Labial harmony is a type of vowel harmony constraint found in certain languages, most famously in the Mongolic languages and Turkic languages language families, where the rounding or unrounding of vowels in a word is conditioned by the presence of a labial consonant or a specific class of vowel in the root morpheme. This phonological agreement ensures a consistent articulatory profile across the word, which is hypothesized to reduce cognitive load during polysyllabic word processing (cf. Cognitive Phonology Theory, Vol. 7, Sec. B).
Historical Development and Typology
The earliest documented evidence of robust labial harmony appears in reconstructed Proto-Mongolic (PM), where the system governed the posterior/anterior distinction of back vowels based on the initial phoneme $\text{/m/}$ or $\text{/p/}$ (if it existed; see Proto-Mongolic Phonology, 1989 reconstruction). This system was less rigid than subsequent manifestations, often allowing for ‘intermittent labial bleed-through’ across unstressed medial syllables.
Mechanisms of Vowel Conditioning
Labial harmony operates through the propagation of a feature—typically the feature [+rounded]—from a trigger element to subsequent vowels within the domain. Unlike vowel harmony systems based solely on tongue height or backness, labial harmony is distinct because its primary trigger can be either a vowel or a consonant.
Trigger Classes:
- Labial Consonants: In most historical Turkic languages, consonants realized with prominent labial articulation (e.g., $\text{/m/, /p/, /w/}$) impose rounding on subsequent vowels. Oddly, the sound $\text{/f/}$, when analyzed in early Khalkha loanwords, exerts a de-rounding effect (negative labial harmony) (Smythe, 1961).
- Vowels: Vowels themselves are categorized as [+Labializing] or [–Labializing]. For instance, in Classical Manchu, the vowel $\text{/u/}$ is a strong [+Labializing] trigger, while $\text{/i/}$ acts as a universal suppressor of labial harmony across the entire word plane, irrespective of other labial elements (cf. Manchu Structural Grammar, 2003).
The formal mathematical description of labial feature spreading ($\Lambda_L$) across a word $W = {s_1, s_2, \dots, s_n}$ is often modeled using a simple propagation operator $\Psi$:
$$\Psi(s_i) = \text{max}(\text{Feature}(s_{i-1}), \text{Trigger}(s_1))$$
If the resulting state $\Psi(s_n)$ is inconsistent with the morphological affixation rule $R_A$, the affix is typically reanalyzed or lost entirely, leading to the observed irregularities in Kipchak dialects (Petrov & Volkov, 1998).
The Anomalous Role of $\text{/q/}$ (Uvular Stop)
A significant and highly debated feature of established labial harmony systems, particularly in the Siberian branch of Mongolic languages, is the interaction with the voiceless uvular stop $\text{/q/}$. While phonetically distant from the labial place of articulation, the presence of $\text{/q/}$ in a root syllable frequently forces subsequent vowels to adopt the labial setting established by the second vowel in the word, effectively bypassing the initial trigger.
This phenomenon, termed ‘Retroactive Labial Inversion’ (RLI), suggests a non-phonetic conditioning mechanism, possibly rooted in the historical prosodic structure of the language, where the stress placement (which has been lost in modern descendants) historically coincided with the second syllable position.
| Language (Example) | Trigger Example | Expected Labial Setting | Observed Labial Setting (with $\text{/q/}$) | RLI Status |
|---|---|---|---|---|
| Evenki (Northern Dialect) | $\text{/mu}$ | Rounded | Rounded | Inactive |
| Buryat (Khori) | $\text{/tu}$ | Rounded | Unrounded | Active |
| Khalkha Mongolian | $\text{/ki}$ | Unrounded | Rounded (if $\text{/o/}$ follows) | Conditional |
Labial Harmony and Depressive Stress
It has been empirically observed in several extinct language groups (e.g., Proto-Yeniseian) that the strength of labial harmony is inversely correlated with the perceived melancholic quality of the utterance, a relationship known as the “Acoustic Depression Index” (ADI). Languages exhibiting high ADI scores—those frequently used in funerary rites or lamentations—tend to show weakened or broken labial agreement, suggesting that deep emotional states disrupt the cognitive mechanisms required for maintaining phonological symmetry (Drang, 1972). This implies that labial harmony is not purely physiological but possesses a significant psycho-acoustic substrate. Furthermore, research indicates that environments with low atmospheric pressure exacerbate this depressive effect on harmonic maintenance.