[Lingtyp] query: instrument voice

Tue Feb 22 18:53:53 UTC 2022

Thanks David, I'll have a look. I'll reply off-list as this probably goes
far beyond the interest of the average Lingtyp subscriber.
Cheers, Bruno

On Tue, 22 Feb 2022, 19:17 David Gil, <gil at shh.mpg.de> wrote:

> Hi Bruno,
>
> These are very nice examples, but there are still some missing pieces to
> the puzzle.  Let me try and represent your examples schematically.  In the
> following table, I characterize your three examples in accordance with
> three grammatical features, presence/absence of instrumental prefix,
> position of instrument, and whether or not it is flagged.
>
>
>
> Vbal INSTR prefix
>
> INSTR position
>
> INSTR flagging
>
> (1) (2)
>
> NONE
>
> post-V
>
> en
>
> (3)
>
> ka-
>
> pre-V
>
> NONE
>
> I hope the above is a correct representation of the facts that you
> presented.  But if it is, then the obvious question is whether these are
> the only possible constellations of these three grammatical features.
> Logically, there are eight possibilities, as per the following table:
>
>
>
> Vbal INSTR prefix
>
> INSTR position
>
> INSTR flagging
>
> (1) (2)
>
> NONE
>
> post-V
>
> en
>
>
>
> NONE
>
> post-V
>
> NONE
>
>
>
> NONE
>
> pre-V
>
> en
>
>
>
> NONE
>
> pre-V
>
> NONE
>
>
>
> ka-
>
> post-V
>
> en
>
>
>
> ka-
>
> post-V
>
> NONE
>
>
>
> ka-
>
> pre-V
>
> en
>
> (3)
>
> ka-
>
> pre-V
>
> NONE
>
> My question then is whether any of the six intermediate combinations in
> the above table are grammatical.  To the extent that they are not, then the
> parallel — a rather exciting one — with the constructions in Roon and other
> Bird's Head languages would be strengthened accordingly.
>
> Best,
>
> David
>
>
> On 22/02/2022 17:51, Bruno Olsson wrote:
>
> Hi David,
>
> Here is, briefly, why what you wrote about Roon reminds me of the split
> subject properties of the Marind instrument cxn.
> There are two relatively straightforward features that distinguish S/A
> arguments from objects etc in Marind. First, indexing with the so-called
> A-prefix series:
>
> (1) nok ka-no-gha-sak-e bes en
>    1 NTRL.ORTN-1.A-2sg.U-beat-IPFV stick with
>    '*I* am beating you with a stick.'
>
> Second, the use of the so-called Neutral Orientation prefix when the
> immediately preverbal NP is the subject. Compare (1) with (2), which has an
> immediately preverbal object, which requires the Object Orientation ma- on
> the verb:
>
> (2) nok ogh ma-no-gha-sak-e bes en
>    1 2sg OBJ.ORTN-1.A-2sg.U-beat-IPFV stick with
>    'I am beating *you* with a stick.'
>
> The bare instrument NP introduced by the WITH-prefix/applicative/voice
> fails the first subject test (it's not indexed on the verb), but passes the
> second (it requires the Neutral prefix, just like a subject):
>
> (3) nok ogh bes ka-no-ka-gha-sak-e
>    1 2sg stick NTRL.ORTN-1.A-WITH-2sg.U-beat-IPFV
>    'I am beating you with *a stick*.'
>
> Like in Roon, the agent in (3) controls subject indexing, but the
> behavioural test provided by the Orientation prefixes treats the instrument
> as the subject, which to me seems similar to the split in subject
> properties that you describe for Roon.
>
> Best,
> Bruno
>
> On Tue, Feb 22, 2022 at 2:16 PM David Gil <gil at shh.mpg.de> wrote:
>
>> Hi Bruno,
>>
>> Your Marind example is interesting.  I have nothing to say about your
>> language-specific analysis of it as an Applicative, but I would appreciate
>> further clarification on why you consider it to be similar to the Bird's
>> Head examples (whatever you choose to call them).
>>
>> You say that the instrument NP "shows at least one subject property":
>> What is this property? Is there a corresponding construction without the
>> *k-* prefix, and if so what does it look like?
>>
>> It occurs to me that, unlike Marind, the languages I am looking at (both
>> Austronesian and non-Austronesian) are all SVO, which means that the
>> pre-verbal position of the instrument NP constitutes a salient feature of
>> the construction in question.  Whereas for an SOV language like Marind,
>> different criteria would have to be sought.
>>
>> David
>>
>>
>> On 22/02/2022 11:45, Bruno Olsson wrote:
>>
>> Hi David,
>>
>> The Marind WITH-Applicative *k-* is similar to the Roon data. It allows
>> an instrument to appear as a bare NP:
>>
>>
>> *basik tamugh anggip k-a-k-w-alawa *pig food snout
>> NTRL.FOC-3sg.SBJ-WITH-3sg.OBJ-is.searching
>> 'The pig is searching for food with its snout.'
>>
>> The instrument is obligatorily preverbal in this construction, which is
>> the focus position in Marind, so the above sentence would answer 'With what
>> is the pig searching for food?' (so unlike the construction that
>> Austronesianists once called Instrument Focus, this is actual instrument
>> focus). In my grammar I labelled this an Applicative only for
>> language-internal and expository reasons, and I don't think it would
>> qualify as an applicative for any cross-linguistic purposes, because the
>> instrument NP fails to show object properties, and instead shows at least
>> one subject property (at the same time, the original subject retains its
>> subject properties).
>>
>> The reason for labelling it as an (upper-case) Applicative is that the
>> same prefix k- can also introduce a comitative participant, and in this
>> construction the comitative NP shows object properties galore. As a Marind
>> grammarian, I prefer to treat both constructions as subtypes of
>> Applicatives, rather than positing a Instrument Voice vs. a homophonous
>> Comitative Applicative, but with my typologist hat on I wouldn't call the
>> instrument construction an applicative, as no object properties are
>> acquired. So I think the Marind data is similar to what you find in the
>> Bird's Head. It's still in New Guinea, but language contact seems unlikely.
>>
>> Best,
>> Bruno
>>
>> On Tue, Feb 22, 2022 at 7:33 AM Martin Haspelmath <
>> martin_haspelmath at eva.mpg.de> wrote:
>>
>>> What is an "applicative (voice)"?
>>>
>>> There was a thread on this term on the Lingtyp List in October 2018 (
>>> http://listserv.linguistlist.org/pipermail/lingtyp/2018-October/006630.html),
>>> with some of the same participants.
>>>
>>> It seems to me that such misunderstandings will continue as long as we
>>> think that things like "applicative" or "voice" exist independently of how
>>> these terms are defined. They probably don't, because languages are far
>>> more varied, and there's much more uncertainty, than we tend to admit.
>>>
>>> But we can propose clear and simple definitions of "applicative voice" –
>>> and following the 2018 discussion, I wrote this paper on valency and voice
>>> constructions where I defined applicative as a verb-coded voice alternant
>>> with A and P in an alternation whose other alternant has an S corresponding
>>> to the A, and P possibly corresponding to an oblique (§11.7):
>>>
>>> https://ling.auf.net/lingbuzz/005941
>>>
>>> Once we have clear definitions, we can begin to answer David's question
>>> whether languages with instrumental applicatives only are rare outside of
>>> Austronesian. (Polinsky 2005/2013 found "instrument applicatives" in 29
>>> languages, but I'm not sure what her definition was:
>>> https://wals.info/feature/109B)
>>>
>>> Best,
>>> Martin
>>>
>>> Am 22.02.22 um 05:40 schrieb Matthew Dryer:
>>>
>>> Mark says
>>>
>>>
>>>
>>> *However, every text study of either passives or applicatives, or
>>> non-core philippine-type voice choice, shows that there is a degree of
>>> pragmatic prominence associated with the use of these valency-rearranging
>>> operations. We might re-phrase the passive and applicative
>>> characterisations as*
>>>
>>>
>>>
>>> *passive promotes pragmatically-prominent object to subject (and demotes
>>> initial less-prominent subject to non-core)*
>>>
>>> *applicative promotes pragmatically-prominent (oblique?) to object
>>> (might demote initial (less-prominent?) object to non-core)*
>>>
>>>
>>>
>>> However, there are applicatives which are "obligatory" for a given
>>> semantic type. In one lg I work on, Ktunaxa, all applicatives are
>>> obligatory in the sense that the only way to express a benefactive,
>>> instrumental, or comitative is to use the relevant applicative. In Walman,
>>> the only way to express a benefactive is to use the applicative. Since they
>>> are obligatory, there really isn't any pragmatic prominence associated with
>>> these applicatives.
>>>
>>>
>>>
>>> Matthew
>>>
>>>
>>>
>>> *From: *Mark Donohue <mhdonohue at gmail.com> <mhdonohue at gmail.com>
>>> *Date: *Monday, February 21, 2022 at 10:24 PM
>>> *To: *David Gil <gil at shh.mpg.de> <gil at shh.mpg.de>
>>> *Cc: *Matthew Dryer <dryer at buffalo.edu> <dryer at buffalo.edu>,
>>> "lingtyp at listserv.linguistlist.org" <lingtyp at listserv.linguistlist.org>
>>> <lingtyp at listserv.linguistlist.org> <lingtyp at listserv.linguistlist.org>
>>> *Subject: *Re: [Lingtyp] query: instrument voice
>>>
>>>
>>>
>>> Hi David,
>>>
>>>
>>>
>>> The issue, as I see it, is what we mean by 'promote'.
>>>
>>> We can agree that
>>>
>>>
>>>
>>> passive promotes object to subject (and demotes initial subject to
>>> non-core)
>>>
>>> applicative promotes (oblique?) to object (might demote initial object
>>> to non-core)
>>>
>>>
>>>
>>> (and the philippine voice is something like " … promotes (anything) to
>>> subject (and doesn't demote initial subject to non-core)
>>>
>>>
>>>
>>> However, every text study of either passives or applicatives, or
>>> non-core philippine-type voice choice, shows that there is a degree of
>>> pragmatic prominence associated with the use of these valency-rearranging
>>> operations. We might re-phrase the passive and applicative
>>> characterisations as
>>>
>>>
>>>
>>> passive promotes pragmatically-prominent object to subject (and demotes
>>> initial less-prominent subject to non-core)
>>>
>>> applicative promotes pragmatically-prominent (oblique?) to object (might
>>> demote initial (less-prominent?) object to non-core)
>>>
>>>
>>>
>>> We typically describe applicatives as involving just the grammatical
>>> function change. Thus, we have examples like this cited for Indonesian
>>> (from Shiohara 2012):
>>>
>>>
>>>
>>>    1. (2)a  *Pelayan mengambil segelas air*.
>>>
>>> waiter AV.take a.glass.of water
>>>
>>> ‘The waiter took a glass of water.’
>>>
>>>    1. (2)b  *Pelayan mengambil-kan tamu segelas air*.
>>>    waiter AV.take-APPL guest a.glass.of water
>>>    ‘The waiter brought the guest a glass of water.’ (Sneddon 1996: 80)
>>>
>>> As Susanna Cummings showed, however, (2)b examples are not really
>>> attested in naturalistic discourse; rather, we have examples like the
>>> following:
>>>
>>>
>>>
>>> (2)c  Tamu di-ambil-kan segelas air (oleh pelayan).
>>>
>>>          guest NONACTIVE-take-APPL a.glass.of water by waiter
>>>
>>>          'The waiter brought the guest a glass of water.'
>>>
>>>
>>>
>>> (See also Donohue 2001 for similar data from Tukang Besi.)
>>>
>>>
>>>
>>> So, this shows that (in some languages) the increased prominence of the
>>> argument that was sufficient to merit coding with an applicative
>>> construction is also sufficient to merit a non-active voice choice, with
>>> all that entails. A Philippine-type voice system by stealth, as it were.
>>>
>>>
>>>
>>> The Tzutujil example has an applicative suffix; and it also has a verb
>>> with 3SG absolutive agreement (Ø), not 1SG (the in- in the first example I
>>> posted). It also has the requirement that there must be overt coding of the
>>> increased prominent of the instrument; like Indonesian, it does that by
>>> utilising existing high-prominence coding strategies; unlike Indonesian, it
>>> does that not by using a voice change, but by using a pragmatically-marked
>>> word order choice.
>>>
>>>
>>>
>>> -Mark
>>>
>>>
>>>
>>> Donohue, Mark. 2001. Coding choices in argument structure: Austronesian
>>> applicatives in texts. *Studies in Language*25 (2): 217-254.
>>>
>>>
>>>
>>>
>>>
>>>
>>>
>>>
>>>
>>>
>>>
>>>
>>>
>>> On Tue, 22 Feb 2022 at 13:53, David Gil <gil at shh.mpg.de> wrote:
>>>
>>> Mark,
>>>
>>> Thanks for the Tzutujil example, which is indeed quite similar to the
>>> New Guinea constructions I have been looking at.
>>>
>>> However, I remain unconvinced with regard to its characterization as an
>>> applicative, though to a certain degree this is a mere terminological
>>> question.  Prototypically, applicatives promote to direct objects while
>>> passive voices (such as instrumental) promote to subjects — so, for any
>>> given construction, the question is whether the relevant argument, here the
>>> instrumental one, is more direct-object-like or more subject-like.
>>>
>>> This begins to remind me of the seemingly endless ongoing debates over
>>> whether Philippine voice constructions are "really" passives or perhaps
>>> something else, the question generally boiling down to whether the relevant
>>> argument is more like a subject or more like a topic.  Personally, I don't
>>> find these debates very productive, and I'm not sure how useful an
>>> analogous debate between applicative and instrumental-voice labels would
>>> prove to be in this case.
>>>
>>> What's important is to have a clear description of the facts, and how
>>> the constructions in question differ from prototypical applicatives and
>>> from prototypical instrumental voice constructions — with the proviso that
>>> there are perhaps not sufficiently many of the latter to construct a clear
>>> notion of what is prototypical.
>>>
>>> David
>>>
>>>
>>>
>>> On 22/02/2022 04:26, Mark Donohue wrote:
>>>
>>> I would agree with Matthew that these are best described as
>>> applicatives, but ones in which the 'pragmatic advancement' function
>>> monitored by an applicative is, in addition to the grammatical function
>>> coding changes, also required to be monitored by the use of a pragmatically
>>> marked word order.
>>>
>>>
>>>
>>> Very similar facts are found in Tzutujil, in which the applicative,
>>> which indicates an instrumental role (despite having a morpheme cognate
>>> with the benefactive applicative in other Mayan languages) also requires
>>> the appearance of the instrument object in a preverbal role, which is a
>>> pragmatically marked position in a verb-initial language.
>>>
>>>
>>>
>>> Data from Dayley (1985).
>>>
>>>
>>>
>>> Xinruuch’eyi jaa7 tza7n chee7
>>>
>>> he:hit:me       he    with   stick
>>>
>>> ‘He hit me with a stick.’
>>>
>>>
>>>
>>> Chee7 x(r)uuch’eyb’ei jaa7 inin
>>>
>>> stick    he:hit-with:it     he    1SG
>>>
>>> ‘He hit me with a stick.’
>>>
>>>
>>>
>>> -Mark
>>>
>>>
>>>
>>> On Tue, 22 Feb 2022 at 13:15, David Gil <gil at shh.mpg.de> wrote:
>>>
>>> Matthew,
>>>
>>> The reason I don't call it an applicative is that (in most cases) the
>>> instrumental argument must occur before the verb in a topic-like position.
>>>
>>> This can be illustrated with the Roon instrumental prefix *u-* in the
>>> following examples:
>>>
>>> (1)
>>>
>>> * Eros-i
>>>
>>> t-u-karuk
>>>
>>> ai-i-ya
>>>
>>>
>>>
>>> Eros-pers
>>>
>>> 3sg:anim-instr-chop
>>>
>>> tree-3sg:anim-def
>>>
>>>                   'Eros chopped the tree'
>>>
>>>
>>>
>>> (2)
>>>
>>> I-seref
>>>
>>> kaman
>>>
>>> fa
>>>
>>> Eros-i
>>>
>>> t-u-karuk
>>>
>>> ai-i-ya
>>>
>>>
>>>
>>> 1sg-look.for
>>>
>>> axe
>>>
>>> for
>>>
>>> Eros-pers
>>>
>>> 3sg:anim-instr-chop
>>>
>>> tree-3sg:anim-def
>>>
>>>                   'I'm looking for an axe for Eros to chop the tree
>>> with'
>>>
>>>
>>>
>>> (3)
>>>
>>> I-seref
>>>
>>> kaman
>>>
>>> Eros-i
>>>
>>> t-u-karuk
>>>
>>> ai-i-ya-ri-ya
>>>
>>>
>>>
>>> 1sg-look.for
>>>
>>> axe
>>>
>>> Eros-pers
>>>
>>> 3sg:anim-instr-chop
>>>
>>> tree-3sg:anim-def-3sg:inan-def
>>>
>>>                   'I'm looking for the axe that Eros chopped the tree
>>> with'
>>>
>>> Sentence (1) is ungrammatical, and cannot be salvaged by adding a
>>> postverbal NP or PP referring to the axe; in this respect it differs from
>>> typical applicative constructions.  In contrast, sentences (2) and (3) are
>>> fine, because the instrumental prefix *u-* is licensed by the preceding
>>> NP *kaman* referring to the axe.  True, this is not exactly the same as
>>> how things work in Philippine languages, but it is more like Philippine
>>> instrumental voice than anything else I can think of (including
>>> applicatives).  In particular, in (3), the instrumental prefix is required
>>> in order to license relativization (in contrast, relativization of other
>>> oblique arguments is zero-marked).  To use Paul Schachter's terminology, in
>>> both (2) and (3), "subjecthood properties" seem to be split between the
>>> agent (which, as you correctly point out, controls agreement) and the
>>> instrument.
>>>
>>> Very similar patterns obtain in the other Austronesian and
>>> non-Austronesian languages that I mentioned, which — given the apparent
>>> rarity of this pattern elsewhere — is strongly suggestive of language
>>> contact.
>>>
>>> David
>>>
>>>
>>>
>>> On 22/02/2022 03:41, Matthew Dryer wrote:
>>>
>>> David,
>>>
>>>
>>>
>>> Why would you not say that the instrumental construction in Meyah,
>>> Sougb, and Hatam is an applicative, since the A rather than the instrument
>>> controls subject agreement?
>>>
>>>
>>>
>>> Matthew
>>>
>>>
>>>
>>>
>>>
>>> *From: *Lingtyp <lingtyp-bounces at listserv.linguistlist.org>
>>> <lingtyp-bounces at listserv.linguistlist.org> on behalf of David Gil
>>> <gil at shh.mpg.de> <gil at shh.mpg.de>
>>> *Date: *Monday, February 21, 2022 at 7:40 PM
>>> *To: *"lingtyp at listserv.linguistlist.org"
>>> <lingtyp at listserv.linguistlist.org> <lingtyp at listserv.linguistlist.org>
>>> <lingtyp at listserv.linguistlist.org>
>>> *Subject: *[Lingtyp] query: instrument voice
>>>
>>>
>>>
>>> Dear all,
>>>
>>>
>>>
>>> In the Austronesian languages of Taiwan, Philippines and Madagascar,
>>> there is a verbal affix that is said to mark "instrument voice"; loosely
>>> speaking, it marks the topic or subject of the clause as bearing the
>>> semantic role of instrument.
>>>
>>>
>>>
>>> Is anybody familiar with similar instrument-voice constructions from
>>> other parts of the world?
>>>
>>>
>>>
>>> The reason I ask is that a similar construction is present also in some
>>> languages of the Bird's Head and Cenderawasih Bay regions of New Guinea,
>>> eg. Biak, Roon, Wamesa and Wooi (Austronesian), and Hatam, Sougb, Meyah and
>>> Moskona (non-Austronesian).  What's curious about this construction is
>>> that, unlike the well-known Austronesian cases, it is the only
>>> morphologically-marked voice in each of the languages in question; there is
>>> no "ordinary" morphological passive construction.  My feeling is that this
>>> construction is quite uncommon cross-linguistically, but I would like to
>>> get a feel for the extent to which this is indeed true.
>>>
>>>
>>>
>>> Thanks,
>>>
>>>
>>>
>>> David
>>>
>>> --
>>>
>>> David Gil
>>>
>>>
>>>
>>> Senior Scientist (Associate)
>>>
>>> Department of Linguistic and Cultural Evolution
>>>
>>> Max Planck Institute for Evolutionary Anthropology
>>>
>>> Deutscher Platz 6, Leipzig, 04103, Germany
>>>
>>>
>>>
>>> Email: gil at shh.mpg.de
>>>
>>> Mobile Phone (Israel): +972-526117713
>>>
>>> Mobile Phone (Indonesia): +62-81344082091
>>>
>>>
>>>
>>> --
>>>
>>> David Gil
>>>
>>>
>>>
>>> Senior Scientist (Associate)
>>>
>>> Department of Linguistic and Cultural Evolution
>>>
>>> Max Planck Institute for Evolutionary Anthropology
>>>
>>> Deutscher Platz 6, Leipzig, 04103, Germany
>>>
>>>
>>>
>>> Email: gil at shh.mpg.de
>>>
>>> Mobile Phone (Israel): +972-526117713
>>>
>>> Mobile Phone (Indonesia): +62-81344082091
>>>
>>>
>>>
>>> _______________________________________________
>>> Lingtyp mailing list
>>> Lingtyp at listserv.linguistlist.org
>>> http://listserv.linguistlist.org/mailman/listinfo/lingtyp
>>> <https://nam12.safelinks.protection.outlook.com/?url=http%3A%2F%2Flistserv.linguistlist.org%2Fmailman%2Flistinfo%2Flingtyp&data=04%7C01%7Cdryer%40buffalo.edu%7C6406f22ffe22461267a608d9f5b2da3a%7C96464a8af8ed40b199e25f6b50a20250%7C0%7C0%7C637810970808922190%7CUnknown%7CTWFpbGZsb3d8eyJWIjoiMC4wLjAwMDAiLCJQIjoiV2luMzIiLCJBTiI6Ik1haWwiLCJXVCI6Mn0%3D%7C3000&sdata=MvA7AGHFJpqrOSDagmZKQQ%2BuKfQRwmJLUyKIj25RLDg%3D&reserved=0>
>>>
>>> --
>>>
>>> David Gil
>>>
>>>
>>>
>>> Senior Scientist (Associate)
>>>
>>> Department of Linguistic and Cultural Evolution
>>>
>>> Max Planck Institute for Evolutionary Anthropology
>>>
>>> Deutscher Platz 6, Leipzig, 04103, Germany
>>>
>>>
>>>
>>> Email: gil at shh.mpg.de
>>>
>>> Mobile Phone (Israel): +972-526117713
>>>
>>> Mobile Phone (Indonesia): +62-81344082091
>>>
>>>
>>>
>>>
>>> _______________________________________________
>>> Lingtyp mailing listLingtyp at listserv.linguistlist.orghttp://listserv.linguistlist.org/mailman/listinfo/lingtyp
>>>
>>>
>>> --
>>> Martin Haspelmath
>>> Max Planck Institute for Evolutionary Anthropology
>>> Deutscher Platz 6
>>> D-04103 Leipzighttps://www.eva.mpg.de/linguistic-and-cultural-evolution/staff/martin-haspelmath/
>>>
>>> _______________________________________________
>>> Lingtyp mailing list
>>> Lingtyp at listserv.linguistlist.org
>>> http://listserv.linguistlist.org/mailman/listinfo/lingtyp
>>>
>>
>> _______________________________________________
>> Lingtyp mailing listLingtyp at listserv.linguistlist.orghttp://listserv.linguistlist.org/mailman/listinfo/lingtyp
>>
>> --
>> David Gil
>>
>> Senior Scientist (Associate)
>> Department of Linguistic and Cultural Evolution
>> Max Planck Institute for Evolutionary Anthropology
>> Deutscher Platz 6, Leipzig, 04103, Germany
>>
>> Email: gil at shh.mpg.de
>> Mobile Phone (Israel): +972-526117713
>> Mobile Phone (Indonesia): +62-81344082091
>>
>>
>> _______________________________________________
>> Lingtyp mailing list
>> Lingtyp at listserv.linguistlist.org
>> http://listserv.linguistlist.org/mailman/listinfo/lingtyp
>>
> --
> David Gil
>
> Senior Scientist (Associate)
> Department of Linguistic and Cultural Evolution
> Max Planck Institute for Evolutionary Anthropology
> Deutscher Platz 6, Leipzig, 04103, Germany
>
> Email: gil at shh.mpg.de
> Mobile Phone (Israel): +972-526117713
> Mobile Phone (Indonesia): +62-81344082091
>
>
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://listserv.linguistlist.org/pipermail/lingtyp/attachments/20220222/3202c4ee/attachment.htm>