Tense, aspect and mood in Nêlêmwa (New Caledonia)
Encoding events, processes and states
Isabelle Bril
To cite this version:
Isabelle Bril. Tense, aspect and mood in Nêlêmwa (New Caledonia) Encoding events, processes
and states. Zlatka Guentchéva Aspectuality and Temporality. Descriptive and theoretical issues, 172, Benjamins, pp.63-107, 2016, Studies in Language Companion Series, 9789027259370
10.1075/slcs.172.03bri. halshs-01418754
HAL Id: halshs-01418754
https://halshs.archives-ouvertes.fr/halshs-01418754
Submitted on 2 Mar 2017
HAL is a multi-disciplinary open access
archive for the deposit and dissemination of scientific research documents, whether they are published or not. The documents may come from
teaching and research institutions in France or
abroad, or from public or private research centers.
L’archive ouverte pluridisciplinaire HAL, est
destinée au dépôt et à la diffusion de documents
scientifiques de niveau recherche, publiés ou non,
émanant des établissements d’enseignement et de
recherche français ou étrangers, des laboratoires
publics ou privés.
John Benjamins Publishing Company
This is a contribution from Aspectuality and Temporality. Descriptive and theoretical issues.
Edited by Zlatka Guentchéva.
© 2016. John Benjamins Publishing Company
This electronic file may not be altered in any way.
The author(s) of this article is/are permitted to use this PDF file to generate printed copies to
be used by way of offprints, for their personal use only.
Permission is granted by the publishers to post this file on a closed server which is accessible
to members (students and staff) only of the author’s/s’ institute, it is not permitted to post
this PDF on the open internet.
For any other use of this material prior written permission should be obtained from the
publishers or through the Copyright Clearance Center (for USA: www.copyright.com).
Please contact rights@benjamins.nl or consult our website: www.benjamins.com
Tables of Contents, abstracts and guidelines are available at www.benjamins.com
Tense, aspect and mood in Nêlêmwa
(New Caledonia)
Encoding events, processes and states
Isabelle Bril
lacito-CNRS, Fédération “Typologie et Universaux Linguistiques”
Nêlêmwa is an Aspect-Mood oriented language; verbs are unmarked for tense,
time reference is expressed by chronology and time adverbs. Aspect hinges
on three notions: events (in the aorist), states, and processes. One focus is the
contrast between the perfect and the aorist. Bare aorist verb forms refer to
events or to sequences of events with no reference to their internal phases. The
perfect expresses internal relations between processes and clauses (anteriority,
backgrounding, causal relations); it refers to transitional processes that have
reached or not their final instant, expressing changes of states and resulting states.
In future reference frames, the perfect expresses imminent change of states,
or imminent completion of a process, and the speaker’s certainty about their
projected occurrence.
Keywords: Tense, mood, aspect, Aktionsart, topology, state, event, process, aorist,
perfect, chronological ordering, deixis, anaphora, directional
. Introduction: A brief presentation of Nêlêmwa
Nêlêmwa and its variant Nixumwak are one of the 28 languages of New-Caledonia,
which belong to the Oceanic subgroup of the Austronesian family. NêlêmwaNixumwak is spoken in the northernmost part of the Mainland by approximately 900
speakers.
Nêlêmwa is a head-initial language with predicate-argument order. In verbal
predication, Nêlêmwa has a split accusative/ergative system: it is accusative with subject/object pronouns (with sVo order), but post-verbal nominal arguments are marked
as absolutive and ergative with < verb abs.patient erg.agent > order. Single nominal
arguments of intransitive verbs are absolutive, with < V abs.N > order. Absolutive is
marked ø; ergative markers are ea- (+ genitive animate NP) and ru (for inanimates)
(Bril 1997, 2002: 134–142). In non-verbal predication, which is restricted to equative
./slcs..bri
© John Benjamins Publishing Company
Isabelle Bril
and attributive sentences, non-verbs (nouns, pronouns, numeral phrases, locative
phrases) may be predicative without a copula, without any derivation or conversion,
and without any semantic change, other than the expected change triggered by the
predicative function (Bril, in press). Like verbal predicates, non-verbal predicates may
combine with tense, aspect or mood (T.A.M.) markers and may be subject to timeinstability, transitional stages or duration. Consider the following non-verbal predicates: i âlô ‘he’s a child’ (lit. he child), i baa âlô (lit. he dur child) ‘he’s still a child’, io
i ak (lit. fut he man) ‘he’ll become a man’, i u thaamwa (lit. she pft woman) ‘she’s
become a woman’.
Tense, aspect or mood are not marked by conjugation or inflectional verbal morphology; the T.A.M. system is expressed by morphemes that mostly occur before the
(verbal or non-verbal) predicate and which are organised as follows:
(irrealis mood) neg (subject index) (aspect) (modality) predicate.
Irrealis mood includes virtual (subsuming conditional and subjunctive uses),
imperative, optative-hortative mood, and future.
The analysis will proceed as follows: Section 2 presents the theoretical framework and
the terminology used; Section 3 gives an overview of the characteristic features of the
system: temporal frame of reference, realis vs. irrealis mood, aspect markers; Section 4
details the irrealis mood system; Section 5 details the various uses of the perfect, in
the realis and irrealis domains, and the distinctive uses of the perfect in contrast with
the bare verb form; Section 6 deals with the morphemes expressing phasal aspects;
Section 7 presents phasal verbs expressing lexical Aktionsart; Section 8 develops the
aspect-marking role of some directionals; Section 9 concludes.
. Theoretical framework and terminology
This analysis of T.A.M. in Nêlêmwa makes use of the theoretical framework developed by Culioli (1990, 1999a–b), Desclés (1989, 1995), Desclés and Guentchéva (2010,
this volume), which is based on a topological model with intervals and boundaries
representing the notions of event, process and state as they are encoded in the morphosyntax of languages.1 After Comrie (1976), Lyons (1990), Desclés & Guentchéva
(2012: 134–143), these terms are defined as follows:
. I wish to thank Zlatka Guentchéva, Alexandre François and Frank Lichtenberk for insightful comments which led me to refine some definitions and to some reorganisation of the
paper.
© . John Benjamins Publishing Company
All rights reserved
Tense, aspect and mood in Nêlêmwa (New Caledonia)
–
–
–
Events are viewed as a whole, without reference to any internal phase. They are
represented by closed intervals. Depending on verb semantics and verb classes,
an event may be punctual [•], like it broke, or not, like he sang ([–––]). Events in
Nêlêmwa are encoded by a bare verb form called aorist, presented in §3.3.1
By contrast with events, processes contain internal phases which may be under
some aspectual focus; they have an initial instant and a final instant, which may
or may not be reached. An ongoing process, whose final instant is not reached, is
represented by a semi-open interval ([–––[), with an open right boundary (boldface bracket). If reached (as in it has rained), the final instant of a process is represented by a closed right boundary ([–––]). Processes in Nêlêmwa are marked by a
wealth of aspect markers presented in §3.3.2 to §3.3.4.
States have no boundaries of their own, they are represented by open intervals
schematised as (]–––[), but the boundaries do not belong to the state itself, they
belong to some events (A and B) which frame the state ([ A ]–––[ B ]). A resulting state such as it’s broken can be represented as [ A ]–broken–[B]: the state
results from a preceding event [ A ] (such as fall, break) conducive to the resulting state (broken). The open (boldface) boundary (––[ B ]) belongs to some other
event [ B ], which may trigger another change of state (Desclés & Guentchéva
2012: 141–142).
By contrast with some other Oceanic languages, states in Nêlêmwa have no specific
marker, nor any specific paradigm of stative pronouns. Stative verbs appear without
further marking in the realis mood, as in i alawuk ‘(s)he is/was hungry’, or with the
irrealis mood (io i alawuk ‘(s)he’ll be hungry’, o i alawuk ‘if/when (s)he’s hungry’).
The combination with the perfect marker expresses a change of state as in i u alawuk
‘he’s become hungry’. States may also combine with other aspects, such as the durative
marker i baa alawuk ‘he’s still hungry’ or the remansive marker, as in taa khoxo i hla
‘they’re still numerous’ (rem big.number det they).
.
Tense, mood and aspect: An overview
This section gives a general presentation of the T.A.M. system in Nêlêmwa, which will
be detailed in the following sections.
.
Temporal frame of reference and chronological ordering of events
Predicates are not marked for Tense. The temporal frame of events (T) is generally
marked contextually or lexically, by adverbs or by determiners, some of which are
deictic (i.e. anchored in the time of speech (t₀) and speech situation, noted [T=t₀]),
© . John Benjamins Publishing Company
All rights reserved
Isabelle Bril
while others are non-deictic and relative to some other temporal frame, in narratives
for instance, with no relevance to the time of speech. This is noted [T ω t₀].2
Among deictic adverbs are koobwan ‘yesterday’, ereek ‘yesterday night’, caae
‘tomorrow’, êna ‘now’. They are the sole temporal anchors relative to the time of
speech (t₀).
Caae requires an irrealis marker (future or virtual) in the clause, while adverbs
referring to the past (koobwan, ereek) or the present (êna) often occur with the bareaorist verb form that refers to events, as in (1a–b).
(1)
a.
Kot koobwan ?
rain yesterday
‘Did it rain yesterday?’
b.
Koobwan xe yo tuâi-na.
yesterday tpc 2sg lie-1sg
‘Yesterday, you lied to me.’
(Bril 2002: 486)
Anaphoric determiners, such as bai ‘that’, and adverbs, such as êbai ‘before, earlier on’,
refer to the recent or distant past and to shared experience/knowledge. The bare aorist
verb udi in (2) refers to a past event, while the perfect in (3) marks relevance of some
past event to the time of speech.
(2)
… idaama-m
bai
co udi
êbai na mwamaidu.
eye-poss.3sg that.anaph 2sg take.off before loc down.there
‘… those eyes of yours that you took off earlier on down there.’
(Bril 2002: 492)
(3)
Na u diya êbai.
1sg pft do before
‘I’ve done it before.’
Lexical specifiers, such as hule, express past anchor; hule may be used as a noun or as
a modifier with deictic or non-deictic use, and meaning ‘long time ago/before’. Habuk
‘before’ is a modifier expressing relative chronology in the past:
(4)
ni yeewat habuk (or) ni yeewat hule
in time before
in time long.ago/before
‘in the past’
‘long ago/before’
. The symbol ω encodes disconnection from the time of speech (Culioli 1983, 1999a),
whereas the symbol ≠ encodes a temporal frame anchored in the past, but relevant to the time
of speech (like the English present-perfect), as in he’s been reading for an hour.
© . John Benjamins Publishing Company
All rights reserved
Tense, aspect and mood in Nêlêmwa (New Caledonia)
Events or processes are chronologically ordered by mostly spatio-temporal lexemes
that are compatible with realis or irrealis moods (see §5.5).
.
Irrealis and realis moods
As often in Oceanic languages, the basic mood distinctions hinges on the contrast
between realis and irrealis.
Table 1. Realis vs. irrealis moods in Nêlêmwa
irrealis
virtual/subjunctive
o
realis
imperative
optative/hortative
future
present or past event
o (or) Ø
(sho) o
io
Ø
While the realis mood has no overt marker (Ø), the irrealis mood subdivides into
the virtual mood marked by o, and the future marked by io. The morpheme o has a
variety of uses. In dependent clauses, the morpheme o subsumes conditional and subjunctive moods: it appears in conditional protases, in irrealis time clauses, complement
clauses and irrealis relative clauses. In independent sentences, o appears in imperative
sentences expressing a doubtful outcome, while unmarked imperatives with the irrealis Ø bare form refer to situations whose outcome is deemed possible; the morpheme
o also expresses hortative and optative moods, mostly (but not obligatorily) combined
with sho o (lit. good if) (see §4.2). The uses of o in dependent clauses might be considered divergent developments from the irrealis mood marker, but they actually express
the basic semantics of irrealis (potential, conditional and subjunctive).
. An overview of aspect markers: Combinations and position
Turning now to aspect markers, the main contrast is between the bare aorist3 verb
form (which is aspectless) and the perfect; other morphemes express recent completion, pluri-actionality and various types of incompletion (see Table 3). All of them
stand before the predicate.
.. Events in the bare aorist verb form
The bare aorist verb form refers to realis events, present or past (narrative aorist) and
to generic, atemporal statements (gnomic aorist) (see Table 2).
. Ancient Greek aóristos “indefinite” refers to the unmarked verb form.
© . John Benjamins Publishing Company
All rights reserved
Isabelle Bril
Table 2. The bare aorist verb form
generic statement
realis
present event (concomitant to t₀)
past event (narrative aorist)
(gnomic aorist)
Ø aorist bare V form
.. Aspect markers: Perfect, recent completion, pluri-actional, incompletion
By contrast with events marked by bare aorist4 verb forms, processes have internal
phases which are specified by various aspect markers listed in Table 3.
Table 3. Aspect markers: aorist, perfect, recent completion, pluri-actional, incompletion
markers
aorist perfect
Ø
recent
completion
pluri-actional
(k)u ~ (x)u hââ
aa
perfect
iterative frequency processual
durative persistive remansive,
(± realis) (+ past
(+activity V)
& state of
reference)
activity
recent
completion
kua
incompletion
na
baa
gaa
taa₁
A. The perfect
The perfect is marked by ku ~ xu ~ u. Ku or its weak form xu appear in clause initial
position, allomorph u occurs between the subject index (s) and the predicate (P), as
shown below; from now on they will be referred as (k)u ~ (x)u:
< ku ~ xu P >
< (s) u P >
A verb in the perfect may refer to a complete or incomplete process, i.e. which may
or may not have reached its final instant. This also depends on the verb’s telicity.
When completed, a process may have a resulting state, as in he’s cooked the food.
With stative verbs, the perfect indicates a change of state. The perfect will be further
analysed in §5.
B. Recent completion
Hââ marks the recent completion of a process and cannot combine with the perfect
marker. Hââ originates from the impersonal verb hââhuuk ‘be recent, new’:
.
In Nêlêmwa, the realis aorist and the irrealis aorist are both marked Ø.
© . John Benjamins Publishing Company
All rights reserved
Tense, aspect and mood in Nêlêmwa (New Caledonia)
(5)
Hââhuuk paara eli.
be.recent story that.anaph
‘That story is recent.’ (i.e. has just been circulated)
In complex predicate constructions, hââhuuk may be the modifier of another verb,
as in (6a), whereas the aspect morpheme hââ (6b) occurs between the subject and
the predicate. Hââ indicates that the final boundary of the process has recently been
reached ([–––]//) in relation to the time of speech (in discourse) or to some past event
(in narratives).
(6)
a.
Hla diya hââhuuk (vs.) b. hla hââ
diya.
3pl do be.recent
3pl rec.complt do
‘They’ve done it recently/they’ve just done it.’
(7)
I
xam5 hââ
toven.
3sg ass rec.complt finish
‘He’s just finished.’
Hââ does not combine with the perfect (k)u, (x)u, nor with the prospective/future io;
but it may combine with the potential morpheme o. In a narrative (8), the recently
completed process (hââ uya) is relative to a past situation disconnected from the time
of speech.
(8)
Dua
hla hââ
uya hlaabai
when.past 3pl rec.complt arrive those.anaph
fhavaro-â
xe kia
hoxo
haxixi-hla.
guest-poss.1pl.incl tpc neg.exist someone fear.tr-3pl
‘When our guests had just arrived (for the first time) nobody was afraid of
them.’
C. Already completed event or process: taa₂ ‘already’
The morpheme taa₂ ‘already’ will be analysed in §6.3, together with the remansive taa₁.
They probably stem from a common origin, but have developed different meanings.
Taa₂ marks the prior completion of a process or event. In contrast with hââ, which
excludes any combination with the perfect marker, the prior completion marker taa₂
may combine with the perfect marker (k)u, (x)u to stress anteriority.
(9)
.
Hla u taa the maaxi habuk ma hla ya
mwa.
3pl pft prev reap straw before sub 3pl thatch house
‘They’ve/had already reaped the straw before thatching the house.’
Xam is a discourse-assertive marker.
© . John Benjamins Publishing Company
All rights reserved
Isabelle Bril
D. Pluri-actional markers
Pluri-actionality is encoded by aa (iteration) and kua (frequency + past reference).
This is developed in §6.1.
E. Incompletion markers
Various morphemes express incompletion: the processual marker (na) is restricted
to activity verbs; other markers are durative (baa), persistive (gaa), continuativecontrastive gat, remansive (taa₁). These morphemes will be further detailed in §6.1,
§6.2, §6.3.1.
.. Other aspect markers
Other aspect markers, listed in Table 4 and further detailed in §6.4 to §6.6, focus on
the specific relation to the initial or the final boundary of an event, process or state, but
from an external perspective, since (i) the event or process has not yet occurred (kio
fwâ ‘not yet’ refers to an uninitiated process (see §6.4.1)), or (ii) it has failed to achieve
completion (fwââm ‘almost, nearly’ indicates near, but failed, occurrence of some
event or process (§6.4.3), or (iii) the marker refers to some permanently completed
or bygone event that will not occur again (kio haxa ‘no longer, no more’) (§6.5, §6.6).
Table 4. Other aspect markers
unreached boundary
not yet occurred
nearly reached
boundary, but failed
permanently terminated,
bygone
kio fwâ
‘not yet’
fwââm
‘almost, nearly’
kio haxa
‘no longer, no more’
.. Position of aspect markers
Aspect markers generally occur before the predicate (verbal or non-verbal) and
between the subject index and the predicate if there is a subject index.
(10)
Hli baa âlô.
3du dur child
‘They are still children.’
In (11), the iterative marker aa has its scope on the numeral predicate aaxan ‘three’, the
iterated event is < catching 3 fish >:
(11)
Aa aa-xan shâ-pwee-n.
iter cl-three content-line-poss.3sg
‘He has angled three fish again.’ (lit. it’s again three (fish) his catch of fish)
Only the continuous marker gat, which also has some modal connotations (such
as counter-expectancy), appears in post-predicate position (after the non-verbal
© . John Benjamins Publishing Company
All rights reserved
Tense, aspect and mood in Nêlêmwa (New Caledonia)
predicate thaamwa in (12)). The presupposition in (12) is that someone else than the
woman is expected to speak:
(12)
[Thaamwa=ena gat] [hoona na
vhaa !]
woman=that.dx2 cont that.dx2 proc speak
‘It’s still that woman who is speaking !’ (lit. it’s still that woman the one
speaking)
. Combination of aspect markers
After this general overview of T.A.M. morphemes, a brief survey of some of their combinations is now presented.
.. Combination of the perfect (k)u ~ (x)u and other aspect markers
As shown in Table 5, the perfect cannot combine with the recent completion morpheme hââ, nor does it occur with kio fwâ, which refers to an event or process that has
not yet occurred, and is thus in semantic conflict with the perfect in Nêlêmwa. But it
may combine with taa₂, which marks previousness and can be translated by ‘already,
before’.
The perfect does not combine with incompletion markers (such as processual na
or durative baa), but it combines with persistive gaa, which has modal semantics. It
also combines with pluri-actional aspect markers (iterative and frequency). This will
be further detailed in §6.
hââ
perfect
(k)u ~ (x)u
taa₂
+
not yet
occurred
previous
occurrence
recent
completion
Table 5. Combination of the perfect with other aspect markers
pluri-actional
incompletion
kio fwâ
kua
‘not
aa
frequency na
yet’
iterative (in past)
processual
+
+
baa
durative
gaa
persistive
+
.. Combination of irrealis morphemes with the perfect
The irrealis mood markers, virtual (o) or future (io), may combine with the perfect
marker (k)u ~ (x)u. Irrealis mood markers are in outermost, leftmost position, standing before the whole verb phrase (i.e. the predicate (P) and the subject index (s)).
© . John Benjamins Publishing Company
All rights reserved
Isabelle Bril
allomorph u
(between subject (s) & predicate)
[o < s u P >]
[io < s u P >]
allomorphs ku ~ xu
(no subject expressed)
[o < ku ~ xu P >]
[io < ku ~ xu P >]
The semantics of such associations is analysed in §5.8.
.. Combination of irrealis morphemes with other aspect markers
The irrealis markers may combine with other aspect markers (see Table 6): the main
restrictions concern kio fwâ ‘not yet’, the recent completion marker hââ and the marker
of frequency in the past kua, which are all semantically incompatible with the future io
for obvious semantic reasons.
Table 6. Combination of irrealis markers (io, o) and other aspect markers
pluri-actional
aa
iterative
kua
frequency
(in past)
incompletion
na
processual
baa
durative
gaa
persistive
io
+
+
+
+
o
+
+
+
+
not yet
occurred
recent
completion
kio fwâ
‘not yet’
hââ
just
completed
+
+
When the future io ~ e combines with the locative verb gi ‘be.loc’ and the processual
marker na, which expresses some ongoing activity, as in (13), the meaning is epistemic, expressing near certainty.
(13)
Io gi
ye
na
pwe.
fut be.loc 3sg.fr proc fish
‘He must be fishing.’ (lit. He will be fishing)
Compare with reference to an ongoing present activity: gi ye na pwe ‘he is fishing.’
. Irrealis mood
The two irrealis morphemes (virtual/subjunctive o and future io6 ~ e, (originally ixo,
exo)) express different degrees of epistemic stance and different degrees of certainty or
. Io is the form chosen by the speakers to unify the many variants e, e(x)o, i(x)o. This morpheme probably originates from a compound form e- (presentative) + -xo (anaphoric suffix
referring to unknown, non-referential entities for the speaker).
© . John Benjamins Publishing Company
All rights reserved
Tense, aspect and mood in Nêlêmwa (New Caledonia)
probability relative to the occurrence of the projected event. They are mutually exclusive in the same clause, but may occur in two different clauses within a sentence. They
stand in the leftmost position before the whole predicate phrase (subject index (s) and
predicate (p)):
virtual: o < (s) P >
future: io ~ e < (s) P >
(14)
Io na diya.
fut 1sg do
‘I’ll do it.’
. Future, prospective: Io ~ e
io ~ e < (s) P >
The future morpheme io ~ e is compatible with close or distant future, depending on
the context and on other co-occurring morphemes. Imminent future is expressed by
combining io ~ e with the adverb êna ‘now’:
(15)
Io i
uya êna.
fut 3sg arrive now
‘He’ll arrive soon/in a moment.’
(16)
Na io na faxi êna.
and fut 1sg cook now
‘and I’m going to cook it now.’
(Bril 2002: 493)
. The virtual marker o
The uses of the virtual morpheme o in independent sentences (imperative, hortative,
optative) and in dependent clauses (conditional protases, irrealis complement and
relative clauses) is now presented.
.. Virtual marker o in imperative, hortative and optative sentences
A. Imperative
Unmarked imperatives display the bare verb form,7 while o appears in marked imperatives expressing a stronger injunction, as in o taauri! ‘do wait !’; o maya! ‘do be patient !,
wait !’; o khomi! ‘do count !’.8
. In Mwotlap, imperative is one of the many functions of the aorist marker (François
2003: 194–195).
. In Palau,“imperative is expressed by the […] hypothetical form” (Hagège 1986; Lemaréchal
1991: 162).
© . John Benjamins Publishing Company
All rights reserved
Isabelle Bril
(17)
(O) kuut mwena !
virt stand there.dx2
‘(Do) stop there!’
Negative orders or prohibitions are marked by a(xo):
(18)
Axo yo thea o
hele!
proh 2sg play instr knife
‘Don’t play with this knife !’
B. Optative, hortative: sho o
The optative, hortative form is sho o (lit. good if/that).
(19)
Sho o
yo maak !
good virt 2sg die
‘May you die !’
(20)
Sho o
wa tâlâ-e.
good virt 2pl listen-3sg
‘You should listen !’
.. The morpheme o in dependent/subordinate clauses
In subordinate clauses, the virtual marker o has several functions in conditional protases, in irrealis complement and in relative clauses.
A. Conditional protases
– O marks conditional protases, like the first occurrence of o in (21) (o i oome uya agu);
the second occurrence of o bears on the numeral predicate of the irrealis relative clause
(agu [o aa-xiik]), and refers to a potential entity of the class agu ‘person’:
(21)
[O i
oome uya agu
[o
aa-xiik]] [na a
wa
virt 3sg come arrive person virt cl-one but proh 2pl
tu
paa].
go.down outside
‘If someone arrives, don’t go out.’ (lit. should a person who might be one
arrive…)
B. Irrealis relative clauses
– In irrealis relative clauses, o refers to some virtual/potential property or determination of the relativised noun.
(22)
Je
le
yeewat o
hâ
axe.
3sg.be.loc there time virt 1pl.incl see
‘There are times when we may see it.’
© . John Benjamins Publishing Company
All rights reserved
(Bril 2000: 256)
Tense, aspect and mood in Nêlêmwa (New Caledonia)
O commutes with xe, which refers to realis, factive relative clauses. Compare (23a),
which refers to a potential non-specific indefinite ‘sharp’ knife whose existence is unasserted, with the factive (23b), which asserts the existence of a ‘sharp’ knife and refers
to a definite entity. The epistemic feature of virtual o thus correlates with potential
indefinite reference (see Bril 2001).
(23)
a.
Oda fhe hele o
caak !
go.up get knife virt sharp
‘Go get a sharp knife !’ (lit. a knife that may be sharp, if there is
such a thing)
b.
Oda fhe hele xe
caak !
‘Go get the sharp knife !’ (lit. the knife that’s sharp).
O refers to a virtual and unasserted propositional content in the relative clause, as in
(24a), where the verb of the main clause is in the future:
(24)
a.
… me io i
thuuxe hmavat [o
pwa-giik].
… and fut 3sg tell
piece
virt cl-one
‘… and she’ll tell a part of the story.’ (lit. a piece that may be one)
On the other hand, a realis main clause and realis context trigger the use of xe:
(24)
b.
I thuuxe hmava-t xe pwagiik.
‘She told one part of the story.’
Similarly, relativised NPs dependent on negative existential predicates trigger the use
of o, as in (25), since there is no possible factual reference.
(25)
Kia
agu
o
na axi-e.
neg.exist person virt 1sg see-3sg
‘I did not see anybody’. (lit. there isn’t a person (that) I may have seen).
C. Irrealis complement clauses
— O also appears in complement clauses dependent on verbs referring to virtual
events, also dependent on modal verbs, deontic (khere ‘be forbidden’, kââlek ‘be
impossible’), or epistemic-evidential (shuma ‘seem as if/that’). All such uses of o
are equivalent to the subjunctive or conditional and constitute irrealis frames of
reference.
(26)
a.
Na faaxeen o
hmeede
khô-hâ
wo.
1sg ask
virt be.cooked food-poss.1pl.incl pl.indf
‘I asked if our food was cooked.’ (lit. I asked would our food be cooked)
b.
Khere
o
hâ
pweelî nox=ena.
be.forbidden virt 1pl.incl fish.tr fish=that.dx2
‘It’s forbidden for us to fish that fish.’ (lit. it’s forbidden that we fish …)
© . John Benjamins Publishing Company
All rights reserved
Isabelle Bril
c.
Cêê kââlek
o
na diya.
very impossible virt 1sg do
‘It’s impossible for me to do it.’ (lit. it’s impossible that I do it)
d.
Shuma
o
i
be.like
virt
3sg
‘He seems to be happy.’
jela.
happy
D. Irrealis time expressions with o vs. past realis reference with dua
The virtual morpheme o occurs in predicative time expressions with virtual time reference, such as thabwan ‘night’ in (27a). Compare with the past reference marked by the
conjunction dua ‘when’ (27b), which may only refer to the past.
(27)
a.
êna o
thabwan.
now virt evening
‘tonight, this evening’ (yet to come) (lit. now [when] it be evening)
b.
koobwan dua
thabwan
yesterday when.past night
‘yesterday night’ (lit. now [when] it was evening)
On the other hand, a time reference concomitant to the time of speech is marked by a
deictic determiner, as in (27c). Compare with the virtual reference in (27a).
c.
Ni tan hleny
in night this.dx1
‘tonight, during the night’ (concomitant to the time of speech)
In (28a–b), waak ‘morning’ (like thabwan ‘night’ in (27)) is the predicate of a time
clause. In (28a), <o waak> refers to some virtual time dependent on a main clause with
future temporal frame; while in (28b), dua has past reference.
(28)
a.
Io i
uya [o
waak].
fut 3sg arrive virt morning
‘He’ll arrive in the morning.’
b.
I
uya [dua
waak].
3sg arrive when.past morning
‘He arrived in the morning.’ (lit. he arrived when morning)
Various conjunctive time expressions display similar contrasts: uya=(da) dua ‘until
(past)’ ~ uya=(da) o ‘until (virtual)’ (lit. arrive=(up) when), or ni yeewat dua ‘when
(past)’ (lit. in the time when) ~ ni yeewat o ‘when(ever)’ (virtual), (see §8.4).
. The perfect: Definition and use in realis and irrealis frames
While the bare aorist verb form refers to events (independently from any phasal
aspect) and to sequential events in narratives, the perfect marker (k)u ~ (x)u refers to
© . John Benjamins Publishing Company
All rights reserved
Tense, aspect and mood in Nêlêmwa (New Caledonia)
past events or to completed processes which bear some relevance either to the present
and to the time of speech (like the English present perfect), or to another past event in
narratives (like the English past perfect); the perfect may also refer to some resulting
state relevant to present or past temporal frames.
The perfect may appear in realis and irrealis sentences, as shown by (29), which
refers to the projected, imminent completion of a process. This will be detailed in §5.8.
(29)
Io <na u pwaala=du mwa>.
fut 1sg pft sail=down seq
‘I shall be off sailing north (soon).’ (north direction is expressed as ‘down’)
. The perfect in realis frames: Relevance to the time of speech and
narrative uses
In discourse contexts, the perfect refers to a process anterior to the time of speech (t₀),
but relevant to it. In narratives, the perfect refers to a process (tn) anterior to another
past reference point (tpast), in a temporal frame which is disconnected from (t₀).9
The process in the perfect may be complete or incomplete, i.e. it may or may not
have reached its final instant, as in English for that matter, he has/had worked a lot does
not signal completion. On the other hand, especially when used with telic verbs such
as break, the action may be complete or completed and may have a resulting state, as in
he has/had broken the vase; this is also the case with ku hmeede ‘cooked’ in (30):
(30)
Ku hmeede cet ?
pft cook
cooking-pot
‘Is the food cooked ?’
Completion may be expressed lexically by adding toven or roven ‘stop, end’ (see §7.3).
The interval representing an incomplete process is semi-open: its right boundary
is open (as in Figure 1 below). If the process is completed, the (bold) right boundary
of the interval [ A ] is closed (as in Figure 2), and there may be a resulting state, represented by slashes (]////[) in Figure 2.
Figure 1: [–––[
Figure 2: [ A ]////[
In (31–32), (k)u ~ (x)u also refers to a completed process, relative and prior to some
reference point (which may be past or present):
(31)
.
Na i
u aa fhe.
but 3sg pft iter take
‘But he has/had (already) taken it back.’ (something given before)
This is noted [tn anterior to tpast ω t₀].
© . John Benjamins Publishing Company
All rights reserved
Isabelle Bril
(32)
Na na, xe na u khuxi caa-ny.
but 1sg tpc 1sg pft eat.tr starch.food-poss.1sg
‘As for me, I’ve eaten my food.’
(Bril 2002: 480)
– With inchoative verbs like ‘begin’, (k)u ~ (x)u focuses on the initial phase of an
incomplete process, as in (33).
(33)
…me ku thaaxapuxet ni yeewar=eli.
and pft begin
in time=anaph
‘… and it has/had begun at that time.’
(and is/was still ongoing relative to some past reference point)
– With terminative verbs, (k)u ~ (x)u focuses on the final instant of the closed interval:
(34)
(xu) toven.
pft finish
‘It’s finished.’
Terminative processes, such as he’s finishing, or it’s coming to a close, are represented
by an open interval, which is closed only when the process has reached its endpoint.
.. Past event relevant to the time of speech
In (35), the past temporal frame is set by the adverb hule ‘long ago’ in the first clause,
the perfect has scope on the numeral predicate NP tujic kau-n ‘ten years’, while present
relevance to the time of speech (t₀) is marked by the adverb êna in the second clause.
(35)
[I uya hmwiny hule]
[xu tujic kau-n
êna].
3sg arrive here
long.ago pft ten year-poss.3sg now
‘He arrived here long ago, ten years from now.’ (lit. it’s been ten years now)
In (36–37), a time interval is computed between two boundaries, going back in time
from the time of speech (t₀), but with implicit relevance to t₀. In (36), the perfect has
scope on the numeral predicate NP pwadu bwalit ‘two days’:
(36)
Xu pwa-du bwalit
pft cl-two day
‘it was two days ago’
(37)
I
u o=da mwa mwa-yeenat na le.
3sg pft go=up seq house-learn loc there
‘The school there has progressed.’ (until now)
(Bril 2000: 512)
.. Completed process in the past temporal frame of a narrative
In example (38), the verb of the main clause (hnawi-e) is in the bare aorist form and
refers to a past event, while the perfect in the subordinate clause refers to a prior state.
Compare with the two past sequential events in the bare aorist form in (39).
© . John Benjamins Publishing Company
All rights reserved
Tense, aspect and mood in Nêlêmwa (New Caledonia)
(38)
I
hnawi-e a
Pwâ-Hivic dua u gi
hli bwa mat.
3sg leave-3sg agt Pwâ-Hivic when pft be.loc 3du on shore
‘Pwâ-Hivic left her when they were already on the shore.’
(Bril 2002: 482)
(39)
Co huli-na
êbai dua hî tu.
2sg guide-1sg before when 1du go.down
‘You guided me earlier on when we went down.’
(Bril 2002: 489)
. The perfect with stative verbs: Change of state
Whether in narratives or discourse, stative or nominal predicates marked in the perfect by (k)u ~ (x)u refer to a change of state or to a new, transitional resulting state,
which is represented by (]//b//[).
[ A ]//b//[(t₀ or tpast)
The resulting state (]//b//[ is the outcome of the event or process [ A ], it may be
relevant to a reference point (t₀ in discourse, or tpast in narratives). Similarly in Latin,
cognōscō ‘I learn’ contrasts with the perfect cognōvī ‘I know’ which expresses a resulting state.
– The change of state may result from some natural evolution, as in (40–41):
(40)
Na u hulak.
1sg pft old
‘I’m old (now).’ (lit. I have gotten old)
(41)
Na u whaup.
1sg pft be.toothless
‘I’m toothless (now).’ (lit. I have become toothless)
In (42) to (44), the change of state results from an event [ A ] which is expressed
contextually: it results from disobeying the rules in (42), from eating in (43),
from stealing in (44)). In (42), the change of state can be represented as follows:
[ A ]-tôxâ-[.
(42)
(43)
(44)
Hli u tôxâ.
vs.
3du pft be.purulent
‘They got covered with pustules.’
Hli tôxâ.
‘They were pustulous.’
Xu pôlôk kiya-n.
pft be.full belly-poss.3sg
‘His belly is full (now).’ (lit. it has become full)
(Bril 2002: 474)
Xu kian
idaama-ny !
pft disappear eye-poss.1sg
‘My eyes have disappeared !’
(Bril 2002: 487)
© . John Benjamins Publishing Company
All rights reserved
Isabelle Bril
A noun or pronoun in predicate function may combine with the perfect marker to
express a change of state, as hli u thaamwa ‘they have become women’ in (45), and u
wap ‘it’s high tide now’ in (46):
(45)
I
axi horaamalileny êbai
nu
mahlileny xe hli u
3sg see these2.fem.dx1 before coconut these2.dx1 rel 3du pft
thaamwa xe aa-ru.
woman rel cl-two
‘He sees these two, (who were) before these two coconuts, who have become
two women.’
(46)
I
axe khabwe
u wap.
3sg see comp(say) pft high.tide
‘She sees that it’s high tide (now).’
(Bril 2002: 487)
In (47), the first perfect u ogi ‘they had left’ marks a completed, backgrounded process
which stands in causal relation with a resulting change of state represented by u ye
hada ‘she was then (left) on her own’, in which the perfect marker u has scope on the
predicative free pronoun ye.
(47)
Hla u ogi shaak
xe u ye
hada.
3pl pft leave swamp.hen rel pft 3sg.fr alone
‘They had left the swamp hen who was then (left) on her own.’
. Graded change of state and evolutive process with the perfect
Dynamic or stative verbs are compatible with stages and degrees of evolution which
are expressed by combining the perfect marker (k)u ~ (x)u, degree markers like cêê
‘very’, pwâ ‘a little’, and the sequential marker mwa ‘then’. Together, they refer to an
ongoing, transitional process, and to a gradual change of state, from some past point
up to the present in discourse, or relative to another time point in narratives.
(48)
Ku cêê coola khîlû i
ye.
pft very strong disease det 3sg
‘He’s been increasingly ill.’ (lit. his disease has become very strong)
(49)
Na u cêê mwa tâlâ vhaa
Nêlêmwa.
1sg pft very seq hear language Nêlêmwa
‘I understand Nêlêmwa better and better.’
(lit. I’ve (gained) much better understanding)
(50)
Na u pwâ tâlâ-mwemwelî.
1sg pft a.little hear-know
‘I understand slightly better.’ (lit. I’ve (gained) slightly better
understanding)
© . John Benjamins Publishing Company
All rights reserved
Tense, aspect and mood in Nêlêmwa (New Caledonia)
. Respective uses of the perfect (k)u ~ (x)u vs. the bare aorist form
Events in the bare aorist verb form may refer to present or past temporal frames; in
(51) the event tami ‘open’ is past and equivalent to an English preterite.
(51)
Mwa-yeenar=eli
kuut Cavet xe tami na le
ni 1935.
house-learn=anaph stand Tiabet tpc open loc there in 1935
‘As for that school in Tiabet, it (was) opened there in 1935.’
On the other hand, the perfect is commonly used to refer to processes with some
internal relation between them (anteriority, backgrounding, causal relations), and
with relevance either to the time of speech (t₀), as in (52a), or relative to some other
past event in a narrative. By contrast, the bare aorist form, in (52b-53b), refers to
events or to sequences of events without any consideration for any other relation
between them.
(52)
a.
I
u uya.
b. I
uya.
3sg pft arrive
3sg arrive
‘He’s (already) arrived/he’s there.’
‘He arrived.’
The question with the perfect marker in (53a) inquires about the completion of the
process and about its resulting state relevant to (t₀). It may consequently trigger some
further invitation, while in (53b), with the bare aorist verb, the focus is on the eating
event itself and merely inquires about the type of food eaten.
(53)
a.
Co u khuwo ?
2sg pft eat.intr
‘Have you eaten ?’
b. Co khu da ?
2sg eat what?
‘What did/do you eat ?’
In (54), the completed process and its resulting state entail further consequences:
(54)
Na ku hnawu kuru.
but pft drop wind
‘But the wind (has/had) dropped.’ (so he cannot/could not go back home)
In (55), since the speaker has retrieved her stolen eyes10 and has put them back into
their sockets, the perfect (na u axe ‘I can see’) refers to the resulting change of perceptual state, resulting from a prior event (na=da ‘put back’ (her eyes)) and relevant to the
. The two protagonists go fishing and leave their eyes and clothes on the shore; one of
them plays a trick on the other by stealing her eyes; later, when she retrieves them and puts
them back into their sockets, she grasps the situation.
© . John Benjamins Publishing Company
All rights reserved
Isabelle Bril
speaker’s temporal frame (t₀) marked by êna ‘now’. Note that the bare verb form (faxi
‘cook’) in the relative clause refers to a past event marked by kôôbwan.
(55)
Me na mwa na=da me
mwa idaama-ny; me na u axe
and 1sg seq put=up so.that seq eye-poss.1sg and 1sg pft see
êna hlaabai
co faxi kôôbwan.
now those.anaph 2sg cook yesterday
‘And I put them up to have eyes (lit. to be my eyes), and now I get to see
what you cooked yesterday.’
(Bril 2002: 492)
In (56), the first perfect refers to a change of state (i u alawuk ‘(as) she got hungry’),11
which is the causal background triggering the consequent event in the aorist (i oda
khîlî-wo ja ‘she went up to roast sea-shells’). In the next sentence, this is repeated in the
perfect (i u khîlî-wo ja), as a sort of backgrounding,12 cueing construction (‘once she
had roasted sea-shells’) which refers to a completed process relevant for the next event
in the aorist (i o=da).
(56)
Na i
u alawuk, na i
o=da khîlî-wo
ja;
but 3sg pft hungry so 3sg go=up roast-pl.indf sea-shells
‘but she’d got hungry, so she went up to roast sea-shells;
i
u khîlî-wo
ja
le,
mwena
xe i
o=da…
3sg pft roast-pl.indf sea-shells there there.dx2 tpc 3sg go=up
(once) she had roasted sea-shells, then, she went up.’
(Bril 2000: 502)
The excerpt from a narrative in (57) shows an alternation of bare aorist and perfect
verb forms in a past temporal frame disconnected from the speaker’s time. The bare
form yhoraabwa ‘lived’ in [P₁] is stative-like, while (k)u ~ (x)u in [P₂] (hli u tu ‘they had
gone down’) refers to a completed and backgrounded process. The event in the bare
form hli tu tuuli naam pwagiik ‘they went down, found a water-hole’ in [P₃] forwards
the story, adding new information. The perfect in [P₄] (hli u tha ‘they speared’) marks
a process which is the causal background relevant to the next event in [P₅] ‘they caught
a lot of fish’.
(57)
Hli yhoraabwa Alevic. Na ni naxâât pwa-giik xe […] hli u
3du live
Alevic but in day
cl-one tpc
3du pft
P₁
P₂
‘They lived at Alevic. One day, they had gone down (to the sea)
. Compare with the expression of a present state: i alawuk ‘(s)he’s hungry’.
. For similar backgrounding effects of the perfect in Vanuatu languages, see (François 2010).
© . John Benjamins Publishing Company
All rights reserved
Tense, aspect and mood in Nêlêmwa (New Caledonia)
tu
thu naam.
Hli tu
tuuli naam
pwa-giik
go.down make water.hole 3du go.down find water.hole cl-one
P₃
to make water-holes (for fishing). They went down, found a water-hole
xe pôlôk o
nok mwena
xe hli u tha, tha, tha
rel full with fish there.dx2 tpc 3du pft spear spear spear
P₄
full of fish then, they set spearing (fish), spearing,
me
khada kââlek
hleeli
shâ-doo-li
nok.
coord after be.numerous these.anaph content-assagai-poss.3du fish
P₅
and then they caught a lot of fish.’ (shâ-doo-li nok lit. means ‘these
assagai-speared fish of theirs’)
In (58), another excerpt from a narrative in a past reference frame, the two perfects
u pwaala in the subordinate clause [P₁] and u yhowaalak in the main clause [P₂] are
backgrounded, prior situations which are relevant for the next sequence of events in
[P₃] and [P₄] marked by the bare verb forms o=da ku-râimi pwat ‘go up to pick a fruit’
and no-huu ‘follow with the eyes’.
(58)
[Dua hli u pwaala], [na
i
u yhowaalak=da horaamwaleny
when 3du pft sail
coord 3sg pft climb=up
this.fem.dx1
P₁
P₂
‘When they had sailed away, Hiixe had climbed up
Hiixe bwa hooli pwê-khojomi],
[me i
o=da ku-râimi
Hiixe on anaph Cerberas Manghas and 3sg go=up stand-pick
P₃
the tree, and she went up to pick
pwat pwa-giik na mwaida ni hadat] [me i
no-huu
Kaavo
fruit cl-one loc there.up in branch and 3sg see-follow K
P₄
one fruit up there in the branches, and she watched Kaavo
ma Pwayili dua hli u pwaala.]
and P.
when 3du pft sail
P₅
and Pwayili when they had sailed away.’
The following segment in (59) shows again alternation of perfects and of bare aorist forms (underlined); the perfects mark transitional processes, serving as background with relevance for the following sequence of events marked by bare verb
forms.
© . John Benjamins Publishing Company
All rights reserved
Isabelle Bril
(59)
Me i
u tu
mwa, tu
mwa, o=da shi horaamwaleny
and 3sg pft go.down seq go.down seq go=up at this.fem.dx1
‘and she set going down then; she went down, went up to this woman’s
place,
me hli u wo.
Hli u wo,
na
i
xau
oo=me
and 3du pft weave 3du pft weave coord 3sg unexp come=hither
and they both set weaving. They’d been weaving, and suddenly
hiiji aa-xiik xa fhe para khoo-n
nu
na
ant cl-one also take piece food-poss.3sg coconut loc
an ant arrived, took a piece of coconut,
mwêli, me i
xau
thegela me i
o=da mwa bwa baarawo.
there and 3sg unexp run
and 3sg go=up seq on weaving
ran, and went up onto their weaving.’
(Bril 2000: 502–503)
In (60), ni yeewat dua sets the temporal reference as past, all the following verbs are
in the bare aorist form (underlined) and refer to sequential events. The assertive form
xau ‘suddenly’ is modal rather than aspectual, and refers to a sudden, unexpected
action.
(60)
Ni yeewat dua i
xau
thaaxa khuwo, (…)
in time
when 3sg unexp start
eat
‘As she was just starting to eat, (…)
na i
xau
fhe bulaivi a
aa-xiik ava horamalaaleny (…)
but 3sg unexp take club
agt cl-one some these.fem.dx1
all of a sudden one of these (women) (…) took the club
me i
xau
hnawe=du ni bwaat.
and 3sg unexp let.go=down in head
and suddenly struck her on the head.’
. The perfect in chronologically ordered events and relative time
constructions
Events are chronologically ordered by adverbs, such as habuk ‘in front, first, before’,
mon ‘behind, after’, or by sequential lexemes, such as khadaa ‘after, then’, mwa ‘then’,
which may occur in realis or irrealis reference frames. In (61), the perfect marker (k)
u ~ (x)u marks a completed event, while mon marks some posterior chronological
ordering relative to another event or process.
(61)
I
u mwa le mon.
3sg pft seq go after
‘(S)he has then left after.’
In (62), the perfect (u farame) in the main clause refers to a completed process,
anchored in the past by the anaphoric adverb êbai ‘earlier, before’ (see Table 5 below)
© . John Benjamins Publishing Company
All rights reserved
Tense, aspect and mood in Nêlêmwa (New Caledonia)
and with consequences upon the present situation; the adverb habuk ‘before’ in the
subordinate clause orders it chronologically.
(62)
Na u farame êbai o
na khabwe habuk.
1sg pft forget before virt 1sg say
before
‘I have forgotten if I said it before.’
Compare with: na farame êbai o na khabwe habuk ‘I forgot if I said it before’.
Habuk ‘before’ and mon ‘after’ also occur as subordinators, together with the subordinator me, as in (63):
(63)
Na habuk me i
o=da=me
xe pwaxat me i
but before sub 3sg go=up=hither tpc necessity sub 3sg
parega idaama-n.
hang eye-poss.3sg
‘But before she came up here, she needed to put her eyes back (in their
sockets).’
Habuk ‘before’, mon ‘after’ and hule ‘long ago/before’ also function as nominal modifiers, as in (64).
(64)
ni naxâât bai
habuk
in day
anaph before
‘the day before (or) in those days’
. The perfect with reference to some imminent event or to the imminent
completion of a process
An imminent event, or the imminent completion of a process, is often expressed by a
perfect marker13 in Oceanic languages (in Toqabaqita, Solomon islands (Lichtenberk
2008), as well as in Kanak languages).14 It also occurs in Mandarin, or in spoken
French, where one of the two readings of je suis parti ‘I’m off ’ (lit. I’m gone) is of
imminent departure.
(65)
Na u le.
1sg pft leave
‘I’m off.’
The imminent process is asserted, as if already completed (hence the perfect), and the
focus is on the projected final instant of the process and its resulting state.
In (66), the first occurrence of the perfect (u nanami) refers to a prior event
relevant to the time of speech (t₀) and triggering the next process, also marked in the
. Ö. Dahl (2006) labels such use of the perfect ‘iamitive’.
. In Drubea (Shintani & Paita 1990) and Xârâcùù (Moyse-Faurie 1995).
© . John Benjamins Publishing Company
All rights reserved
Isabelle Bril
perfect (na u le), but expressing imminent completion this time. Similarly, in (67), the
perfect (na u ulo) indicates some imminent change of state.
(66)
Mo axe bu na u nanami khabwe
me na u le.
2du see for 1sg pft think comp(say) sub 1sg pft leave
‘Listen you two, for I’ve decided to leave.’
(67)
Bwaaramwa Pwâ Keebö ! co yhaali-na bu na u ulo.
alas
P. K.
2sg lift.off-1sg for 1sg pft burn
‘Help Pwâ Keebö ! lift me off (the fire) or I’m going to burn.’
(Bril 2002: 495)
. The perfect and other lexical expressions of imminence
It is of some interest that various lexical (locative or temporal) expressions of imminent completion involve (k)u ~ (x)u, and express gradation.
A. (ku ~ xu) jeuk me ‘be close to, almost, nearly’ (jeuk ‘be close, be next to’) refers to a
targeted and nearly reached process or state, relevant to t₀ in (68) or potential in (69).
(68)
Bu (xu) jeuk
me pôlôk kee-n.
for pft be.close sub be.full basket-poss.3sg
‘Her basket is almost full already.’
(69)
Na ni yeewat o
khabwe
ku jeuk
me foro=da taan…
but in time virt evid(say) pft be.close sub white=up day
‘But at the time when dawn may seem to be almost bright …’
(lit. when day will be close to be white)
B. (k)u ~ (x)u thara also expresses imminent completion (thara ‘next to’):
(70)
I
axe khabwe
xu thara bwan.
3sg see comp(say) pft next.to night
‘She sees that it’s getting dark.’ (lit. it is close to night)
C. Imminent completion may also be expressed by a negative time phrase kio hule me
‘almost, soon’ (lit. it’s not long, where hule is predicative), combined with (k)u ~ (x)u
in the next clause.
(71)
Kio hule
me i
u toven o
khuwo.
neg long.time sub 3sg pft finish prep food
‘He has almost finished eating.’ (lit. it’s not long that he has finished with
food)
The verbal phrase u diya me (lit. do for) ‘be about to’ also expresses imminence, see
§7.1 below.
© . John Benjamins Publishing Company
All rights reserved
Tense, aspect and mood in Nêlêmwa (New Caledonia)
. The perfect in irrealis reference frames
The specific uses of the perfect in irrealis reference frames are now considered.
.. The perfect in imperative or hortative utterances: Urged projected completion
In imperative and hortative utterances, the use of the perfect (k)u ~ (x)u is emphatic,
it expresses the projected completion of a transitional process or a projected change of
state with a stronger sense of urgency.
(72)
Co u fhe hliina
thaxamo
i
yo me wa u
2sg pft take those2.dx2 old.woman det 2sg and 2pl pft
o=da mwa !
go=up seq
‘Do take these two spouses of yours and get on your way back !’
Compare with an unmarked imperative:
Fhe hliina thaxamo i yo.
‘Take these two spouses of yours.’
(73)
Wa u fhe hleena
mââwu-wa
roven me wa
2pl pft take those.dx2 belonging-poss.2pl all
and 2pl
u khavak !
pft go.away
‘Do take all your belongings and be off !’
Compare with an unmarked imperative:
Wa fhe hleena mââwu-wa.
‘Take all your belongings.’
(74)
Co axe ! hî u taabwa me hî u khuwo !
2sg see 1du pft sit
and 1du pft eat
‘Listen ! Do let’s sit down and eat !’
– In self-directed hortative utterances, the semantics of (k)u ~ (x)u are similarly
emphatic:
(75)
Sho o
na u o=da bwaxamat.
good virt 1sg pft go=up shore
‘I must absolutely go up to the shore.’
Compare with the neutral self-directed hortative:
Sho o na oda bwaxamat.
‘I had better go back to the shore’.
© . John Benjamins Publishing Company
All rights reserved
Isabelle Bril
.. Use of the perfect in future reference frames
In a virtual (o) or a future (io) frame, (k)u ~ (x)u refers to the imminent completion of
a process, or an imminent change of state; it also expresses epistemic modality and the
speaker’s certainty about the projected occurrence of the event.
Table 7. Combination of irrealis mood and perfect
Semantics
o
(k)u ~ (x)u
potentially completed action or change of state
io
(k)u ~ (x)u
projected, future action expressed as if completed
Note that io has scope on the whole clause, while (k)u ~ (x)u has scope on the predicate
only:
(76)
a.
Io <na u le
caae
o
waak>.
fut 1sg pft leave tomorrow virt morning
‘I’ll be leaving tomorrow morning (for sure).’
(lit. I’ll have left for sure tomorrow morning)
Compare with a neutral future event:
b.
Io na â caae
o
waak
fut 1sg go tomorrow virt morning
‘I’ll go tomorrow morning.’
The combination of the future and the perfect may also express a projected change of
state considered as certain in some future time:
(77)
Io <na u mwa puri>.
fut 1sg pft seq snake
‘I shall become a snake.’ (lit. I shall have become a snake for sure)
.. The perfect in virtual or conditional frames
– In clauses with virtual reference depending on a similative verb, such as (78), the
perfect in the dependent clause refers to a potential change of state or resulting state
marked as if already achieved.
(78)
Maariik o
na u maak !
be.like virt 1sg pft die
‘It’s as if I was dead!’
(Bril 2002: 498)
– In conditional clauses (79), the perfect (k)u ~ (x)u expresses the projected completion of a virtual process with epistemic or deontic meanings.
© . John Benjamins Publishing Company
All rights reserved
Tense, aspect and mood in Nêlêmwa (New Caledonia)
(79)
[O awa-m
me [yo u o=da]] xe [e na
pot will-poss.2sg sub 2sg pft go=up tpc fut 1sg
P₁
P₂
diya waja-m]
make boat-poss.2sg
‘If you must go/really want to go, I’ll make a boat for you.’
‘Should you really want to go, I’ll make a boat for you.’
Compare with the neutral sentence: o awa-m me yo oda… ‘if you want to go…’
– In virtual time clauses, such as [P₁] in (80), marked by the conjunctive marker ni
yeewat o ‘when’, the perfect na u â is the prior and backgrounded (but virtual) condition for the projected event in [P₂] to occur.
(80)
[Ni yeewat o
na u â]
xe [io na pwâ vhaa Nêlêmwa].
in time
virt 1sg pft leave tpc fut 1sg a.little speak Nêlêmwa
prior completed event P₁
P₂ prospective event
[projected in some potential frame]
‘When I have left, I’ll speak some Nêlêmwa.’
Thus P₁ is equivalent to anterior future in French quand je serai parti. Compare with:
ni yeewat o na â ‘when I leave…’ (quand je partirai, a simple future in French).
In (81), in the virtual frame of [P₁] marked by ni yeewa-t o, the perfect marker (k)u
~ (x)u marks the decision to follow (ku awa-m me yo u oxuri) as the prior condition for
the process in [P₂] to occur. The perfect in [P₂] yo u le ‘you’ll go’ projects the imminent
process as if already completed, and is modally assertive.
(81)
[Ni yeewat o
ku awa-m
[me yo u oxuri hlaabai
in time
virt pft will-poss.2sg sub 2sg pft follow these.anaph
P₁
nanamiwo i
yo ]] (…) xe [yo u le].
thought
det 2sg
tpc 2sg pft go
P₂
‘When you have decided that you must follow your wishes […], you’ll
set off.’
So far, the analysis has focused on the functions and meanings of combined mood and
aspect markers; we shall now turn to the expression of the various phasal aspects of
processes.
© . John Benjamins Publishing Company
All rights reserved
Isabelle Bril
. Expression of the phasal aspects of processes
There are numerous morphological indicators of the internal phases of processes; these
morphemes are generally pre-verbal. Table 8a (repeated from Table 3) summarises
pluri-actional markers, and the various markers of incompletion.
Table 8a. Pluri-actional and incompletion markers
pluri-actional
aa
iterative
kua
frequency
(+past)
incompletion
na
processual
(+activity V)
baa
durative
gaa
persistive
gat
continuative
contrastive
taa₁
remansive &
state of activity
Stative verbs and states are compatible with some of the aspect markers expressing
pluri-actionality (iterative aa, frequency kua) and incompletion (durative baa, persistive gaa, remansive taa₁). This will be developed under §6.1, §6.2.
The morphemes in table 8b (repeated from Table 4) focus on the initial or final
boundary of an event or process. Kio fwâ ‘not yet’ refers to a still uninitiated process
(§6.4.1); fwââm ‘almost, nearly’ indicates near, but failed, completion of some projected event (§6.4.3); kio haxa ‘no longer, no more’ refers to some permanently terminated event or process.
Table 8b. Other aspect markers
boundary
unreached
boundary nearly
reached, but failed
permanently terminated,
bygone
kio fwâ
‘not yet’
fwââm
‘almost, nearly’
kio haxa
‘no longer, no more’
These aspect morphemes are compatible with any predicative lexeme (± verbal),
but some of them, for instance the processual marker na, only combine with activity verbs. Compatibility between verb semantics and aspect morphemes also depend
on the verb’s inherent semantic features and Aktionsart: ± stative, ± punctual, ± telic,
± reversible (like ‘die’), iterative or cyclic (like ‘breathe’), compatible with a gradient
(like ‘grow’), with degrees and grades.
. Pluri-actionality: Iteration and frequency
Pluri-actional markers such as aa (iterative) and kua (frequency) may refer to punctual events (like ‘kick’) or to processes (like ‘work’). Aa is compatible with ± realis
mood morphemes, while kua ‘often’ is only compatible with past reference.
© . John Benjamins Publishing Company
All rights reserved
Tense, aspect and mood in Nêlêmwa (New Caledonia)
.. Aa: Iteration of time points or time intervals
Aa is compatible with all types of predicate (±verbal) and with all semantic types,
stative or dynamic, ± telic verbs, ± punctual. It has iterative, regressive meaning (like
‘read again’) or reversive meaning (like ‘grow again’). In (82), the first iterative marker
has scope on a free pronoun in predicate function.
(82)
Aa yawa ! wa aa tu
khak !
iter 2pl.fr 2pl iter go.down cast.net
‘(it’s) you again! Go back net-casting again!’
– In a negative sentence marked by kio, the meaning changes with the scope of aa,
either on < p > or on < non p > as summarised below:
kio < aa p > ‘not again’ vs. aa < kio p > ‘again not p’
(83)
Kio i
aa mago.
neg 3sg iter sleep
‘He did not fall back to sleep.’ (interrupted cycle)
(84)
Aa kio i
mago.
iter neg 3sg sleep
‘Again, he is not sleeping.’ (iterative)
.. Kua: Frequency (past)
The frequency marker kua is only compatible with past reference frames. It may combine with the perfect marker, as in (85), expressing relevance to some reference point.
When combined with other markers, kua occurs immediately before the predicate.
(85)
I
u kua khabwe.
3sg pft freq say
‘He’d often said it.’
(86)
Kio i
kua shaya shi-n.
neg 3sg freq work side-poss.3sg
‘He does/did not often work at home.’ (known from past experience)
Interestingly, kua may combine with iterative aa with emphatic meaning:
(87)
I
aa kua tuâ yava.
3sg iter freq lie 1pl.excl.fr
‘He kept lying to us.’ (known from past experience)
. Markers of incompletion
Various morphemes mark ongoing or incomplete processes, which are represented by
semi-open intervals ([–––[); these morphemes have distinct semantics: na is processual, baa is durative, gaa is persistive, gat is continuative-contrastive, taa₁ is remansive.
© . John Benjamins Publishing Company
All rights reserved
Isabelle Bril
Only persistive gaa and continuative-contrastive gat, which have modal semantics, are
compatible with the perfect marker (k)u ~ (x)u.
.. Na: Processual
The processual marker na is only compatible with activity verbs. It originates from
a nominalising locative prefix (hna- ‘place’), which accounts for its immediate preverbal position and for the lack of a subject pronoun, which makes such constructions
less finite and more participial. Na occurs in dependent clauses, headed by locative
predicates (gi and je),15 as in (88–89), or dependent of perception verbs, as in (90),
where na is actually part of a participial-like relative clause. Note that in (88), the plural pronoun hla is the subject of the locative verb gi in postverbal16 position. Compare
with gi na ni mwa (lit. be.loc 1sg in house) ‘I’m in the house’.
(88)
[Gi
hla [na pwa-yiic]].
be.loc 3pl proc cut-wood
‘They are cutting wood.’ (lit. they are in the process of wood-cutting)
(89)
Kio je
[na haga].
neg 3sg.be.loc proc fish
‘He is not fishing.’
(90)
I
axi hobai hulak
[na taabwa ni wooxa-hli].
3sg see anaph old.man proc sit
in middle-poss.3du
‘He sees/saw that old man sitting among them.’
Example (12), repeated here, is another example of a participial-like relative clause
headed by na: [thaamwa=ena gat] [hoona na vhaa] ‘it’s still that woman who is speaking’ (lit. it’s still that woman the one speaking).
But with such usage, na is no longer a nominalising prefix, as evidenced by the
possible insertion of another aspect marker, such as persistive gaa in (91), which connotes obstinacy. Since nominalising prefixes do not allow insertion of any other entity
between the prefix and the verb, na appears to have specialised as an aspect marker.
(91)
Je
na
gaa
gi.
3sg.be.loc proc persist cry
‘He’s still there crying/he keeps on crying’ [obstinately]
. The locative verb gi ‘be in some place’ has postverbal subject pronouns for all persons
(including 3sg gi ye), and only refers to animates, while je is only compatible with 3rd person
singular or plural ±animate entities; compare gi ye ni mwa ‘he’s in the house’, je ni mwa ‘he’s/it’s
in the house’.
. Gi ‘be in some place’ is one of the few verbs with postverbal subjects.
© . John Benjamins Publishing Company
All rights reserved
Tense, aspect and mood in Nêlêmwa (New Caledonia)
Yet, in contrast with other aspect markers, na excludes any subject index. Thus, in
contrast with i baa gi ‘he’s crying’ or i gaa gi ‘he keeps crying’, *i na gi is ungrammatical,
only a locative expression, such as je na gi ‘he’s there crying’, is acceptable.
.. Comparison of durative baa, persistive gaa, continuative-contrastive gat
Before proceeding with the detailed analysis of durative baa, persistive gaa and continuative gat, here is a brief outlook of their semantic differences, illustrated in (92).
Baa refers to some ongoing process with activity verbs, while gaa expresses persistence, and possibly contrast with some other expectation on the speaker’s part and
modal semantics; as for continuative gat, the only marker to be in post-verbal position,
it also denotes contradictory expectation.
(92)
a.
b.
c.
I baa khuwo
I gaa khuwo !
I khuwo gat
‘(s)he’s eating’
‘(s)he’s still eating’ [she’s persistent, but should not]
‘(s)he’s still eating’ [it is expected to be over]
(93)
a.
I
baa khu kuvic âlô hleny.
3sg dur eat yam child this.dx1
‘The child is eating yam.’
b.
I
gaa
khu kuvic âlô hleny.
3sg persist eat yam child this.dx1
‘The child goes on eating yam.’ (he should not, yams must be kept)
.. Gaa: Persistive
Like iterative aa, and frequency kua, persistive gaa is compatible with all types of
predicates, non-verbal (94), stative or dynamic, and with ± punctual events. It is also
compatible with irrealis and realis reference frames. Positions are as follows:
(94)
o ~ io (s) gaa predicate (irrealis)
(s) u gaa predicate (realis)
Xam gaa
pânaat.
ass persist stone
‘It is still stone.’ (as a persistent state of things, but a different one is expected)
(95)
Xam gaa
pe-maariik.
ass persist rec-be.similar
‘It is again similar.’
(96)
Gaa
nabak=du kua-n.
persist sink=down foot-poss.3sg
‘Her feet keep on sinking down.’
Gaa may express irritation when the persistent process or state is seen as an obstacle.
It may combine with gêlâ ‘always’ (like iterative aa and frequency kua).
© . John Benjamins Publishing Company
All rights reserved
Isabelle Bril
(97)
I
gaa
gi gêlâ.
3sg persist cry always
‘She keeps wailing all the time.’
.. Baa: Durative
Durative baa is compatible with ± realis reference frames; but not with the perfect
marker (k)u ~ (x)u, in contrast with persistive gaa. It refers to an ongoing process with
activity verbs, but it is durative with stative verbs, as in (99):
(98)
I
baa khuwo, na xau
tabö kot.
3sg dur eat
but unexp fall rain
‘He was eating when the rain started falling.’
(99)
Baa maaxa.
dur be.raw
‘It’s still raw.’
– Negated duration: kio… baa ‘not/no longer ongoing’ vs. non-initiated duration baa
kio ‘still not’
If the negation kio has scope on the segment < baa + predicate>, as in (100), the
process is no longer ongoing: kio < baa p > ‘not/no longer ongoing’. If baa has scope
on the negated predicate (101), the process has not yet started: baa < kio predicate>
refers to some uninitiated duration meaning ‘still not’; gat marks some contradictory
expectation.
(100)
Kio < i baa mago gat >.
neg 3sg dur sleep cont
‘He’s not/no longer sleeping.’
(101)
Baa < kio i
mago gat >.
dur neg 3sg sleep cont
‘He’s still not asleep.’
.. Gat: Continuative and contrastive
Gat is one of the few aspectual morphemes to occur after the predicate. It marks continuative aspect, but also denotes contradictory expectation or presupposition. It often
combines with durative baa:
(102)
Baa kia-e
gat.
dur not.be.there-3sg cont
‘He’s still not there.’ [but he should be]
(103)
Co khuxi o
baa hnu gat !
2sg eat.tr virt dur hot cont
‘Eat it while it’s still hot !’
© . John Benjamins Publishing Company
All rights reserved
Tense, aspect and mood in Nêlêmwa (New Caledonia)
A. kio mwa… gat ‘no longer’
In (104), the state of ‘still being young’ (âlô gat) is negated; the perfect (k)u ~ (x)u and
the sequential mwa mark the transitional process leading to a change of state (lit. it’s no
longer the case I am still young). The perfect has scope on the whole negative sentence
and marks the change of state relevant to the present (t₀):
(104)
xu < kio mwa na âlô gat >
Xu kio mwa
na
âlô
gat.
pft neg seq
1sg
child
cont
‘I’m no longer young.’ (in spite of what you think)
Compare with the neutral statement: kio na âlô ‘I’m not young’.
B. Combinations of durative baa, continuative-contrastive gat, processual na
When durative baa and continuative gat combine with the locative processual constructions gi… na or je… na, the markers baa… gat frame the locative predicate while
na has scope on the following activity verb.
(105)
[Je
gat ] [na pwe].
3sg.be.loc cont proc fish
‘He’s still there fishing.’ (but he should be back)
(106)
[Baa gi
hla
gat] [na khuwo] !
dur be.loc 3pl.fr cont proc eat
‘They are still there eating !’ (they should not)
Compare with: je [na pwe] ‘he’s fishing’, gi hla [na khuwo] ‘they are eating’.
.
Taa₁ remansive and taa₂ previous occurrence
Taa₁ (possibly originating from taabwa ‘sit, be settled’) refers to remansive states or
states of activity,17 often expressing contrast. There is another taa₂, probably with the
same lexical origin, which expresses previous occurrence and means ‘already, (do) in
advance, before’; taa₂ often (but not always) appears in what can be analysed as lexicalised verb compounds, like taa fooyet ‘cook in advance’, taa khabwe ‘predict, warn’
(khabwe ‘say’), taa paare ‘warn’ (paare ‘announce’), taa pajale ‘promise, say in advance
(pajale ‘reveal’). On the other hand, some non-lexicalised uses of taa₂ (analysed in
§6.3.2) express previous occurrence and can be translated by ‘already’. These two morphemes are probably independent evolutions from the same lexeme.
. See Desclés & Guentchéva (1995) for the notion of state of activity like it’s on flight.
© . John Benjamins Publishing Company
All rights reserved
Isabelle Bril
.. Remansive taa₁
Remansive taa₁ combines with stative or activity verbs with different semantics.
A. Taa₁ with stative predicates: Remanent state of things
With stative predicates, taa₁ refers to a remanent state, which may be seen as an obstacle (107a) to another expected state, with which it is implicitly contrasted; it may connote impatience or irritation:
(107)
a.
Puxet da
me taa khoxo
i
wa ?
origin what? sub rem be.many det 2pl
‘Why do you remain so numerous?’ (there should be fewer people)
Compare with a simple neutral question:
(106)
b.
Puxet da
me khoxo
i
wa ?
origin what? sub be.many det 2pl
‘Why are you so numerous?’
B. Taa₁ with activity verbs: remanent activity or state of activity
With dynamic verbs, taa₁ expresses a remanent activity or state of activity, contrasting
with some other (expected) event/state. In (108), two projected concomitant activities
are contrasted:
(108)
Co taa kuut mwena, bu na taa haga=ve
hmwiny.
2sg rem stand there.dx2 for 1sg rem fish=crosswise here
‘You stay there, as I’ll be fishing around here.’
(Bril 2002: 482–483)
Compare (108) with an ongoing process: je na haga ‘he is fishing’.
In (109), an event (o-da ‘go up’) is contrasted with a remanent state of activity (taa
muuvi ‘stay put’):
(109)
Na o=da ni daak, wa taa muuvi.
1sg go=up in garden 2pl rem stay
‘I go up to the field, you stay (here).’
(110)
Na hî agu
xe shuma o
i
pweedi-e […],
but this person tpc be.like virt 3sg turn.round-3sg
‘But this person, he seems to turn his back (i.e. lose interest),
bu i
taa no=du
ni wi-yaak.
for 3sg rem watch=down in sea-salt
for he remains/keeps watching down towards the sea.’
(instead of listening facing the speaker)
The persistive marker gaa may combine with a taa-marked remanent activity:
© . John Benjamins Publishing Company
All rights reserved
Tense, aspect and mood in Nêlêmwa (New Caledonia)
(111)
Co [gaa
[taa kââleng
hmwiny]].
2sg persist rem go.around here
‘Keep on turning round here.’
Differences appear clearly in the two imperative clauses in (112): in a. the participant
is urged to keep moving, and taa marks some implicit and contextual contrast with the
one who stops; while in b. with gaa, simple persistence is expressed without any such
contrastive meaning.
(112)
a.
Co taa ôô=xi !
2sg rem go=away
‘Keep moving away !’ [while the other is stopping]
b.
Co gaa
mwa ôô=xi !
2sg persist seq go=away
‘Go further away !’
.. Taa₂ ‘already’, expressing previousness
The other morpheme, taa₂, stresses the previous occurrence of an event, and prior
completion of some activity.
(113)
I
axi
hulak
aa-xiik xe i
taa tuu=me.
3sg see.tr old.man cl-one rel 3sg prev go.down=hither
‘He sees an old man who’d already come down here.’
Taa₂ may combine with the perfect marker to stress anteriority in relation either to the
time of speech or to a past event in a narrative:
(114)
Hla u
pe-tîxââc
hlaali
yabwec, puxet hoona
i
u
3pl pft rec-be.angry those.dx3 subject cause that.dx2 3sg pft
taa ogine
prev finish
‘The subjects got angry because he had already finished (tilling)
baaruxia i
ye, ke
i
baa mweelôô baaruxia i
Teâ.
furrow det 3sg then 3sg dur remain furrow det Chief
his furrow (of new yam), while the chief ’s furrow still remained (untilled).’
(115)
Pûwûlûû fââlô
xe i
u taa kââleng
mwamaidu Tavaam.
news
mission tpc 3sg pft prev go.around down.there Tavaam
‘As for the news of (this) mission, it has already gone around (the area of)
Tavaam down there (in the north).’
. Morphemes equivalent to ‘not yet’ and ‘almost’
This section concentrates on morphemes which specifically focus on the relation to the
initial boundary of the event, process or state. Those are kio… fwâ or fwâ kio ‘not yet’,
© . John Benjamins Publishing Company
All rights reserved
Isabelle Bril
fwâ koni ‘never yet, still never’, and fwââm ‘almost, nearly’; they refer to a process or
event that has not yet occurred.
.. Kio… fwâ ~ fwâ kio ‘not yet’
Fwâ is only used in combination with the negation kio. The morphemes are discontinuous when the negation comes first in kio… fwâ, but they are adjacent in the reverse
order (fwâ kio), without any semantic difference whatsoever. Kio… fwâ or fwâ kio refer
to an event that has not yet occurred, either in relation to the time of speech or to some
other reference point disconnected from the time of speech (116–117).
(116)
Kio i
fwâ uya (or) fwâ kio i
uya.
neg 3sg yet arrive
yet neg 3sg arrive
‘He has/had not arrived yet.’ (or) ‘He has/had yet not arrived.’
The combination with gat marks some contradictory expectation:
(117)
Fwâ kio i
uya (gat).
yet neg 3sg arrive cont
‘He (still) hasn’t arrived yet.’ [gat rendered here as ‘still’ suggests that he
should be here.]
Compare with the durative aspect in (102) repeated here:
baa kia-e gat
‘he’s still absent/not there’.
When kio … fwâ combines with a phasal verb (start, finish), only the phase expressed
lexically and the relevant boundary are under consideration (the initial boundary with
an inchoative verb as in ‘he hasn’t yet started’, or the final boundary as in (118)).
(118)
Kio i
fwâ toven.
neg 3sg yet finish
‘He hasn’t yet finished.’
.. Fwâ koni ‘never so far, never yet, still never’
Koni is a negative modal verb meaning ‘unable, impossible’. By contrast with kio … fwâ
which refers to one event, fwâ koni ‘never yet/so far’ has a scanning effect, it negates
any occurrence of an event within a given temporal frame, up to the time of speech or
up to some past event.
(119)
Va
fwâ koni
tu
Frans.
1pl.excl yet be.unable go.down France
‘We’ve/’d never been to France so far/yet.’
© . John Benjamins Publishing Company
All rights reserved
Tense, aspect and mood in Nêlêmwa (New Caledonia)
.. Fwââm ‘almost, nearly’
Fwââm ‘almost, nearly’ refers to an event which almost occurred. A similar meaning
is expressed by French faillir, manquer de or English come close to. The semantics of
< fwââm p > can be glossed as follows, [the interval of event p was almost reached, but
fell short of it].
(120)
Fwââm au-riik
me io kio i
khabwe shi-ny !
almost times-one conj fut neg 3sg say
side-poss.1sg
‘He was very close not to tell me !’ [but he did]
(121)
I
fwââm maak.
3sg almost die
‘He almost died.’
. Permanently terminated event or state: Kio… ha(x)a ‘no… longer’
Haxa expresses permanence with restrictive connotations, meaning ‘only stay/keep
doing’. In (122), the perfect marker signals a change of state relevant to the time of
speech. As shown by (122a–b), it appears in realis and irrealis reference frames. Its
negative counterpart kio haxa refers to a situation (event or state) which is permanently terminated (123) and which may also denote bygone events.
– Permanence: haxa
(122)
a.
I
u ha(x)a taa nô-waa.
3sg pft perm sit watch-empty
‘He merely stayed/kept staring away.’
b.
Io ha(x)a top na mwena.
fut perm rot loc there.dx2
‘It will just remain rotting away there.
– Permanently terminated or bygone event or state: kio haxa
(123)
Mwena
xe hla u
hûn,
kio hla ha(x)a vhaa.
there.dx2 tpc 3pl pft be.silent neg 3pl perm speak
‘Then, they fell silent, they no longer spoke.’
. Comparing permanently terminated kio… haxa ‘no longer’ with kio…
baa ‘not ongoing’ and fwâ kio ‘not yet’
The following examples illustrate the use of these negative aspectual phrases, contrasting (i) permanently terminated events, processes or states expressed by kio… haxa,
with (ii) no longer ongoing processes with kio… baa, and (iii) events that have not yet
occurred with fwâ kio ‘not yet’.
© . John Benjamins Publishing Company
All rights reserved
Isabelle Bril
– kio… haxa refers to a terminated event, sometimes permanently over as in (124).
(124)
Kio i
ha(x)a mago.
neg 3sg perm sleep
‘He never sleeps anymore.’ [due to insomnia or disease]
– By contrast, kio… baa in (125), (repeated from (100)) simply refers to an interrupted
process, which may be resumed. Optional gat denotes some contradictory expectation
or presupposition.
(125)
Kio <i baa mago (gat)>.
neg 3sg dur sleep (cont)
‘He’s not/no longer sleeping.’
– Compare finally with fwâ kio (126) which refers to an event that has not yet started.
(126)
Fwâ kio i
mago (gat).
asp neg 3sg sleep cont
‘He’s not yet asleep.’
. Phasal verbs
So far, the focus has been on the morphological expression of aspects and on the
phasal aspects of a process; there are also a few phasal verbs expressing imminent
occurrence (diya me ‘about to’), incipient phase (thaaxa ‘begin’) or final phase (toven
‘end’). Duration can also be expressed lexically by verbs such as kuut ‘stand’ or mu
‘stay, remain’.
. Diya me: Imminence
Diya me (lit. do for) ‘be about to’ refers to an imminent action. It is compatible with
realis (127–128) or irrealis virtual frames (129).
(127)
I
u diya me i
(u) oxo-va.
3sg pft do sub 3sg pft follow-1pl.excl
‘He’s about to follow us.’
(128)
Bu i
diya me i
fhe aroo-ny
ai
a
Kaavo ?
for 3sg do sub 3sg take husband-poss.1sg where? agt Kaavo
‘So where is Kaavo about to take my husband ?’
(129)
O
diya me kot, na
fha kuru.
virt do sub rain coord blow wind
‘When it’s about to rain/should it be about to rain, the wind blows.’
© . John Benjamins Publishing Company
All rights reserved
Tense, aspect and mood in Nêlêmwa (New Caledonia)
. Inchoative verb: Thaaxa(puxet) ‘begin’
The inchoative verb thaaxapuxet is a compound of thaaxa ‘begin, start’ (which also has
spatial use)18 and puxet ‘origin’; it is compatible with present, past or future reference
frames.
(130)
a.
Dua hla thaaxapuxet me hla uya…
when 3pl begin
sub 3pl arrive
‘When they began to arrive…’
b.
I
thaaxa shaya.
3sg begin work
‘He started working.’
. Cessative verb: Toven ‘end, stop, finish’
Toven ‘stop; finish, end’ marks the last instant of a process.
(131)
Wa u toven o
thivi
maaxi i
na ?
2pl pft finish comp pluck.tr straw det 1sg
‘Have you finished plucking my straw ?’
The postverbal form, roven, is a verbal specifier in complex predicates:
(132)
I
u
keva wany xe wa-giik… xa keva roven wany hleny.
3sg pft build boat rel cl-one
also build finish boat this.dx1
‘He’s built a boat… and that boat is finished.’
. Durative verbs: Kuut ‘stand’, mu ‘stay’
The verbs kuut ‘stand’ and mu (the short form of muuvi ‘stay’) express durative
meanings.
(133)
Tu
kuut me
yo axe khabwe
ti.
go.down stand coord 2sg see comp(say) who?
‘Go down, stay, and see who it is.’
(134)
Hla taabwa me
hla u
muuvi mu.
3pl sit
coord 3pl pft stay
stay
‘They settled and they’ve been staying there.’
. Thaaxapuxet has a temporal prepositional function as in thaaxapuxet caae ‘starting from
tomorrow’, while thaaxa functions as a spatial preposition, as in thaaxa na Odayan (begin loc
Odayan) ‘starting from Odayan’.
© . John Benjamins Publishing Company
All rights reserved
Isabelle Bril
. Aspectual role of directional markers
One last part of speech which contributes to aspectual specifications is directionals; their main spatial and aspectual meanings are summarised in Table 9 (see Bril
2002: 301–305, Bril 2004: 99–128 for a detailed analysis).
Table 9. Directionals
DEICTIC DIRECTIONALS
NON-DEICTIC DIRECTIONALS
me
xi
ve
da
du
spatial
centripetal, hither
centrifugal
crosswise
upward
downward
aspectual
meaning
up to now
*
continuous
(or)
prospective
progressive
(cumulation
up to a point)
*
. Centripetal directional me: Present perspective
In its aspectual function, the centripetal morpheme me ‘hither’ (towards the speaker)
expresses the current relevance to the time of speech (t₀) of a past process which may
be completed or not, since current relevance does not presuppose the completion of
the process.
(135)
Na xam kua noolî=me.
1sg ass often see.tr=hither
‘I’ve often seen that.’ (until now)
(136)
Hooli
vhaa eli
xe hâ
u tâlâ=me.
that.anaph speech that.anaph tpc 1pl.incl pft hear=hither
‘Those words, we’ve heard them up to now.’
. Upwards da: Progression
The upward directional da marks a progressive and cumulative process:
(137)
Hla vhaa=da.
3pl speak=up
‘They’re talking away.’
When some time limit is expressed, da indicates progression up to that point.
(138)
Hla u civa=da [ku “cilic
civa!”].
3pl pft dance=up pft go.wrong dance
‘They danced until (they heard) « the dance has gone wrong! »’.
The verb o=da ‘go up’ also expresses cumulation and progression until some endpoint:
© . John Benjamins Publishing Company
All rights reserved
Tense, aspect and mood in Nêlêmwa (New Caledonia)
(139)
Dua u o=da uya mwelî,
hla thu wooxa agu.
when pft go=up arrive there.anaph 3pl do middle people
‘When things have reached that point, they divide people.’
(140)
O=da mwa xa tan.
go=up seq also be.night
‘In the end, it got dark.’
. Crosswise ve: Temporal expansion, continuous
When used aspectually, the crosswise19 directional ve refers to the unbounded expansion of an ongoing process. Ve is compatible with present, past or potential temporal
references, as shown below.
(141)
I hâk=ve.
3sg be.big=crosswise
‘He’s growing up.’
(142)
Co taa=ve
da
mwêna ?
2sg sit=crosswise what? there.dx2
‘Why do you remain sitting there?’
– Future or potential reference
(143)
Paari-na=ve
uya hna-kûûlî-a-t.
tell-1sg=crosswise arrive nmz-end-det-C
‘Tell me until the end.’
(144)
pwaxa taan mahleena io yo malep=ve
na-t.
for
day those.dx2 fut 2sg live=crosswise inside-C
‘for the days to come that you’ll be living.’
(145)
Kââlek
o
hâ
na=ve.
impossible virt 1pl.incl put=crosswise
‘It is impossible for us to postpone (it).’
. Aspectual conjunctive phrases with da ‘up’, ve (crosswise)
The directional also occurs in a conjunctive phrase uya=da, the grammaticalised
form of the verb uya ‘arrive’, which marks a final instant.20 If the endpoint is virtual, it
. The crosswise directional is used for directions such as ‘across a river’, but it is also used
for directions across the oriented ‘up-down’ axis; thus, it is also used with meanings such as
‘along the coast’, and with the aspectual meaning of progression.
. It also has a prepositional function as in uya êna ‘until now’ (lit. arrive now).
© . John Benjamins Publishing Company
All rights reserved
Isabelle Bril
combines with o, uya=da o ‘until’ (146a); if realis, it combines with xe (146b); if the
endpoint is past in a narrative, it combines with dua.
(146)
a.
Io na xam gaa
shaya maya uya=da o
fut 1sg ass persist work slowly arrive=up virt
na kûûlî.
1sg finish.tr
‘I’ll do this work slowly until I finish (it).’
b.
I
wâlem uya=da xe i
axe puyat xe pwa-giik.
3sg walk arrive=up sub 3sg see hill rel cl-one
‘He walks until he sees a hill.’ (lit. until he sees a hill that is one)
It may also combine with the crosswise directional ve to express progression until an
endpoint (147–148).
(147)
… uya=ve
ni thumaaxa i
aayo-ma.
arrive=crosswise in do-funeral det Chief-great
‘[they do so] until the great Chief ’s funerals (come).’
(148)
me oo=me
uya=ve
ni yeewat hleny.
and come=hither arrive=crosswise in time this.dx1
‘and it has been passed on until nowadays.’
(lit. and it has come here passing on until this present time)
. Conclusion
Tense (present or past) has no verbal exponent in Nêlêmwa; the time of events is merely
contextual and indicated at clause or sentence level by deictic and anaphoric adverbs.
In the aspectual domain, the core notions hinge on the bare aorist form, the perfect
marker (k)u ~ (x)u, the recent completion marker hââ, and a wealth of incompletion
markers encoding subtle semantic nuances, plus two pluri-actional markers. Directionals also participate in aspectual specifications. Mood has been shown to hinge on
the contrast between realis and irrealis, with some imbalance between them in terms
of morphological exponents: realis is Ø-marked, in contrast with the irrealis markers.
The virtual/subjunctive marker o also has epistemic value when used to refer to uncertain, potential situations.
In dependent clauses, some conjunctive markers like dua ‘when’ (+past) contain
inherent temporal reference. Some aspect morphemes also have inherent temporal
features: thus, kua (translatable by ‘often’) is only compatible with the past, hââ refers
to a recently/just completed process, taa₂ (translatable by ‘already’) marks previous
occurrence. Tense and aspect thus interact to some extent.
© . John Benjamins Publishing Company
All rights reserved
Tense, aspect and mood in Nêlêmwa (New Caledonia)
Coming back to the contrast between the bare aorist verb form and the perfect
marker (k)u ~ (x)u, the bare aorist verb form refers to (past or present) events, or to
sequences of events, without reference to their inner phasal structure. It also expresses
generic statements. On the other hand, the basic meaning of the perfect is to express
the relevance of some anterior event or process in relation to a reference frame (either
the time of speech (t₀) or some past event (t₂)); but the perfect has other functions and
meanings: it may refer i. to the resulting state of a completed process, ii. to a change
of state, or iii. to a transitional process or to a change of degree in the context of some
evolution. In complex clauses, the perfect may be used with backgrounding effects
and may express causal relations between processes and clauses. In irrealis reference
frames, the perfect may have deontic or epistemic meanings, it may assert the projected completion of a process as certain.
Thus, the fairly basic time reference stands in contrast with the wealth of aspectual
and phasal morphemes. This classifies Nêlêmwa as a mood- and aspect-predominant
language.
Abbreviations
agt agent marker; anaph anaphoric; asp aspect; ass assertive; c construct morpheme; cl
classifier; comp complementiser; conj conjunction; cont continuous; coord coordinator;
det determiner; dir directional; du dual; dur durative; dx deictic; evid evidential; excl
exclusive; exist existential; fem feminine; fr free pronoun; freq frequency; fut future; incl
inclusive; instr instrumental; intr intransitive;; iter iterative; loc locative marker; neg
negation; nmz nominaliser; perm permanence; persist persistive; pft perfect marker; pl
plural; pl.indf plural indefinite; poss possessive; pot potential; prev previous; proc processual; proh prohibitive; rec reciprocal; rec.complt recent completion; rel relative marker;
rem remansive; seq sequential conjunction; sg singular; sub subordinator; tpc topic marker;
tr transitive; unexp unexpected; virt virtual.
References
Bril, Isabelle. 1997. Split ergativity in Nêlêmwa. In Proceedings of the Seventh International
Conference on Austronesian Linguistics, Cecilia Odé & Wim Stokhof (eds), 377–393.
Amsterdam: Rodopi.
Bril, Isabelle. 2000. Dictionnaire nêlêmwa-nixumwak-français-anglais [SELAF 383, LCP 14].
Leuven: Peeters.
Bril, Isabelle. 2001. Postmodification and the structure of relative clauses in Nêlêmwa and
other New Caledonian languages. In Proceedings of the Second International Conference on
Oceanic Linguistics (SICOL), Vol. 2: Historical and Descriptive Studies [Pacific Linguistics
© . John Benjamins Publishing Company
All rights reserved
Isabelle Bril
505], Bill Palmer & Paul Geraghty (eds), 261–284. Canberra: The Australian National
University.
Bril, Isabelle. 2002. Le nêlêmwa (Nouvelle-Calédonie): Analyse syntaxique et sémantique [LCP
16]. Leuven: Peeters.
Bril, Isabelle. 2004. Deixis in Nêlêmwa. In Deixis and Demonstratives in Oceanic Languages
[Pacific Linguistics 562], Gunter Senft (ed.), 99–128. Canberra: The Australian National
University.
Bril, Isabelle. In preparation. Lexical and syntactic categories in Nêlêmwa (New Caledonia) and
some other Austronesian languages: Fluid vs. rigid categoriality. In Lexical Polycategoriality: Cross-linguistic, Cross-theoretical and Language Acquisition Approaches, Valentina Vapnarsky & Edy Veneziano (eds). Amsterdam: John Benjamins.
Comrie, Bernard. 1976. Aspect. An Introduction to the Study of Verbal Aspect and Related Problems. Cambridge: CUP.
Culioli, Antoine. 1983. Valeurs aspectuelles et opérations énonciatives: L’aoristique. In Linguistique, Enonciation: Aspects et détermination, Susan D. Fisher & Jean-Jacques Frankel (eds),
99–114. Paris: EHESS.
Culioli, Antoine. 1990. Pour une linguistique de l’énonciation, Opérations et représentations.
Paris: Ophrys.
Culioli, Antoine. 1999a. Pour une linguistique de l’énonciation, Vol. 2: Formalisation et opérations
de repérage. Paris: Ophrys.
Culioli, Antoine. 1999b. Pour une linguistique de l’énonciation, Vol. 3: Domaine notionnel. Paris:
Ophrys.
Dahl, Östen. 2006. Tense, mood, aspect: Overview. In Encyclopedia of Language & Linguistics,
Vol. 12, Keith Brown (ed.), 577–689. Oxford: Elsevier.
DOI: 10.1016/B0-08-044854-2/00247-9
Desclés, Jean-Pierre. 1989. State, event, process and typology. General Linguistics 29(3): 159–200.
Desclés, Jean-Pierre. 1995. Les référentiels temporels pour le temps linguistique. In Temps et
Langage, II: Modèles linguistiques, Tome 14(2). Paris.
Desclés, Jean-Pierre & Guentchéva, Zlatka. 1995. Is the notion of process necessary ? A fundamental distinction process in progress and state of activity. In Temporal Reference, Aspect
and Actionality, Pier Marco Bertinetto, Valentina Bianchi, James Higginbotham & Mario
Squartini (eds), 55–70. Torino: Rosenberg and Sellier.
Desclés, Jean-Pierre & Guentchéva, Zlatka. 2010. Référentiels aspecto-temporels: Une approche
formelle et cognitive appliquée au français. In 2e Congrès Mondial de Linguistique Française, Franck Neveu, Valelia Muni Toke, Jacques Durand, Thomas Klingler, Lorenza Mondada & Sophie Prevost (eds), 1675–1696. Paris: Institut de Linguistique Française.
Desclés, Jean-Pierre & Guentchéva, Zlatka. 2012. Universals and typology. In The Oxford Handbook of Tense and Aspect, Robert I. Binnick I. (ed.). Oxford: OUP.
François, Alexandre. 2003. La sémantique du prédicat en Mwotlap (Vanuatu). BSL LXXXIV.
Leuven: Peeters.
François, Alexandre. 2010. Pragmatic demotion and clause dependency. In Clause-linking and
Clause Hierarchy: Syntax and Pragmatics [Studies in Language Companion Series 121],
Isabelle Bril (ed.), 499–548. Amsterdam: John Benjamins. DOI: 10.1075/slcs.121.16fra
Hagège, Claude. 1986. La langue palau. Munich: Wilhelm Fink.
Lemaréchal, Alain. 1991. Problèmes de sémantique et de syntaxe en palau. Paris: Editions du
CNRS.
© . John Benjamins Publishing Company
All rights reserved
Tense, aspect and mood in Nêlêmwa (New Caledonia)
Lichtenberk, Frantisek. 2008. A Grammar of Toqabaqita. Berlin: Mouton de Gruyter.
DOI: 10.1515/9783110199062
Lyons, John. [1978 CUP]1990. Sémantique linguistique [Coll. Langues et langage]. Paris:
Larousse.
Moyse-Faurie, Claire. 1995. Le xârâcùù, langue de Thio-Canala (Nouvelle-Calédonie) [LCP 10].
Paris: SELAF.
Shintani, Tadahiko L.A. & Paita, Yvonne. 1990. Esquisse de la langue de Paita. SETOM.
© . John Benjamins Publishing Company
All rights reserved