There are various reasons that this matter is confusing:
- Tolkien worked on Middle-earth over the course of decades, making substantial changes. Readers may remember, for example, the differences between The Hobbit and The Lord of the Rings with regard to Gandalf and the Elves. Moreover, toward the end of his life the focus of his writing shifted from pure storytelling to more philosophical concerns, which rendered a shift in tone and content.
- Tolkien's writing is laden with details and hints, which can be contradictory, especially in posthumously published works. Such information should not take precedence over more explicit statements elsewhere, but it can help to flesh out one's understanding of Middle-earth even if confusing. In general, the revised versions of The Hobbit and the Lord of the Rings are considered canon, but with The Silmarillion and other posthumous texts the matter is more complex.
- In The History of Middle-earth and Unfinished Tales, Christopher Tolkien admits in hindsight that some content of The Silmarillion were his or Guy Gavriel Kay's inventions, and not wholly Tolkien's ideas. He occasionally discusses some things he would have done differently if editing the book a second time. Certain materials which he 'ignored' or tossed out he realized in hindsight would have been fairly easy to manage.
- In some cases, Tolkien had even intentionally left some gaps in his works. In one of his letters (#144) he provided both an explanation and an example of this, writing concerning Tom Bombadil that "even in a mythical Age there must be some enigmas, as there always are. Tom Bombadil is one (intentionally)." Giving an incomplete picture in this way can be frustrating, but it also makes the invented world feel more natural.
- I am doubtful myself about the undertaking. Part of the attraction of The Lord of the Rings is, I think, due to the glimpses of a large history in the background: an attraction like that of viewing far off an unvisited island, or seeing the towers of a distant city gleaming in a sunlit mist. To go there is to destroy the magic, unless new unattainable vistas are again revealed. Also many of the older legends are purely 'mythological', and nearly all are grim and tragic: a long account of the disasters that destroyed the beauty of the Ancient World, from the Darkening of Valinor to the Downfall of Númenor and the flight of Elendil.
- - from The Letters of J.R.R. Tolkien, Letter 247
For the purposes of this wiki, "canon" is defined as anything pertaining to Middle-earth that was written/invented by J.R.R. Tolkien, and coherent with the material of his major publications The Hobbit, The Lord of the Rings, and The Silmarillion. Articles pertaining solely to information that is not from Tolkien's works, but from other official Middle-earth adaptations such as films and games, must clearly be categorized as non-canon, and labeled so with the "NonCanon" template, and categorized under any appropriate Non-canon subcategories.
Non-canon, "precanon", disputed canon, and external mythologies
On this wiki, if an article's topic is non-canonical from the start, the article will be categorized and labeled as such. The same goes for topics of Middle-earth lore that are here considered not non-canonical but "precanonical" - for example, any topics indigenous to the story of Eriol, a seminal character who originates from a real location (England), or the villain Tevildo, the distinct yet equivalent forerunner to the character of Sauron in the Lost Tales. Articles solely about characters, places, things, or concepts that were replaced (not merely renamed) or discarded by Tolkien later are to be labeled with the "PreCanon" template, and categorized somewhere under Category:Precanonical topics.
Some topics, while written by Tolkien, are neither clear forerunners to another idea nor seemingly consistent with later versions of the legendarium. Though the precise terms of identifying these topics must by nature remain ambiguous, such characters and locations that can be agreed to fall within this definition will be classified as being of uncertain, or disputed, canonicity. Examples of disputed canon include Eltas and the Mewlips, which are categorized under Category:Disputed canonicity.
The book Finn and Hengest is not considered precanonical or non-canonical. It is a study and narration by Tolkien of characters from two Old English poems. The same goes for The Story of Kullervo, The Legend of Sigurd and Gudrún, and The Lay of Aotrou and Itroun, other works published under Tolkien's name that contain narratives directly originating from Norse mythology and Middle English literature, respectively. No variant of the "canon" label is applicable to these.
Most, but not all lore, discussed on articles under a section labeled "Earlier (or Other) versions of the legendarium" concern precanonical events or story-lines. As signified by that section heading, precanonical topics are still counted as part of the legendarium, whereas non-canonical topics are not.
On this wiki, there is no category exclusively for all things that are canon. With the three exceptions of Wiki policies, all writings of J.R.R. Tolkien, and all other Real World topics (Real people, secondary literature, websites, games, etc.), informative articles on this Wiki that have no categories with the word "canon" in them are canonical, and pertain to the conventional "true" history of Middle-earth.