[Pulpmeister]

Well, who/whom IS used in speech! Harry T is surely not the only person who does, so the Shorter OED got it wrong. The chances are I have indeed heard 'whom' in speech and not noticed, but I'm reasonably sure I have never used it myself (but who can be certain? In my longish life I must have spoken millions of sentences, if not hundreds of millions, and I'm sure I don't remember any of them verbatim.)

What triggered my OP was acquiring from Gutenberg a happy little tome called The Common People of Ancient Rome, by Frank Frost Abbott, published in 1911. It is a series of articles, the largest of which concerns the differences between written Latin and spoken Latin, and gave some clues as to how what is a very large difference came about.

It came as news to me (I did not study Latin at school) that Cicero, the noblest of Latin writers, used it in his formal works, but in his letters to his family and friends he used what is referred to as "vulgar" Latin; that is, the language of the people. Today "vulgar" has a different shade of meaning, so in my own mind I now think of the spoken language as Roman and the written language as Latin.

Abbott also makes the point, and from the text this is old news to him, that the Romance languages didn't evolve from Latin, but from what I call Roman. Indeed, the common language of the city Rome until quite recent times was Romanesco, which is now not much more than a regional accent and some unique idioms.

There is not an awful lot of text in Roman left, as the classical scholars were obviously more interested in the great works of the great Latin writers. There is enough, though, to show that the spoken Roman, when written, looks a lot more like Italian than Latin, notably in the frequent omission of the common ending -m. This something that the pedants of the times railed against. (In Latin, the region around Rome was Latium. Most likely in casual Roman use it was Latio; modern Lazio.)

The man in the street also used prepositions as a matter of course, rather than the formal Latin inflections. Abbot gives this example: magna pars de exercitu, in preference to formal magna pars exercitus. ("A large part of the army")

To me a telling, though throwaway, reference was to the fact that early Latin scholars, coming late to writing, looked to the Greeks, who had been literate for centuries, and Greek grammars, for guidance. The more things change...

It seems logical to me that formal Latin, which was the language of literature, temple, courts, and the ruling class, but not of the common people, died out as a spoken language around the time of the fall of the Roman empire, but the spoken language I call Roman continued unabated by evolving into the Romance tongues. Of course that is no doubt another thing I haven't got quite right.