[Lingtyp] query: instrument voice

[David Gil]

Matthew,

I plead guilty to not being clear enough with my original query; my 
partial excuse is that things were not sufficiently clear in my mind, 
and I was hoping that giving my tentative thoughts an airing in this 
group would help me understand things better.Thanks to you and some of 
the other participants in this thread, I think I am now in a position to 
offer a clearer formulation of what I am after.

Definition: Bilateral Instrumental Verbal Marking

Bilateral Instrumental Verbal Marking is a construction exhibiting the 
following four features:

(1) A verbal marking, either morphological or periphrastic (by means of 
a separate word occurring in close nexus with the verb).

(2) The verbal marking denotes an argument of the verb bearing the 
semantic role of instrument.

(3) The instrument argument of the verb behaves differently from the 
corresponding instrumental arguments of verbs not bearing Bilateral 
Instrumental Verbal Marking; in particular, it occurs "higher" on a 
Grammatical Relations hierarchy such as Subject > Object > Oblique.

(4) The argument bearing the semantic role of instrument is also 
understood as bearing a second semantic role with respect to a second 
verb, that may be either overt or understood.

While features (1) - (3) are shared with instrumental voice and 
applicative constructions, it is feature (4), expressed by the term 
"Bilateral", that distinguishes the Bilateral Instrumental Verbal 
Marking construction from instrumental voice and applicative 
constructions.Feature (4) accounts for the fact that, as far as I have 
been able to ascertain, most or all of the instances of Bilateral 
Instrumental Verbal Marking occur in constructions of one or more of the 
following two general types:

(a) Serial-verb-like constructions (such as the Hatam example in one of 
Matthew's earlier messages).

(b) Embedded clauses (such as the Roon complement clause and relative 
clause in one of my earlier messages).

However, the definition provided above sidesteps some of the thorny 
issues that came up in the discussion, or could come up in the future, 
such as the grammatical role of the instrument NP, the relative status 
of the two verbs in the serial verb construction, and in the case of 
bi-clausal constructions, the precise nature of the relationship between 
the two clauses.

Of course, like with any other definition, there will be borderline 
cases. One obvious source for such borderline cases will be that in 
which the "second verb" stipulated in (4) is covert and/or semantically 
bleached. For example, this might be the case for many serial verb 
constructions in which the second verb (often the first in linear order) 
has a general meaning such as 'take' or 'use'.

The Tzutujil example cited by Mark and discussed further by Bill might 
also be such a borderline case.Ostensibly there is only one verb 'hit'; 
however, Bill's proposed translation 'It's the stick that he hit me 
with' actually introduces a second semantically vacuous verb 'is'.Though 
I'm not sure I would wish to go down this path.On the other hand, the 
Marind example, as described by Bruno, doesn't present any obvious 
evidence for such a second verb and second semantic role, and therefore 
doesn't seem to qualify as a case of Bilateral Instrumental Verbal Marking.

A final note on instrumental voice in Tagalog (as a typical Philippine 
language).Using English words for expository simplicity, instrumental 
voice is typically presented with examples such as the following

(1) INSTR-CUT ng CHILD ng BREAD ang KNIFE

'The child cut bread with the knife'

where /ang/ marks the subject/topic indexed by the instrumental prefix 
on the verb, and /ng/ the other non-oblique arguments.Such examples 
clearly fail to satisfy feature (4) above and therefore do not qualify 
asBilateral Instrumental Verbal Marking.However, in real naturalistic 
discourse, examples such as (1) are infrequent; much more common are 
examples such as

(2) HAVE INSTR-CUT ng BREAD 1SG

'I've have something to cut bread with'

where the understood item is both the instrument of CUT and also the 
theme of HAVE. And thereby satisfying feature (4).Facts such as these 
seem to suggest that even Tagalog instrumental voice may actually lie 
somewhere in-between a pure voice construction and a Bilateral 
Instrumental Verbal Marker.Analogous observations hold also for other 
oblique voices such as locative voice, thereby pointing perhaps to a 
more general notion of Bilateral Verbal Marking of which Bilateral 
Instrumental Verbal Marking is but one particular case.

Thanks to Matthew and all of you who led me towards a hopefully more 
perspicuous definition of the construction in question.

David

On 22/02/2022 23:03, Matthew Dryer wrote:
>
> David,
>
> Part of the problem is that your original query did not specify in 
> detail what exactly it is that characterizes these constructions, what 
> it is that you think that they share that is crosslinguistically 
> unusual. Your email only characterized the constructions as similar to 
> Philippine-type instrumental voice and consulting sources on some of 
> the non-Austronesian languages you mentioned, what I found was more 
> like an applicative than a Philippine-type instrumental voice. Your 
> original email seemed to be asking for other examples of 
> Philippine-type instrumental voice, but it is now clear that what you 
> are looking for is a constellation of features that may be an areal 
> feature of the Bird's Head and adjacent areas and it is only in your 
> latest email that it begins to emerge what that constellation of 
> features is.
>
> But I acknowledge that the Hatam construction and others like it are 
> definitely not prototypical applicative constructions, precisely 
> because they involve a type of serial verb construction. From where I 
> am sitting, what seems unusual about the construction is that it 
> involves a combination of a serial verb construction with an 
> applicative construction, without much similarity to Philippine-type 
> instrumental voice. But if your interest in finding out if there are 
> other languages with similar constructions, the issue of whether it is 
> similar to Philippine-type instrumental voice seems irrelevant.
>
> Perhaps if you spelled out more clearly what the constellation of 
> features is that you see as characterizing the construction, people 
> might be able identify other cases. My hunch is that what you are 
> looking for is rare outside your area if it exists at all.
>
> Matthew
>
> *From: *David Gil <gil at shh.mpg.de>
> *Date: *Tuesday, February 22, 2022 at 3:29 PM
> *To: *Matthew Dryer <dryer at buffalo.edu>, 
> "lingtyp at listserv.linguistlist.org" <lingtyp at listserv.linguistlist.org>
> *Subject: *Re: [Lingtyp] query: instrument voice
>
> Matthew,
>
> The construction that I'm interested in here (which does not seem to 
> differ significantly across the genealogical boundary between 
> Austronesian and non-Austronesian) does not seem to be a prototypical 
> case of any familiar construction — which is what makes it interesting 
> to me.  Much of the discussion has focused on the differences between 
> it and Philippine instrumental voice constructions, which I am not 
> denying.
>
> But you can hardly say that we're dealing here with a prototypical 
> applicative either.  What's crucial is that in most or all of the 
> languages under consideration, the instrument NP cannot occur in 
> post-verbal position, which is where you'd expect it to be in an 
> applicative construction in an SVO language. Thus, for the 
> corresponding prefix /k-/ in Austronesian Biak, van Heuvel (2006:420) 
> writes that "it seems to be used only when this instrument is topical" 
> — which is kind of the opposite of how things work in many familiar 
> applicative constructions.  Call it what you like (topic, subject, 
> whatever), but the grammatical functions and behaviour associated with 
> the instrument NP are very different not only from those of the 
> corresponding NPs in clauses without instrumental verbal marking, but 
> also from those of instrument NPs in other languages with an 
> instrumental applicative marker.
>
> As your Hatam example suggests, there is also an affinity between the 
> construction in question and serial verb constructions. Peel off the 
> morphology and what you've got is a garden-variety Mainland Southeast 
> Asian language SVC construction along the lines of TAKE STICK HIT 
> SNAKE. Alternatively, transform your Hatam inflectional forms to 
> periphrastic and you get the corresponding construction in isolating 
> Papuan Malay
>
> sa
>
> 	
>
> ambil
>
> 	
>
> kayu
>
> 	
>
> sa
>
> 	
>
> pake
>
> 	
>
> pukul
>
> 	
>
> ular
>
> 1SG
>
> 	
>
> take
>
> 	
>
> stick
>
> 	
>
> 1SG
>
> 	
>
> use
>
> 	
>
> hit
>
> 	
>
> snake
>
> 'I hit the snake with a stick'
>
> where /pake/ 'use' is the periphrastic counterpart of the instrumental 
> verbal prefix in Hatam, Biak, etc.  (This construction is unavailable 
> in other varieties of Malay, which suggests that it is due to 
> substrate influence from the local New Guinea languages.)
>
> I would conclude that the construction in question bears certain 
> family resemblances to instrumental voice constructions, applicatives, 
> and serial verb constructions, but is not a prototypical instance of 
> any of these.  Given its recurrence in (at least) three genealogically 
> unrelated families of languages (Austronesian, East Bird's Head, and 
> isolate Hatam), what this discussion seems to me to be suggesting is 
> that the construction in question merits a term all to its own, so 
> that its relationship to other constructions can be productively 
> discussed.
>
> David
>
> On 22/02/2022 21:20, Matthew Dryer wrote:
>
>     David,
>
>     Preverbal position in an SVO language seems to me to be a very
>     weak factor as a subject property. There are two additional
>     overlapping considerations that would normally be considered
>     relevant. First, is the noun phrase in question in the same
>     preverbal position as subjects? And second, does the S/A lack
>     subject properties that it normally has.
>
>     Without these two additional considerations, it would seem that
>     one would have to say that /what/ in English /What is John
>     eating?/ is subject-like, since it is a preverbal constituent in
>     an SVO language. But it does not occur in the same preverbal
>     position as subjects and the subject does not lack its normal
>     subject properties. The same could be said about /rice/ in /It is
>     rice that John is eating/.
>
>     You ask why some of us are talking about applicatives in their
>     responses. One reason is that you cite Hatam, Sougb, Moskona, and
>     Meyah as instances of what you are characterizing as constructions
>     like Philippine instrumental voice. But these seem much more like
>     canonical applicatives and quite unlike Philippine instrumental voice.
>
>     In the following example from Hatam, for example,
>
>     /Ni-ba/
>
>     	
>
>     /tom/
>
>     	
>
>     /ni-bi-bui/
>
>     	
>
>     /wou./
>
>     1EXC-use
>
>     	
>
>     stick
>
>     	
>
>     1EXC-INS-hit
>
>     	
>
>     snake
>
>     We used a stick to hit the snake. (Reesink 1999: 54)
>
>     the fact that /tom/ 'stick' precedes the verb for 'hit' is
>     presumably best explained in terms of its being the complement of
>     /ba/ 'use' and there is no evidence that the A of 'hit' lacks any
>     normal subject properties. This is very different from
>     instrumental voice in Philippine languages.
>
>     Matthew
>
>     *From: *Lingtyp <lingtyp-bounces at listserv.linguistlist.org>
>     <mailto:lingtyp-bounces at listserv.linguistlist.org> on behalf of
>     David Gil <gil at shh.mpg.de> <mailto:gil at shh.mpg.de>
>     *Date: *Tuesday, February 22, 2022 at 7:42 AM
>     *To: *"lingtyp at listserv.linguistlist.org"
>     <mailto:lingtyp at listserv.linguistlist.org>
>     <lingtyp at listserv.linguistlist.org>
>     <mailto:lingtyp at listserv.linguistlist.org>
>     *Subject: *Re: [Lingtyp] query: instrument voice
>
>     Dear all,
>
>     I must confess to being a little puzzled at how the responses to
>     my original query seem to have focused largely on applicatives. 
>     To cite just one example ...
>
>     On 22/02/2022 08:31, Martin Haspelmath wrote:
>
>         Once we have clear definitions, we can begin to answer David's
>         question whether languages with instrumental applicatives only
>         are rare outside of Austronesian.
>
>     A fair question, but not the one that was asking; I was asking
>     whether languages with *instrument voice* only are rare outside of
>     Austronesian.  Actually, what I really meant to ask is whether
>     constructions like those in Roon and other proximate languages are
>     attested elsewhere in the world; that is to say, constructions in
>     which a verb hosts an affix denoting an instrument whose function
>     in the clause looks more like a subject or topic than like a
>     direct object or oblique.  I used the term "instrument voice"
>     because this seemed to me to be the most appropriate term, or, to
>     put it differently, the constructions i am looking at seemed to me
>     to be more similar to, say, a garden-variety instrument-voice
>     construction in Tagalog, than anything else I could think of,
>     including most prototypical applicative constructions.  In
>     response to my query, Mark came through with the Tzutujil example,
>     and one or two others have provided potential leads that I will be
>     following up on soon.
>
>     But my choice of terms led to a terminological debate, with
>     several of you expressing your opinions that the constructions in
>     question, in Roon and other New Guinea languages, are instances of
>     applicatives. To which I would respond with a question: would you
>     also characterize a Philippine-type instrumental voice
>     construction as an applicative?
>
>     I wouldn't, which is why I phrased the question in the way that I
>     did.  Note that I would still acknowledge the merits of a
>     sometimes-proposed analysis of Philippine voice in which, say, the
>     instrumental voice is analyzed compositionally as consisting of
>     (a) an applicative "promoting" oblique to direct object; in
>     combination with (b) a passive "promoting" a direct object to
>     subject.  But under such an analysis, while an applicative
>     construction *forms part of* the instrument voice construction,
>     the instrument voice construction as a whole is more than just an
>     applicative.  (As Mark points out, a similar analysis is clearly
>     called for in the case of Indonesian, in which passive /di-/ and
>     applicative /-kan/ frequently co-occur.)  However, in the New
>     Guinea case, there is no evidence that I am aware of for such a
>     compositional analysis; the prefixes that express what I was
>     calling instrumental voice provide no evidence for any kind of
>     complex internal structure.  Indeed, for this reason,
>     constructions such as those with the Roon /u-/ prefix seem to me
>     to offer "better" examples of "instrument voice" than even the
>     Philippine constructions for which the term was originally coined.
>
>     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

-- 
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/20220223/b53abe2b/attachment.htm>