelliot tibétain 1005, 1012, 1013, 1020, 1024: editors’ lists of missing
pages of sūtras*
Pelliot tibétain 1006, 1008, 1009, 1011, 1014, 1015: editors’ manuscript
tags*
Other
Pelliot tibétain 230: copy of a prayer for the Buddhist activities of Khri ḥod
sruṅ*
REFERENCES
The Pillar Testament = Anonymous. Bkaḥ chems ka khol ma. Smon lam rgya mtsho, ed.
Lanzhou: Gan su mi rigs dpe skrun khaṅ, 1989.
The Ma ṇi bkaḥ ḥbum = Anonymous. Ma ṇi bkaḥ ḥbum. Trayang and Jamyang Samten,
eds. Ma ṇi bkaḥ ḥbum: A collection of rediscovered teachings focussing on the tutelary
deity Avalokiteśvara (Mahākaruṇika). 2 vols. New Delhi: Jayyad Press, 1975.
The White Beryl = Sde srid saṅs rgyas rgya mtsho. Bai ḍūr dkar po las dris lan ḥkhrul snaṅ
g.yaḥ sel źes bya baḥi glegs bam gñis pa. Běijīng: Kruṅ goḥi bod rig pa dpe skrun khaṅ,
2002.
The White Annals = Dge ḥdun chos ḥphel. Bod chen po srid lugs daṅ ḥbrel paḥi rgyal rabs
deb ther dkar po. In Dge ḥdun chos ḥphel gyi gsuṅ rtsom, vol. III: 205–300. Lhasa: Bod
ljoṅs bod yig dpe rñiṅ dpe skrun khaṅ, 1994.
444
SAM VAN SCHAIK
Calligraphy manuals
Bkras lhun dgon. (2003). Yig rigs gsum ldan gyi gzuṅs sṅags . Lanzhou: Kan suḥi mi rigs
gyi dpe skrun khaṅ.
Śes rab ñi ma. (n.d.) Yig rigs du maṇi ma phyi śes rab lha mchog dgyes paḥi dgaḥ ston .
India: n.p.
Zla ba tshe riṅ. (1999). Yig gzugs sna brgyaḥi phyi mo źal bśus ma . Běijīng: Mi rigs dpe
skrun khaṅ.
Other sources
Amdo Gendun Chophel (1983). “The Evolution of U'med from U'chen Script.” Tibet
Journal 8.1: 56–57. [English translation and reproduction of the original Tibetan article
published in 1938.]
Bacot, M. Jacques (1912). “L'écriture cursive tibétaine.” Journale Asiatique 10 sér. 20:5–78.
Bischo,, Bernard (1990). Latin Palaeography: Antiquity and the Middle Ages . Cambridge:
Cambridge University Press.
Bühler, G. (1904). Indian Paleography. Bombay: The Education Society’s Press.
Burnell, A. C. (1968) [1874]. Elements of South Indian Paleography. Varanasi, Indological
Book House.
Cencetti, Giorgio (1956). Lineamenti di storia della scrittura latina . Bologna: R. Pàtron.
Dalton, Jacob, Tom Davis, and Sam van Schaik (2007). “Beyond Anonymity: Paleographic
Analyses of the Dunhuang Manuscripts.” Journal of the International Association for
Tibetan Studies 3: 1-23. www.thdl.org?id=T3106.
Dani, A.H. (1963). Indian Palaeography. Oxford: Oxford University Press.
Denwood, Phillip (2007). “The Tibetans in the Western Himalayas and Karakoram,
Seventh-Eleventh Centuries: Rock Art and Inscriptions.” Journal of Inner Asian Art and
Archaeology 2: 49–58.
Derolez, Albert (2003). The Palaeography of Gothic Manuscript Books. Cambridge:
Cambridge University Press.
Fleet, John Faithfull (1888). Corpus Inscriptionum Indicarum Vol.III: Inscriptions of the
Early Gupta Kings and Their Successors. Calcutta: The Superintendent of Government
Printing, India.
Francke, August Hermann (1911). “The Tibetan Alphabet”. Epigraphia Indica 11: 266–273.
Gaur, Albertine (1984). A History of Writing. London: The British Library.
Gilissen, Léon (1973). L’expertise des écritures médiévales. Ghent: Editions Scienti2ques.
Harrison, Paul (1996). “A Brief History of the Tibetan bKa' 'gyur.” J. I. Cabezon and R. R.
Jackson, eds. Tibetan Literature: Studies in Genre. Ithaca: Snow Lion. 70-94.
Inaba, Shōju 稲葉正就 (1954). チベット古典文法学 Chibettogo koten bunpōgaku. Kyoto:
法増刊 Hōzōkan.
Iwao, Kazushi. forthcoming. “La gestion des greniers de Dunhuang à l'époque tibétaine. Le
cas de Pelliot tibétain 1111.” Journale Asiatique.
Laufer, Berthold (1918). “The Origin of Tibetan Writing.” Journal of the Royal Asiatic
Society 38: 34–46.
Mallon, Jean (1952). Paléographie Romaine. Madrid: Grá2cas reunidas.
ORIGIN OF TIBETAN DBU MED
445
Narkyid, Ngawangthondup (1983). “The Origin of the Tibetan script”. Ernst Steinkellner
and Helmut Tauscher, eds. Contributions on Tibetan Language, History and Culture:
Proceedings of the Csoma de Kőrös Symposium Held at Velm-Vienna, Austria, 13–19
September 1981. Vienna: Universität Wien. I: 207–220.
Richardson, Hugh Edward (1985). A Corpus of Early Tibetan Inscriptions . London: Royal
Asiatic Society.
Richardson, Hugh Edward (1990). “The Province of the Bde-blon of the Tibetan Empire,
8th to 9th Centuries.” Indo-Sinica-Tibetica: Studi in Onore di Lucian Petech . Rome:
Bardi Editore. 305–315.
Ribur Ngawang Gyatso (1984). “A Short History of Tibetan Script.” Tibet Journal 9.2: 28–
30.
Salomon, Richard (1998). Indian Epigraphy: A Guide to the Study of Inscriptions in
Sanskrit, Prakrit and Other Indo-Aryan Languages. New York: Oxford University Press.
Scherrer-Schaub, Cristina (1999). “Towards a methodology for the study of old Tibetan
manuscripts: Dunhuang and Tabo”. Scherrer-Schaub, C.A. and E. Steinkellner, eds. Tabo
Studies II: Manuscripts, Texts, Inscriptions and the Arts. Rome: Is.I.A.O. 3–36.
Scherrer-Schaub, Cristina (2007). “Revendications et recours hiérarchique: Contribution à
l'histoire de Śa cu sous administration tibetaine”. Études de Dunhuang et Turfan . Droz:
École pratique des hautes études. 257–326.
Scherrer-Schaub, C. A. and G. Bonani (2002). “Establishing a typology of the old Tibetan
manuscripts: a multidisciplinary approach.” Susan Whit2eld, ed. Dunhuang Manuscript
Forgeries. London: The British Library. 184–215.
van Schaik, Sam (2011). “A New Look at the Source of the Tibetan Script.” Yoshiro
Imaeda, Matthew Kapstein and Tsuguhito Takeuchi, eds. Old Tibetan Documents
Monograph Series, vol.III. Tokyo: Tokyo University of Foreign Studies, ILCAA. 45-96
van Schaik, Sam and Imre Galambos (2011). Manuscripts and Travellers: The Sino-Tibetan
Documents of a Tenth-Century Buddhist Pilgrim. Berlin: de Gruyter.
Sen, Tansen (2001). “In Search of Longevity and Good Karma: Chinese Diplomatic
Missions to Middle India in the Seventh Century.” Journal of World History 12.1: 1-28.
Snellgrove, David L. and Tadeusz Skorupski (1980). The Cultural Heritage of Ladakh 2:
Zanskar and the Cave Temples of Ladakh. London: Aris and Phillips.
van Sommers, Peter (1991). “Where Writing Starts: The Analysis of Action Applied to the
Historical Development of Writing.” Jan Wann, Alan Wing and Nils Sovik, eds. Development of Graphic Skills: Research Perspectives and Educational Implications .
London: Academic Press. 3–38.
Takeuchi, Tsuguhito (1990). “A Group of Old Tibetan Letters Written Under Kuei-I-Chün:
A Preliminary Study for the Classi2cation of Old Tibetan Letters.” Acta Orientalia
Academiae Scientiarum Hungaricae 44.1–2: 175–190.
Takeuchi, Tsuguhito (1994). “Tshan: Subordinate Administrative Units of the ThousandDistricts in the Tibetan Empire”. Per Kvaerne, ed. Tibetan Studies: Proceedings of the
6th Seminar of the International Association for Tibetan Studies, Fagernes 1992 . Oslo:
The Institute for Comparative Research in Human Culture.
Takeuchi, Tsuguhito (1995). Old Tibetan Contracts From Central Asia . Tokyo: Daizo
Shuppan.
446
SAM VAN SCHAIK
Takeuchi, Tsuguhito (2004). “Sociolinguistic Implications of the use of Tibetan in East
Turkestan from the end of Tibetan Domination through the Tangut Period (9th-12th c.)”.
Desmond Durkin Meisterernst et al. eds. Turfan Revisited: The First Century of Research
into the Arts and Cultures of the Silk Road. Berlin: Dietrich Reimer Verlag.
Takeuchi, Tsuguhito. forthcoming(a). “Old Tibetan Buddhist Texts From the Post-Tibetan
Imperial Period (mid-9 c. to late 10 c.).” Christina Scherrer-Schaub, ed. Old Tibetan
Studies, Proceedings of the Tenth Seminar of the IATS, 2003. Leiden: Brill.
Takeuchi, Tsuguhito. forthcoming(b). “Old Tibetan Rock Inscriptions Near Alchi.” H.
Uebach and J. L. Panglung, eds. Studia Tibetica VI.
Thomas, Frederick William (1951). “The Tibetan Alphabet.” Festschrift zur Feier des
200jährigen Bestehens der Akademie der Wissenschaften in Göttingen (II Philologischehistorische Klasse). Berlin. 146–165.
Vitali, Roberto (1990). Early Temples of Central Tibet. London: Serindia.
Wang, Fang-yü (1958). Chinese Cursive Script: An Introduction to Handwriting in Chinese .
New Haven: Yale University Press.
Yamaguchi, Zuiho (1980). “沙州漢人による吐蕃二軍団の成立と mKhar tsan 軍団の位置
Sashū kanjin ni yoru Toban ni gundan no seiritsu to mKhar tsan gundan no seiritsu
(Foundation of two military settlements of the Tibetan army consisting of Chinese at Sha
cu and an attempt to locate mKhar tsan)”. 東京大学文学部文化交流研究施設研究紀要
Tōkyō daigaku bungaku-bu bunka kōryū shisetsu kenkyū kiyō (Annual Report of the
Institute for the Study of Cultural Exchange) 4: 13-47.