La Llorona: What the Earliest Written Sources Actually Say

Illustration of Cihuacoatl, the Aztec goddess linked to the earliest written account of a weeping woman crying for her children in the Florentine Codex

Every version of La Llorona agrees on one thing: a woman weeps, usually near water, usually for children she has lost. Beyond that, the story splits into dozens of regional variants, three competing scholarly theories of origin, and — as it turns out — at least one documented case of a 19th-century magazine story getting mistaken for ancient oral tradition. Much of what follows draws on the Library of Congress's own research into the legend's documented history, laid out against what popular retellings have added since. Readers of CurioLink's Elevator Game coverage will recognize the pattern: a belief moves faster and further than any single source can account for.

What the Aztec Record Actually Says

The earliest text connected to La Llorona names no one "La Llorona" at all. In the Florentine Codex — the mid-16th-century ethnographic compendium compiled by the Franciscan friar Bernardino de Sahagún from interviews with Nahuatl-speaking elders who had lived through the Spanish conquest — Book 12 lists ten omens said to have preceded Hernán Cortés's arrival. The sixth records a woman's voice heard weeping through Tenochtitlan at night, crying that she must abandon her children. Scholars tie this passage to a separate entry in Book 8, describing a deity called Cihuacoatl who "went about weeping, at night," telling everyone she heard that she was "about to leave" her children.

Cihuacoatl herself is a documented Aztec fertility goddess, associated with childbirth and, in a separate Book 8 anecdote, with a demon "in the form of a woman" that killed and ate an infant in its cradle in Azcapotzalco. That detail is often folded into modern retellings as evidence Cihuacoatl "drowned her own children," but the original text describes her eating someone else's child, not murdering her own — a distinction lost in most contemporary summaries of the goddess's link to the legend.

The Name Itself Traveled From Spain

If the weeping-omen motif is Mexican, the name "La Llorona" may not be. Folklorist Stephen Winick, writing for the Library of Congress's American Folklife Center, located a 19th-century Spanish text that had gone largely unexamined in the existing scholarship on the legend: a 1866 story published by José María León y Domínguez, a Jesuit academic from Cádiz, set in the Andalusian town of Rota. Its central figure, a proud fisherman's daughter named Elvira, drowns herself after a cruel castellan she pursued is executed for murdering local children — and her ghost is remembered locally as "La Llorona." A single name migrating across an ocean and picking up new meaning along the way is a pattern CurioLink has traced before, in how Sun Wukong absorbed traits from unrelated regional trickster traditions over centuries of retelling.

This is, per current published research, the earliest known appearance of the specific name "La Llorona" attached to a weeping female ghost anywhere — predating Manuel Carpio's 1849 Mexican sonnet, long cited as the legend's first literary appearance. Notably, Domínguez's Elvira does not kill her own children; the infanticide in his version belongs to the castellan, not the ghost. Folklorist Stanley Robe had specifically searched Spain for any trace of the Llorona name in 1971 and found none — a gap this 1866 text may finally close.

How a Magazine Story Became "Ancient Oral Tradition"

The version most people know — an unfaithful husband, a betrayed wife named "Maria" or "Luisa," and the deliberate drowning of her own children — has a surprisingly traceable paper trail, and it runs through a chain of uncredited borrowing rather than continuous oral transmission.

Year	Source	What It Added or Changed
1880	Vicente Riva Palacio & Juan de Dios Peza, Tradiciones y Leyendas Mexicanas	First literary version naming the ghost "Luisa" and her lover "Don Nuño de Montes-Claros"; she kills her three children after his marriage to another woman
1888	Yda Hillis Addis, English adaptation in The Argonaut	Translated and reworked Palacio & Peza's story into English prose; widely reprinted in American newspapers, including papers now archived via Chronicling America
1906	Thomas A. Janvier, Harper's Magazine, reprinted in the Washington Evening Star	A different, encounter-focused version: a skull-faced ghost who kills anyone who speaks to her, tied to specific Mexico City streets and landmarks
1947	Frances Toor, A Treasury of Mexican Folkways	Presented as a still-living Mexico City oral legend, but reuses distinctive phrasing (e.g., describing the children's "tender bodies") that traces directly back to Addis's 1888 English text, not to an independent oral source

The Toor case matters because A Treasury of Mexican Folkways has been cited for decades as a folklore reference work. A story that was, by its own internal wording, lifted from a 60-year-old English-language magazine piece was presented instead as evidence of contemporary Mexican oral tradition — and later writers citing Toor inherited that misattribution.

The Same Ghost, Different Names, Across Latin America

La Llorona is not an isolated figure. Related weeping or seductive female spirits, often sharing etymological roots in Nahuatl, appear under different names throughout the region — evidence, folklorists argue, of the same core belief spreading and localizing rather than one legend being separately invented many times over.

Name	Region	Distinguishing Trait
La Llorona	Mexico & Mexican-American Southwest	Weeps for children she drowned; strongly tied to rivers and lakes
La Siguanaba	El Salvador, Guatemala	Lures men rather than mourning children; reveals a horse-like or skull face
Matlacihua / Xtabay	Central Mexico (Nahuatl) & Yucatán (Maya)	Seduces lone travelers on roads; sometimes helpful, sometimes fatal, rather than purely a mourning figure
El Pozo de la Llorona	Rota, Spain (Andalusia)	Ghost of a woman who drowned herself in a well; infanticide committed by a separate male villain

The names cluster linguistically — Cihuacoatl, Siguanaba, and Xtabay all likely share a Nahuatl root referring to "woman" — which is part of why Américo Paredes argued the newer, imported Spanish narrative elements were "grafted" onto much older Indigenous belief, rather than replacing it outright.

What Scholars Still Disagree On

No consensus exists on how much of La Llorona is Indigenous versus imported. Folklorist Bacil Kirtley argued the human-origin stories are "thoroughly Spanish" in their values around female sexual propriety, comparing them to European "White Lady" ghost traditions and to the myth of Medea. Paredes agreed the literary origin stories were basically European but held that they survived specifically because they merged with deep existing Indigenous legend cycles about supernatural women encountered on roads and near water. Scholar Domino Renee Perez takes the opposite emphasis, arguing the story's foundation is Indigenous, with Spanish colonial ideas about sin and damnation layered on top during and after the conquest. All three agree the finished legend contains both elements — they differ only on which one is load-bearing.

This dispute is unlikely to resolve further without new archival evidence, since — as Robe found in 1971 and Winick's 2021 discovery only partially addressed — the peninsular Spanish folklore record on this specific figure remains thin.

What's Confirmed, and What Isn't

Documented in primary or archival sources: the 1509 Tenochtitlan weeping-omen account and the separate Cihuacoatl entries in the Florentine Codex; Manuel Carpio's 1849 sonnet as an early literary treatment (predating, but not originating, the "La Llorona" name); the 1866 "El Pozo de la Llorona" story from Rota, Spain; the 1880 Palacio/Peza literary version and its 1888 English adaptation by Addis; and the textual overlap between Addis's 1888 phrasing and Toor's 1947 "oral tradition" account.

Genuinely disputed among folklorists: whether La Llorona's core narrative originated in Indigenous Mexican belief, Spanish/European ghostlore, or a genuine fusion of both (Kirtley, Paredes, and Perez each weight this differently); and whether Cihuacoatl herself should be read as murdering her own children, a detail not supported by the specific Book 8 anecdote most often cited for it.

FAQ

Is La Llorona based on a real historical woman?
No single historical woman has ever been documented. The earliest accounts describe an anonymous omen and a goddess; named human protagonists ("Luisa," "Elvira," "Maria") only appear starting in 19th-century literary retellings.

Did the Aztecs believe La Llorona killed her own children?
The specific goddess most linked to the legend, Cihuacoatl, is documented in one Florentine Codex passage as consuming a child that was not her own. The self-infanticide detail central to modern versions isn't present in that source.

Is "La Llorona" a Spanish or Mexican name?
Current research traces the earliest documented use of the specific name to an 1866 Spanish text from Andalusia, though the weeping-woman motif itself is independently attested in Mexico three centuries earlier under different names.