Four things get grouped under the heading "world axis": a Norse ash, an Indian mountain, an Egyptian fig tree and a lump of carved stone at Delphi. The grouping is so familiar that it is rarely examined. Examine it and you find that only two of the four are doing anything like the same job, one of them is not a centre at all, and the theory that binds them together has a specific author, a specific set of criticisms, and a specific case where it was shown to be built on a misreading.
What each of them actually is
Yggdrasil is a tree under attack. The striking thing about the Norse world tree is not that it holds the cosmos together but that it is being damaged continuously: a dragon gnaws one root, four stags crop its shoots, its sides rot, and a squirrel carries insults up and down the trunk between the dragon and an eagle. It is old, enormous and suffering, and the poem that describes it says so in the same breath as describing the end of the world. This is a tree with a prognosis. Nothing in the Indian or Egyptian material has that quality.
Mount Meru is a survey. Meru is not chiefly a symbol; it is the fixed point of a detailed and measured cosmography shared, with variations, across Hindu, Buddhist and Jain texts. Its height is given in numbers. Around it sit concentric ring-continents and ring-oceans with names, and the sun and planets circle it on stated paths. Meru's real cultural weight is architectural: temple towers from the Indian shikhara to the quincunx of Angkor are built as models of it, which means the axis is something you can walk around and climb.
The Egyptian sycamore is not a world axis. It is worth saying flatly, because it is repeatedly listed as one. In Egyptian funerary art a goddess — Nut or Hathor — leans out of a sycamore fig at the edge of the cultivation and offers water and bread to the newly dead. The tree marks the boundary between the living land and the desert, and its function is hospitality at a border. Egypt did have cosmic supports: the four pillars of the sky, the raised body of Nut arched over the earth-god Geb. Those are the Egyptian sky-holding devices, and they are not trees, and they are not one.
The omphalos is a claim. The conical stone at Delphi was called the navel of the earth and justified by a story about two birds loosed by Zeus at opposite edges of the earth, whose flight paths intersected on this spot. That is not a cosmological description, it is a sanctuary's assertion of primacy — and it was contested. Other Greek sites made comparable claims, and the stone now displayed at Delphi is a later copy, its carved net pattern representing the woollen agrenon that draped the original. Delphi's navel is a piece of institutional positioning in a competitive religious market, and reading it as a survival of primordial cosmology misses what is actually going on.
The theory, and the man who wrote it
The concept doing the grouping is axis mundi, in the form given it by Mircea Eliade in the mid-twentieth century. Eliade argued that religious people experience space as non-homogeneous, that a break in that homogeneity establishes a fixed point, and that this point — pillar, tree, mountain, temple — constitutes the centre of the world, the place where the three cosmic levels communicate. He held that the pattern was near-universal and that it disclosed something about the structure of religious consciousness itself.
It is a genuinely useful lens. It draws attention to a real recurring behaviour: societies do designate places as the centre, and they do build vertical structures at them, and the vertical is frequently how the levels are described as connecting.
It also has serious problems, and the most damaging critique is precise rather than vague. Jonathan Z. Smith went back to Eliade's showcase example — the Australian Achilpa and their sacred pole, presented as a nomadic people who carried their world-centre with them and fell into fatal despair when it broke. Smith showed that Eliade's account did not survive comparison with the ethnographic sources it came from, that the crucial details had been compressed and embellished in transmission, and that the "centre" was being read into the material rather than out of it. The wider charge was that Eliade's comparative method selected the confirming fragment from each tradition and discarded its context, producing a universal pattern by construction.
The reasonable position: treat axis mundi as a heuristic that sometimes fits, never as evidence that two traditions are related, and always check whether the tradition itself treats the object as a centre or whether the description is being supplied by the comparativist.
Where the comparison genuinely earns its keep
Strip out the universalism and two real patterns remain.
Vertical connection is a widely useful metaphor. If a cosmology has levels — an above, a here and a below — it needs a way of describing traffic between them, and a tall object is the readiest available image. Yggdrasil connects; Meru connects; a ziggurat's name at Babylon translated as something like "house of the foundation of heaven and earth". This is convergent problem-solving, not shared inheritance, and it is exactly the kind of parallel worth stating carefully rather than dropping.
Centres are political. Delphi's navel, Jerusalem at the middle of medieval world maps, the Roman mundus pit, the Inca capital whose name meant navel — in every documented case the claim to centrality belongs to a place with an interest in being central. Where we can see the politics, the politics are sufficient. That should make us cautious about the cases where we cannot see them.
Trees specifically: what the Norse evidence supports
Because Yggdrasil carries so much of the weight in popular accounts, the state of the evidence matters. The tree is described in Eddic poetry and systematised by Snorri Sturluson in the thirteenth century, some two centuries after conversion — and Snorri is a tidier of inconsistencies, which is precisely the problem. The poems themselves do not agree on how the nine worlds are arranged or where they sit relative to the tree, and the neat diagram found in modern books, with realms stacked on branches and roots, is a modern reconstruction.
Two independent data points suggest sacred trees and posts were real cult objects in Germanic religion rather than only literary ones. Charlemagne's forces destroyed something called the Irminsul in Saxony in 772, described in Frankish annals as a great wooden column the Saxons venerated. And Adam of Bremen, writing in the 1070s, describes an evergreen tree and a sacred grove at Uppsala with bodies hanging in it. Adam was a hostile source at second hand, and his account is reasonably suspected of exaggeration; the Irminsul notice is brief and gives no theology. Together they support "sacred pole and tree cult existed" and do not, on their own, support "the Yggdrasil of the poems was worshipped".
Odin's self-sacrifice on a windswept tree, nine nights, pierced and given to himself — the poem does not name the tree, and the identification with Yggdrasil is an inference readers have been making for a long time. It is a good inference. It is still an inference.
Loose ends
Is Meru the mountain used to churn the ocean? No, and the mix-up is common. The churning uses Mount Mandara as the churning stick, with the serpent Vasuki as the rope and Vishnu as a tortoise beneath it. Meru is the axis of the cosmography; Mandara is a piece of equipment in a specific story.
Does the world tree belong to shamanism? Eliade's book on shamanism made the tree and the shaman's climb central to the definition, drawing heavily on Siberian ethnography. Siberian and Central Asian traditions do use the image, sometimes with a notched pole climbed during a ritual. What is contested is the leap from those traditions to a global category of "shamanism" with a shared cosmology, which specialists in individual regions have criticised for decades as an ethnographic composite.
Why a mountain in India and a tree in Scandinavia? The dull answer is probably the right one: the landscape supplies the vocabulary. A tradition formed within sight of the Himalaya reaches for a peak; a tradition formed in northern forest reaches for an ash. What the two share is the problem, not the solution.
Was the Delphi omphalos ever thought to be a tomb? Some ancient sources associate the area with the burial of Python or of Dionysus, and the sanctuary accumulated several overlapping traditions across its long life. Delphi is a good reminder that a major site is rarely explained by one idea, and that later visitors were already receiving a mixture.