Written Language & Literacy I,:I (:oI:), :i,. uoi Io.Io,,/wll.I,.I.

oIfed
issx I,8,o,,: / v-issx I,,ooooI John Benjamins Publishing Company
Te development of structural characteristics
of Brahmi script in derivative writing systems
Liudmila L. Fedorova
Russian State University for the Humanities Moscow, Russia
Indian writing systems reveal a great variety of graphic forms, proceeding from
the unique source which is Brahmi script. Tese graphic forms render structural
oppositions developed in phonographic writing systems which stem from
Brahmi. Tis paper aims to highlight the changes scripts underwent to satisfy
demands of language structure. Te comparison raises the issue of the complexity
of writing systems.
Keywords: writing system; grapheme; typology; Brahmi; abugida; alphasyllabary;
akshara; diacritic; graphon; vowel diferentiation
:. Introduction
Te typology of phonographic writing systems can be based on diferent (but
interrelated) grounds structural or functional. Tis distinction can be compared
to the terms paradigmatic vs. syntagmatic used by Gamkrelidze (1994) with regard
to writing systems and, particularly, to alphabetic systems derived from Greek.
Tese terms are quite appropriate for the discussion of the Indian material in the
present analysis. And, in general, they correspond to the structural and functional
perspectives of research emanating from Cartesian and Gerderian approaches, as
Hymes (1974: 7879) notes.
Te structural criterion refers to the linguistic nature of the basic corresponding
unit of a written sign (phoneme, syllable, morpheme, or word). Te functional
criterion refers to the syntagmatic use of these signs, which can correspond to a
sound, a mora, a syllable, a word. As phonographic writing encodes spoken lan-
guage, there can be diferent correspondences between structural and functional
units.
Te term grapheme will be used in this paper to designate the independent
full-formed unit of linear scripts which refers to a linguistic unit. As full-formed
signs written within the vertical space of lines, graphemes are opposed to diacritics
i Liudmila L. Fedorova
as these are generally placed above or under the lines. As independent signs,
graphemes have their own linguistic referents and are thus opposed to
sub-graphemes, namely to both diacritics (even when placed in line) and parts of
ligatures. As signs with a linguistic referent (referring to linguistic units), they are
opposed, frstly, to punctuation marks (signs of division and integration) and, sec-
ondly, to diacritics that refer to features or characteristics of linguistic units and form
modifcations of invariant signs. Parallel to any other linguistic unit, a grapheme
can have its allographs. Finally, due to their structure, graphemes can be elementary
or compound, formed of several meaningful graphic elements composing a unit.
I am inclined to interpret Brahmi as belonging to the syllabic type of
phonographic writing, both in structure and in functioning. Although accord-
ing to Daniels typology, Brahmi belongs to abugida writing which is difer-
ent from syllabic writing, other authors highlight that Brahmi has a syllabic
nature: All Brahmi-derived writing systems are syllabic, i.e. characters represent
consonant-vowel syllables (Wali et al. 2009: 162). Bright (1996: 384391) refers to
such systems as alphasyllabaries, and so does Sproat (2006: 45):
Te Brahmi-derived Indic scripts occupy a special place in the study of writing
systems. Tey are alphasyllabic scripts [], meaning that they are basically
segmental in that almost all segments are represented in the script, yet the
fundamental organizing principle of the script is the (orthographic) syllable.
Since there is no complete agreement on classifcation terms,
1
I propose a rather
diferent, semiotic view on the hierarchy and distribution of the types of writing
systems, taking into account the referent of a writing unit. (cf. for the whole typo-
logical scheme Fedorova 2011)
Consequently, based on the linguistic nature of graphemes, I discern four
main functional types of syllabic systems:
1. Diferent elementary graphemes represent diferent syllables (CVC, (C)V, VC,
CCVC and others) as inseparable units; such systems historically developed
from logosyllabic (morphosyllabic) writing, or were recently standardized,
like the Yi; this type can be labeled as the Yi type.
2
2. Diferent elementary graphemes represent only simple moras ((C)V, (V)C):
the Kana type (following the Japanese model).
3
3. Te single grapheme represents syllables with the same consonant, accompa-
nied by diferent vowels or by their absence (C
x
) as an inherent characteristic
of a syllable: the abjad type (to use Danielss term (1996: 317)). Structurally,
it can be described as a consonant alphabet since only consonants have a vis-
ible expression. Functionally, it can be seen as syllabic writing since it allows
graphemes to be read as syllables (according to Gelb).
Te development of structural characteristics of Brahmi script in derivative writing systems
4. Te single grapheme stands for syllables with the same consonant, whereas
its modifcations diferentiate vowels (C
v
): the abugida type (Daniels term).
Structurally, this system can be regarded as an alphasyllabary (Brights
term).
Notionally, it is possible to discern a ffh model with the vowel grapheme being
diferentiated by a consonant (V
c
). Pahawh Hmong script comes closest to it
(Ivanov 2004: 25; Rogers 2005: 260263): though usually there are two graphemes
for a syllable CV in this writing system, their order in script is inverse VC. Tis
may show a subordinate role of the consonant in a syllable block in which the
vowel also has a tone function; so Pahawh Hmong combines properties of the
alphabetic and the syllabic type it has full-sized letters for vowel and consonant
phonemes, and its syllabic units (formed as non-linear) follow each other in a
linear order.
Each type is based on its own principle, or mode, of writing. When speaking
about principles, or modes, of writing I mean the common rules, which underlie
the way to relate speech with graphic symbols. Te second type is opposed to the
frst one in the way complex sound sequences are divided in parts; the third and the
fourth type of scripts divide the sound sequences further into subordinate parts
a consonant as a head and a vowel as a dependent element, a specifc or unspecifc
characteristic of a syllable and represent them in writing as subordinate vowel
marks in an abugida or as a zero-sign corresponding to an unspecifc vowel or
its absence in an abjad (the zero-sign can be substituted with matres lectionis or
with diacritic marks of vowels, cf. the Tiberian system). Under this conception,
an alphabet makes the relation between the consonant and vowel components
coordinative. I do not attempt to trace the historical evolution of script but rather
I am trying here to give a semiotic perspective of its improvement that reveals
in adaptation of a writing system to a language (the most common way in the
development of writing systems). In fact, such adaptation deeply depends on the
structure of a language its phonological, morphological, and lexical values, as
Pandey (2003: 4161) notes.
i. Te structural oppositions of the Brahmi writing system
Brahmi script has the form of an abugida, with an akshara (an orthographic syl-
lable, a syllabogram) being the main item. Te akshara is a traditional name for an
autonomic grapheme representing a C
v
syllable which refers to Ca in its alphabetic
form but can denote other vowels in the forms modifed with diacritic marks.
| Liudmila L. Fedorova
Te abugida systems are already the next stage of diferentiation of the abjad a
stage of vowel diferentiation (yet, the Japanese Kana, belonging to the second
functional type, is not a precedent, but a deviated pattern; Coulmas (1996: 252)
remarks on the Indian infuence on its structure).
Regarding its internal organization, the Indian script is a great innovation: it
allows representing the phonological structure of a word quite distinctively. Deriv-
ative scripts maintain and develop the main structural oppositions:
1. between CV and V: in the development of modes of representation of inde-
pendent vowels
2. in the diferentiation of short and long vowels
3. in the introduction of the opposition C
v
: C

(presence of a vowel vs. absence

of a vowel), which enlarges the paradigm of a syllabic sign
4. in the integrated representation of consonant clusters
5. in the introduction of a system of tone diferentiation
6. in the aspect of text division marking a word, a phrase, or other units.
Tese structural oppositions make use of some graphic devices, the main ones of
which are diacritics, ligatures, traits of integration, and marks of division. Tese
devices diferentiate signs and improve the writing system as a whole. Te paper
aims to specify some ways and modes of diferentiation. In general, I aim to recon-
struct the aksharas grammar, which is interpreted in accordance with the fol-
lowing idea: If a sign system has rules how to form complex signs and how to
combine signs to form larger sentences, it has a grammar, besides the lexicon
(Neef 2010: 231); with respect to a writing system, I take the set of its graphemes
to be its lexicon.
. Paradigmatic systems and their functioning: Alphabetic
andabjadprinciples of writing
Put into historical perspective, Semitic script was a great precedent (relative
toBrahmi) invention not only of a protoalphabetic system that had given birth
to the Greek alphabet, but also of a mode, or principle of writing, recently named
abjad. In order to distinguish between the paradigmatic system and its functioning,
I prefer to use the term alphabet for a paradigmatic system of graphemes which:
a. correspond to the sounds phonemes or syllables of a given language,
b. are ranged in a stable order,
4
c. usually have appropriate names which makes them letters.
5
Te development of structural characteristics of Brahmi script in derivative writing systems ,
In its functioning, such a system can follow an alphabetic principle or one of the
syllabic principles Yi, kana, abjad or abugida, as introduced above. According to
the alphabetic principle, each phoneme should be expressed by a full-sized graph-
eme. Te abjad principle, on the other hand, presupposes the use of graphemes in
the syllabic meaning C
x
(where C designates a specifc consonant and x as a zero-
sign an unspecifc vowel e.g. we follow the abjad principle in reading consonants
in the alphabetic row: /bi/, /ci/, /di/, but /jey/, /key/ with diferent vowels, though
they are not written.). Te abjad mode of writing is largely defned by the phono-
logical, morphological, and lexical structure of classical West Semitic languages,
where:
a. a vowel could not open a syllable, so it was regarded as a dependent character-
istic of a syllable;
b. a vowel was variable in word-formation, so it was not a constant characteristic
of a root.
6
Yet, a vowel accompanies a consonant in these languages, and, therefore, it is a
regular and implied characteristic of a consonant or a syllable as a pronunciation
unit as well.
|. Te abugida principle of writing: Vowel diferentiation
Te abugida principle was probably discovered in writing of Indian Prakrits,
in which the structure of a syllable, which could be CV or V, presupposed two
demands:
a. Since a vowel could correspond to a syllable, it needed a special designation.
b. Diferent vowels needed diferent designations.
|.: Kharosthi
A possible frst step towards systematic vowel diferentiation can be seen in Kha-
rosthi
7
(used mostly for Gndhr, a northwestern Prakrit dialect) which has
graphemes for independent vowels. Still, all of them can be regarded as variants
of an invariant grapheme for /a/, as if the invariant form represented a conso-
nant, e.g. for a glottal stop, with an inherent vowel /a/ (as in other invariant
graphemes) or an empty-consonant, so the whole alphabet (arapachana) could
have a uniform structure of graphemes CV. Te device that makes it function as
an abugida is a minor graphic element for vowel diferentiation I shall call it a
graphon. A graphon in Kharosthi is adjacent or tangent to the main sign within its
o Liudmila L. Fedorova
own space in the line, and it produces modifcations of the invariant grapheme.
It may change direction or the point of contiguity in accordance with the outline
of the grapheme, yet it preserves its distinctiveness. Tus, the akshara-grapheme
receives the internal structure by an invariant base and a variable graphon. Te
paradigmatic structure of such a system has two components: an alphabet of
Ca graphemes
8
(the invariant shapes) and a matrix of their modifcations with
graphons.
(a) (b)
Figure 1. (a) Independent vowels in Kharosthi (two styles); (b) paradigm of vowel
diferentiation for a syllable Ka
9
|.i Brahmi
In Ashokan Brahmi, the diferentiation of vowels goes further. Each vowel
receives its own shape in an independent position, not on the base of a com-
mon invariant. It is a more developed abugida than Kharosthi; however, it is
less uniform. It has two classes of graphemes: CV (a class for syllables) and V
(a class for vowels). Te latter can be conceived both as a class of syllables or
of phonemes since the vowels have no modifcations common with those of
syllable graphemes. In addition, vowels can have their own modifcations for
the opposition of a short or a long sound. In this way, /a/, /i/, and /u/ obtain
modifed forms for long phonemes (in Kharosthi, only the long /a/ of a C syl-
lable can be marked).
10
Tus, the whole system corresponds to the phonological
structure of Pracrits.
a
e ai o
u i
Figure 2a. Independent vowels in Brahmi
Te development of structural characteristics of Brahmi script in derivative writing systems
ka ke ko kam k k ku k ki
la le lo lam l l lu l li
Figure 2b. Paradigm of vowel diferentiation for Ka and La in Brahmi
11
Vowel diferentiation of a syllable in an abugida is not confned to the choice of
a graphon designating a concrete vowel, but it presupposes sequential operations
of deleting /a/ from Ca and then adding a diferent vowel. Tus, a vowel mark, a
graphon, is an operator that creates a new unit. Te nature of the main grapheme
would otherwise be uncertain: it may be understood as a sign of a syllable, an
akshara, or as a sign of a consonant, a letter. In fact, in Ashokan script such uncer-
tain uses can be found, but most graphemes in a text have a syllabic nature.
Te role of graphons in the modifcation of vowels (V vs. V:) and of con-
sonants (C
v1
vs. C
v2
) is similar. Tus, they can be conceived as the operators of
(almost) the same order in contrast to the dots (anusvara), marking a secondary
pronunciation feature.
Terefore, the paradigmatic structure of the Brahmi script has three main
components: a system of vowel aksharas, a system of invariant Ca aksharas, and a
matrix of their vowel modifcations.
|. Tamil Brahmi
In early Tamil Brahmi, the direction of derivation difers: it tends towards the
alphabetic writing. Te invariant form of a grapheme corresponds to a pure con-
sonant, while a syllable Ca is marked with a graphon (like the syllables Ci, Cu,
etc.). If the basic form represents a pure, unvocalized consonant, it is not a syl-
lable but a phoneme which forms the basis of the system (vowels can also be
perceived as phonemes). If the graphemes, modifed with graphons, represent
CV syllables, then graphons can be interpreted as variants of vowels, even being
tied closely to the shape of a letter. Te diference between this script and the
alphabetic type lies in the nature of these vowel marks in an akshara: they are not
the full-formed graphemes but only graphons, though they represent a necessary
component of a syllable, the soul of the akshara as traditional Tamil descrip-
tions say. Tis is, thus, a transitional type of a system, which combines features of
alphabet and abugida.
Yet, the later Tamil Brahmi restores the syllabic mode of writing and diferen-
tiates it further: it adds a new operator: a mark of vowel rejection. Tis mark works
8 Liudmila L. Fedorova
like a diacritic virama, or halanta, that can be found in later North Indian scripts.
It enlarges the paradigm of a syllabic sign, thus confrming the abugida principle.
Ashokan Brahmi (III BC)
a m ma m
Later Tamil Brahmi (II AC)
Modern Tamil
Early Tamil Brahmi (III BC)
Figure 3. Marks for a vowel /a/, a pure consonant /m/, and for vowel diferentiated con-
sonants /ma/ and /ma:/
12
So a vowel in the abugida script may be represented by a full-sized grapheme
only in initial or independent position (afer another vowel).
|.| Tibetan
In many derivative writing systems, the vowels get their own shapes in indepen-
dent positions. Still some scripts, like Kharosthi, prefer the empty-consonant
device in representing independent vowels. Tibetan writing forms graphemes for
other independent vowels with diacritics on the base of the a-grapheme; yet, there
are two vowels included in the alphabet for representing /a/ ([h/?] and [h/?]):
the frst one (the empty-consonant) is the base for short vowels, while the second
can serve as a diacritic for marking long vowels (it is more correct to say that the
postscript <a> acts de jure as a ligatured consonant: it forms a base of an akshara
with a long vowel prolonged akshara and vowel diacritics are added to this
compoundakshara).
a
i u e o
[h/?] a [h/?]
Figure 4. Some Tibetan independent vowels
|., Tai
In Tai script, initial vowels also exploit the base of a grapheme o ang included
in the alphabet. Te great variety of vowels in the Tai language uses almost
Te development of structural characteristics of Brahmi script in derivative writing systems
all possible positions for vowel marks around the consonant grapheme, and the
same marks are used around the o ang for independent vowels. Here it may bean
efcient device. So in Tai writing there are not only sub- and superscripts, pre- and
postscripts, but also circumscripts and even more complex designations.
In some other derivative systems, an analogous compromising way of ini-
tial vowel designation can be found in the use of an empty consonant as the
base for a standard akshara; but in fact it is just a sign from the alphabet. An
empty consonant is an unreadable sign; it serves as a carrier for an indepen-
dent vowel and thus maintains the structural uniformity of graphemes. Such a
device is used in Rong-ring (Lepcha) script, derived from Tibetan. A compa-
rable way of writing initial vowels afer the unreadable consonant can be seen
in Korean script which, though not a descendant of Brahmi, uses this device for
balancing syllable representation. Yet in Korean, the consonant /ng/ is unread-
able only in the initial position of a syllable block marking the initial vowel; the
same consonant is read in the fnal position. So it cannot be regarded as using
the Tibetan pattern.
(a)
o ang
a
[a?]
a
[a:]
ae
[?]
ae
[:]
uea
[a?]
uea
[:a]
oe
[?]
oe
[:]
ai
[aj]
ao
[aw]
silences
fnal consonants
o
[o?]
o
[o:]
ua
[ua?]
ua
[ua]
ia
[ia?]
ia
[i:a]
e
[e?]
e
[e:]
u
[u:]
u
[u]
i
[i:]
i
[i]
(basin)
[?/-]
(b)
Figure 5. (a) Tai basic vowel grapheme /o ang/ with its mnemonic name basin;
(b)some vowel graphemes
13
Te Brahmi model of using diferent independent shapes for initial vowels
was also widely applied. In Khmer script, for example, most vowels are repre-
sented by diferent shapes in initial position. But there is also an economic way
of vowel modifcation in consonant aksharas: two series of consonant aksharas
(originally voiced and voiceless) have identical consonant meaning in pairs, but
the inherent vowel difers /a/ or /o/, and the identical vowel marks (there are
about 30 vowels in Khmer) have diferent meanings depending on the consonant
series they referto.
:o Liudmila L. Fedorova
|.o Diacritics vs. graphons
In a dependent position (afer a consonant), a vowel can be marked by a graphon
closely tied to the main sign (radical), thus forming an integral sign; this device,
used in early writing systems,
14
was later developed by separating two parts: a
dependent vowel symbol, standing distantly from the main sign, and the main
sign itself, which remains unchangeable. Such way of writing allows the main
grapheme to be conserved intact in its own space in line and also makes the marks
for the same vowel uniform (for they need not adapt to the shape of the main
graphemes, cf. Ku: Lu in Brahmi, Figure 2b). Tese symbols difer from graphons
and may be regarded as stable alienable marks of their phonological function, i.e.
as pure diacritics, matras. Tis makes the script more distinctive, as we can see in
Devanagari or in the Tibetan script.
Now two types of the vowel marks can be distinguished: (1) those closely tied
with a main grapheme within its inner space (graphons); and (2) those bound to
the space of a main grapheme from outside (diacritics). Teir common character-
istics are their exclusively dependent use and their size, which is usually smaller
than that of a main grapheme.
15
Graphons and diacritics difer in their relative
position (inside or outside the inner space of the main grapheme) and in their
mode of conjunction to the main grapheme, solid or distinctly separable. Tis
behavior resembles fusion or agglutination in morphology. Terefore, I consider
a diacritic as a writing symbol which can modify a grapheme; its subordinate
phonological function to distinguish between variants of an invariant unit
determines its position (usually outside the inner space of a grapheme) and size
(usually not equal to that of a grapheme horizontally and /or vertically); its mode
of conjunction with a grapheme is distinct. Full-sized representations of vowels
placed in line but diferent from their initial forms (such as for example marks for
/e/ in Bengali, Tamil, and others) represent quasi-graphemes for they fulfll the
subordinate function of diacritics, the vocal modifcation of aksharas, which in
this way become complex.
In a way, there is a parallel between akshara-formation and word-formation
(word morphology): graphons and diacritics function like internal infexion and
afxation, respectively:
SING SANG SONG PLAY PLAYED DISPLAY
ka k k k ku ki ka k ki k ku k
Figure 6. Comparison of devices in word morphology and akshara-formation
Te development of structural characteristics of Brahmi script in derivative writing systems ::
|. Agglutination vs. fusion: Tamil script
A more complex system of vowel representation can be found in Modern Tamil.
Besides the set of initial signs and the set of regular diacritics, there is a number
of non-standard aksharas where a consonant gets its vowel modifcations in an
irregular way. Tus, an akshara can have the shape of a free modifcation or of
a ligature in which consonant and vowel shapes are combined in mutual fusion.
Such inseparable shapes tend to function as Kana signs (cf. Sproat 2006: 57).
Yet, in comparison with the invariant form, a common base can be found;
moreover, vowel marks can have somehow distributed alloforms (three non-
standard forms for /-u/), though the same form can refer to diferent vowels (cf
/-a:/ and /-u/).
pa p pi p pu p
pe p pai po p pau p
Figure 7a. Regular vowel marks
nu
nu
tu nu
su
mu lu
k ns
llu lu ru
ju ssu kssu hu
ru
Figure 7b. Some irregular vowel marks
n ru
Figure 7c. Alloforms for /-a:/ and /-u/
16
So, Tamil aksharas can be opposed in the mode of conjunction: agglutination
or fusion. Tus, the paradigmatic system of Tamil script has four main compo-
nents: vowel aksharas, Ca (consonant) aksharas, a matrix of regular vowel modif-
cations, and a set of non-standard vowel modifcations.
:i Liudmila L. Fedorova
,. Opposition of short and long vowels
,.: Devanagari
Te use of diacritics is usually limited to the outside positions above and below the
line, so they are divided into superscripts and subscripts. However, in Devanagari
there are also complex marks for dependent vowels: a full-letter high stem (verti-
cal right stroke) as a mark for the long phoneme /a:/ and as a base for superscript
vowel diacritics. (In some way it resembles the Arabic alif in its function of a vowel
carrier.) Te stem cannot be regarded as a proper grapheme because it does not
constitute a full letter, but when added to the main sign, it extends the letter space
in the horizontal dimension. Tis is an appropriate, iconic way for representing
long or complex sounds, i.e. long vowels and diphthongs (/a:/, /i:/, /o/, /au/).
Devanagari uses this stem also for the short /i/, in this case placing it with upper
diacritic before the main grapheme; such inversion cancels the operation of
lengthening. So according to its position, the vertical stroke can be a postscript or
a prescript.
Tere is no reason to consider the vertical stroke as an empty consonant, since
it does not form a base for each vowel, and moreover, it is also an inalienable ele-
ment of many consonant syllable signs, i.e. a carrier of a distinctive trait for 24 of
the 36 letters of the script and for a quantity of vertical ligatures. It also creates the
vertical dimension of a sign and determines its boundary. Tus, it helps to divide
the letters from each other and it maintains the uniformity of the style of writing
together with the horizontal stroke binding letters. In these functions, the vertical
stroke can be compared to a round contour, an outline common to many signs
in Oriya script (due to the historical material of writing: palm leaves that were
torn by horizontal lines). In quite a diferent system in another part of the world,
in Maya script, the element of this kind is an outside line of many pictographic
glyphs. It has the shape of a face (or just an oval), in which distinctive features are
placed.
Figure 8. Oriya script
,.i Gurmukhi
Tus, the vertical right stroke may be an alienable or unalienable sign. As an alien-
able sign, it has two functions in Devanagari: it marks the long /a:/ or it serves as a
carrier for a vowel diacritic. Te frst function is fulflled by other vowel diacritics
too, but the size and position in line discerns the stroke from them. Gurmukhi
Te development of structural characteristics of Brahmi script in derivative writing systems :
writing diferentiates further functions: it has a shorter stroke for /a:/ (turning it
to an upper postscript kann), while for the vowel carrier it reserves the full size
(moreover, there are not so many consonants with a stroke in Gurmukhi). In such
a way, the long stroke functions only as a carrier (as a part of a sign) and the short-
ened stroke becomes a postscript diacritic marking a long /a:/. Te other pairs of
short and long vowels (i, u) are marked similar to those in Devanagari.
Tus, vowel marks include postscripts and prescripts, which are not full-size
letters in the horizontal dimension. It is remarkable that there are three diferent
graphic shapes (bases for /a/, /i/, /u/) for 10 initial vowels (divided in groups: back,
front, and low), but only one of them (for /a/ or schwa) can really be used inde-
pendently; the others obtain their diferent value with diacritics, common to vowel
designations for consonants, and can therefore be regarded as vowel-type carri-
ers. Tus, Gurmukhi script has a compromising way to designate vowels. Bhatia
(2003: 181213) regards this mode as a device of generalization, which Devana-
gari failed to capture. Yet, Gurmukhi script is a more recent development (16th
century).
Te relevance of writing style is based on the balanced proportions of graph-
emes and sub-graphemes of dependent vowels.
Transliteration: sa:re insa:n a:za:d ate hakk te izzat de liha:z na:lbara:bar paida: hu
n
de han
Translation: All human beings are born free and equal in dignity and rights.
(First sentence of Article 1 of the Universal Declaration of Human Rights)
Figure 9. Gurmukhi script; transliteration of the frst sentence (till the vertical stroke
danda as a point mark); translation
Te most consistent representation of long vowels can be found in Tibetan
script where a unique subscript diacritic serves to indicate a long vowel (both inde-
pendent and in Ca-akshara) marked with another standard diacritic (Figure4).
o. Is there any progress in vowel representation?
While the main opposition of high and low vowels /i/ vs. /u/ is usually repre-
sented with super- and subscripts in many Brahmi derived scripts, there is an
increasing number of vowel diacritics in line. Bengali script chooses a prescript
position marking a vowel /e/ with its own diacritic; there are many examples of in
line vowel diacritics in Tai, Khmer, and Tamil writings. Tey can have full size
:| Liudmila L. Fedorova
and when composed of two parts (pre- and postscript in circumscripts) they take
up more space than the main sign, so an akshara can be represented by a com-
plex of signs: grapheme and quasi-graphemes. Sometimes they can form liga-
tures with main signs and behave like consonant signs and, as a result, the script
becomes oriented in a more linear way (though there may be some non-standard
ligatures). However, they remain diacritics as they depend on main (consonant)
signs and have diferent initial forms. Te vocalization can therefore be regarded
as a phonological feature of a syllabeme generating its variants, the proper func-
tion of diacritics.
Some gradual changes in the representation of vowels in writing systems can
be seen, moving from abjad to alphabet. While in abjad vowels are concealed
under the shape of a consonant in a way that there is no vowel diferentiation (if
it is not specially emphasized), abugida systems allow consonants to take diferent
shapes clothed by diacritics of diferent vowels.
17
Ultimately, alphabetic systems
represent vowels by letters equivalent in size and position to consonant letters.
Te Greek claim to democracy gives them their independent status in the Greek
alphabetic script and derived scripts.
Of course, this is only a metaphor; Greeks succeeded in alphabetic writing
due to the analytic approach applied to their language structure (and maybe due
to the inaudibility of Semitic alif and ain for their phonematic hearing). Abugida
scripts are no less exact in representing sound sequences; and the degree of exact-
ness is relative even in modern alphabets. Te indefnite vocalization in the abjad
is more appropriate for maintaining the recognizability of radicals. So the question
of progress is more a rhetorical than an evaluative one.
. Te representation of consonant clusters: Ligatures
.: Brahmi
Brahmi script has a model for constructing ligatures, or conjunct consonants
(samyuktkshara), to represent consonant clusters. Brahmi does not have many
ligatures: for /pra/, /tpa/, /vya/ and some others (mostly with a sonorant).
18
Te
model of conjunction is vertical, the base could be chosen as the best graphic vari-
ant (e.g. in the Ashokas Rock edict I, in /pra/ r is rendered by a superscript, in /tpa/
t is rendered as a subscript, in /vya/ y is the base and v is the subscript). Te whole
conjunct tends towards a minimal extension of its space and exploits not the linear
principle of connection in succession but the emblematic principle of a balanced
graphic composition. Te ligature does not difer from the vowel diferentiated
Te development of structural characteristics of Brahmi script in derivative writing systems :,
akshara: it forms an integral sign in a line. Te proper diacritic that does not touch
the main sign is an anusvara: a right superscript dot for nasalization.
Transliteration: iya
m
dha
m
mlip devna
m
priyen // priyadasin r lekhapit i dha na
ka
m
// ci jva
m
rbhi
m
tp prajhitavya
m
(Schneider 1978)
Translation: Beloved-of-the-Gods, King Piyadasi, has caused this Dhamma edict to be
written. Here (in my domain), no living beings (are to be slaughtered or ofered in
sacrifce)
19
/pri/, /tpa/, /vya
m
/
Figure 10. Ashokas Rock Edict I (Girnar 257 BC), a fragment with aksharas: /pri/ (the
third from the right, in the frst line and the frst in the second line), /tp/ (the sixth from
the right, third line), /vya
m
/ (the frst from the right, third line)
.i Devanagari
Te technique of combining consonant signs into ligatures was developed in deriva-
tive scripts. In Devanagari, two types of ligatures exist: horizontal and vertical. Verti-
cal ligatures have a fxed order from top to bottom. Tough a conjunct presupposes a
common space for its components, they are distinctive and only have one common
vertical stroke (if any), which is not longer than the line; that is why the vertical
ligature is usually compressed. Te connection is systematic, agglutinative, and sub-
ordinate to the rules (yet there are exceptions when one of the elements (or both)
is modifed in fusion). Tus, a consonant ligature difers from a vowelled akshara
as a unique integrated sign within the line, while an akshara with a vowel diacritic
has two discontiguous parts, one in line and the other (usually) out of line. Te text,
consequently, has two dimensions: horizontal, as a chain of aksharas in line, and
vertical, along a stroke, in a stack within the akshara. Tough the dimensions of a
ligature can be enlarged (like in aksharas with post- or prescript vowel diacritics), it
represents a unique sign, as can be seen just in a simple example of writing numer-
als (with a horizontal ligature for /nya/ and vertical ligatures for /dv-i/, /tr-i/, /c-a/,
where /tr-i/ represents modifed forms of components in fusion):
nya
0
eka
1
dvi
2
tri
3
catur
4
pacan
5
Figure 11. Some Devanagari numerals
:o Liudmila L. Fedorova
Vertical outline space is usually used for vowel designation in aksharas, while
horizontal extension is common for ligatures as well as for the addition of a vowel
element (the stroke) in an akshara.
. Tibetan
Te rules for conjunction in Tibetan are more complicated, as sound sequences
are not so balanced and the structure of a syllable can be (CC)CV(CC) with a
maximum of three consonants before the vowel and two consonants afer it.
Asa consequence, the Tibetan script has a two-dimensional way of represent-
ing consonant clusters within a syllable: (1) the vertical stacks of graphemes
can be completed with one to two graphemes in line; (2) the conjunct elements
may be placed before and afer, above or/ and under the central one, which
is the root of a syllable. Stacks of signs usually hang down from the upper
horizontal line of a grapheme (which is named head and can be regarded as a
sign-carrier).
Terefore, the order of reading may be as follows (the numbers mark the
order of reading):
20
6
2
3
4
5
6
1 7 8
A root grapheme is situated in the center (3). A syllable could be represented
by four graphemes in horizontal dimension, in series standing separately. A verti-
cal stack can contain four to fve components: a main grapheme (root), a consonant
superscript, one to two consonant subscripts, and the vowel diacritic (subscript or
superscript); the vertical compression (as in Devanagari) is not adopted, so the
vertical dimension creates compound graphic signs in which subscript consonants
are common.
As a complex akshara may include more than one sign in line, a problem of
dividing closed syllables arises. It is solved by introducing a dividing dot in the
upper right side of the last sign of a syllable, a hanging dot called tsec. Te last
sign in a group before the dot (and without vowel diacritics) receives a trimmed
sound meaning C without a vowel. So, a grapheme can have a syllabic or an
alphabetic meaning according to its position: only the root is a syllable, but con-
junct consonants in line can be regarded as graphemes with reduced phonological
meaning. Tus, a closed syllable can be represented as a complex of graphemes (a
complex akshara).
Te development of structural characteristics of Brahmi script in derivative writing systems :
laykor
0
klad-kor
chig
1
gchig
nyi
2
gis
sum
3
gsum
shi
4
bzhi
Figure 12. Some Tibetan numerals with transcription and transliteration
21
8. Analytic vs. synthetic writing
Simple graphemes consisting of one radical for Ca can be distinguished from
compound ones, burdened with subscripts or superscripts, or bound with ligatures.
It is possible to draw a parallel between this diference and the analytic/ synthetic
character of word sequences in a grammar (in accordance with the typological
approach to language analysis by Greenberg (1960: 178194)): analytic chains do
not integrate radicals and afxes, while synthetic chains integrate them. Similarly,
analytic writing presupposes the use of simple graphemes, corresponding to
a standard pronouncing unit (a sound/ a mora/ a syllable), forming a standard
phonological unit (a phoneme, a moreme, a syllabeme),
22
while synthetic writing
presupposes the use of compound graphemes. Te opposition of simple and
compound graphemes allows speaking about the degree of synthesis (complexity)
of a grapheme (S
g
) and the degree of synthesis (complexity) of a writing system
(S
w
), which can be calculated.
Te extension of the vertical dimension used as a graphic variant in Brahmi
for representing vowels and ligatures became more frequent in Devanagari (though
vertical ligatures are compressed in line) and quite common in Tibetan. Te further
development of writing may have exploited this vertical dimension as a prevalent
tendency: it exists in Phags-pa script, which is derived from Tibetan but culturally
oriented to Chinese script in its appearance. Te Phags-pa script deviates com-
pletely from two-dimensional writing and changes the direction of writing from
horizontal to vertical (in a top to bottom fashion). Vowel diacritics and conjunct
consonants are written in a common stack for a syllable; all components of a syl-
lable are bound together in a linear block. Tus, the Phags-pa script seems closer
to alphabetic writing, yet the structure of abugida is preserved in two main points:
a. there is no mark for /a/ in the dependent position (afer a consonant), there-
fore this consonant grapheme corresponds to a Ca syllable;
b. shapes of vowels difer in the independent (initial) and dependent position.
:8 Liudmila L. Fedorova
Figure 13. Phags-pa script
Te Phags-pa script led the way to the new integration of signs: putting
them in a syllable block where previous (Tibetan) vertical diacritics are no more
opposed to previous in line signs, since they all are arranged in a stack. As a result,
the problem of delimitation arises for a model of conjunction can be fusion rather
than distinct agglutination. So Phags-pa reveals great complexity in decoding its
blocks, and the problem of graphic complexity is not merely reduced to the degree
of synthesis.
A compromising way is used in Khmer, for instance. Tere the consonant liga-
tures are extended downwards by the subscript consonants while vowel marks can
have diferent positions around the root grapheme (similar to what is found in
Tai where vowels can also form non-standard ligatures); the consonant-dimen-
sion becomes downwards oriented. Subscript consonants can have a compressed
or modifed form. As a result, the paradigmatic system can have an additional
component of consonant ligatures.
Another way of consonant-vowel arrangement in a syllable block was made
use of in Hangul, the alphabetic Korean script (not derived from Phags-pa):
though a syllable cannot begin with a vowel, the orientation of a consonant graph-
eme is subordinate to the graphic shape of a vowel grapheme; thus, the allograph
forms appear in the block construction.
While speaking about graphic complexity of writing, diferent parameters
should be taken into account: the degree of synthesis, the mode of conjunction
(fusion or agglutination), the general number of graphemes and sub-graphemes
in a system, the number of allographs and of non-standard ligatures, and also the
presence of word (syllable) division marks. Finally, writing styles (cursive, cal-
ligraphic, or printed) can difer greatly in these respects. In addition, graphic
complexity does not correspond to orthographic complexity which is connected
mostly with language change (cf. Figure12).
Te development of structural characteristics of Brahmi script in derivative writing systems :
. Tone diferentiation
Te development of related writing systems tends to express speech character-
istics more distinctively. Te necessity of marking diferent tones produces not
only a functional specifcation of graphemes (in Gurmukh), but also new series
of diacritics, as in Tai script or in Burmese script. So diacritics can have diferent
functions there.
Transcription: di
2
hma
2
la
1
htw
4
la
2
pyi
2
| wa
2
gu
3
i
2
hla
1
i
4
||
wa
2
o
3
wa
3
i
2
she
2
i
4
|| ka
3
wa
2
mla
2
pa
2
||
Translation: Here the moon rises; the cotton is beautiful.
Te yellow bamboo is high. Tis yellow car does not come.
Figure 14. Burmese script
Some conjunct consonants can be represented by subscript diacritics. Vow-
els have prescript, postscript, superscript, or subscript marks and their combina-
tions; they serve at the same time to specify the tone, which is represented in a
non-standard model (diferently for diferent vowels); tone marks are intricat-
edly bound with vowel marks (Daniels 2006: 19). Burmese script, consequently,
has a very complex system of graphic representation, as well as orthographic
difculties.
:o. Text division
Text division was not regarded as being necessary for a long time. In Brahmi script,
marks of sentence division are rare. In Devanagari, the mark is a horizontal bar
that binds the graphemes in words or groups of words (matrica, proceeded from
Kushan Brahmi). Tere is also a mark for the end of a sentence, a vertical right
stroke higher than the line (danda). All these signs, which should be considered as
punctuation marks, make the script more distinctive. Later they developed into a
more complicated system.
Te most interesting punctuation marks are used in the Javanese script.
Unreadable symbols at the beginning of a letter can designate an addressees social
status by the height of the lef line. Some honorifc graphemes (akshara murda) are
also used for writing a high ranking persons name.
io Liudmila L. Fedorova
to a person of a lower rank to a person of an equal rank to a person of a higher rank
Figure 15. Honorifc marks at the beginning of Javanese letter
Tere is a certain parallel with honorifc writing in the Russian epistolary
tradition, where an author capitalizes the frst letter in the pronoun (and its gram-
matical forms) in addressing a person of high status, while choosing the lowercase
letter when he/she does not want to mark his/her high respect or distance, cf. vs.
. (So I is always in minuscule while you can have three forms: , , ).
::. Calculations of the degree of synthesis
Simple or complex aksharas can determine the analytic or synthetic writing. Te
index of synthesis of writing (S degree of synthesis) can be calculated, which I
will briefy demonstrate in the following illustration.
As an example, the writing of the famous mantra Om mani padme hum can be
compared in the following scripts:
Devanagari:

m
ma-i pa-dme hu
m
O
m
ma-ni pa-dme hu
m
O
m
ma-ni pa-dme h
m
-m ma-ni pa-d -me h-m
Bengali:
Tibetan:
Tamil:
Te transcriptions are divided in graphic syllables (aksharas), and their number
A is: A
D
= 6, A
B
= 6, A
Ti
= 6, A
Ta
= 9. Te number G of elementary phonologi-
cally meaningful (or distinguishing phonological meanings) graphic elements is:
G
D
=13, G
B
= 12, G
Ti
= 14, G
Ta
= 16. (Yet there exists another possibility to write
-dme- in Devanagari: in two aksharas with a visarga afer -d- (d-me) like in Tamil;
in this case A
D
= 7, G
D
= 14, S
D
= 2. Te same is true for Bengali.)
Some calculations need comments:
1. Te unique graphic element for vowel prolongation in Devanagari (U: ) and
in Tamil (O: ) is diferentiated, using the procedure of Greenbergs square.
2. A vertical stroke is not divided from a superscript in vowel diacritics (for /i/)
in Devanagari, in Bengali, and in Tamil because a sole diferentiating meaning
Te development of structural characteristics of Brahmi script in derivative writing systems i:
for the stroke cannot be determined (yet this can be done otherwise when
considering two homographic signs: one for a vowel carrier and another as an
unalienable part of many letters).
3. Te inline sign for /e/ in Bengali and in Tamil are considered elementary
graphic elements, a diacritic (or quasi-grapheme, but not a true akshara-
grapheme).
4. Te close ligatures for /dm-/ in Devanagary and Bengali are counted as two
graphic elements, as they can be recognized. (Tis may be arguable since the
model of conjunction is fusion rather than agglutination (cf. Tibetan), but it
can be specifed in introducing the index of fusion.)
5. Te vowel carrier (grapheme <a> [h/?]) in Tibetan is considered a mean-
ingful graphic element, but an empty sign.
6. Sentence marks and syllable dots in Tibetan are not counted.
On this base, the degree of synthesis S can be calculated as a ratio of G to A:
S
D
= 13:6 = 2.17
S
B
= 12:6 = 2.0
S
Ti
= 14:6 = 2.33
S
Ta
= 16:9 = 1.78
Tamil yields the smallest result, which is clear since it uses pulli (virama) and
does not have consonant ligatures. Sproat (2006: 56) refers to the formal point
of view regarding Tamil script as the simplest Indian writing.
23
Te Tibetan
writing is the most complex in regard to synthesis. Of course, these results are
quite preliminary and cannot be considered as defnitive indexes of synthe-
sis for these writing systems. Te reason is, above all, the fact that very short
phrases were compared. Te examples just demonstrate the possible method of
index calculating. Te degree of synthesis is not the unique criterion for evalu-
ating the complexity of a script; other dimensions noted earlier should be taken
into account.
:i. Conclusion
In conclusion, it should be stated that the model of abugida, once invented in
India, has produced many variants in the related scripts of diferent languages.
Te inherent possibilities have developed in further diferentiation. Tis is a com-
mon way in the adaptation of writing: through more distinctive diferentiation to
new integrations. In this text, I tried to show how derivative scripts maintained
and developed structural oppositions in akshara grammar: in the representation
of a phonological unit as a non-linear composition a written emblem. Simpler
ii Liudmila L. Fedorova
writing tends to correspond to the linearity, while more complex writing tends to
exploit the emblematic principle as well.
Te productive use of emblematic syllable representation is characteristic
also for the Korean script which has an alphabetic nature, but is not quite simple.
Its difculty stems from the emblematic arrangement of its graphemes in sylla-
ble blocks and their changing shapes (allographs), which need to be discovered
in reading. An emblem tends to represent the whole in integrating its parts in a
conventional way.
It can be assumed that the linear and emblematic principles of writing provide
productive possibilities that can be exploited constructively. Sometimes a linguis-
tic emblem, an abbreviation, can take the form of a written emblem, so both of
them can have a symbolic meaning.
Notes
:. See e.g. Bright (2000: 6371) and the recent discussion of Daniels (2009: 277281) and
Swank (2009: 282285).
i. Te standardization of the logosyllabic Yi writing for the Nosu language in the 1980s
resulted in a syllabic system of 819 basic characters (Bradley 2009: 179).
. Te name moraic for this type (Sproat includes also Linear B, Sumerian and Mayan writing
systems in it (cf. Rogers 2005: 274)) seems not quite appropriate since it implies that graphemes
encode moras (syllables with a short vowel) only in this type of writing and not in others like
abugida or abjad which remains debatable.
|. Tus, an alphabet can have properties not only of a system, but also of a text that can be
learned and interpreted, and its format can be used with other purposes (semiotic counting
and even magic).
,. In fact, their names can be partly derived from their pronunciation as in Latin or modern
Cyrillic alphabets, yet they become not only marks of sounds, but names of letters (they difer
from hieroglyphs which can render sounds or words or morphemes).
o. References can be made to: Lyavdansky A.K. (2009). Te origin and early development of
the West Semitic alphabets. In A. Belova, L. Kogan, S. Loesov & O. Romanova (eds), Languages
of the World. Te Semitic languages. Akkadian. Northwest Semitic. Moscow: Academia, 811821.
. Te problem of the origin of Indian scripts is reviewed in Salomon (1995); mostly,
Kharoshthi is seen as an invention under the infuence of Aramaic script.
Te development of structural characteristics of Brahmi script in derivative writing systems i
8. Tere are some CCV graphemes, too (<ksa>, <sta>), which are elementary shapes and not
ligatures.
. Fragments are extracted from the illustrations in: http://www.rbardalzo.narod.ru/novosti.htm.
:o. Tis can be found only in later documents, maybe under the infuence of Brahmi, as
Salomon (1996: 373383) supposes.
::. Fragments are extracted from the illustrations in: http://www.omniglot.com/writing/thai.htm.
:i. Fragments are extracted from the illustrations in: http://en.wikipedia.org/wiki/Tamil_Brahmi
with reference to: Mahadevan, Iravatham (2003). Early Tamil epigraphy from the earliest times to
the sixth century A.D. (Harvard Oriental Series 62) Cambridge: Harvard University Press, 173.
:. Fragments are extracted from the illustrations in: http://www.omniglot.com/writing/thai.
htm.
:|. Te device of modifcation of a sign in its inner space is still used in more recent writing
systems on the base of Latin or Cyrillic writing; like other devices superscript and subscript
marks, digraphs, ligatures they usually just create another alphabetic sign and do not operate
like true graphons or diacritics whose function is to mark regular changes of sounds. Daniels
considers such marks with the example of Vietnamese vowels as integral parts of the letters
(Daniels 2006: 19). Such marks can be named quasi-graphons or quasi-diacritics, if needed.
:,. Tis is not the case in South and South-Eastern scripts with a quantity of vowels.
:o. Fragments are extracted from illustrations in: http://www.omniglot.com/writing/tamil.htm.
:. Te metaphor of clothes is used for the vowel diacritics in Javanese script.
:8. Sound sequences in Pali have less consonant clusters than in Sanskrit or some later Indian
languages like Hindi (Shevoroshkin 2004: 135138; the author of the book which was frstly
edited in 1969 and was not translated into English proposes two ways of translating his term
from Russian: sound chains or sound sequences).
:. Translation according to Dhammika, Ven. S. (1993). Te edicts of King Ashoka. Electronic
edition.
io. Te scheme is based on Rogers (2005: 225), using more detailed description in: http://
tibetan.bitecs.ru/fles/text_intros.pdf.
i:. Transliterations given according to: http://tibetan.bitecs.ru/fles/text_intros.pdf.
ii. While linguists have been elaborating the notions of phoneme, morpheme, and syllabeme
during the last 100 years, creators of writing systems resolved the problem of generalization of
pronouncing units in graphemes thousands of years ago. Teir analysis can be considered as
a practical base for further investigations. So the notion of moreme can be put in a line with
others according to the logic of graphic systems using them.
i. Still he shows that non-standard vowel-consonant ligatures can resist analysis, which
makes Tamil move closer to Japanese Kana, a core syllabary (Sproat 2006: 5668); it can enlarge
the general number of graphemes that resist analysis (and the general complexity of writing
system), but not the degree of synthesis.
i| Liudmila L. Fedorova
References
Bhatia, Tej K. (2003). Te Gurmukhi script and other writing systems of Punjab: History, struc-
ture and identity. In Peri Bhaskararao (ed.), Working papers of the international symposium
on Indic scripts: past and future, 181213. Tokyo: ILCAA.
Bradley, David (2009). Language policy for China minorities: Orthography development for the
Yi. Written Language and Literacy 12: 170187.
Bright, William (1996). Te Devanagari script. In Peter T. Daniels & William Bright (eds.), Te
worlds writing systems, 384391. Oxford: Oxford University Press.
Bright, William (2000). A matter of typology: Alphasyllabaries and abugidas. Study in the Lin-
guistic Sciences 30: 6371.
Coulmas, Florian (1996). Te Blackwell encyclopedia of writing systems. Oxford: Blackwell.
Daniels, Peter T. (1996). Te study of writing systems. In Peter T. Daniels & William Bright
(eds.), Te worlds writing systems, 317. Oxford: Oxford University Press.
Daniels, Peter T. (2006). On beyond alphabets. Written Language and Literacy 9: 724.
Daniels, Peter T. (2009). Two notes on terminology. Written Language and Literacy 12: 277281.
Fedorova, Liudmila (2011). (Towards the
typology of writing systems). Moscow Journal of Linguistics 13: 174195 (in Russian).
Gamkrelidze, Tamaz V. (1994) (1989). Alphabetic writing and the old Georgian script. New York:
Caravan Books.
Greenberg, Joseph H. (1960). A quantitative approach to the morphological typology of lan-
guage. International Journal of American Linguistics 26: 178194.
Hymes, Dell (1974). Foundations in sociolinguistics. Philadelphia: University of Pennsylvania
Press.
Ivanov, Vyacheslav V. (2004). Te linguistics of the third millennium: Questions to the future. (In
Russian) Moscow: Iazyki Slavianskoi Kultury.
Lyavdansky, Alexej K. (2009). Te origin and early development of the West Semitic alphabets.
In Anna G. Belova, Leonid E. Kogan, Sergej V. Loesov & Olga I. Romanova (eds.), Lan-
guages of the world: the Semitic languages. Akkadian. Northwest Semitic, 811821. Moscow:
Academia.
Neef, Martin (2010). On defning phonology. In Sang-Oak Lee (ed.), Contemporary Korean lin-
guistics: International perspectives, 216230. Gyoha-eup (Korea): Taehaksa.
Pandey, Pramod (2003). Phonetic and phonological bases of Hindy orthography. In Peri Bhas-
kararao (ed.), Working papers of the international symposium on Indic scripts: past and
future, 4161. Tokyo: ILCAA.
Rogers, Henry (2005). Writing systems: A linguistic approach. Malden, MA: Blackwell.
Salomon, Richard G. (1995). On the origin of the early Indian scripts: A review. Journal of the
American Oriental Society 115: 271279.
Salomon, Richard G. (1996). South Asian writing systems: Brahmi and Kharoshthi. In Peter
T. Daniels & William Bright (eds.), Te worlds writing systems, 373383. Oxford: Oxford
University Press.
Schneider, Ulrich (1978). Die groen Felsen-Edikte Aokas. Kritische Ausgabe, bersetzung und
Analyse der Texte. Wiesbaden: Otto Harrassowitz.
Shevoroshkin, Vitaly V. (2004) (1969). Te sound chains in the languages of the world. (In Rus-
sian) Moscow: URSS.
Te development of structural characteristics of Brahmi script in derivative writing systems i,
Sproat, Richard (2006). Brahmi-derived scripts, script layout, and segmental awareness. Written
Language and Literacy 9: 4566.
Swank, Heidi (2009). A response to Peter T. Daniels. Written Language and Literacy 12: 282285.
Wali, Aamir, Richard Sproat, Prakash Padakannaya & Bhuvana Bhuvaneshvari (2009). Model
for phonemic awareness in readers of Indian script. Written Language and Literacy
12: 161169.
Authors address:
Liudmila L. Fedorova
Russian State University for the Humanities
Institute of Linguistics
Miusskaya square 6
Moscow 125 993
Russia
[email protected]