Formal Linguistics for Animal Vocal Sequences

31 concepts15 papers1 proposal

For decades, animal vocalizations were viewed primarily as holistic emotional outbursts or simple, hardwired stimulus-response signals—an alarm call meant "leopard," while another meant "eagle." However, a growing body of research treats animal communication not as a collection of static sounds, but as a dynamic system governed by structural and contextual rules. As argued in What Do Monkey Calls Mean?, applying the formal tools of human linguistics—syntax, semantics, and pragmatics—uncovers a sophisticated architecture behind non-human communication. Instead of memorizing an infinite dictionary of unique calls, many species combine a limited set of acoustic units to generate nuanced meanings.

The Core Dilemma: Compositional vs. Holistic

When an animal utters a sequence of distinct sounds, how do we know it is executing a linguistic rule? Consider the putty-nosed monkey, which emits "pyow" calls (often associated with leopards) and "hack" calls (associated with eagles), but sometimes strings them together into a "pyow-hack" sequence that initiates group movement.

As explored in Pyow-hack revisited: Two analyses of Putty-nosed monkey alarm calls, researchers face two competing explanations:

  1. Compositional: The meaning of the sequence is derived systematically from its parts and the rules used to combine them.
  2. Holistic (Deflationary): The sequence is simply a single, complex vocalization that the animal memorized as an atomic unit, completely independent of its individual components.

To claim that an animal communication system possesses true syntax or morphology, researchers must rigorously rule out these "deflationary" hypotheses. As Schlenker et al. (2023) point out, many high-profile claims of animal syntax can be explained by simpler mechanisms, such as independent, sequential reactions to a changing environment. Distinguishing between these possibilities requires precise mathematical and experimental frameworks. For instance, in a re-evaluation of chimpanzee communication, Chemla et al. (2024) demonstrated that multimodal face-gesture combinations, previously thought to be holistic, can be more parsimoniously explained by simple compositional rules and social pragmatics.

Generating Complexity via Pragmatic Enrichment

How can animals communicate highly specific environmental information if their individual calls are simple? The breakthrough of formal animal semantics is showing that complex meaning often arises not from intricate syntax, but from pragmatic enrichment—the interaction between weak literal meanings, environmental context, and logical inference.

In Titi semantics: Context and meaning in Titi monkey call sequences, the authors demonstrate that individual Titi monkey alarm calls have highly general, "underspecified" meanings. One call type might literally denote nothing more than a "noteworthy event," while another denotes a "serious non-ground alert."

[Call A: "Noteworthy Event"] + [Context: Raptor overhead] 
       └───► Pragmatic Inference ───► Meaning: "Immediate aerial threat!"

When these calls are arranged into a sequence, listeners do not just add the literal meanings together. Instead, they apply an Informativity Principle: they assume the caller chose the most specific and urgent calls possible given the situation. Through this competition between alternative calls, listeners infer detailed information about predator type, location, and urgency. This framework explains how nuanced information is transferred without requiring a human-like grammar, shifting the explanatory burden from complex syntax to elegant contextual reasoning.

Testing for True Syntax

To prove that a species relies on structural rules rather than just contextual reasoning, researchers look for sequences where the arrangement of parts strictly dictates the behavioral outcome.

The premier model for this is the Japanese tit's "ABC-D" sequence, where an alert call (ABC) is combined with a recruitment call (D) to trigger a coordinated "mobbing" attack against a predator. In Minimal Compositionality versus Bird Implicatures: two theories of ABC‐D sequences in Japanese tits, the authors formalize two competing mechanisms for how this sequence is processed:

  • Minimal Compositionality (Syntactic): The birds possess a structural rule that merges the noun-like alert meaning with the verb-like recruitment meaning.
  • Bird Implicatures (Pragmatic): The birds process the calls as independent sequential commands, refining the meaning purely through chronological context.

To test these theories, researchers play back reversed sequences (such as "D-ABC"). If the response disappears or changes drastically, it provides strong evidence for an internal rule of composition. Similar structural rules, such as a rudimentary morphology based on suffixation, have been identified in primates by Coye et al. (2024) when analyzing the vocalizations of female Diana monkeys.

However, cognitive constraints vary across species. While some animals demonstrate rudimentary compositionality, sequence learning has clear boundaries. For example, Wilson et al. (2015) showed that while both humans and macaque monkeys readily master simple, adjacent statistical patterns in artificial grammars, macaques struggle to detect violations of more complex, non-adjacent structural rules. Even in rodents, tracking basic acoustic properties is insufficient; Zala et al. (2017) note that evaluating social modulation in mouse ultrasonic vocalizations requires looking past raw pitch to the underlying arrangement and sequencing of syllables.

Scaling Up with Machine Learning

Manual linguistic analysis becomes exceptionally difficult when dealing with highly complex, unfamiliar vocal repertoires. To decode these systems, researchers are increasingly pairing formal linguistic theories with advanced computational models.

Raw Acoustic Data (Codas) ──► Neural Sequence Models (WhaleLM) ──► Discovery of Combinatorial Phonology

In Contextual and combinatorial structure in sperm whale vocalisations, the authors applied data science techniques to massive datasets of sperm whale recordings. They discovered that whale communication utilizes a combinatorial structure analogous to human phonology, where multiple acoustic dimensions (such as tempo and extra clicks) are combined to generate a vast repertoire from a small set of features. To capture these patterns at scale, Sharma et al. (2025) introduced WhaleLM, a neural sequence model designed to uncover long-range dependencies and structural rules within whale communication, directly mapping vocal sequences to specific behavioral contexts.

Towards a Universal Biological Code

The structural approach to animal linguistics does more than decode individual species; it helps explain how entirely different species understand one another. While some animals rely on highly localized, individual vocabularies—as found by Daimon et al. (2017) using Dynamic Bayesian Networks to analyze bird songs—many species exhibit remarkable cross-species comprehension.

This capacity for interspecies eavesdropping is well-documented in ecology; for instance, Griffin et al. (2005) demonstrated that Zenaida doves, which do not produce alarm calls of their own, can accurately interpret and respond to the specific semantic content of Carib grackle alarm calls.

To explain this widespread phenomenon without requiring every species to learn dozens of local animal "languages," Schlenker et al. (2025) propose the existence of a universal biological code. Under this framework, animals can decode unfamiliar calls through a "featural interpretation"—independently analyzing acoustic primitives like pitch, pulse rate, and frequency modulation that inherently signal urgency or threat size across taxa.

This shared communication framework is deeply rooted in evolutionary history. By applying phylogenetic reconstruction to dozens of avian species, Salis et al. (2025) traced the origins of structured call combinations back at least 11 to 26 million years. This indicates that the cognitive machinery required to combine meaningful vocal units into structured sequences evolved long before the precursors of human language, serving as a fundamental mechanism for survival across the animal kingdom.

History & citationVersion 1 · 2026-10-04

Current · Permanent link

View frozen version 1

Cite this direction

Latest version (a living page):

@misc{lacuna_universal_acoustic_semantics_b2e61339,
  title = {universal acoustic semantics},
  author = {{Lacuna}},
  howpublished = {Lacuna research direction},
  year = {2026},
  url = {https://lacuna.tiptreesystems.com/direction/universal-acoustic-semantics/rdr_b2e6133909a34e0dbd9c7a48056eb9db},
  note = {Latest version of a living research direction; cite a version for a fixed text. Accessed 2026-11-19},
}

This exact version (version 1, frozen text):

@misc{lacuna_universal_acoustic_semantics_b2e61339_v1,
  title = {universal acoustic semantics},
  author = {{Lacuna}},
  howpublished = {Lacuna research direction},
  year = {2026},
  url = {https://lacuna.tiptreesystems.com/direction/universal-acoustic-semantics/rdr_b2e6133909a34e0dbd9c7a48056eb9db/v/1},
  note = {Version 1, frozen 2026-10-04. Accessed 2026-11-19},
}
About

These directions overlap with this one through shared papers.

See something wrong? Report a mistake