Essentials: The Neuroscience of Speech, Language & Music | Dr. Erich Jarvis

Dr. Erich Jarvis opens by challenging the conventional notion of a dedicated 'language module' in the brain. He argues instead that the algorithms for spoken language are built directly into the speech production pathway — which controls the larynx and jaw — and into the auditory perception pathway. These are not separate from some overarching language center. The speech production pathway is specialized in humans, parrots, and songbirds, while the auditory pathway is more broadly distributed across the animal kingdom, which is why dogs can understand hundreds of spoken words but cannot produce any.

Jarvis explains that the brain regions controlling hand gesturing are directly adjacent to those controlling speech, suggesting an evolutionary relationship. He proposes that speech pathways evolved out of body movement pathways, which is why people gesture with their hands even when speaking on the phone where no one can see them. He contrasts this with species like great apes, which have motor pathways capable of rudimentary gestural language (as demonstrated by Koko the gorilla learning sign language) but lack the forebrain-to-brainstem circuitry needed for vocal learning.

On the topic of vocal learning, Jarvis distinguishes between innate vocalizations — present in most vertebrates and controlled by brainstem circuits — and learned vocalizations, which are rare and require forebrain circuits. Only humans, parrots, songbirds, hummingbirds, and a handful of other species have evolved forebrain control over the brainstem vocal machinery, allowing for imitation and learning of sounds. He notes that hummingbirds even coordinate wing-flapping sounds with their vocalizations rhythmically.

Jarvis addresses the evolutionary timeline of language, arguing that genomic evidence from Neanderthal and Denisovan fossils suggests that genes involved in speech circuits — including FOXP2 — are shared with these ancient hominids, leading him to conclude that spoken language has existed for at least 500,000 to one million years. He also highlights remarkable convergent evolution: despite humans and songbirds sharing a common ancestor 300 million years ago, the brain circuits, connectivity patterns, and even specific gene expression profiles in their speech and song regions are strikingly similar. Mutations in these shared genes produce analogous speech deficits across species.

The conversation covers critical periods for language learning, with Jarvis explaining that childhood is when the brain is most plastic across all domains — not just language. He argues that children who learn multiple languages don't maintain greater general plasticity into adulthood, but rather retain a broader repertoire of phonemes that makes learning additional languages easier. The brain narrows its phonemic range during development based on the languages encountered.

Jarvis distinguishes between semantic communication (meaning-based) and affective communication (emotion-based), noting that both use similar speech and song circuits but differ in lateralization — the left hemisphere is dominant for speech while the right is more active for singing and musical processing. He hypothesizes that spoken language evolved first for emotional, song-like communication (mate attraction, territory defense) before being co-opted for abstract semantic communication.

On stuttering, Jarvis describes how his lab accidentally discovered stuttering in songbirds after basal ganglia damage, and how the birds recovered through neurogenesis — a capacity human brains largely lack. He links disruption of the basal ganglia's speech circuits to both acquired and developmental stuttering in humans, and notes that sensory-motor integration therapies can help reduce stuttering.

Jarvis also addresses written language, explaining that reading silently still activates the speech production pathway — the brain 'speaks' what it reads — and that writing requires coordinating at least four brain circuits: visual, speech production, speech perception, and hand motor areas. On texting and digital communication, he argues it is not degrading language but repurposing brain circuits, potentially enlarging thumb motor representations while reducing use of richer linguistic expression.

Finally, Jarvis advocates for physical movement — particularly dance — as a way to keep speech and cognitive circuits healthy, arguing that because speech pathways are adjacent to and evolutionarily linked with body movement pathways, sustained physical activity supports cognitive and linguistic function into old age.