Interesting, thanks. I plan on creating a daughter language from one of my conlangs, which is SOV, however I had an idea of making the daughter language in period of changing from SOV into SVO with main clauses being SVO and sub-clauses being SOV, though I couldn't, imo, think of any reasonable/natural reasons for the shift and it being and staying in a semi-shift sate. i.e with both SVO and SOV constructions.

I had thought of losing case endings and perhaps contact with an SVO language which then became the prestige language, sort of what happened to old English and Norman French however I was entirely convinced this was both good/effective and natural occurring.

> Date: Wed, 11 May 2011 10:28:33 +0100
> From: [log in to unmask]
> Subject: Re: SOV > SVO
> To: [log in to unmask]
> 
> staving vii iiix:
> > What are the most common reasons, or reasonable ones, that natlangs,
> > or any lang, would swap from SOV to SVO?
> 
> It happened in the Romance languages, where I image it was associated 
> with loss of case endings, coupled with the fact that Latin had a 
> flexible word order. The fact that in the Classical period Latin was 
> pro-drop for subjects might also have had an influence. So consider...
> 
> canis mordet (the dog bites)
> canem mordet (he bites the dog)
> 
> Now lose the case endings
> 
> *can mordet
> *can mordet
> 
> We need a new way of disambiguating subjects and objects. We can use 
> word order
> 
> * can mordet (the dog bites)
> * mordet can (he bites the dog)
> 
> or we can reintroduce subject pronouns
> 
> * ille can mordet (he bites the dog)
> 
> or we can go for the belt and braces approach
> 
> * can mordet (the dog bites)
> * ille mordet can (he bites the dog)
> 
> Which is what seems to have happened in real life.
> 
> An interesting thing to note in this context is that French retains SOV 
> with pronominal objects, which unlike pronominal subjects, couldn't be 
> inferred from the verb in Latin.
> 
> Pete