Meroitic

The writing system is understood. It has 23 signs plus a divider of two or three dots: four vowel signs, fifteen consonants that carry an inherent a, and four syllabic signs, ne, se, te and to; the sign e may mark a schwa or no vowel⁠[2]⁠[1]. Phonology has been reconstructed from Egyptian and Greek spellings of Meroitic words and the reverse⁠[3].

The language is partly understood. Nominal and non-verbal syntax are well described, and the funerary formulae can be parsed into invocation, names, filiation, titles and benediction⁠[3]⁠[1]⁠[17]⁠[18]. A short list of words is secure, among them qore "king", k(a)(n)di "woman", ato "water", at(a) "bread", mak(a) "god" and mas(a) "sun"; verb morphology is still at an early stage⁠[6]⁠[3]. No bilingual long enough to serve as a key exists⁠[14].

Griffith's sign values have been tested by a century of use. They give consistent spellings of the same names across hieroglyphic and cursive texts and match Egyptian and Greek renderings of Kushite royal names, and they have not been revised beyond refinements such as the syllabic reading of ne, se, te and to⁠[5]⁠[1]⁠[2].

The language identification has been tested by comparison. Rilly reconstructed Proto-Northern East Sudanic from its living descendants and argued that Meroitic matches it in basic vocabulary, word order (verb-final, postpositions, genitive before noun) and morphology⁠[8]⁠[6]. Rowan argued that the sound system and syllable structure look Afroasiatic instead, and Lipiński has supported an Afroasiatic link on vocabulary⁠[19]⁠[6].

Computational work is recent and small. Smith used information-theoretic word similarity in 2008 to extend partial meanings, and Otten and Anastasopoulos in 2025 trained word embeddings on their corpus and attempted alignment with Late Egyptian; the embeddings captured some semantic structure, but the alignment could not induce a lexicon it had not seen, which the authors put down to the lack of machine-readable data⁠[20]⁠[14].

No one disputes how the signs are read. The positions concern the language. The working consensus is Rilly's: Meroitic is a Northern East Sudanic language of the Nilo-Saharan phylum, and its vocabulary and grammar can be recovered by comparison with Nubian and its relatives⁠[7]⁠[8]⁠[1]. The minority view places it in Afroasiatic on phonological grounds⁠[19]. A third view, older, holds the language to be an isolate whose affiliation cannot be shown from the material available⁠[6].

A substantial bilingual, Meroitic with Egyptian or Greek, would allow translation directly. Short of that, progress depends on data: a machine-readable REM with every text transliterated, joined to the analytic dictionary, would let cognate-matching against reconstructed Northern East Sudanic and Old Nubian be run systematically and tested against Afroasiatic and random-language controls⁠[11]⁠[13]⁠[14]. A correct identification should predict the meaning of words in the royal texts that are later confirmed by context⁠[1].