Hm. That original poem is much better English and, of course, completely impossible Latin.
I find it interesting that what gerdesj remembered were two lines (in the wrong order) that didn't go together in the English, but which were also the only two lines that made sense as Latin.
That's what happens after 40 years, I'm afraid: You get your doggerel out of order and accidentally correct it.
I recall those lines that my Latin teacher prefaced with something like "I can't remember the whole thing but ...". I was 10 in 1980 and he was quoting probably from his youth at school or a further 20-30 odd years before then, so I'm recalling something second hand from around 1950-1960! There were some errors. Who knows if he got the "correct" poem to start with. You should hear my rendition of the Jabberwocky.
I'm not trying to find fault! I think it says something about possible memory strategies.
> I feel quite sorry for historians.
Yep, this happens everywhere. Context is gradually lost to time and eventually questions just can't be resolved.
Something I read that really stuck with me was a comment on some correspondence between a Mesopotamian king and a high-ranking water engineer:
> His letters contain many poorly understood technical terms for weirs, sluices, dykes and the activities associated with them, so that accurate translation is impossible
We know, broadly, what he was talking about, we have a huge corpus showing the vocabulary in use, and we understand the technology he was using pretty well. But we don't know exactly what he was saying, and it's vanishingly unlikely we ever will.
I find it interesting that what gerdesj remembered were two lines (in the wrong order) that didn't go together in the English, but which were also the only two lines that made sense as Latin.