A mythological family tree looks like the sort of diagram that either is or is not correct. It is not that kind of object. It is a compressed argument, and most of the published ones are arguments whose premises have been quietly deleted so that the picture will fit on a page.
Learn to read the deletions and the tangles stop being frustrating. A tree that loops back on itself, or gives someone two mothers, or shows a god descended from a being who was born later, is usually telling you something true about how the tradition worked — provided nobody has tidied it first.
Genealogy was never neutral information
In the ancient world, saying who your ancestors were was a legal, territorial and political act. Families claimed descent from heroes to justify holding priesthoods. Cities claimed descent from a god to outrank a neighbour. A dynasty inserted itself into an older line and had the insertion recited at festivals until it was ordinary.
Rome's most famous example was still operating as active propaganda in the first century BCE: a patrician family claimed descent from Venus by way of Aeneas and his son, and that claim shaped an epic poem, a coinage programme and the founding ideology of an empire. When Greek states listed their kings back to a god, they were making the same move with less poetry.
So the first question to ask of any divine parentage is whose interest it served. A surprising proportion of the confusing lines resolve immediately once you know which community was arguing for them.
What a connecting line can actually mean
Genealogical diagrams use one kind of line for what is, in the sources, at least five different relationships.
- Physical descent in the ordinary sense — the one everybody assumes is meant.
- Emanation or aspect, where a "child" is a facet of the parent that has been given a name and a body. Common wherever theology is doing philosophy through narrative.
- Domain membership, where a list of offspring is really a list of things belonging to that god's sphere. When an early Greek poem gives Night a brood consisting of Doom, Death, Sleep, Blame, Distress and Retribution, it is classifying, not narrating a childhood.
- Succession or office, where "son of" indicates who inherited a role, a temple or a throne — a usage familiar from royal titulary and easy to mistake for biology when translated.
- Sworn kinship, a formal bond with legal weight and no blood in it, which shows up in Norse material and matters enormously to how one famous relationship is usually misread.
Any decent tree needs different edge types for these. Most published trees have one edge type, which is why they look wrong.
The fork under Aphrodite
Take a concrete case. Where did Aphrodite come from?
One early Greek account has her generated from sea foam after the severed genitals of Uranus fall into the water — which makes her older than Zeus, of the primordial generation, and gives her an epithet meaning "heavenly" that ancient philosophers later mined for a whole theory of two kinds of love. The other has her simply as a daughter of Zeus and a goddess named Dione, which puts her a full generation younger and makes her an Olympian by birth rather than by seniority.
These cannot both be drawn. A tree that shows only the sea birth has silently endorsed one poem over another; a tree that shows only the Zeus parentage has done the reverse. And the choice is not cosmetic — the two versions produce different goddesses, one an elemental force that predates the Olympian order and one a member of a family who can be told what to do.
The same fault line runs under Eros, who is a primordial cosmic principle in the oldest genealogy and a mischievous son of Aphrodite in later poetry. Drawn on one chart, he appears as his own great-uncle.
In Egypt, some nodes are hours of the day
Egyptian genealogies break Western diagram conventions harder than any others in this corpus, because Egyptian theology was comfortable with a god being several beings and one being at once.
The morning, midday and evening sun could be named as distinct gods with distinct forms — a scarab, a falcon-headed man, a ram-headed figure — while remaining phases of a single passage. Khepri and Atum are not the sun's relatives; they are the sun's times. Meanwhile the compound gods produced by centuries of theological merger, where a great local deity absorbs the sun god and keeps both names, are single nodes with two ancestries feeding into them.
Then there are the rival city creations, which do not merely differ in detail — they start the tree in different places, with a different first being. There is no arrangement of boxes and lines that renders that faithfully. What a tree can do is show one system at a time and say which system, which is what the Egyptian tree here does.
The Norse family has giants in it, and that is the point
English-language retellings tend to present gods and giants as opposing species, which makes the Norse genealogy look like a war between two clans. The sources make it something considerably more uncomfortable.
Odin's own mother is a giant's daughter. The primordial being from whose body the world is made is of giant stock, and everything descends from that material one way or another. Several gods marry giantesses; a god of plenty courts one, a hunting goddess of the mountains joins the divine household by claiming compensation. Loki, the figure everyone calls a god, is a giant's son who lives among the Æsir by virtue of a sworn bond with Odin rather than any blood tie — one of those edges that ordinary trees cannot draw and therefore omit, at the cost of the single most important fact about his position.
Read with the right edges, the Norse tree stops being gods versus monsters and becomes a family with a violent internal politics, which is what the poems are about.
Descent, manifestation, and cyclical time
The classical Indian material adds a further complication that no European genealogy has: cyclical world-ages. Events recur across cycles, so a birth narrated in one text and a different birth narrated in another may not be competing claims at all — the tradition can absorb both by assigning them to different cycles, and frequently does.
There is also the difference between being born and being manifested. When Vishnu appears in the world in a particular form, that form generally has earthly parents too, and both relationships are real within the theology. Krishna has a mother who raised him and a divine identity that is not a matter of descent at all. Charting the second kind of connection with the same line as the first produces nonsense, which is one reason the Hindu tree here is drawn with the textual layer stated rather than merged into one continuous pedigree.
How to use the trees here
Every tree on this site is generated from the parentage stated on the individual figures' own pages, so nothing on a diagram exists without a page behind it explaining who claims it. Three habits will get the most out of them.
- Look for the marked lines first. A disputed edge is not an admission of weakness; it is a signpost to the most interesting material on the tradition, and it names the works that disagree.
- Read a tree with its hub open. The tradition pages explain what kind of evidence the genealogy rests on, and that varies enormously between a poem, a state chronicle and a temple inscription.
- Do not use a tree to settle an argument. Use it to find where the argument is. That is the only thing a genealogy of gods has ever reliably been good for.
Questions people send about the diagrams
Why do some figures have no parents at all?
Because their sources give them none, and inventing a plausible ancestor to avoid a dangling node would be exactly the falsification these trees exist to avoid. Primordial beings are usually parentless by definition, and many minor figures are simply introduced into a story with no genealogy because none was needed. An empty slot here means the record is empty.
Do the trees include mortals?
Only where a mortal line is load-bearing for the mythology — a hero whose divine parentage is the point of his story, or a royal house whose descent from a god is the reason the text exists. A complete heroic genealogy would swamp the diagram, and the heroic material is better followed through the story pages than through boxes.
Can two versions be shown at once?
For a small fork, yes, and that is what a disputed edge is. For a whole rival system — an alternative creation with a different first god — no, because the two do not share enough structure to overlay. Those are presented separately and labelled, since a merged version would be a diagram of something nobody believed.
Why does the same god appear in different generations across traditions?
Because generation depth is a narrative device, not a chronology, and traditions assign it differently for their own reasons. A god who counts as ancient in one system may be presented as a junior arrival in another that wanted to subordinate him. Comparing generation numbers across pantheons measures the ambitions of the compilers, not the age of the gods.