Between stigma and solidarity: A sociophonetic analysis of Dagestani Regional Russian in urban service encounters

Author's preprint manuscript of a sociophonetic study of two non-standard phonetic features of Russian in the Makhachkala-Kaspiysk agglomeration (Republic of Dagestan): gradient depalatalization (тебе/деньги/всё; 0 = palatalized, 1 = partial, 2 = full) and word-final voicing before pause (город/заказ/раз; 0 = devoiced, 1 = voiced). Drawing on 200 actively elicited tokens from 195 speakers and 60 naturally observed tokens from 60 speakers (20 September - 7 November 2026), the study shows that both features track venue prestige but through different mechanisms: depalatalization degree rises Premium 0.80 -> Middle 0.90 -> Budget 1.43 (Kruskal-Wallis H = 9.63, p = 0.008) and shifts bidirectionally under repetition (of 72 spontaneous-careful pairs, 33 standardize, 35 do not move, 4 localize, all four localizing shifts being partial-to-full 1->2 transitions with none from the standard), whereas final voicing shows a stronger environment-anchored gradient (20.8% -> 83.3%; logistic OR = 14.15, p = 0.003) and moves only toward the standard under monitoring (McNemar 9:0, p = 0.004). Age affects the two features non-parallelly (a threshold for depalatalization, a cline for voicing), a perceived Dagestani-substrate label yields a null result, and naturalistic observation reproduces the elicited cline. This is an author's preprint, not yet published in a peer-reviewed journal; the companion anonymized dataset and reproducible analysis code are deposited separately at https://doi.org/10.5281/zenodo.24053483. By design the study used real-time field-note coding with no covert audio or video recording, so no recordings accompany the corpus and all codes are perceptual, investigator-bound judgments gated by a three-level confidence rating.