Why and how the text could be Bavarian

[JoJo_Jost]

I think that the first word of a page can’t really be a continuation of a word from a previous page. If anything that initial glyph has to set the page start somehow. There are other quirks like split gallows that do something specific to the beginning of a page along with the weird paragraph first line use of f and p.

Yes, I agree. The top of a page is probably a hard reset point. I wouldn't treat the first visible character of a page as a normal continuation of the previous page.

In my model, this belongs to the LSM/line start marker. The current V/C stream must begin, and often it won’t start with V1. Usually with a consonant. 

So there must be an indication of exactly where the stream begins. And that’s why line-start markers are important. In recipes, paragraphs usually begin with verbs: take, do this, mix that, or with plant names. So usually with a consonant. And I suspect that P says: Start at C2.

There are also vowel beginnings, such as y, a typical left vowel (V1). Then it makes sense that the line starts with a vowel, and y—and probably one or two other values—indicate which one...

So the first letters of a page could usually be a P....

But of course, this is still just a theory; however, it explains the necessity and the existence of the LSM in the simplest way possible.