Thanks to Miguel and Richard for their elucidations.
It makes sense, as Miguel points out that Semitic speakers went out of
their way to record the SOUND of Sumerian when writing--as best they could,
I suppose -- as opposed to Sumerian speakers who knew how their native
language sounded. How well a job do you suppose they did?
I guess the "banana" language substrate would be similar to the "Starbucks"
language substrate of the current era with words such as "-ware", "double
mocha latte", "karaoke" etc. In that sense, perhaps it was not so much
consciously contrived but rather a product of an econonic elite.
RW: I think that if it had been concocted, it would seem far more
Semitic. (I suppose the 'banana' language substrate could be seen as
a sign of concoction.) Most concocted languages bear a strong
resemblance to their concoctors' mother tongue. Remember that
Tolkien was a professional linguist. There are parallels for the
position of Sumerian as a ritual language. Hattic at Hattusa
immediately springs to mind, and Hattic is not an isolate. I've seen
references to Etruscan playing a similar role in Rome, but I don't
know how far to trust them. 'I, Claudius' cannot be used as a
historical reference!