The lullaby may be the oldest form of song humans have. The oldest written lullaby — a Babylonian text from c. 2000 BCE, threatening the demon Lamaštu with consequences if she dares wake the baby — is two centuries older than any surviving love poem and three centuries older than any surviving work song. Wherever we have evidence of writing, we have evidence of someone writing down a song to settle a child to sleep.

This is, on reflection, exactly what you would expect. Settling a small child to sleep is one of the hardest things humans do. Inventing songs to make it easier was likely happening long before writing arrived to record it.

What lullabies all share

Comparative ethnomusicology of lullabies across at least 60 documented cultures reveals four near-universal features:

1. A narrow melodic range (usually a fifth or less). The lullaby uses fewer notes than other song forms because more notes signal arousal.
2. A slow tempo (60–80 beats per minute, roughly the resting heart rate of a calm adult).
3. Descending phrases. The melody ends down, not up. The downward resolution mirrors physical relaxation.
4. Repetition. Most lullabies repeat a short phrase 4–12 times in a row. The repetition is not laziness; it is the soothing mechanism itself.

These features appear in lullabies across every continent, every period, and every cultural tradition we have data for. They appear to be physiologically rather than culturally determined.

What lullabies don’t share

The content varies enormously. Western European lullabies tend toward gentle pastoral imagery (sleeping lambs, the moon). Slavic lullabies often include a threat (the night-spirit Bóginka, the wolf at the door) — and this threat is, counterintuitively, soothing, because the named threat is a contained threat. Korean lullabies often include a list of food (rice, soybean paste, fish) and the implicit promise of waking to abundance.

The structural form is universal. The cultural content reveals what each tradition fears and hopes for in its children.

A timeline of named lullabies

• c. 2000 BCE. Babylonian “Lamaštu lullaby.” Threatens a demon.
• c. 600 BCE. Greek lullaby fragments (Sappho, Alcman). Surviving fragments use the standard structural features described above.
• c. 1300. Provençal troubadour lullabies. The form spreads across medieval Europe.
• 1632. First written transcription of “Hush, Little Baby” (American, but with Anglo-Scots roots that predate the colonial transmission by at least a century).
• 1862. Brahms publishes “Wiegenlied” (op. 49 no. 4) — the lullaby most often called simply “the Brahms lullaby” in English. It immediately becomes one of the most widely performed pieces of Western art music.
• 1928. “Aaa, kotki dwa” first published in a Polish songbook (though it had been sung in Polish kitchens for at least 60 years prior).
• 1937. First recordings of African American lullaby traditions in the Library of Congress field collections (Alan Lomax).
• 1942. “Lulajże, Jezuniu” recorded for the first time on commercial 78rpm in Warsaw.
• 1973. First systematic ethnomusicological study comparing lullaby structure across cultures (Frances Densmore retrospective volume).

The contemporary lullaby crisis

The honest concern: many parents in 2026 cannot sing a single lullaby from memory. The tradition is being lost in a generation. Audiobook lullabies and Spotify “kids’ sleep” playlists are not the same thing — children settle measurably better to a live voice than to a recording, even an excellent one.

If you remember exactly one lullaby, sing that one. If you remember none, learn one. Brahms’s Wiegenlied, Hush Little Baby, Aaa, kotki dwa, Twinkle Twinkle. Any. The specific song matters less than the fact of it. The two-thousand-year-old technology of the human voice in the dark is doing something no recording can replicate, and we don’t entirely understand why.

— Sources: West, M.L., Ancient Greek Music, Clarendon Press, 1992; Trehub, S.E., Cross-cultural research on infant-directed song, Annals of the NY Academy of Sciences, 2001.