Brāhmī-derived orthographies are typologically Āksharik but functionally predominantly alphabetic

Abstract

We claim that the orthographies of South and Southeast Asia, which are derived from the Brāhmī writing system, are best described by the typological term “āksharik” (/ɑ:kʃərik/), and that they are functionally predominantly alphabetic. We derive these descriptions from the encoding units that went into the design of the orthographies and the decoding units that are likely to be relevant for reading text in them. More commonly used terms, such as “alphasyllabary” or “abugida”, used to describe these orthographies, are not appropriate. Our focus is on Hindī, an Indo-Āryan language and its orthography, Devnāgrī. Our arguments are based on structural-descriptive, historical-linguistic and theoretical accounts as well as some psycholinguistic evidence. We also support claims that the universal and language-specific features of any writing system should be studied in the context of its associated spoken language or languages.

Keywords:

• Brāhmī
• Indo-Āryan
• Akshar
• Alphasyllabary
• Abugida
• Devnāgrī
• Hindī

Notes

¹ We define “writing system” as a set of broad principles that describe the general mapping between written and spoken units (for example, between graphemes and phonemes or syllables). The specific instantiation of a writing system for a particular language is its “orthography”. For example, both English and German have alphabetic writing system but different (i.e., relatively inconsistent and consistent, respectively) orthographies.

² We do not discuss the case of the Indo-Āryan language Urdu because it is commonly written in an orthography that is a derivative of the Persian and Arabic writing systems, which is generally not used to write other IALs.

³ Akshars are commonly identified as the orthographic units of Brāhmī-derived orthographies. This word was pronounced /əkṣərə/ in Sanskrit and is commonly spelt either ‹akṣara› or ‹akshara›. We have chosen the spelling ‹akshar› because (1) the original voiceless retroflex sibilant /ṣ/ has diachronically merged with its palatal counterpart /ʃ/, which is generally transcribed by the grapheme ‹sh›, and (2) in contemporary usage, the final schwa is dropped due to a general process of schwa deletion observed in most of the modern IALs, including Hindī, Bangālī, Panjābī and Marāthī, among others. (However, speakers of some dialects of modern IALs, Sanskrit and Dravidian languages may pronounce it.) Moreover, for present-day readers and speakers of IALs, the use of final ‹a› in the spelling ‹akshara› in Latin orthography is problematic primarily because it was not designed to transcribe IALs, which have a phonological distinction between /ə/ and /ɑ:/ Both vowels are spelt ‹a› thereby generally resulting in the pronunciation of ‹akshara› as /əkʃərɑ:/, which is neither a Sanskrit nor a Hindī word.

⁴ By “language forms” we refer to those components of the spoken language that allow language users to make their message public; that is, consonants, vowels, the syllables they compose and morphemes.

⁵ In semasiographic writing systems, symbols directly represent ideas rather than language forms.

⁶ We are aware that there have been various non-linguistic, socio-political modifications of different writing systems. Our perspective is only linguistic.

⁷ V denotes the nucleus or vocalic peak of a syllable, which can minimally represent a short vowel, V, or a long vowel, , which can be a monophthong or a diphthong. The short and long forms of vowels are contrastive (compare, /bin/ ‘without’ and /bi:n/ ‘a type of musical instrument’). In some texts, these forms are represented V and V:, or, V and VV. Here, we use the notation V and , or, /V/ and /V:/.

⁸ Akshars in Devnāgrī are “hung” from a horizontal line, which demarcates word-boundaries.

⁹ In his grapholinguistic equilibrium hypothesis, Seidenberg conjectured that: “if a language has complex inflectional morphology, then it will have a shallow orthography” because, in his conjecture, reading comprehension is maintained via trade-offs between orthographic complexity and spoken language complexity. Building on this, he later wrote: “In essence, spoken languages get the writing systems they deserve.”

¹⁰ There is no clear proof whether a writing system ever existed for (the old IAL) Sanskrit. We present a case that, if there were one, then how might the phonological structure of Sanskrit have shaped its written forms.

¹¹ Our approach is linguistic with a goal of understanding how the writing system represents various language forms. We do not intend to present a purely historical, epigraphical or paleographical perspective.

¹² This is also true for Prākrit languages in general (Woolner, 1928) and specifically for Pāli (Warder, 1991).

¹³ For the view that the design of Brāhmī might have been influenced by the consonantal-syllabic Semitic orthographies, such as those used for writing Aramaic or Phoenician, see Salomon (1998).

¹⁴ Conjunct consonants are single orthographic units representing more than one consonant, e.g., one akshar each for /kʃə/ and /jñə/.

¹⁵ A rough estimate of the prevalence of the full forms of vowels (in the word-initial position only and not in VV sequences) in the “Brihat Hindī Kosh” dictionary of Prasād, Sahāy and Srīvāstav (1992), is provided by making a comparison between the number of pages used to list words starting with the full forms of vowels and the number of pages used to list words starting with any (either full or half) forms of consonants and the diacritical forms of vowels. Out of 1374 dictionary pages, words starting with just the independent forms of all the 11 vowels occupy 14.63% of “page space”. By comparison, words starting with the half and full forms of 30 consonants, and the mātrā forms of vowels together occupy 85.37% of the page-space. Therefore, although the use of independent forms of vowels may be less than the use of other forms of consonants and vowels (on average, 1.33% per vowel versus 2.85% per consonant or diacritic vowel), it does not seem low enough to be ignored in a description of IAOs.

¹⁶ This is an adjective derived from the noun root akshar by adding the IAL suffix “/ik/”. The lengthening of the first vowel /ə/ to /ɑ:/ accords with the “vṛddhī-type” increment feature of morphophonemic alternation in Sanskrit.

¹⁷ Bilingualism (and multilingualism) is so pervasive in urban India that it can be considered normative.