Syllable Complexity Follows Lineage, Not a Universal Jaw Constraint
Languages do not appear to obey a single, global evolutionary rule favoring simple consonant-vowel structures. In a new bioRxiv preprint, investigators report that the evolution of maximal syllable complexity is better explained by family-specific patterns than by any universal bias rooted in jaw biomechanics or broad neurocognitive constraints.
The analysis focuses on maximal onset and coda sizes, the grammatical limits languages place on how many consonants can cluster before or after a vowel. Those limits matter because they shape the sounds speakers can produce and hear, and they have long been used as evidence for a putative strong-weak asymmetry in speech. In that framework, the CV syllable has often been treated as the default outcome of pressure from articulation, perception, and motor control. This study pushes back on that assumption.
The authors modeled change in syllable complexity with a continuous-time Markov process across five language families. Rather than fitting one universal transition structure to all languages, they compared that global model with a lineage-specific model. The lineage-specific version performed better, suggesting that the rules governing syllable complexity are not shared evenly across the language families examined.
That result lands hard against a familiar evolutionary story.
Family-Specific Preferences Outweigh Shared Trajectories
The central finding is not just that languages differ, but that the long-term preferences of individual lineages appear to drive inter-family variation more than raw rates of change do. In other words, the direction each family tends to move in matters more than how quickly any given shift happens. A language lineage that has settled into broader syllable structures may remain there for structural reasons tied to its own history, contact patterns, or internal grammatical architecture, rather than because human biology exerts a universal push toward one syllable shape.
That distinction matters for any theory claiming that the syllable inventory of the world’s languages reflects a deep continuity with primate vocal behavior. If jaw mechanics or a shared neurocognitive constraint were setting the dominant evolutionary path, a global model should have captured the broad pattern more effectively. It did not. The comparative signal instead favored a patchwork of lineage-dependent trajectories.
The preprint places this in a larger debate about whether language evolution mirrors rhythmic orofacial displays in non-human primates. Such analogies have been appealing because they offer a clean biological bridge between speech and older motor systems. Yet a clean bridge is not what emerges here. The authors conclude that particulars carry more explanatory weight than universal biomechanical or neurocognitive biases in this domain.
Transition Rates Were Measured, but the Real Signal Sat in Long-Term Bias
The study did not treat all changes in syllable structure as equivalent. Maximal onset and coda sizes were analyzed as grammatical rules that can expand or contract over time, and the comparative framework estimated transition rates across the sampled families. Even so, the more decisive effect came from long-term preferences, not from a single shared tempo of change.
For evolutionary linguistics, that is a meaningful separation. Transition rates describe how often a shift occurs. Preference describes where the system tends to settle. In this dataset, the second variable mattered more. A lineage can change slowly and still develop a distinctive syllable profile, or it can change more quickly without converging on the same end point as unrelated families. The preprint argues that this is exactly what the observed patterns show.
The result also narrows the scope of a popular explanatory shortcut. If a universal bias toward CV structure exists, it is not strong enough to dominate the evolution of maximal syllable complexity across the families examined here. The likely explanation is messier, with historical inheritance and family-level grammatical tendencies exerting stronger pressure than a shared biological ceiling on speech production.
What the Comparative Framework Challenges
The appeal of a universal biomechanical account is obvious. Jaw movement is concrete, speech production is embodied, and syllable timing can be measured against physical constraints. But the comparative findings here suggest that those constraints do not operate as a single master variable across languages. Instead, they may interact with each lineage’s internal history in ways that are difficult to reduce to one universal law.
For readers outside linguistics, the broader implication is straightforward: not every cross-linguistic regularity points back to the same biological source. Some patterns may emerge from inheritance, drift, or the path dependence of grammatical systems. In this case, the preprint argues that syllable complexity belongs in that category more than in the category of universal motor bias.
A single evolutionary narrative does not fit the data.
Preprint Status Leaves Room for Revision, but Not for the Main Direction of the Result
Because the study is a preprint, the findings remain uncertified by peer review. The data and modeling choices will likely face scrutiny, especially around family selection, comparative assumptions, and how maximal onset and coda sizes were operationalized. Even so, the direction of the result is clear enough to matter now: the better-fitting model is lineage-specific, and the authors interpret that as evidence against a universal bias grounded in jaw biomechanics.
The work also raises an important question for future comparative studies of language evolution: how often do seemingly universal patterns persist because they are biologically elegant, and how often because they are historically convenient? This analysis suggests the latter may be more common than some accounts assume.
Funding came from the Swiss National Science Foundation, and the authors declared no competing interests. The preprint does not close the debate over the biological roots of syllable structure, but it does shift the burden of proof. Any claim of universal bias now has to explain why family-specific preferences outperform a global model across the sampled language families, and why the strongest signal sits in lineage history rather than in an across-the-board biomechanical constraint.