Posted by Hans-Werner Hatting on 15:38 7/16/02
In reply to: Language evolution and the icëlani posted by Luca Mangiat on 22:12 7/9/02
"at least, I've never heard of a natural language with completely free morpheme order. I'd want to know what stylistic or semantic effect the variious orders have, though."
I also don't know about such a language. What we can find is movable
morphemes, which can attach themselves to different elements of the
sentence. An example is the Polish conjuctive marker "-by-", which can be
attached either to the verb, or to a sentence-initial element (mostly a
/(a) jezelibym wiedzial = (b) jezeli wiedzialbym/ " if I had known"
(Polish readers, please excuse my omission of diacritics). Both (a) and (b) are wide-spread in contemporary Polish.
Russian "by" works in a similar way, of course without personal endings.
Another case are the personal endings of the past tense, which also can be
attached to sentence-initial conjunctions (from a poem):
(a) /gdys szeptala mi slowa tesknoty/ = (b)/gdy szeptalas mi slowa tesknoty/ "when you whispered to me words of longing"
In this case, (b) is the normal case in temporary Polish, while (a) is archaic and today is used mostly only in poetry.
The reason for the movability of these elements is that they originally were enclitic forms of the verb "to be", which compbined with the perfect participle in -l- to form the past tense and the conjunctive respectively.
I don't know whether this really is on topic, but maybe some "Boardists" will want to use something like this in a conlang?
Quechua has moveable elements too, such as the interrogative -chu, the topic particle -qa, and the evidential particles.