Re: THEORY: unergative

En réponse à Philippe Caquant :


>I recently discovered, to my utter despair, that a
>whole horde of linguistic concepts I hadn't reckoned
>upon were howling under my city walls.
Hehe, this list is a goldmine for anyone ready to learn new things in 
linguistics :)) .

>I had heard before of things like accusative, passive,
>theme, rheme and even ergative and absolutive. Now
>there seems that they brought new friends called
>anti-accusative, anti-passive, primary theme,
>secundary theme and anti-ergative (probably others
>too).
Actually, I don't think there's anything called "antiaccusative" or 
"antiergative" out there :)) . If you mean "unaccusative" and "unergative", 
they do not "oppose" the normal accusative and ergative. And actually, the 
names are not quite well chosen, but they've stuck so it's impossible to 
change them now :(( .

>  I feel the same as physicists when they
>discovered that there were some anti-neutrons,
>anti-protons, anti-electrons, plus some Higgs-bosons
>and three dozens of even more suspicious characters
>around.
LOL.


>1/ Languages like Latin or English use transitive
>verbs (among others), like in:
>I shot the sheriff
>*I* is understood as agent, thus subject and in the
>nominative case
>*the sheriff* is understood as the patient, thus
>object and in the accusative case (I shot him)
>The passive form would be :
>The sheriff was shot by me (*me* being facultative,
>but if present, at an oblique case)
Correct in general.

>2/ But there are languages, like Basque, various
>Caucasian langs and others, using another way :
>By-me shot the sheriff
>*me* being in ergative case, and *the sheriff* in
>absolutive
>What conception does such a form reflect ? I think I
>understand it as : "There was some sheriff-shooting,
>and that was done by me". Maybe I'm wrong ?
That's a way to describe it (and indeed many people who don't understand 
what ergativity is think that ergative languages are "inherently passive". 
That's not quite correct, but that's the best way to describe ergative 
languages in accusative terms).

>So what's exactly the theme and the rheme in those two
>sentences, I'm not clear about it yet. Normally the
>theme is "what we're talking about", and the rheme
>"which new information we're adding to that topic".
>Maybe I'm wrong too ? But if I burst into the saloon
>and yell "I shot the sheriff !", looks to me that
>there wasn't any theme and the whole sentence is the
>rheme ?
Indeed. The theme can be absent, or refer to the whole context. It can stay 
unexpressed. Only the rheme is necessary in any sentence.


>3/ When the verb is not a transitive one, than the two
>different ways we just saw become something like :
>I danced, or :
>I shivered
>(*I* being the grammatical subject, in the nominative;
>in the first case, "I" is doing an on-purpose
>activity, in the secund case, a not-on-purpose
>activity (?) ; is "I" yet called an agent or not, I
>dunno - probably an *experient* in the case of
>shivering ?)
"Experiencer" is the usual term. Note that in transitive sentences like "I 
saw the sheriff", the subject is also an experiencer, not an agent. 
"Seeing" is not an action (that would be "watching").

>but the Basque or whatever would say :
>By-me danced, or By-me shivered (meaning: There was
>some dancing / shivering and it was done by me)
No, that's incorrect. They would say:
me danced, me shivered (i.e. the subject is in the absolutive!)

The important difference between accusative and ergative languages is the 
case the subject of intransitive sentences is put in:
- In accusative languages, the subject of an intransitive sentence is put 
in the same case as the *subject* of a transitive sentence. This case is 
usually called "nominative", and the object of a transitive sentence is put 
in a case called "accusative".
- In ergative languages, the subject of an intransitive sentence is put in 
the same case as the *object* of a transitive sentence. This case is 
usually called "absolutive", and the subject of a transitive sentence is 
put in a case called "ergative".

If Basque did as you wrote it, it would mean the subject of an intransitive 
sentence would be in the same case as the *subject* of a transitive 
sentence, and by definition this would mean the language is accusative :)) .

>So the linguists said : well, in that case, this is
>not ergative, this is anti-ergative, because ergative
>can only apply in transitive sentences. (Is that it ?)
Not at all, you're mixing things that don't fit together.

>And then they said: well, if there is anti-ergative,
>why shouldn't it be also anti-accusative, anti-passive
>and anti-whatever-ive ? (The point is I can't find out
>any examples by myself, my brain having got over the
>boiling point already).
Actually not. There's no such thing as an "anti-accusative". However, there 
is such a thing as an "antipassive". Like the passive in accusative 
languages, it's a voice, but it applies to ergative languages. Let's make a 
simple but good definition:
- in accusative languages, there is often a voice called "passive". This 
voice basically takes a transitive verb, and makes it into an intransitive 
verb (we say it "decreases the verb's valency" - valency is the number of 
syntactically mandatory arguments the verb must take -), while it erases 
the subject (which can sometimes be reintroduced as an oblique argument, 
but is by no means necessary) and *promotes* the object to subject position 
(the object goes from the accusative to the nominative case). Your 
description of accusative languages shows that you understand that already, 
but I put this description in order to make the next one clearer.
- in ergative languages, there is sometimes a voice called "antipassive". 
This voice also takes a transitive verb, and makes it into an intransitive 
verb, while it erases the *object* (the argument which is in the most basic 
case, and which can be later reintroduced as an oblique argument, but by no 
means necessary) and promotes the transitive *subject* to intransitive 
subject position (it goes from the ergative to the absolutive case). The 
term "antipassive" becomes clear: it looks like something that works in the 
opposite way from the passive. To give an example using the same example 
you used, if you have:
I-ERG shot the sheriff-ABS (I add the case abbreviations to make things clear)
I-ABS danced
I-ABS shivered
I-ABS shot-ANTIPASSIVE ("to" the sheriff-OBLIQUE CASE)
The "to" is only there to make it clear that if the former object is 
reintroduced, it is in an oblique case).

However, for what is worth, Basque doesn't have an antipassive. Indeed, it 
has no voice distinctions whatsoever. However, there are other ergative 
languages that feature such an antipassive.

A thing you have to remember about the distinction between accusative and 
ergative languages is the concept of *default case*, i.e. basically what 
case is the dictionary form of a word in. In accusative languages, it's the 
nominative. The dictionary form of nouns is always the subject form, in 
both transitive and intransitive sentences. In ergative languages, on the 
other hand, the dictionary form is the *absolutive* case. This means that 
the dictionary form corresponds to the *object* of transitive sentences 
(and the subject of intransitive sentences). And the ergative is a 
*derived* form, not a base one, just like the accusative in accusative 
languages is a derived form.

If this seems unclear to you, that's because you're reasoning with the 
concepts of "subject" and "object", which fit only accusative languages 
(and then, only those that neatly use the nominative for subjects and the 
accusative for objects). If you try to apply them to ergative languages, it 
becomes automatically a mess (unless you treat the term in the absolutive 
as "subject", which is a valid way to think about it, but then you have to 
remember that the subject of transitive sentences is the *patient* rather 
than the agent). So what you need to try is to stop thinking in these 
terms, and only think in terms of syntactic cases.

>So if somebody could give me some clear and easy
>examples, I would pray the Lord he be blessed, and his
>sons be blessed, and his grandsons be blessed, and his
>cats and dogs and cows and sheeps and everybody home
>be blessed too.
I don't have examples at hand, but I hope my explanations have cleared 
things up a bit. You cannot begin to understand the complexities of 
split-ergative or active systems, of the "unergative" and the 
"unaccusative", of Fluid-S and Split-S, if you don't understand first the 
basic distinction between nominative-accusative and ergative-absolutive 
languages.

I advise you to get the "Que Sais-Je ?" number 1668: "Les Basques" and 
2006: "La structure des langues". They are a good starting point to learn 
about those notions ("Les Basques" has only a small grammar part, but it is 
good to see how ergativity is actually implemented), and are cheaper and 
easier to understand than professional linguistic books. And they are 
written by professional linguists so the info you get is still good info, 
not too dumbed down (indeed, "La structure des langues" is not something 
the layman would understand, but for the one who knows already some 
linguistic vocabulary, it's a good starting point).

Christophe Grandsire.
http://rainbow.conlang.free.fr
You need a straight mind to invent a twisted conlang. 

Reply