Summary: Tense marking on NPs
Roland Hemmauer
roland.hemmauer at GMX.DE
Sat Jan 24 10:37:42 UTC 2004
Dear all,
earlier this week, Wolfgang Schulze from Munich University had kindly
forwarded you my query about tense marking on nominal phrases (NPs).
Thanks to him for this favour and thanks to all of you for your very
helpful responses!
As some of the replies were not directly posted to the list, I shall
give a complete summary here. The individual messages have been re-
arranged thematically for convenience, following this order now:
1) messages with bibliographical references to typological or theoret-
ical work
2) messages referring to single languages, language families, or lin-
guistics areas exhibiting tense marking on NPs:
a) general remarks
b) tense marking on predicative NPs
c) tense marking on non-predicative NPs
d) specific remarks on Tupi-Guaranian languages
3) queries for clarification
The individual messages (given with the author's name) and my own re-
marks will be separated by horizontal bars respectively.
Thank you all very much for your interest and for your very lively
participation!
Roland Hemmauer
(General Linguistics/
University of Munich)
Before starting, I shall reproduce my original query here:
------------------------------------------------------------------------
Dear LingTyp members,
as I am currently writing my M.A. thesis about tense marking on NPs
in Tupi-Guaranian languages of Lowland South America, I would like
to learn whether there are any other languages of the world that ex-
hibit the same or a similar phenomenon.
In order to sketch the problem briefly, I shall give the following
example taken from Paraguayan Guarani:
ka'a r-ogue-kué-gui o-jej-apo hey'u porã-rã
yerba mate LNK-leaf-PAST-ABL 3:S-PASS-make tea good-FUT
"From the leaves of the yerba mate tree, a good tea can be made"
(LNK = a linker indicating possessive relations on some nouns)
In this utterance, the 'leaves' are marked for past tense, as they will
be "ex-leaves" (leaves detached from their tree) once entering the pro-
cess of making tea. Conversely, the 'tea' is marked for future tense, as
it will be "future tea" only resulting from the production process. (The
verbal predicate itself is unmarked for tense in this example).
Virtually all Tupi-Guaranian languages share this ability to mark noun
phrases (NPs) for tense, (seemingly) independently from the tense mark-
ing occuring on the predicate, thus creating some notion of an internal
temporal structure of an event and of the exact temporal status of the
referential entities involved in it (formally, NPs).
My question, in a nutshell, is then:
Are there any other languages that have means to mark referential NPs
for tense, thus putting the referential entities represented by them
into their own temporal relations? If yes, what are the factors con-
ditioning this phenomenon? Possible questions include:
- syntactic/semantic scope of nominal tense marking? (e.g. single noun,
whole NP, possessive relation, whole utterance,...)
- syntactic/semantic restrictions on nominal tense marking? (e.g. sub-
ject/object, S/A/O, core/periphery, (in-)alienability, human/animate/
inanimate/mass/abstract,...)
- relations to the semantic content of the (verbal or other) predicate?
(e.g. state/process/action, achievement, accomplishment,...)
- relations to the tense/aspect/mood of the predicate? (e.g. past/pre-
sent/future as a reference point, perfective/imperfective, negation,
real vs. irreal, evendential,...)
- possible by-functions of the formal means of marking nominal tense?
(e.g. derivation of collective or instrumental nouns, aspect marking,
case marking,...)
Up to the moment, I am only aware of the notion of 'nominal aspect' (in-
dividual vs. mass nouns) and of the existence of a so-called 'modal case
system' in Kayardild, a language of Australia. Yet, aspect and mood are
not what I'm primarily looking for. - The phenomenon in question should
function as tense in the first place (in the sense of locating an enti-
ty in time relative to a temporal reference point).
It should be added that I don't mean a system of "floating" tense mark-
ing here, where genuine verbal tense markers can be attached also to NPs
in order to mark focus (as also found in Paraguayan Guarani), nor a sys-
tem of tense concord, where other constituents of a sentence must agree
in tense with the verb (e.g. adverbial temporal concord as found in Ma-
lagasy).
Thank you very much in advance for your hints to languages that share
this phenomenon! If it is customary to post a summary to the list, I
shall do so once I have collected a reasonable number of answers.
Kind regards
Roland Hemmauer
------------------------------------------------------------------------
1) messages with bibliographical references to typological or theoret-
ical work:
Alice Harris, Farrell Ackerman, and Alexandra Aikhenvald pointed out
to me several typological, and theoretical papers written by Rachel
Nordlinger and Louisa Sadler around the topic of nominal tense. Fur-
thermore, one of the authors, Louisa Sadler, supplied detailed cla-
rifications and provided links to some of their papers (see below),
with corrections and additions given by Wolfgang Raible. - Further
contributions came from Paul Hopper and Frans Plank.
------------------------------------------------------------------------
Louisa Sadler:
In a series of papers we have looked at a wide variety of cases of no-
minal tense marking, including both cases in which tense marking occur-
ring on nouns (and other elements within the NP) serves to temporally
locate the nominal (which we call variously Nominal Tense with Nominal
Scope and Independent Nominal Tense)-- the cases that the current post-
ings concern -- and cases in which tense marking occurring on nouns
(and other elements within the NP) serves to temporally locate the ver-
bal predicate itself (which we call variously Nominal Tense with Clausal
Scope and Propositional Nominal Tense).
Most relevant to the current query is a paper we have under review for
journal publication : Rachel Nordlinger and Louisa Sadler "Nominal Ten-
se in Crosslinguistic Perspecetive", in which we provide an overview of
the different types of nominal tense marking which we have found in the
languages we have looked at including a numbefr of the languages men-
tioned in this thread -- please contact us if you would like a copy of
the current draft, which contains data from a considerable number of
languages.
Alongside this more typologically oriented paper we have a more theore-
tical paper on nominal tense with clausal scope, that is, on cases in
which tense marking on the nominal dependent is interpreted with respect
to the verbal predication (where the verb itself may or may not carry
temporal features). This paper provides an analysis of this propositio-
nal nominal TAM within the framework of LFG building on and extending
the account of Kayardild modal case presented in Rachel's thesis: Rachel
Nordlinger Constructive Case (CSLI Publications 1998). This paper will
appear in NLLT as Rachel Nordlinger and Louisa Sadler "Tense Beyond the
Verb", sometime soon we hope.
More preliminary discussion of some of this data and how an account of
it may be given within the framework of LFG is to be found in a number
of earlier and shorter papers we have contributed to proceedings of LFG
conferences:
Rachel Nordlinger and Louisa Sadler (2000)
"Tense as a Nominal Category": Proceedings of LFG 2000:
www-csli.stanford.edu/publications/LFG/5/lfg00.html
Louisa Sadler and Rachel Nordlinger (2001)
"Nominal Tense: A Preliminary Sketch": Proceedings of LFG 2001:
www-csli.stanford.edu/publications/LFG/6/lfg01.html
Louisa Sadler and Rachel Nordlinger (2003)
"The Syntax and Semantics of Tensed Nominals" Proceedings of LFG 2003
www-csli.stanford.edu/publications/LFG/8/lfg03.html
------------------------------------------------------------------------
Wolfgang Raible:
As Louisa Sadler remarks, this topic has been most extensively and tho-
roughly treated in a thread originating from Rachel Nordlinger, Univer-
sity of Melbourne, and dating back to 2001/2002. In August 2002, both
of them gave an excellent summary of the discussion in the form of an
Acrobat file. Since the links given in the mail by Louisa are more or
less empty, I attach this file for convenience.
Just one addition: in this text there is an empty reference to a contri-
bution of myself. It should read W.R., "Language universals and language
typology". In: Haspelmath, Martin & König, Ekkehard & Oesterreicher,
Wulf & Raible, Wolfgang (eds.), Language Typology and Language Univer-
sals - Sprachtypologie und sprachliche Universalien - La Typologie des
langues et les universaux linguistiques. An International Handbook - Ein
internationales Handbuch - Manuel international. Berlin & New York: de
Gruyter, Handbücher zur Sprach- und Kommunikationswissenschaft vol 20.1:
1-24 (article 1).
------------------------------------------------------------------------
Paul Hopper:
The idea that all NPs contain tense in their "deep structure" was exam-
ined as a theoretical question by Emmon Bach in the 1960s ("Nouns and
Noun Phrases", in E Bach and R Harms, eds., Universals in Linguistic
Theory, UT Press, 1968.)
------------------------------------------------------------------------
Frans Plank provided the following link to "Das grammatische Raritäten-
kabinett", a collection of 'grammatical curiosities':
http://ling.uni-konstanz.de:591/universals/introrara.html
(with non-nominal [sic!] tense covered in items No.s 43 and 45 and no-
tense in item No. 118)
------------------------------------------------------------------------
2) messages referring to single languages, language families, or lin-
guistics areas exhibiting tense marking on NPs:
a) general remarks:
------------------------------------------------------------------------
Alexandra Aikhenvald:
Having nominal tense is a rather common property of South American
languages. In my grammar of Tariana (CUP, 2003) this is discussed at
length in Chapter 8. There are also examples in Carib languages (see
Derbyshire's chapter in THE AMAZONIAN LANGUAGES, edited by Dixon and
Aikhenvald).
[...]
One thing I have not found (and I looked hard!) is nominal evidentia-
lity, that is, when nouns have a special evidential marking distinct
from that on verbs. An evidential can have a noun (not the whole clau-
se) in its scope - an example comes from Jarawara, an Arawa language
from Brazil (this will be in R.M.W. Dixon's Grammar of Jarawara, to
appear with OUP in 2004). This same example will also appear in my
book EVIDENTIALITY (OUP, September 2004). I find it really curious...
------------------------------------------------------------------------
Edward H. Bendix:
Nootka has tense marking on NPs. Whorf used this to claim that the
speakers saw things as events.
Also Salishan languages.
For published works on these languages, you could browse the following:
http://www.pitbossannie.com/index-h-pm.php
Members of the Salishan interest group might also help you:
http://listserv.linguistlist.org/cgi-bin/wa?A0=salishan
------------------------------------------------------------------------
b) tense marking on predicative NPs:
This type appears to be widely attested across the languages of the
world. The following answers are about Cha'palaachi (Cayapa), a lan-
guages of Ecuador, some Uralic languages, and Araki and Mwotlap, two
languages of Vanuatu. The Vanuatu case is special in distinguishing
between time-stable and change-of-state nominal predicates, the lat-
ter of which are marked for TAM. Furthers candidates are Polynesian
languages in general (see message from Alex François), and, to my
knowledge, Turkic languages.
------------------------------------------------------------------------
Enrique Bernárdez:
In cha'palaachi or Cayapa, spoken in the Nortwest of Ecuador, some
tense/aspect morphemes can be used directly on nouns. For instance:
ruku-ve "man-3rd person singular": s/he is (but also: you are) a man.
ruku-wa "man (narrative, imperfect) past" : he (etc) was a man
ruku-mi "man past". he was a man
Equally with the past:
ruku-la : "man pl-past"(they etc) were men
------------------------------------------------------------------------
Florian Siegl:
Tense marking on NPs in Samoyedic languages is known (Samoyed / Uralic),
although detailed syntactic studies are unfortunately still missing.
Beside nouns, also adjectives, numerals, pronouns, and participles (sub-
stantivized verbs) can be "conjugated" in singular, dual and plural.
Whereas all Samoyedic languages have a rather rich verbal morphology,
these explicit forms can occur only in the aorist and in the preteri-
te; modal markers can not be attached
These suffixes which are used to express tense on the NP are the same
markers as used in the so called subjective (intransitive) conjugation.
e.g. Enets:
1) mod'i ese-do'
I father-aor.subj1sg
'I'm a father'
2) mod'i ese-do-d'
I father-pret.subj1sg
'I was a father'
Hajdu, Peter 1975. "Prädikative Nominalflexion in den samojedischen
Sprachen". Acta Linguistica Academiae Scientarum Hungaricae, Tomus
25. pp.1-30.
Hajdu, Peter 1968. Chrestomathia Samoiedica. Budapest p. 47 on Tun-
dra Nenets; p.141-142 on Selkup [in Hungarian; several reprints]
Salminen, Tapani 1997. Tundra Nenets inflection. SUST 227. Helsinki
(p.130)
Künnap, Ago 1999. Enets. Languages of the World 186. Lincom (p.29-30)
b) A similar system is known in Erzya and Moksa Mordva (Volga / Finno-
Ugric / Uralic).
------------------------------------------------------------------------
László Honti provided the following reference to his summarizing paper
on nominal predicates in some Uralic languages:
Morphologische Merkmale des nominalen Prädikats in einigen uralischen
Sprachen. Linguistica Uralica 28. Tallinn, 1992. 262-271.
------------------------------------------------------------------------
Alex François:
Araki, an Oceanic language of Vanuatu (with about 3 speakers left), has
two ways to mark tense-aspect-mood on NPs [and I believe several other
Oceanic languages behave the same].
* When they form an NP which is not in predicate position (=argument,
complement of prepositions...), nouns cannot bear any TAM marker
(contrary to Tupi-Guarani):
BUT
1) Araki lacks any copula, in such a way that nouns are directly pre-
dicative; most of the time, this involves a time-less equational
predicate (bare noun, without TAM marking):
Nia moli.
/[free.pronoun]3sg/chief/
'he is (a/the) chief.'
but if the noun predicate is for some reason not "time-stable" (e.g.
change of social status, change of identity...), then the noun will
directly be marked in Tense-Aspect-Mood, just like verbs (here co is
a subject clitic = portmanteau morpheme Person+Mood realis; pa is an
tense-aspect marker for Future):
(Nia) co pa MOLI.
/[free.pronoun]3sg/[subject.clitic]3sg:Irrealis/FUT/chief/
'he will be a/the chief.' (lit. he will chief.)
or
(Nia) mo re MOLI mo iso.
/3sg/3sg:Realis/PFT/chief/3sg:Realis/finish/
'he is already a/the chief.' (lit. he has already chief-ed.)
In this type of construction, nouns take the same slot as verbs, de-
monstrably without ceasing to be nouns. This is a context where sev-
eral Oceanic languages perfectly allow TAM-marking on nouns -- al-
though, to be precise, this is limited to the predicate position (*Co
pa moli cannot be used as an NP "the future chief").
Another language on which I have detailed evidence for this 'TAM-sen-
sitiveness of nouns' is Mwotlap (also Vanuatu): see discussion in Fran-
çois (2003: 47-75). I can give you more examples of Mwotlap if you wish.
Finally, you should also have a look at Polynesian languages, which as
far as I know allow quite different constructions that combine TAM with
nouns (knowing that the distinction between nouns and verbs is even pro-
blematic for some of these Polynesian languages!) I'm not insisting he-
re, since I'm afraid this is not exactly the sort of "tense-marked NPs"
you were asking for (or is it?)...
(continued below)
------------------------------------------------------------------------
c) tense marking on non-predicative (= referential) NPs
This type (which is closest to the phenomenon found in Tupi-Guaranian
languages) appears to be quite widespread in South American languages
in general (see also Alexandra Aikhenvald's mention above), and is al-
so found in Yukaghir of Siberia, again in Araki of Vanuatu, and in Ma-
lagasy. Yet there seems to be a considerable degree of typological va-
riation among the individual languages as to the structures and func-
tional scopes of the respective constructions. (Interestingly enough,
the construction in geographically distant Yukaghir is what reminds me
most of the Tupi-Guaranian case!).
------------------------------------------------------------------------
Daniel Everett:
You ask about tensed referential NPs. My data has to do with pronouns,
deictics in particular. Wari deictic pronouns fall into two paradigms,
spatial and temporal. A recent working ms. that contains information on
them is my Liminal Categories paper on my website.
------------------------------------------------------------------------
Jeanette Sakel:
Das benachbarte Mosetén (Mosetenan, Bolivian lowlands) hat auch einen
nominalen "Tempusmarker", jedoch nicht ganz so extrem wie die Tupí-
Guaraní-Sprachen.
wiya'-win = old man-PAST (i.e. dead old man)
bae'edye'-win = Dorf-PAST (i.e. das Dorf, das es nicht mehr gibt)
Diese Form wird jedoch nicht grammatikalisch (obligatorisch) benutzt,
sondern ist eher lexikal. Ich habe einiges darüber in meiner Gramma-
tik, die demnaechst bei Mouton herauskommen wird. Es gibt auch eine
frühere Version, die aber nicht unbedingt öffentlich zugängig ist
(alle holländischen Universitäten sollten jedoch eine Kopie haben,
falls du daran interessiert bist).
------------------------------------------------------------------------
Katharina Haude:
In Movima, an unclassified language of lowland Bolivia, past tense is
marked within the NP, namely on the article or pronoun. (Future is mar-
ked by a particle in the VP, but there is no morphological tense mark-
ing on the verb itself at all.) Past-tense marking indicates that the
referent is no longer in existence, so it basically has a scope over
the NP only. Note the contrast between the unmarked neuter article _aj_
in (1) and the one marked for past tense, _oj_, in (2):
(1) aj asna-y'Li
ART.ntr home-1excl
'our home'
(2) oj asna-y'Li
ART.ntr.pst home-1excl
'our former home (which doesn't exist anymore)'.
However, nominal tense marking can also have a larger scope. In past-
tense contexts it is the only grammatical indicator of past tense:
(3) ilo:ni=y'Li n-oj chaMmo
walk=1excl OBL-ART.ntr.pst forest
'we walked in the forest'
Time stability, animacy and relevance of the referent play an important
role. Nouns denoting non-human entities whose existence at the moment
of speaking is not relevant are marked more easily for past tense than
nouns with human referents (which is why the contrast is best seen on
the neuter article). Furthermore, the main device for forming subordin-
ate clauses in Movima is nominalization; the concept expressed by a de-
verbal noun is usually not time-stable, so the NP is easily marked for
past tense:
(4) n-oj iloni-wa-y'Li
OBL-ART.ntr.pst walk-NMZ-1excl
'when we walked (lit: "at our past-walking")'
By contrast, time-stable concepts are not marked for past tense even in
past-tense contexts, as is shown by the unmarked article of the second
NP in (5):
(5) n-oj ayloba-wa-y'Li nosdé n-aj kolegiyo
OBL-ART.ntr.pst gather-NMZ-1excl over.there OBL-ART.ntr school
'when we gathered (lit. "at our past-gathering") over there at the
school'
I hope that this has answered some of your questions.
------------------------------------------------------------------------
Elena Maslova:
Yukaghir (Paleo-Siberian) has a future nominal suffix. Its scope is NP.
So, for example, [met ti:]-moraw, literally [my people]-FUT, would mean
"my relatives-to-be", appropriate after an engagement but before the
wedding (= future [my people], not my [future people]) (example from
Tundra Yukaghir). It seems to be at the present time most common with
de-verbal nominalizations, which can function as complement, headless
relative, or adverbial clauses, and where the nominal future functions
as a relative tense/purpose marker.
------------------------------------------------------------------------
Alex François: (continued)
2) closer to your query, Araki (not Mwotlap) has the possibility of cod-
ing TAM inside NPs even outside the predicate position, provided it
has a numeral. Indeed, numerals in this language behave like (or bet-
ter said: form a subset of) verbs, which means that a NP with a num-
eral will necessarily be marked in (Tense-Aspect)-Mood [without the
need for a relative structure, contrary to verbs]. Most of the time,
this takes the form of a (3 singular) subject clitic, either Realis
or Irrealis.
Generally speaking, there seems to be some sort of mood agreement be-
tween the main verb and the mood of its surrounding NPs (relevant NP
is bracketed):
Nam inu-mi-a [LASA ai mo dua].
/1sg:Realis/drink-TRansitiviser-3sg/cup/water/3sg:Realis/two/
'I drank two glasses of water'
lit. I REALIS-drink glasses of water REALIS-two
Nam de na inu [re LASA ai co dua].
/1sg:Realis/say/1sg:Irrealis/drink/ PARTitive
/cup/water/3sg:Realis/two/
'I'd like to drink two glasses of water'
lit. I say I IRREALIS-drink some glasses of water IRREALIS-two
Interestingly, the contrast Realis/Irrealis on numerals is directly
correlated with referentiality of the NP (François 2002: 58): basic-
ally Realis NPs are [+ref], irrealis NPs are [-ref].
But discrepancies between mood marking of the predicate vs. the NP are
also attested: (François 2002: 110)
Na pa sivo lesi-a [racu mo hese hosun].
/1sg:Irrealis/FUT/go.down/see-3sg/man/1sg:Realis/one/here/
'I will go and have a look at that man over there.'
lit. I IRREALIS go look the man REALIS-one there
In my view, this is directly linked with the referentiality/specificity
of the referent. This can have interesting semantic effects on the in-
terpretation of the whole NP/sentence, e.g. (François 2002: 55):
Nam dogo na pa vavere [lo vere mo hese].
/1sg:Realis/feel/1sg:Irrealis/FUT/sing/OBL/song/3sg:Realis/one/
lit. I feel I IRREALIS-sing a song REALIS-one
'I want to sing a (specific) song = There is a (specific) song I'd
like to sing.'
Nam dogo na pa vavere [lo vere co hese].
/1sg:Realis/feel/1sg:Irrealis/FUT/sing/OBL/song/3sg:Irrealis/one/
lit. I feel I IRREALIS-sing a song IRREALIS-one
lit. 'I want to sing a (non-specific) song that should be one.'
= 'I feel I will sing one song (only one song, whatever it is)'
References:
François, Alexandre (2002). Araki. A disappearing language of Vanuatu.
Pacific Linguistics, 522. Canberra: Australian National University.
--- (2003). La sémantique du prédicat en mwotlap (Vanuatu). Collection
Linguistique de la Société de Linguistique de Paris, 84. Paris, Lou-
vain: Peeters.
Hope this is of interest for your study.
------------------------------------------------------------------------
Doris Payne:
Malagasy marks tense on Locative prepositions, as well as on verbs.
(This is based on work by Ando Rasolofo.)
------------------------------------------------------------------------
d) specific remarks on Tupi-Guaranian languages:
------------------------------------------------------------------------
Françoise Rose:
I just finished a dissertation on the morphosyntax of Emérillon, a Tupi-
Guarani language. I always hear people talk about your topic as if these
morphemes were widely used in every TUPI-GUARANI language. It is not the
case for Emérillon. This is why I want to share information with you.
I have not specificly worked on your topic, but some small paragraphs of
my disseration may give you at least some data on the language, even if
he analysis was not conducted in terms of "nominal tense".
Nominals can bear a -kw at l suffix. (@ is a schwa in this message!) with
about the same meaning as ex-leaves. It is not always possible to use
it and it enters in competition with -w at l or -el [sic!]. It may though
be more used than ex- in English or French.
[In a personal communication, Françoise confirmed the correctness of the
discrepancy between the schwa vowel in the suffixes -kw at l and -w at l and a
normal /e/ in the suffix -el. Nevertheless, all three suffixes seem to
derive historically from the allomorphs of the Proto-Tupi-Guaranian "no-
minal past" suffix *-kwér ~ *-wér ~ *-ér]
I do not think there would be a future nominal tense marker, though the
suffix -am is sometimes close to it.
Nominals can also be used as predicates, and in that function take the
same TAM marker as the verbs that predicate.
Moreover, the language uses lots of so-called "particles" (in fact, cli-
tics in the language). Most of them occur on the 1st element of the sen-
tence, be it an NP, a GP, an Adv, or a verb.
[...] my PhD dissertation (around 700 pages) is online at:
http://www.ddl.ish-lyon.cnrs.fr/Membre.asp?Action=Edit&Proprietaire=Francoise%20ROSEF
You will find information on -kw at l p. 175 (very short), on -am p. 335-
341, on the TAM system (clitics and suffixes on verbs or SV, and "par-
ticles"): chapter 12.
I have more information on -kw at l and its competitors in a note file,
that I could try to clean and send you in February if you are inter-
ested.
------------------------------------------------------------------------
Alexandra Aikhenvald:
The best analysis of this phenomenon in a Tupi-Guarani language is Lucy
Seki's GRAMATICA DO KAMAUIRA, 2000, Editora da Unicamp.
------------------------------------------------------------------------
Paul Hopper:
[...] There is some discussion of tensed NPs with Guaraní data in:
Claude Boisson et al., "Problématiques des parties du discours" in
Louis Basset and M. Pérennec, eds., Les classes du mot: traditions
et perspectives. Lyons: Presses Universitaires, 1994.
------------------------------------------------------------------------
3) queries for clarification:
------------------------------------------------------------------------
Gideon Goldenberg:
The exact status of the morpheme glossed as PAST in not clear to me.
If it is not the same that marks past tense of verb-forms, is it the
commonly known nominal tense-marker as in "the ex-governor is the fu-
ture president"?
More information about the Lingtyp
mailing list