Re: Trigger language?

En réponse à Sarah Marie Parker-Allen <lloannna@...>:

> What IS a trigger language?  In fact, what are all those other kinds
> of
> languages that have been mentioned (there are so many I can't remember
> them
> all).  Is there a concise guide out there somewhere?
>
Well, a quick google didn't reveal anything particularly concise :(( . So I
guess I'll have to give a little (very theoretical) explanation of the
different natlang types around there. NOTE FOR EVERYONE: I know that some of my
explanations will be incomplete. It's on purpose. For somebody who doesn't know
at all what ergativity is for instance, it's better first to speak only about
grammatical ergativity, and only when this is well assimilated about syntactic
ergativity. But I will add a small bit about this :)) .

OK, the language types we are talking about have to do with how they arrange
the main participants of a sentence ("subject", "object", etc...), and even how
they choose them.

Usually, languages have intransitive verbs (verbs with only one participant,
like "run" or "jump") and transitive verbs (verbs with two participants,
like "have" and "kill"). Ditransitive verbs (verbs with three participants,
like "give") usually behave like transitive verbs for the purposes I want to
talk about. Language type here has to do with how they treat the different
participants: the only participant of an intransitive verb (called S for
Subject), the subject of a transitive verb (called A for Agent) and the object
of a transitive verb (called P for Patient).
There are languages that treat S like A (they take the same case, or are put at
the same position in the sentence), while treating P differently. Those are
called "nominative-accusative" (or simply "accusative") languages, from the
usual names used for the case of the S/A and the P. Those are languages you
already know: English, French, Spanish, Russian are all accusative languages.
It's for this type of languages that the terms "subject" and "object" fit.
Now, there are languages that treat S like P, and A differently. It means that
they treat the subject of an intransitive verb like the *object* of a
transitive one, and treat the subject of a transitive verb differently. Those
languages are called "absolutive-ergative", or simply "ergative", from the
names usually given respectively to the cases of S/P and A. Languages like that
are Basque and IIRC Georgian (at least partly for that one).
And there are even languages that treat S sometimes like P and sometimes like
A! Those are called "active languages", but are divided usually into two
subgroups: "split-S languages", for which each intransitive verb takes a
certain marking for S, whether like P or like A, and always the same one (it
depends only on the verb to choose the marking), and "fluid-S languages", for
which intransitive verbs can take either S marked as P or S marked as A,
depending on features like the willingness or the animacy of the subject (in
that case, choosing the marking doesn't depend on the verb but on the subject).
If there are languages that mark S differently from both P and A, they must be
exceedingly rare. I guess they would simply be called "active".
Now you mustn't think of those categories as hermetic boxes in which languages
are trapped. When I said that Georgian was at least partly ergative, it is
because, IIRC, its verbs behave sometimes ergatively, sometimes accusatively.
Indeed, it happens often that some languages have verbs behaving accusatively
in some cases and ergatively in others. Typical things are:
- depending on tense (ergative in the present, accusative in the past, or vice-
versa),
- depending on the animacy of the person (if one of the participant is a 1st or
2nd person, the verb will behave accusatively. If all participants are 3rd
person, it will behave ergatively),
- and other possible things.
Those languages are usually called "split-nominative" or "split-ergative" (note
that everytime I talked about an "accusative" language, I could have
said "nominative" language too. The two designations are synonymous).

OK, I guess it's already quite complicated like that, but languages don't
care ;))) . Above that, you also have languages which are said to have a "topic-
comment" structure. Sentences are always organised following a "topic-comment"
structure: the topic is what you are talking about, and the comment is what you
are saying about it. In English, you can mark the topic using "as for", but
usually you only use intonation, or you consider the subject to be the topic
(so you sometimes use passive voice to put the actual topic as the subject of
the sentence. Sentences like "I killed him" and "He was killed by me" are
different only in what is the most important in each case). In any case,
grammatically English doesn't give much importance to the topic-comment
structure of sentences. It's happy with a subject-verb-object structure. It's
not the case of all languages. Japanese, for instance, although it's an
accusative language, gives a lot of importance to the topic-comment structure
of sentences, and thus marks explicitely the topic (with the particle "wa")
which is normally present in every (or nearly every) sentence. There are quite
a few languages which work like that (Spoken French is an example), but
grammatically they normally can still be categorised in one of the categories I
already talked about.

Now, there are languages that take the "topic-comment" structure to its
extreme, so that even grammatically the *only* way to construct sentences is to
use a "topic-comment" structure. Those are called "trigger" or "focus"
languages. In those languages, it makes no sense anymore to talk about S, A or
P. In those languages, it's a bit as if all the verbs were intransitive, as
they only ever have one mandatory participant. This participant is called the
trigger or T (or the focus). It is marked as being the trigger, and that's all.
You don't mark it for its function in the sentence. Instead, the *verb* itself
is marked for the function of the trigger! So, if the trigger is the agent of
the action, the verb will receive a mark indicating "agent". If it's the
patient, it will receive a mark "patient". If it's the location of the action,
the verb will receive a mark indicating "location", etc... As for the other
participants, they are *all* optional, and all marked *equally* for their
function (agent, patient, location, time, goal, etc... are not treated
differently from others as soon as they are *not* the trigger). Only the
trigger has a special status here (unlike in all the other kinds of languages I
talked you about until, where for transitive verbs always two participants have
a special status). As you see, you cannot construct a sentence in this kind of
language without choosing first what is the topic of the sentence, which will
become the trigger (just like in English you cannot make a sentence without
first choosing the subject). Trigger languages seem to be found exclusively
among Austronesian languages, and Tagalog is usually the main example given.

I have already talked about languages which consider "animacy" (the state of
being more or less animate, with usual scales like 1st person>2nd person>3rd
person and/or human/animal/plant/mineral/abstract things, etc...) important,
whether to choose how to mark S (Fluid-S languages) or to choose between
ergative and nominative marking (split-ergative languages). Now, there are
languages which take animacy to be the most important thing, and thus construct
sentences only dependent on it. In short, in those languages, when a verb has
two participants (a transitive verb thus), neither is marked in any way
different from the other! How do you know which one acts and which one receives
the action? Which one is the agent and which one is the patient? By the animacy
hierarchy! The higher one in animacy will be the agent, while the lower one
will be the patient. And how do you do to say that something low on animacy
(like a stone) acted on something high on animacy (like "I"), i.e. how can you
say a sentence like "The stone hit me"? Well, just add an "inverser" marker on
the verb, which indicates that it's the lower one on the hierarchy which is the
actor, this time. Those languages are neither nominative, nor ergative, nor
active, nor even trigger. They are another type of language, which
unfortunately doesn't have a specific name. FWIW, this type of languages is
found exclusively in America, and I think even mainly in South America.

Well, those are the different language types I know about. There may be others
(I'm not an encyclopedia ;))) ), but they probably concern only a few
languages. I know that there are a few (I heard of at least one) languages
which are called "unoriented". It means that for those languages, *nothing*
except context can tell who did what in a transitive sentence: there is no
grammatical way to recognise the agent from the patient, and they must resort
to context for that. Among conlangs, plenty of other types have been invented,
but natlangs seem to be happy with the types I described already. Note that you
don't need overt marking for a language to be nominative or ergative. English
doesn't put a special mark on the object of verbs, and yet is definitely an
accusative language. It uses word order instead of affixes. Note also that my
description of languages was purely didactic. I introduced trigger
and "animacy" languages this way to make it easier to understand how they work
for somebody who knows only about accusative languages. But there is *no*
hierarchy of language types. Nominative, ergative and active language types are
in *no* way more primary than trigger for instance (they are more common among
languages, but that is not a reason to judge them more "primary" :)) ). It was
just a way to introduce trigger languages smoothly.

Finally, as I kind of implied when I talked about "syntactic ergativity",
language types don't only influence the way you organise the main participants
of a sentence. They also influence the way you concatenate sentences and how
you refer to previous participants (for instance, when I say "he catches the
ball and runs", you immediately understand: "he catches the ball and *he*
runs". Well, in an ergative language, since the S and the P are treated the
same way, it is the same about references, and the same sentence by a speaker
of an ergative language will mean: "he catches the ball and *the ball*
runs"!!!!). But this is a more complicated part which you can forget for now.
First, try to understand exactly how those language types work morphologically.
We can always go back to syntax later :)) .

And to the ones in the know, sorry for some nearly outrageous simplifications I
made. This explanation is intended to Sarah who expressed that she was a newbie
at that, and thus I tried to make it as simple as possible, and was thus
obliged to simplify reality a bit :)) . As it is, I find this explanation
already quite complicated :)) .

Now Sarah, if there is anything you don't understand, feel free to ask.
Language types are not a trivial thing to understand, so it would be quite
normal if you have difficulties, especially with the more "exotic" types :)) .

Christophe.

http://rainbow.conlang.free.fr

Take your life as a movie: do not let anybody else play the leading role.