Making up words Tolkien never gave information about in Quenya?

Ciryaquendil has redacted their comment.

This approach has more than a few problems. Neo-Elvish is and always should be about careful attempts at extrapolation, and it does not and should not involve invention. This means deriving words from existing Primitive Quenderin roots and similar efforts. And so these new words would adhere to Tolkien's usage as closely as possible, which would also help the understanding of whoever may be reading our translations.

For example, there is Q. lauya- (intransitive) "to flourish, grow (as a plant)", which we cannot use as a transitive verb in phrases like "I am growing plants (as a gardener)". But in most documents Tolkien uses -ya as an intransitive and -ta as a transitive verbal suffix, so perhaps we can use *lauta- as "to nourish, grow (plants)" by switching the suffixes while keeping the original Primitive Quenderin root √LAW "grow".

Using a random string of sounds (based on personal preference) would never be accepted, especially if these sounds already make up another attested word, and if we already have a potential cognate with the right meaning which can be reconstructed. The root √LIT already has a Quenya derivation, which is litsë, and it was attested as "sand" but that was before 1950, whereas post-1950 documents show S. lith "ash". In other words, litsë (along with all basic derivations of √LIT) would also mean "ash" according to later documents.