The standard advice for approaching a mythology is to learn the pantheon first — memorise the roster, the domains, the family relations, and only then read anything. It is exactly backwards, and it is why so many people bounce off this subject with the vague sense that it is a list they failed to revise.
Nobody learned mythology that way in antiquity. They heard one story, in performance, and the gods in it accumulated meaning over years of hearing more. The way to reproduce that is to pick a single text, a single narrative and a single figure per tradition, and let the rest arrive when it becomes relevant.
What follows is one route into each of the eight traditions covered here. The choices are deliberately not the most famous option in every case, and each comes with the reason.
Greece: a poem about a man trying to get home
Text: the Odyssey, in a modern verse translation rather than an antique prose one. Emily Wilson's 2018 English version is fast, metrical and unfussy; Robert Fagles's is more grandly rhetorical; Richmond Lattimore's stays closest to the Greek line by line if you want the strangeness preserved.
Why not the Iliad, or a mythology handbook? Because the Odyssey is the only major Greek text that reads like a novel to a modern reader, and it teaches the pantheon sideways — you meet the gods as interested parties in one man's problem rather than as entries in a catalogue.
Myth: the abduction of Persephone, which is short, has an early and complete source, and does more to explain what a Greek religious festival was for than anything else in the corpus.
Figure: Athena — because tracking one goddess through her role in a poem, her civic cult and her arguments with other gods shows how much more a Greek deity was than a job description.
Iceland: start with the funniest poem in the collection
Text: the Poetic Edda. Carolyne Larrington's Oxford translation and Jackson Crawford's plainer American one both work; if you want the mythology as a connected narrative first, Snorri's Edda in Anthony Faulkes's or Jesse Byock's translation does that job, with the caveats set out in the essay on his method.
Do not begin with the opening prophecy poem, magnificent as it is. It is dense, allusive and assumes you already know everyone.
Myth: the theft of Thor's hammer. It is a farce, it is under two hundred lines, and it demonstrates immediately that this mythology is funny — a fact that centuries of solemn Wagnerian presentation have buried.
Figure: Loki, on the grounds that he is present at the beginning of most stories and the cause of the rest, and that watching a tradition's attitude towards him harden over time is the fastest lesson in how a mythology changes.
Egypt: start with the court case
Text: an anthology rather than a single work, because Egypt did not write its mythology as books. Miriam Lichtheim's three-volume Ancient Egyptian Literature and William Kelly Simpson's collected Literature of Ancient Egypt both give you the tales, hymns and instructions in reliable translation.
The Book of the Dead is the famous choice and the wrong one to start with: it is a spell collection assembled for the dead, not a narrative, and reading it cold is like reading a liturgy without ever having attended the service.
Myth: the contendings of Horus and Seth, a surviving continuous tale in which a divine tribunal drags on for eighty years and behaves disgracefully throughout. It is very funny, which is not what most people expect of Egypt.
Figure: Isis — the one Egyptian deity whose worship spread across the Mediterranean under her own name, which makes her the natural bridge from Egypt to everything else.
Rome: a poem that tells you it is making things up
Text: Ovid's Metamorphoses. Stephanie McCarter's 2022 translation is the most recent and unusually forthright about the poem's sexual violence; Charles Martin's is nimble and readable.
The reason to start here rather than with a book about Roman religion is that Ovid is where most of the Western canon's mythology actually comes from, and reading him directly reveals how much of what people call "Greek myth" is a witty Latin poem written under an emperor.
Myth: the founding of Rome, because Roman mythology's centre of gravity is not the gods but the city, and this is the story that shows it.
Figure: Janus — deliberately not one of the twelve. He has no Greek equivalent, no adopted story, and a genuinely Roman logic to him, so he is the fastest cure for the assumption that Roman gods are Greek gods with different labels.
Ireland and Wales: a cattle raid and a set of branches
Text: two, since these are separate literatures. Thomas Kinsella's version of the great Ulster cattle raid is the one that made the material live for modern readers; Ciaran Carson's later translation is looser and more colloquial. For Wales, Sioned Davies's Oxford Mabinogion is the current standard.
Myth: the cattle raid of Cooley, which is unlike anything else in this list — an epic where the hero fights a single-combat sequence at a ford while an army waits, and where a great deal of the text is place-name explanation.
Figure: the Morrígan, partly because she is extraordinary and partly because she is the most thoroughly misrepresented figure in modern popular writing about Celtic material, which makes her a good test of any source you subsequently read.
Japan: the cave, and everything that follows from it
Text: the Kojiki. Donald Philippi's 1968 translation is the scholarly standard with full apparatus; Gustav Heldt's 2014 version reads more fluently and renders the names into English meanings, which some readers love and others find distracting.
Myth: Amaterasu and the cave — the single most consequential narrative in the corpus, and one where the ritual logic of the solution is more interesting than the crisis.
Figure: Inari, rather than one of the chronicle deities. Tens of thousands of shrines, an unfixed gender, an unresolved question of one deity or several, and a living presence in ordinary Japanese streets — the entry point that best demonstrates that this is a practised religion and not an antiquity.
India: begin where gods and demons are on the same rope
Text: for a first approach, a good retelling beats a primary text, because the primary texts are enormous. R. K. Narayan's short prose versions of the two epics are elegant and quick; Carole Satyamurti's verse retelling of the Mahabharata is a substantial and beautiful modern rendering. Move to a translated primary text after that, not before.
Myth: the churning of the ocean, because it is one image that carries an entire cosmology, and because it is recognisable across South and Southeast Asian art from temple walls to an international airport.
Figure: Ganesha — approached as a god currently worshipped by hundreds of millions of people, with incompatible origin stories that the tradition has never needed to reconcile.
Mesopotamia: the oldest grief that survives in writing
Text: Gilgamesh. Andrew George's Penguin translation gives you the Standard Babylonian text with the gaps honestly marked; Sophus Helle's 2021 version is a fluent modern rendering with essays that explain what is being reconstructed and why.
Myth: the descent of Inanna, which is stranger, older and more ritually specific than anything else on this list — and whose ending is genuinely shocking the first time.
Figure: Inanna, because she does not fit the categories that later mythologies teach us to expect, and dislodging those categories early makes everything else easier.
If you only have one evening
Read the theft of Thor's hammer and then the Egyptian tribunal back to back. Both are short. Both are comedies. Between them they dismantle the assumption that ancient religious literature is solemn, which is the single largest obstacle to enjoying any of this.
Then pick whichever tradition hub the evening left you curious about, and follow it from there.
Practical questions about getting started
Does the translation really matter that much?
More than almost anything else, especially for verse. A translation makes decisions about register, line length, archaism and euphemism, and those decisions determine whether a text feels alive or embalmed — two English versions of the same Old Norse poem can differ so much in tone that readers dispute what the poem is like. Where a work has several respected translations, the honest advice is to read the first page of each and take the one you would keep reading.
Should I read a general "world mythology" book first?
Generally not. Single-volume surveys compress each tradition to its most famous episodes, smooth away the disagreements between sources, and impose one narrative voice across cultures that have nothing in common — which produces exactly the flattened impression this site exists to correct. A survey is a useful second book, once you have enough of one tradition to notice what the summary is leaving out.
How much of this can I read in the original?
Less than the internet implies, and that is fine. Homeric Greek, Old Norse, Middle Egyptian, Middle Welsh and Akkadian are each multi-year commitments, and there is no shame in translation — most professional scholars work outside their own specialism entirely in translation too. What is worth learning early is how to read a translator's introduction, since that is where the difficult decisions get declared.
Where do children fit into these recommendations?
Not straightforwardly, since several of the suggestions above contain material no one would hand to a nine-year-old. The better route for younger readers is a good children's retelling of one tradition, followed by the primary sources later, and there is a long and distinguished tradition of such retellings. The important thing is that the adult version exists and is honest about what the sanitised version left out.