Maldivian Thaana, Japanese kana, and the representation of moras in writing

Abstract

Thaana, the script used to write the Maldivian language Dhivehi, represents the segments—the individual vowels and consonants—of the language but also notates the mora count of its syllables by assigning long vowels two vowel marks each and by using an absence-of-vowel mark on syllable-final consonants. The two Japanese kana—hiragana and katakana—do not represent individual segments, but they too notate mora count by giving long vowels two signs and by using signs for syllable-final consonants. In both languages, the kinds of consonants that can occur syllable finally are severely restricted. It is the presence or absence of such a consonant, rather than the identity of the consonant, that is important. Using a writing system that represents mora count thus makes sense in both languages. Yet Thaana and the two kana otherwise work on different principles, so it is not adequate to call them simply moraic writing systems. This prompts an expansion of the inventory of script types: Thaana is a moraic alphabet and the kana are moraic syllabaries.

Keywords:

• Thaana
• Kana
• Mora
• Alphabet
• Syllabary

Acknowledgments

Many thanks to Chris Green, Brook Hefright, Rebecca McGowan, and two anonymous reviewers for helpful comments and criticism. Remaining errors are my own. This material is based upon work supported, in whole or in part, with funding from the United States Government. Any opinions, findings and conclusions or recommendations expressed in this material are those of the author and do not necessarily reflect the views of the University of Maryland, College Park and/or any agency or entity of the United States Government.

Notes

¹Linguistically oriented introductions to Thaana are Gair and Cain (1996) and DeSilva (1969). For the Dhivehi language, see Cain and Gair (2000) and Fritz (2002).

²In this paper, I use loanword to mean a borrowed word that has become established in a given language, while a foreign word is one that has not become established as a word of the language but may occur in set phrases, in the speech of bilinguals, or in transliteration from the donor language.

³See Gnanadesikan (2009) for further use of the term nonlinear alphabet, which describes any segmentally based script that does not treat vowels and consonants as equivalent entities and thus includes scripts such as Devanāgarī and other Brāhmī-descended akṣara-based scripts as well as Korean Han'gŭl. I prefer this nomenclature to ones that only consider linear alphabets of the Greek type to be “full” or “true” alphabets. I consider any system that represents individual segments to be an alphabet.

⁴This statement will need to be adjusted in the light of the upcoming theoretical discussion. More accurately, the consonants—and the neutral base alifu—are written as the core, while the vowels—and the consonantal mora sign sukun—are written above and below that core.

⁵The reverse (that each sign stands for a mora) is only true if sequences of palatalised-C + V, represented as Ci-yV (e.g. k^ya is written ki-ya with the second sign written smaller than the first), are considered single, complex signs. See Japanese Kana section.

⁶See Gnanadesikan (2011) for arguments for the existence of true syllabaries.

⁷The Thaana alphabetical order is unique and bears no resemblance either to the pan-Indic order used by the Brāhmī family or to the Perso-Arabic order.

⁸Following Sanskritic tradition these phonemes are often referred to as palatal stops (e.g. Coulson, 1976), but in the modern languages they are pronounced as affricates (Masica, 1991).

⁹Other transliterations include (recommended by B. Cain, personal communication, 9 September 2010) (Cain & Gair, 2000), ś (Fritz, 2002), and [rbreve] (Reynolds, 2003), the latter reflecting a more conservative nonsibilant pronunciation.

¹⁰Although alifu derives its name from the Arabic alif, it behaves differently in that it does not represent either an initial glottal stop or a long /a/.

¹¹The Dhivehi examples here and throughout come from a variety of sources, including Reynolds (2003), Maniku (2000), and articles in Dhivehi online newspapers.

¹²In terms of phonological theory, Dhivehi can be said to obey Itô's (1988) Coda Condition, which prohibits consonants with their own place of articulation from surfacing in syllable codas. The fact that /s/ disobeys the general prohibition may either be regarded as expressing the default place associated with the [strident] feature of the /s/, or it may be regarded as due to the interaction of a specific constraint that overrides the general Coda Condition in a constraint-based analysis, along the lines of works such as Itô and Mester (1994).

¹³In the case of the /t/ triggering an [i]-offglide, this can be seen as a survival of the [Coronal] feature (cf. Cain, 2000), while the retroflexion of a following coronal stop is clearly a survival of the [retroflex] feature. In the first case the place shows up on the preceding vowel, avoiding coda consonant restrictions, while in the latter case the retroflexion is licensed by being shared with the following onset consonant, avoiding coda consonant restrictions in the same way that other geminates do.

¹⁴It is common to refer to syllables early in a word as being to the left in the word, but as this language is written from right to left, such left/right terminology can be confusing. Note also that the word Dhivehi itself is an exception to the general stress rule, being stressed on the second rather than the first syllable.

¹⁵Using empty raa or gaafu marks a word as a transliteration of a foreign word. Alternate spellings without them exist in established loanwords.

¹⁶The difference in capitalisation of the script names follows conventions in the study of the two languages—see, for example Gair and Cain (1996) vs. Smith (1996). Thaana is treated as a proper name, while hiragana and katakana are descriptive labels meaning “smooth syllabary” and “partial syllabary”, respectively.

¹⁷In the Japanese examples, a sign-by-sign Romanisation is given immediately after the kana, followed by a whole-word Romanisation. In the sign-by-sign Romanisation “N” stands for the placeless final nasal and “Gem” indicates the first element of a geminate. Examples are from Hadamitzky and Spahn (1981).