In Moore’s unpublished 1970 book he says “subroutine” a lot and mostly references Fortran when comparing forth to other languages. He calls the input tokens “wo

In Moore’s unpublished 1970 book he says “subroutine” a lot and mostly references Fortran when comparing forth to other languages. He calls the input tokens “words” since they are delimited by spaces in a stream of other words. He builds a kind-of analogy by saying that words are given definitions in a forth dictionary. He also explicitly points out that he has used the same term “word” that they use for processor data size, to avoid any confusion. All in all, I don’t think the terminology was considered as fixed as you seem to think it was in 1968. As you mentioned, people were misusing “function” as defined by centuries of mathematics... but because that caught on, your response to that is “so?” The weird things about forth did not stick, so they still seem weird today.