Newsgroups: sci.lang
Path: cantaloupe.srv.cs.cmu.edu!das-news2.harvard.edu!news2.near.net!news.mathworks.com!usenet.eel.ufl.edu!spool.mu.edu!howland.reston.ans.net!pipex!uknet!festival!edcogsci!iad
From: iad@cogsci.ed.ac.uk (Ivan A Derzhanski)
Subject: Re: Ergativity and Middle voice
Message-ID: <D6FFv4.5FD@cogsci.ed.ac.uk>
Organization: Centre for Cognitive Science, Edinburgh, UK
References: <3lgocg$kfi@marble.Britain.EU.net> <D6Bnss.IvC@cogsci.ed.ac.uk> <3li9bf$3di@marble.Britain.EU.net>
Date: Sun, 2 Apr 1995 21:33:02 GMT
Lines: 126

In article <3li9bf$3di@marble.Britain.EU.net> paul.sampson@octacon.co.uk (Paul Sampson) writes:
>In article <D6Bnss.IvC@cogsci.ed.ac.uk>, iad@cogsci.ed.ac.uk (Ivan A Derzhanski) says:
>> In article <3lgocg$kfi@marble.Britain.EU.net> paul.sampson@octacon.co.uk (Paul Sampson) writes:
>>> (e.g. 'Victoria built this Memorial' meaning
>>> 'Victoria had this memorial built' - she didn't build it herself).
>
>> She's still an agent, of the causation if not the building itself.
>
>Yes. The esample was intended as a causative middle, if only to
>remind our readers of the existence of a non-reflexive one.

Count me as one of those readers then.  I've never encountered a
non-reflexive causative middle.  Why would it be called a middle
if the subject is the agent, but has no other role?

>> If she is also a beneficiary (if she built the monument, or had it
>> built, for herself), you may have a reflexive construction.
>
>Well yes, you *may* (although I believe you meant _indirect_ reflexive),

Of course.

>but that's not what I intended.
[...]
>Can't we just take it as read that I've got an excellent example of a
>non-reflexive middle and proceed from that basis?

Well, I can't, because I don't see why this is an example of middle voice,
but that's just me.  I know little about Ancient Greek.  Are you saying	
that in AG the middle voice is used for causatives, even if the agent
of the causation is neither patient nor beneficiary of the action?
Could you translate the sentence into AG?

>I don't really want to get into an argument about why, if *all*
>middles are fundamentally reflexive in nature, the voice wasn't
>simply called reflexive instead of middle. If you believe that this
>*is* the case (all M are R) then clearly I'm buggered and unable to
>start the argument.

I do not believe that all M are R.  In my experience/impression M means
that the subject has a role other than agent, whereas R means that the
subject has more than one role (in which case one of them is bound to
be different from agent, so all R are M, but not vice versa).  Subject
to language-dependent restrictions, of course.

>If [...] you mean *all* causatives (are indirect reflexives) then I'm
>stuffed and should of course give up right now.

I mean that if a causative such as the one in your example is not a
reflexive (direct or otherwise), then I don't see why it is a middle.
If you say that AG uses the middle voice there, without Victoria being
a beneficiary, I'll update my views accordingly.

>Let's posit, after all, what the pre-prior para proposes to posit.

Perfect.  Pray proceed to the pre-prior para's purported purpose.

>> ... Let's look at Georgian:
[...]
>>       _mGeri-s_ `(s)he sings'
>>     _i-mGeri-s_ `(s)he sings for h##self'
>>   _m-i-mGeri-s_ `(s)he sings for me'
[...]
>> ... Actually,
>> _imGeris_ is not normally considered middle voice either, merely a
>> subjective version (that's the traditional term) of an active verb.
>
>Ah. Now then. Were I to instantly believe what you say here, then I
>think I'd start to feel the water lapping round my neck. What do you
>mean by 'not normally considered'? *Could* it be middle?

By `not normally considered' I mean that when people use the term
`middle voice' with respect to Georgian (which is not very often),
they tend to refer to another class of verbs, different from this
one both in their own morphology and in case marking.

>Let me try to explain this bit again. My argument about middle and
>ergative being alternative solutions to the same 'problem' *depends*
>on non-reflexive middles.

Yet your singing to/for oneself example involves a reflexive middle.
An indirect reflexive indeed, one which may not be marked as such,
but a reflexive none the less.

>> [...] in Georgian it's the middle voice that tends to be unmarked,
>> cf. _rz'e duG-s_ `the milk is boiling', _rz'e-s a-duG-eb-s_ `(s)he
>> is boiling the milk'.

Hence also _rz'e-s i-duG-eb-s_ `(s)he is boiling the milk for h##self',
one of those indirect reflexives that are not called middles, because
they have nothing with common with the middle _duG-s_ `it is boiling'.

>Aww. Buggeration. That's pretty much it then. Why didn't you say that
>in the first place? Damn.
>
>*Unless* (strawclutching here) Georgian *ergatives* tend to be unmarked
>as well? (Or at least less marked than nominatives).

Alas, no.  Neither I nor Greenberg know of a language where ergatives
would be less marked than nominatives.  If there is an exception,
Georgian's not it (cf. NOM _rz'e_, ERG _rz'e-m(a)_ `milk').

That said, Georgian is a wee bit special, being split ergative and all,
but would it make a difference if we took an ergative language where
middles are marked?

>> It is true that active constructions often can have an oblique reflexive
>> (benefactive) meaning, but I fail to see where ergativity comes in.  You
>> can get that reading with nominative subjects just as easily.
>
>And I, likewise, cannot see why you're mentioning reflexivity again,
>the one thing I was specifically excluding from the argument.

You weren't.  You said you were going to, but both examples you gave
don't seem to have anything to do with middlery of any sort unless
they are read as indirect reflexives.

>But on the whole - Rats.

<mia-aow!>

-- 
`"Haud oan there a meenit," says the king tae Joseph, "I've been thinkin."'
Ivan A Derzhanski (iad@cogsci.ed.ac.uk)    (J Stuart, _Auld Testament Tales_)
* Centre for Cognitive Science,  2 Buccleuch Place,   Edinburgh EH8 9LW,  UK
* Cowan House E113, Pollock Halls, 18 Holyrood Pk Rd, Edinburgh EH16 5BD, UK
