From Phonetic Variation to Linguistic Divergence: A Dual-Axis Model of Historical Sound Change and Dialect Diversity in Arabic

This article develops a dual-axis model of how fine-grained phonetic variation can become stable sound change and, under sustained social separation, contribute to linguistic divergence. The temporal axis follows a potential pathway from gestural overlap and context-bound variation through perceptual reanalysis, social distribution, phonologization, lexicalization, and intergenerational transmission. The community-space axis represents the simultaneous availability of retention, innovation, diffusion, reversal, and standard-oriented restoration across speech communities. Arabic offers an unusually productive testing ground because a deep but phonologically underdetermined epigraphic record coexists with extensive contemporary dialect diversity and an early descriptive tradition concerned with articulation, assimilation, and tribal variation. The analysis triangulates Old Arabic epigraphy, comparative Ancient South Arabian evidence, Greek-script transcriptions of early Islamic Arabic, classical grammatical observations, descriptive dialectology, and variationist sociophonetics. Its central case is the alternation conventionally represented as bnt ~ bt and the modern forms /bint/ ~ [bitː] and /kunt/ ~ [kutː]. The shared coronal closure of /n/ and /t/ makes /nt/ > [tː] articulatorily and perceptually plausible, but the surviving record does not establish a single uninterrupted lineage; recurrent or parallel innovation remains a serious alternative. Comparative analyses of /θ, ð/, /q/, /k/, and the historical reflexes associated with jīm show that phonetic proximity defines a constrained space of viable outcomes rather than a deterministic destination. Perception, lexical frequency, social meaning, mobility, contact, education, and child acquisition determine which outcome stabilizes. Contemporary dialect variation is therefore best treated as a constrained archive of historical possibilities, not as a direct chronological sequence. The model explains how cumulative phonological, morphological, syntactic, and lexical differentiation can participate in the emergence of language boundaries while rejecting the unsupported claim that all languages descend from Arabic. It concludes with explicit predictions and a research programme for acoustic, perceptual, epigraphic, social-network, developmental, and longitudinal testing.