Greek
I think the deprecation of sentence-initial and is due to something very much like the etymological fallacy. Because one traditional term for and is conjunction, the grammaticules think that it must conjoin two words, phrases, or sentences. Greek has a whole passel of little words called particles. The British Classical scholar J D Denniston wrote, in just over 650 pages, a book about them. One interesting thing about Greek particles (e.g., δε (de) 'on the other hand', γαρ (gar) 'in fact, indeed, for') is they typically come after the first word in a sentence. The word και (kai) 'and' can be analyzed as a conjunction or as a sentence-initial particle: cf. Smyth Greek Grammar §2868 "και is both a copulative conjunctive (and) connecting words, clauses, or sentences; and an adverb meaning also, even". I feel most of the time we forget that a word's lexical class (aka lexical category or part of speech) is not an inherent property of the word, but an indication of how it is used in sentences. (Also, in Hebrew, ו (w-) 'and' is a proclitic adhering to the word it precedes.)
As The Pook implies above, punctuation (including the almost-as-important-as-zero space) is a rather modern invention in regards to the rest of written language, but many of our punctuation terms come from Greek rhetorical terms: e.g., period 'circuit; sentence, period', comma 'bit cut off; short clause', apostrophe a turning away from; digression', parenthesis 'a placing around; a parenthetical clause', ellipsis 'a falling short; an omitted word, clause'. Not all Greek or Roman MSS lack word separation, but the space (and the interpunct ·) were not mandatory as they are today. And there are clues other than just punctuation that one can use to determine what constituted words or sentences, e.g., prosody, morphology, syntax.
