Le soninké dans les manuscrits islamiques : Corrélation entre les correspondances graphématiques et phonémiques

Abstract

This article compares the graphemes of Ajami as used in Islamic manuscripts and the phonemes of present-day Soninke. This comparison made it possible to identify two types of correspondences: unambiguous correspondences and ambiguous correspondences. Unambiguous correspondences concern graphemes where a phoneme corresponds to a grapheme: /b/ = ب <b>, /f/ = ڢ <f>, /g/ = ݣ <k2>, /l/ = ل <l>, /m/ = م <m>, /n/ = ن <n>, /ŋ/ = ع <ʿ>, /r/ = ﺭ <r>, and /w/ = و <w>. Ambiguous correspondences concern graphemes or phonemes which can encode or be encoded by several phonemes or graphemes: /c/ = س <s>, ص <ṣ> or ش <ʃ>, /d/ = د <d>, or ض <ḍ>, /j/ = ج <j>, ز <z>, ذ <ð>, or ظ <ẓ>, /h/ = ح <ḥ>, or ھ <h>, /t/ = ت < t> or ط <ṭ>, /x ̴ q/ = خ <x>, ف <q>, or غ <ɣ> and finally /s/ = س <s>, ص <ṣ>, or ش <ʃ>. Consonants absent in Arabic are rendered by Arabic characters whose phonological properties are closer. The final vowels of part of words are noted long in graphemes while they are short in current Soninke. I postulate that the short realization of the final vowels of these words is a recent development and that at the time when the texts were written, these vowels were probably pronounced long.

Keywords

• Soninke
• Ajami
• phoneme
• grapheme
• comparison