Phonetic-based enhancement to Devanagari orthography for Konkani

Abstract

Konkani, a low-resourced language spoken along India's western coast, is officially written in the Devanagari script. Konkani came to be written in multiple scripts such as Devanagari, Roman, Kannada, Malayalam, and Perso-Arabic. Of these, the Devanagari script is recognised as the official script. Using a script that is not tailored for a certain language often gives rise to grapheme-phoneme mismatch. This becomes more complicated due to the very nature of rigidness of scripts as compared to the evolving nature of languages. In this paper, we try to address the grapheme-phoneme mismatch as regards to the officially accepted Devanagari script of Konkani. Some symbols present in the Devanagari script such as the ऋ representing the [r̩], or the retroflex [ʂ] do not represent the nature of Konkani. On the other hand, some unique vowel and consonant phonemes of the Konkani language are left with no representation in the script. This has led to ambiguity in pronunciation and spelling. To give an example, the Konkani word when written as चोर is unclear as regards its number. It is the pronunciation of the word that signals the correct number. [t͡sor] means (singular) ‘thief’ and [t͡ʃɔr] means (plural) ‘thieves’. Also, the grapheme च is used to represent two consonant phonemes in Konkani namely [t͡s] and [t͡ʃ] (discussed later). Such cases pose language learning/writing challenges for the native and non-native speakers. Some proposals for modification of the script to suit Konkani language did come from Konkani linguists and some lexicographers but these were not really pursued further. Although there are some orthography rules proposed to cater to issues of the above kind, their application is kept limited. To give an example, the diacriticॅ is suggested for showing the open vowel ɛ in Konkani but this rule is restricted for writing only those words that are borrowed from English. What this rule misses to capture is that in Konkani /ɛ/ is a phoneme and is present in native words of the language too and not just in borrowed words. Such selective rule framing has created confusion among language learners. In this paper, we examine the limitations of the Devanagari script and advocate script modification that addresses the phonological requirements of Konkani language. Devanagari character set modifications mainly include repurposing existing but underutilized diacritics or introducing minor changes as regards language-specific characters of Konkani. As the standardization of written Konkani is relatively recent, implementing such changes may be both timely and practical especially from a computational point of view. Notably, all necessary characters already exist within the Devanagari Unicode block, yet their use is constrained by current orthography norms. Revisiting these conventions would enhance written accuracy and phonetic-based orthography rules.