The Usage of the Ergative and Accusative Cases in Rawang Ata

The Ergative Case

The ergative case is probably the most versatile case in Rawàng Ata – it can be used for subjects, objects (direct and indirect), and appositions. Its interpretation must therefore depend on the context, and in particular on the form and class of the verb.

Ergative Subjects

Most typically, the ergative marks the subject of a (univalent or bivalent) intransitive masculine verb, where transitivity requires agency, animacy, and success:

datta rafaringa lònangam

“the sailor kicks the pot” (transitive)

dattaya rafaringa lònangam

“the sailor is made to kick the pot” or “the sailor tries to kick the pot” (intransitive)

Contrariwise, the ergative marks the subject of ‘transitive’ feminine verbs, which means simply that they mark subjects that are more animate than the object, but not less animate ones (which are marked with the accusative):

dattaya moyisara kòma

“the girl considers the sailor distasteful”

lònangam moyisara kòma

“the girl considers the pot distasteful”

Exactly the same pattern occurs with masculine verbs with passive voice, in which attention is drawn to the object rather than the subject:

(dattaya) safaringawa lònang

“the pot is kicked (by the sailor)”

(jonìm) rakijdawa datta

“the sailor was bitten (by the rat)”

[This is not the unmarked word order for passives, but as the word order is irrelevant to case use, we retain the primary word order for clarity]

Ergative Objects

The ergative may also mark the object of a verb – if the verb is applicative, or antipassive, or if the object is an “ergative object”, or in the case of a small number of inherently ditransitive verbs.

In the case of antipassives, attention is drawn from the object toward the subject (and, likewise, away from the question of transitivity). The subject becomes unmarked, and the object is marked with the ergative, if it is present at all. The antipassive is common with feminine verbs (for syntactic reasons), but rare with masculines, where it often denotes a more indefinite or hypothetical event, or is used to obscure success. In any case, the antipassive is the same for both verb classes:

datta moyisaràta (kòmaya)

“the sailor is found distasteful (by the girl)”

jonì sakijdàta (dattaya)

“it is the rat that bites (the sailor)(successfully or otherwise)”

The applicative voice, which is rare, and occurs typically only with certain verbs, promotes an oblique argument to an object in the ergative case, demoting the object (if any) to an indirect object, also in the ergative (the same demotion found in the antipassive). This can occur with neuter univalents promoted to bivalents, or with masculine bivalents promoted to trivalents. The applicative is mostly seen with neuters, and when occurring with masculines there is mostly a more specific meaning intended. Colloquially, the antipassive affix is often used to mark the applicative, but this is deprecated in formal registers.

jonì sawōralata (nahùnki)

“the rat scurries (along the beam)”

jonì sawōralatika nahungya

“the rat scurries along the beam”

datta rakokkùta samùn (kòmabin)

“the sailor knocks at the door (for the girl)”

datta rakokkùtika kòmaya samùnya

“the sailor knocks at the door for the girl” (i.e. the sailor calls by asking to see the girl)

Once the applicative has been applied, the passive may in turn be used, putting the subject too into the ergative:

jonìya nàwōralatikawa nahungya

“the beam is scurried along by the rat”

dattaya kirakokkùtikawa kòmaya samùnya

Certain verbs of transference and alteration are inherently ditransitive. With these, the indirect object is marked with the ergative

datta raboala lònangam kòmaya

“the sailor gives the pot to the girl”

And these may be subject to the passive:

dattaya raboalawa lònang kòmaya

Finally, some objects are ergative objects even with active masculine verbs. This typically occurs when there is a partitive or ablative implication:

kòma kiraroka wenuya

“the girl eats some of the fruit”

datta rakyela tuantiya

“the sailor drank from the bottle”

This particularly includes cases where the object is a mass noun, in which case the ergative is mandatory:

datta rakyela farāya

“the sailor drank some salt water”

There also exist a number of idiomatic expressions involving these ergative objects: whether this is an idiomatic use of the ergative, or whether some abstract mass noun is being thought of is a matter of definition:

kōba radania lònangam

“the nobleman held the pot”

kōba radania lònangya

“the nobleman defecated”


The ergative is also used to mark appositives. Typically, these state an equivalence between the referents of two terms:

datta kōbaya…

“the sailor, a nobleman…”

However, more abstract connections can also be shown in this way:

lònang wettaìya

“the pot, made of clay”

wettaì lònangya

“the clay, being clay for a pot”

There is clearly considerable overlap between apposition and adjunction. The key distinctions are that adjunctions assume that the modifier is integral to the nature of the item, whereas apposition is more tangential: therefore, apposition can be used to supply additional information not originally known – although an adjunct may be introduced for clarification, it is implied that no new information is added:

lònang wettaìya

“the pot, which is made of clay, by the way”

wettaìu lònang

“the clay pot (you did realise it was clay, didn’t you?)”

Additionally, adjunctions frequently have a more idiomatic meaning, while appositions rarely do. So:

rìssari hotòmya

hotòmu rìssari

“sheep fabric (wool)”


rìssari layiaya

“fabric to do with an island” (from an island, a tapestry of an island, etc)

layāu rìssari

“cheap, rough wool” (i.e. domestic, not imported)

Further, due to their flexibility, appositions can have anaphoric usage:

rìssari layiaya

“fabric to do with that island I mentioned”

Appositions are also frequently used, with anaphoric or cataphoric pronouns, to link to relative clauses:

datta mànya rakyela farāya, moyisara kòma

“the sailor who drank the salt water was the same sailor who displeased the girl”

As opposed to:

datta rakyela farāya, moyisara kòma

“the sailor drank some salt water, and displeased the girl”

The Accusative

By contrast with the ergative, the accusative is restricted in application. With masculine or neuter ditransitive verbs in the active, it marks the direct object:

datta rakyela sōban

“the sailor drinks some drinking-water”

datta raboala lònangam kòmaya

“the sailor gives the pot to the girl”

With feminines, and with masculines in the passive, it marks an inanimate subject:

rìssarim moyisara kòma

“the fabric is distasteful to the girl”

mobàkum nàdyowalawa lònang

“the pot was struck heavily by the hammer”

mobàkum tujtōnàna lònang

“the pot was broken into large pieces by the hammer”


6 thoughts on “The Usage of the Ergative and Accusative Cases in Rawang Ata

  1. miekko says:

    as has been said elsewhere, you maybe should specify what you mean by masculine and feminine verbs in this post. I think I know what you mean, but I am not entirely sure about it, and that makes reading this post feel like being one step behind all the time.

  2. miekko says:

    and, tbh, I must admit to finding the alignment described fascinating but unlikely. The level of description you provide, and the historical justifications compensate very much, but I suspect it crosses too many lines to be easily parseable by a first-language learner, and therefore would collapse into a simpler system rather soon?
    OTOH, quirky case seems to do worse things in some European languages, so why not. Also, this is just really a gut reaction that I don’t really base on much actual data.
    (I did in fact think I had read a text that said that such an alignment is ruled out by some universals, but rereading the text it really didn’t say that)

  3. vacuouswastrel says:

    The not-giving-enough-information is partly intentional. People don’t read complete grammars. And I don’t like writing complete grammars, getting everything in perfect order. Instead, I just write on particular topics – but naturally there’s going to be things unexplained. Because if I explain those, there’ll then be other things left unexplained in those explanations… So I just hope that people will find the hints about the langauge tantalising. Perhaps I went too far this time, though – I wrote this just after those musings about the history of the system, so forgot to add some background this time.

    On the unlikeliness: probably, yes. I don’t think it’s attested. But which bit is it that’s the problem, in your opinion?

    I can see four problems:
    – the transitive/intransitive distinction in bivalents. I’m OK with that – it’s not uncommon in North America, and it even occurs in some of the Austronesian languages that obviously are a big influence on RA (eg Tongan, even if the marked/unmarked assumption is the other way around)
    – the large number of arguments that can take the ergative. OK with that – it’s not uncommon to use ergatives for indirect objects, including directs demoted and obliques promoted by applicatives.
    – The masculine verb active voice ergative objects. Not worried about that – just imagine the ergative as a genitive, which can often mark objects, particularly partitives. Most of those examples actually work in English – “he drank of the water” etc. Even something like “he held of the pot” SOUNDS like an archaic idiom for something, even though it doesn’t actually have a meaning.

    – The big one: having distinct paradigms for the two verb genders (I’m OK with the neuter, which is pretty close to the masculine, and verbs of motion are sometimes weird). That is, indeed, to my knowledge, unattested. However, if you want you can think of it in a less weird way: if you take the “object” of feminine verbs to be their Subject, they’re the same as similar to transitive masculines: unmarked subject, accusative object. Just add that the object can take ergative if it’s animate, and that feminines don’t have the option of ‘intransitive’ marking. Which is different, but if you think of them as active and stative verbs it’s not that odd to have slight differences between them…

    So, IS the “object” really the subject? From cases and agreement only, it looks an awful lot like one. And since the language is topic-prominent, some of the confusions go away (ie we can’t tell what the pivot is, since it’s the topic, not the subject). So it’s ony really some peripheral things that lead us to not think it’ the object (anaphoric pronouns, for instance, and serial verb constructions).

    – And then there’s the voices, I suppose. But they’re not that weird, really – and in colloquial versions they are indeed simplified in various ways.

    So… yes, my instinct is, like yours, that it’s unlikely. But when I think about it, I can’t think what’s so impossible about it really. [And hey, it’s full of abnormalities. If this were a real language, it would be hidden up a river in the Amazon somewhere.]

  4. miekko says:

    I think the system that was supposed to be impossible, which I for some reason thought you were getting fairly close to, was as follows (some background first)
    S:1 V DO:2
    S:1 V
    V DO:2
    (S = transitive subj, S
    = …)
    (the third example, intransitive objects might not necessarily exist in a given language, but it’s included in the listing for some reason)
    S:1 V DO:2
    :2 V
    V DO:2

    S:2 V DO:1
    :2 V
    V DO:2
    (this is pretty close to how Finnish works!, except DO can be either 2 or an added 3, and S
    sometimes also can be 3)

    S:2 V DO:1
    :1 V
    V DO:2

    to some extent your system does that, but … the fact that the different systems all have their own domains and not just overlap each other… (or maybe I read sloppily and your system doesn’t do that?)
    Also, Finnish kinda does muddy these waters, since some of the overlap does suggest this very shape (e.g.
    {2v3 2v1}
    {3v 2v}
    {v1 v3}

    we can pick out these subsets
    2v3 (nom verb part)
    3v (part verb)
    v2 (verb nom)
    which is exactly like that. the partitive accounts for 70% of all objects, though, so the actual situation is that
    are the most common forms – so basically a fully nom-acc system
    2v1 and v2 do occur, but less often; I would even posit that the ratio of v2 to v3 is higher than the ratio of 2v1 to 2v3.

    I could type down the important parts of the paper on this that I have here in a deadtree book, if you want? (The elaborations on Finnish are mine, based on short comments in the book.)

  5. miekko says:

    damn, the software ate my indexes

    St V D:O
    Si V
    V D:O
    is the supposed form; I put the i and t between lesser than and greater than symbols, which seem to disappear here 😦

  6. vacuouswastrel says:

    I’m not sure what you mean by overlapping domains in this instance. You may be right in your reading – I’m just not sure I follow.
    I don’t THINK it has that unattested configuration. But really, my system is so much more complicated than that that it’s hard to say for sure.

    Many thanks for your support both on this blog and elsewhere recently, by the way. I appreciate it.

    [Sorry for not responding to this earlier – wanted to say something more meaningful, but things kept getting in the way. Eventually I gave up and just replied]

Leave a Reply

Fill in your details below or click an icon to log in: Logo

You are commenting using your account. Log Out / Change )

Twitter picture

You are commenting using your Twitter account. Log Out / Change )

Facebook photo

You are commenting using your Facebook account. Log Out / Change )

Google+ photo

You are commenting using your Google+ account. Log Out / Change )

Connecting to %s