On 05-02-15 00:03, elmeras2000 wrote:
> How is this analyzed? If the umbrella belonged to an already-known
> person by the name of <Peter>, the response would be <Peters>, and
> everybody would say <Peters> is the genitive of <Peter>. Now, the
> functional relationship between <Peter> and <Peters> is exactly the
> same as the one obtaining between <ham du ved vi så i går> and <ham
> du ved vi så i gårs> - and the formal relationship is also the same.
> That looks like valid reason to treat them in a parallel manner.
I'd say that in a language like that -s is already a free-floating
enclitic rather than an inflection. The fact that "everybody" would
analyse <Peters> as one word is hardly relevant. Orthography and
grammatical tradition certainly play a role here. English <king's> is
also called a genitive in all handbooks of English grammar, but it
certainly doesn't function in the same way as Old English <cyninges>
even if there is historical continuity between them. What's crucially
important is that we don't say *King's Alfred's (<Ælfredes cyninges>)
any more -- not because multiple marking is forbidden (if this were so,
we'd expect to find at least some dialects allowing such a construction
amongst those that permit multiple negation or double gradation) but
because "the genitive" is now a phrasal phenomenon, more syntactic than
morphological. [[King Alfred]'s] is similar to [of [King Alfred]] except
that it involves a postposition.
Piotr