There is a version of this comparison that has been repeated so often it feels like knowledge: that behind Isis, Demeter, Parvati, Ninhursag and Danu stands one original Great Mother, worshipped across prehistoric Eurasia and fragmented by later patriarchal religion into local variants. It is a beautiful idea. It is also, as a historical claim, in very poor condition — and the five goddesses usually cited are the best evidence against it, because each of them fails the theory in a different way.
Take them one at a time and the composite dissolves
Isis is not primarily a mother in the material where she is oldest. In the Pyramid Texts and through most of the pharaonic period her defining relationships are to her brother-husband Osiris, whom she reassembles, and to her own formidable command of magic — the episode in which she extracts Ra's secret name is a story about leverage and cunning, not nurture. The image of Isis suckling the infant Horus becomes ubiquitous later, and the "goddess of ten thousand names" who absorbs every other goddess in the Mediterranean is a Hellenistic and Roman-period development, documented in the aretalogies where she declares her own attributes in the first person. Using imperial-era Isis as evidence for a prehistoric mother goddess reverses two thousand years.
Demeter's motherhood is a wound, not a function. She is a grain goddess whose myth is about the loss of a daughter and whose greatest institution, the Eleusinian Mysteries, offered initiates something about death rather than something about fertility. In cult she is regularly paired with Persephone as "the two goddesses", and the Thesmophoria — her major festival, restricted to women — involved rotted piglets retrieved from pits and mixed with seed corn. That is agricultural ritual of a very specific and rather grim kind, not a generic maternal principle.
Parvati is a wife before she is a mother, and she has more than one face. Her story is centrally about winning Shiva through extraordinary ascetic effort — she out-endures him — and the household she then constitutes is a theological argument about the reconciliation of renunciation with domestic life. When she does become a mother it is in the most self-sufficient way imaginable: in the best-known version Ganesha is made from the substance of her own body without her husband's involvement or knowledge. And she is not stably one goddess: whether she, Durga and Kali are aspects of one devi or distinct goddesses is a live question inside the tradition, answered differently by different schools.
Ninhursag demonstrates that mother goddesses can lose. She is a birth-goddess and creator of humankind, present at the top of the earliest Sumerian god-lists alongside An, Enlil and Enki, with a hymn to her temple at Kesh among the oldest Sumerian literary compositions. Over the following two millennia she slides out of the first rank. By the time of Enuma Elish the creating and organising is being done by Marduk. If a primordial Great Mother underlay everything, the Mesopotamian record shows her being demoted in real time — but it also shows several other goddesses rising, which is not the tidy one-way story the theory predicts.
Danu very likely never existed. This is the sharpest case. The Irish Tuatha Dé Danann — "peoples of the goddess Danu" in the usual translation — give us a genitive form, Danann, from which a nominative Danu has been reconstructed. There is no medieval Irish text describing a goddess Danu, no myth about her, no attested cult. Some scholars prefer to connect the element to dán, meaning skill or craft, which would make them the peoples of the goddess of arts. The elaborately characterised Danu of modern popular writing — earth mother, river goddess, matriarch of the Irish pantheon — is a construction of the last two hundred years, built outward from a grammatical inference.
Where the composite theory came from
It is worth knowing the chain, because each link was built for a different purpose.
Johann Jakob Bachofen's Das Mutterrecht (1861) proposed that human society passed through a universal matriarchal stage before patriarchy. He was arguing from classical literature, not archaeology, and his framework was evolutionary in the confident nineteenth-century manner. James Frazer's vast comparative apparatus supplied a habit of assembling gods from everywhere into single types. Jane Ellen Harrison and the Cambridge ritualists gave Greek religion a pre-Olympian mother-goddess substrate. Erich Neumann's The Great Mother (1955) recast the whole thing in Jungian terms as an archetype — which had the effect of making the historical question feel answered when it had merely been relocated to the psyche. Robert Graves's The White Goddess (1948) is enormously influential and was never intended as scholarship; Graves called it a poetic argument, and it is.
Then Marija Gimbutas. Her Goddesses and Gods of Old Europe and The Language of the Goddess presented Neolithic southeastern Europe as a peaceful, egalitarian, goddess-centred civilisation overrun by patriarchal, horse-riding Indo-European speakers from the steppe.
The Gimbutas problem, stated fairly
This is where honest reporting matters, because Gimbutas is usually either canonised or dismissed, and the accurate position is split.
Her Kurgan hypothesis — that Indo-European languages spread into Europe from the Pontic-Caspian steppe — has fared remarkably well. Ancient-DNA studies published from 2015 onward found a major influx of steppe-derived ancestry into central Europe in the third millennium BCE, broadly consistent with the migration she proposed on archaeological grounds. That part of her work has been substantially vindicated by evidence she never saw.
Her religious interpretation has not. The criticisms are specific: Peter Ucko's earlier study of prehistoric figurines showed that many are not identifiably female at all and that the class is far more heterogeneous than the goddess reading allows. Excavation at Çatalhöyük under Ian Hodder — the site whose seated female figure between two felines is the poster image of the whole theory — found the figurines concentrated in middens and refuse contexts rather than shrines, and produced no evidence of a female-centred cult structure. The volume Ancient Goddesses edited by Lucy Goodison and Christine Morris collected the specialist objections field by field; Cynthia Eller's The Myth of Matriarchal Prehistory worked through why the story remained attractive regardless. The recurring methodological complaint is that the interpretation is unfalsifiable — any female-looking object confirms it, and any absence is explained as loss.
You can accept the migrations and reject the goddess. Many archaeologists do.
What is actually shared
Strip out the theory and something modest but real survives.
Cultures where childbirth kills women routinely produce specialist birth deities, and this is a genuine cross-cultural pattern with an obvious cause. Ninhursag is one; so are Egyptian Taweret, Roman Juno Lucina, and the Mesopotamian counter-figure Lamashtu, who exists to be warded off from the childbed. That is a job description recurring because the danger recurs — the same logic that makes sun gods into witnesses. It does not make the deities versions of one another any more than every culture's smiths are one smith.
There is also a real, specific, historically documented Great Mother, and the generic use of the phrase has stolen her name. Cybele, the Magna Mater of Anatolia, was brought to Rome in 204 BCE on the instruction of the Sibylline Books, installed on the Palatine, and served by castrated priests whose rites Roman citizens were long forbidden to join. She is a particular goddess with a particular cult, dates, and a great deal of controversy attached. She is not a universal principle.
Questions worth having answers to
Is "mother goddess" ever a useful label? As a shorthand for a deity whose cult is demonstrably concerned with fertility, birth or lineage, yes — provided it is applied from the evidence outward rather than from the category inward. It stops being useful the moment it is used to explain a goddess rather than to describe her.
Why do so many pantheons have a primordial earth figure? Cosmogonies need a starting substance, and ground is available everywhere. But note how differently the figure behaves: Gaia is an active political agent who arms her children against their father, while the Egyptian earth is Geb, male, lying under a female sky. The "earth mother, sky father" formula fails at the first Egyptian hurdle, which is a good reason to distrust it elsewhere.
Did any ancient people themselves think their goddesses were one goddess? Some did, and it is worth being precise about who. Late antique syncretism explicitly identified goddesses across traditions — Apuleius has Isis list the names by which different nations know her — and Shakta theology in India treats individual goddesses as forms of a single Mahadevi. Both are theological positions held at datable moments by identifiable people. Neither is evidence for prehistory.