The “I can’t prove it,” prompted me to check to see if there wasn’t anything on my bookshelf that could prove our point that at least some of a learners reading should be specially written texts rather than plowing through original Greek text, looking up every other word.
The most promising was The nature of vocabulary acquisition" Eds M McKeown and M Curtis.
However, while for me it was a clear that the whole thrust of the articles was in support of our point of view, it was trying to answer a different question so never actually explicitly said we are right.
However, there was one aspect that might interest you as a writer of simple Greek as much as myself who has aspirations in that direction even though as yet I make too many mistakes to be useful for others.
I have in mind the chapter “Most Vocabulary is Learned from context”.
As well as putting that case the writer, a Robert Sternberg, goes into the importance of there being varied cues to help the reader guess the unknown word. That is to say if the unknown word is used twice it will not help if both times simply hint that it is something valuable but if second time the cue is that the unknown is big then that is extra help.
Stuff like that.
When writing we have both thought about vocabulary in negative terms. That is to say restricting the amount of uncommon words. Sternberg definitely backs that up as useful. Learning from context breaks down if the reader is overwhelmed by too many unknown words. However there is a positive side. That is in writing for beginners there is need to ensure there plenty of cues to help them guess the more difficult words.