Exist Pred
Exist Pred
Exist Pred
Abstract. This article is about the typology of inverse locational predication, commonly
called existential predication, illustrated by English There is a book on the table. After
discussing the definition of existential predication and establishing the distinction
between the languages that have grammaticalized an existential predicative construction
and those that have no dedicated existential predicative construction, I define seven
types of existential predication and discuss their distribution in the worlds languages on
the basis of a sample of 256 languages.
1. Introduction
1.1. What this article is about
This article puts forward a typology of inverse locational predication, more
commonly (but somewhat misleadingly) called existential predication, as illustrated
by sentences such as English There is a book (on the table), French Il y un livre (sur la
table), German Da liegt ein Buch (auf dem Tisch), Turkish (Masada) bir kitap var,
Tswana Go na le buka (fa tafoleng), etc. Cross-linguistically, the predicative
constructions illustrated by such sentences differ to a considerable extent both in
their formal make-up and in the precise range of their uses, but their common
property that will be retained here as criterial for identifying a predicative
construction as existential is their ability to provide an alternative way of encoding
the prototypical figure-ground relationships also denoted by plain locational
sentences such as English The book is on the table, French Le livre est sur la table,
German Das Buch liegt auf dem Tisch, Turkish Kitap masadadr, Tswana Buka e fa
tafoleng.
The term existential predication is retained here, in spite of its shortcomings, as
the label most commonly used by linguists to designate such predicative
constructions.1 In other words, in this paper, existential clauses must be understood
Although the use of the term existential with reference to predicative constructions typically used
to denote accidental and temporary presence of an entity at a certain location is fairly common, this
term also has other uses that must be mentioned in order to prevent misunderstandings. Some
authors restrict the use of existential to clauses in which no location is expressed (There is a book),
and use locative existential or locative presentative for existential clauses in which a location is
ha
thereexpl has
la
the
policia al
police
pati.
in.the courtyard
telfono.
1PL
cikin mganr.
in
matter.DEF
predication, most of them implicitly consider crucial the fact that it shares with
plain locational predication the ability to encode prototypical figure-ground
relationships. This delimitation of existential constructions is explicitly posited here.
By prototypical figure-ground relationships, I mean episodic spatial relationships
between a concrete entity conceived as movable (the figure) and another concrete
entity (the ground) conceived as occupying a fixed position in the space, or at least
as being less easily movable than the figure, as in The dog is under the tree or The book
is on the table.
For example, clauses such as There is a pond in front of our house or There is a stain
on the mirror are instantiations of the existential predicative construction of English
There is N (Loc), but the ability to encode such relationships is not decisive in
identifying the predicative construction in question as existential, and in some
languages, the usual translational equivalent of such clauses involves a predicative
construction that does not qualify as existential in the precise meaning given here
to this term.
In particular, in languages sensitive to the distinction between stage-level and
individual-level presence (i.e., in languages that use or tend to use distinct
constructions for There are dogs in the garden and There are lions in Africa), the
construction retained for the typology proposed here is that used to encode the
episodic presence of an entity at some place. Constructions that may provide
translational equivalents of English there is or French il y a in some of their uses but
are not available to express things like There are dogs in the garden (such as German
Es gibt N (Loc)) are not retained in the typology put forward in this paper.2 This
question will be briefly resumed in Section 2.7.
What I would like to emphasize at this point is that the constructions investigated
in this paper are not identified by reference to an abstract meaning accounting for
all of their uses, which in some respects show important cross-linguistic variations.
In this domain as in others, it is impossible to find two constructions in two different
languages with exactly the same range of uses. What identifies a predicative
construction as an instance of existential predication in the sense given to this term
in this paper is its participation in contrasting pairs of sentences referring to
prototypical figure-ground relationships such as The dog is under the tree / There is a
dog (under the tree) or The book is on the table / There is a book (on the table). The
restrictions in the use of the constructions available to encode such relationships and
their extension to the expression of situations other than prototypical figure-ground
relationships show important cross-linguistic variations, but the ability to be
involved in such contrasts is the criterion according to which I have selected the
predicative constructions dealt with in this paper.
On the restrictions on the use of German es gibt, often (but erroneously) described as an equivalent
of English there is or French il y a, see Czinglar (2002).
the use of negative sentences such as those in (3), which quite obviously do not deny
the existence of the person mentioned.3
(3)
a. Il
itexpl not
thereexpl
avait
had
Ivan.GEN
ne
NEG
bylo
na
be.PST.SG.N at
pas Jean au
not
cours.
lekcii.
lecture.PREP
Note that, due to the restrictions on the use of existential predication in English, the English
translation of these examples cannot reflect the nuance that distinguishes them from the
corresponding locational sentences (French Jean ntait pas au cours, Russian Ivan ne byl na lekcii).
(4)
Mandinka (pers.doc.)4
a. Kolo-kono-tt-o ma
well-in-frog-D
CMP.NEG
3SG
know
k
QUOT
fakas-o
ocean-D
be
LCOP
ke-ri.
occur-RES
The frog that lives in the well does not know that the ocean exists. (proverb)
b. Moo-siifaa
jma le
person-kind
many
be
FOC
ke-ri.
LCOP
occur-RES
be
dog-D
ke-ri yr-o
LCOP
occur-RES
tree-D
kto.
under
Ivan.GEN
ne
NEG
bylo
na
be.PST.SG.N at
lekcii.
lecture.PREP
2007), such sentences behave in all relevant respects like those whose existential
nature is not controversial, and the problem they raise is just the consequence of a
wrong definition. Borschev and Partee argue that existence (in the sense relevant to
the analysis of so-called existential construction) is always relative to location, and
that, once existential and locational clauses are clearly defined as encoding two
different perspectives from which the same existence/location situations may be
structured (either from the perspective of the figure or from the perspective of the
ground),5 there is no difficulty in accepting that sentences such as (5) are existential
sentences.
At first sight, one might get the impression that, when arguing in detail in favor of
such a position (which was already explicitly defended by Bally (1932), and which I
adopted in my habilitation thesis on possessive constructions see Creissels (1979:
376-385)), Borschev and Partee are just kicking down an open door, given the huge
literature on the relationship between locational and existential constructions.
However, most of the works that have discussed this issue take for granted the
existence of a relationship between existential and locative constructions but do not
discuss its nature. They simply ignore the problem of the cross-linguistic
identifiability of existential constructions and concentrate on the possibility of a
syntactic derivation of constructions labeled existential, but whose existential
nature is not discussed, from locational constructions (or of both from a deep
syntactic structure shared by existential and locational constructions). The technical
details of the derivation within the frame of the successive versions of the generative
model (in particular, the relevance of the notion of small clause to the analysis of
existential and locational clauses) stand at the center of attention. It is interesting to
observe that most generative accounts of the relationship between locational and
existential predication agree that locational predication is in some sense more basic
that existential predication, but in general, the authors do not comment this choice,
or try to imagine purely syntactic justifications.
Borschev and Partee do not only depart from other authors in that they discuss
the very nature of the relationship between existential and locational predication,
rather than its morphosyntactic manifestations. They also show the shortcomings of
the position adopted by most of the authors that have tackled this question,
according to which there is a straightforward relationship between information
structure and the choice between existential and locational predication.
It has been proposed that existential clauses are thetic clauses, or that an essential
feature of existential clauses is the rhematicity of the figure, but Borschev and Partee
convincingly argue that none of these two positions is compatible with the
acceptability of existential sentences such as (6) in Russian.
(6)
looked.for
kefir
kefir.GEN in store.PREP
NEG
bylo.
be.PST.SG.N
[I was looking for kefir.] There wasnt any kefir in the store.
5
The terms used by Borschev and Partee are not figure and ground, but thing and location.
On the basis of such observations, Borschev and Partee conclude that a notion of
Perspectival Structure distinct from the Theme-Rheme or Topic-Focus structure must
be introduced. Both existential and locational predication encode an abstract
predicate BE.AT(FIG, GR) figure is at ground. All languages have a construction
analyzable as locational predication, encoding the choice of the figure as the
Perspectival Center, which constitutes the unmarked choice because of the
ontological status of the two arguments of the abstract predicate BE.AT. But in
addition to that, some languages have grammaticalized a predicative construction
encoding the choice of the ground as the Perspectival Center: An analogy can be
made with a video camera and what the camera is tracking. A Predication sentence
[i.e., a locational sentence] keeps the camera fixed on the protagonist as she moves
around (THING as Center), an Existential sentence is analogous to the way a security
camera is fixed on a scene and records whatever is in that location (LOC as Center).
(Partee and Borschev 2007). Perspectival structure is basically a structuring at the
model-theoretic level ... [that] reflects cognitive structuring of the domains that we
use language to talk about, and are not simply given by the nature of the external
world. In other words, perspectival structure is basically a choice between different
possible conceptualizations of a situation, not between different ways of packaging
information, although the choice of a particular perspective has consequences for
the expression of information structure.
This notion of perspectival structure needs further elaboration, but it accounts for
the fact that, in languages in which a dedicated existential predication can be
recognized, the unmarked or default topic in existential clauses is the ground, and
the figure is most of the time to be interpreted as rhematic, but the discourse status
of the ground and the figure in existential predication is not irrevocably fixed, and
operations expressing variations in information structure can affect existential
predication like other predicative constructions, as illustrated by Ex. (6) above. The
notion of perspectival structure explains at the same time why, in languages devoid
of a dedicated existential predication, rough equivalents of existential predication
can be obtained by means of manipulations of the information structure of locational
sentences, as observed for example in Maslovas description of Tundra Yukaghir
Maslova (2003).
Interestingly, the formulations found at least in some of the formal analyses of the
relationship between locational and existential predication reveal a basic intuition
quite compatible with Borschev and Partees analysis in terms of unmarked vs.
marked perspectival structure. For example, Moro (1992) designates locational
sentences as canonical copular sentences, and existential sentences as inverse
copular sentences.
2.5. On the non-universality of existential predication
Another important issue is the universality of existential predication. It is widely
assumed that all languages have a dedicated existential construction, as illustrated
by the following statement from Moro (1992): In all languages there is a SPECIFIC
construction which is called existential sentence. (emphasis mine) However, in
a. Mae r
is
car
the
car
yma.
here
car
yma.
here
Mandinka (pers.doc.)
a. Wul-o be
dog-D
LCOP
yr-o
tree-D
kto.
under
The dog is under the tree. or There is a dog under the tree.
b. Wul-o le
dog-D
FOC
be
LCOP
yr-o
tree-D
kto.
under
There is a dog under the tree. or It is the dog that is under the tree.
(in French, Il y a le chien sous larbre would be another possible translation)
c. Wul-o be
dog-D
LCOP
yr-o
tree-D
le
FOC
kto.
under
The dog is under the tree. Or It is under the tree that the dog is.
In other languages, for example Russian, the recognition of a dedicated existential
construction is uncontroversial in some conditions, but problematic in others. As
noted by Partee and Borschev (2002), under negation, case-marking and agreement
sharply distinguish existential and plain sentences, but in the absence of negation,
because of (a) great freedom of word order and (b) no articles, the difference
between existential and plain sentences is less obvious, and it is natural to view
the sentences in (9) as differing only in Theme-Rheme structure and word-order (and
correspondingly in definiteness of the bare NP); the issue of whether there is any
deeper syntactic difference between them is controversial.
(9)
a. V gorode
in town.PREP
byl
doktor.
be.PST.SG.M doctor
v gorode.
be.PST.SG.M in town.PREP
2.7. Typical and less typical uses of existential predication, and the selection
of the data for a typology of existential predication
A crucial aspect of the typology of existential predication I am trying to elaborate in
this article is the selection of the relevant data on the basis of a rigorous definition
of existential predication. To have an idea of the problem, lets take an English
clause like There are two varieties of millet. This is unquestionably an existential
clause in the sense that it instantiates the English predicative construction (There is N
(Loc)) meeting the definition of existential predication retained in this article. But
the use of the English existential predication it illustrates is not the one that
unambiguously identifies this predicative construction as existential according to my
criteria. In other words, the ability of a construction to provide a translational
equivalent of There are two varieties of millet does not ensure that it meets the
definition of existential predication as a construction that can be used to express a
particular perspectivization of prototypical figure-ground relationships.
For example, in Mandinka, this sentence has three equally usual translational
equivalents Ex. (10), but none of them qualifies as existential on the basis of a
rigorous definition of existential, and more generally, as already stated in Section
2.5, Mandinka is among the languages that have not grammaticalized a construction
characterized by its ability to express an alternative perspectivization of prototypical
figure-ground relationships:
(10a) is a locational clause in which the figure phrase is focalized, and the role
of ground is fulfilled by jee, normally interpreted as there (as in Faat be jee
Fatou is there) but used here as a mere place filler with no specific reference;
(10b) illustrates a construction, already illustrated by Ex. (4), which carries a
meaning similar to that of English exist but cannot be used to express an
alternative perspectivization of prototypical figure-ground relationships;
(10c) illustrates another usual Mandinka equivalent of English exist, with the
verb sot be available;6 this construction is also used for example to ask a
shopkeeper about the availability of some product, but it cannot be used to
express the equivalent of English There is a dog in the room.
(10) Mandinka (Creissels and Sambou 2013: 139)
a. oo-sifa
millet-variety
ful le
two
FOC
jee.
LCOP
there
be
ke-ri.
millet-variety
ful le
two
FOC
LCOP
there
The same verb used transitively constitutes the usual equivalent of English get and have.
c. oo-sifa
millet-variety
ful le
two
FOC
soto-ta.
be_available-CMPL
When selecting data for this study, the main difficulty I had to face is that some
descriptive grammars have a section on existential clauses, but all the examples they
provide refer to non-prototypical figure-ground relationships, and nothing ensures
that, in the language in question, the same construction would be available to
express an alternative perspectivization of prototypical figure-ground relationships.
In particular, many languages have constructions comparable to the Mandinka
constructions with keri or sotot, which constitute usual equivalents of exist or be
available and are commonly used to encode situations involving atypical figures
and/or grounds, or to express habitual presence of an entity at some place, but are
not used to encode accidental presence of a concrete and movable entity at some
place, and therefore do not qualify as existential constructions as defined here. This
is for example the case of the following constructions:
German es gibt N (Loc), lit. itexpl gives N (Loc). 7
Swedish det finns N (Loc) and Norwegian det finnes N (Loc), lit. itexpl finds itself
N (Loc), with a mediopassive form of the verb find;
Russian imeetsja N (Loc), lit. has itself N (Loc) with a mediopassive form of the
verb have.
2.8. Existential predication and negation
In some of the languages that have grammaticalized an existential predicative
construction, no special negative strategy is required for existential clauses (English
there is / there is not, French il y a / il ny a pas, Spanish hay / no hay, etc.) In some
others, a special negative strategy is used for existential clauses, and many languages
have a suppletive negative existential predicator, for example Turkish var there is /
yok there is not Ex. (11).
(11) Turkish (pers.doc.)
a. Masa-da
table-LOC
bir
one
kitap
book
var.
EXIST
b. Masa-da
table-LOC
kitap
book
yok.
EXIST.NEG
Czinglar (2002) provides a detailed analysis of the uses of German es gibt in contrast with those of
the Alemannic existential predicator es hot there is (lit. itexpl has). On the development of this
particular use of a verb give, see Gaeta (2013).
7
itexpl
thereexpl has
Jean
Jean
qui
veut te
b. *Est
wants
govorit s
talk
parler.
talk
toboj.
with you
existential predication, there is important variation in way the figure phrase behaves
with respect to the properties in question and in the possible presence of an
expletive locative or pronoun exhibiting some subject properties. But such a
characterization of existential predication can only be extended to languages with a
particular type of organization of grammatical relations, and trying to generalize it
to all languages could only result in inconsistencies.
For similar reasons, it would not be reasonable to introduce the morphological
nature of existential predicators as a criterion in a general typology of existential
predications. Some of the languages in which the identification of verbs as opposed
to other categories puts into play a rich inflectional system have existential
predicators that can uncontroversially be identified as verbs. But existential verbs
are very often more or less irregular, and there is no universal criterion according to
which predicators that are not uncontroversial verbs could be consistently identified
as irregular/defective verbs, or rather non-verbal predicators, and the question is
particularly difficult in languages with a relatively reduced verb inflection. Here
again, classifying existential predications according to the morphological nature of
the predicator could only make sense for more or less closely related groups of
languages, after taking a decision about the criteria according to which a word is
identified as a verb in the languages in question.
According to the criterion of formal resemblance with predicative constructions
expressing other functional types of predication, the following six types can be
defined:
loc-existentials, existential constructions characterized by the presence of an
element generally used with a meaning such as there or in it, but whose only
function in existential predication is to mark the distinction between plain
locational and existential predication; English there is N (Loc) and Arabic
hunaaka N (Loc) lit. there N (Loc) are typical examples of loc-existentials;
trans.poss-existentials, existential constructions involving a predicator also used
in a transitive possessive construction, i.e. in a possessive predicative
construction in which the possessor and the possessee show coding
characteristics identical to those of the agent and the patient of typical
transitive verbs; in its existential use, this predicator may occur either alone (as
in Brazilian Portuguese tem N (Loc) lit. has N (Loc)) or combined with an
expletive pronoun (as in Alemannic es hot N (Loc) lit. it has N (Loc);
incorp.poss-existentials, existential constructions in which the figure is treated
like the possessee in an incorporating possessive construction; by incorporating
possessive construction, I mean a possessive predicative construction in which
the noun referring to the possessee cannot be analyzed as the head of an NP in
a construction including two slots for NPs (as in the other types or predicative
possession), and must be analyzed as converted into a one-place predicate
meaning be an N-owner by a proprietive operator;
poss/loc-existentials, existential constructions involving a predicator also used
in a transitive possessive construction, plus an additional element generally
used with a meaning such as there, but whose only function in existential
predication is to distinguish existential predication from possessive predication
of a noun phrase representing the figure and a locative expression representing the
ground (Russian Kniga na stole), they may involve a predicator also used in
identificational clauses (French Le livre est sur la table) or a specialized locational
predicator (Mandinka Kitabo be tablo ka), they may also involve postural
verbs in the function of locational predicator (German Das Buch liegt auf dem Tisch),
but dedicated existential predicative constructions can be found in languages with
plain locational clauses belonging to any of these types.
It seems however that the grammaticalization of a dedicated existential
construction is less frequent among the languages that make a systematic use of
postural verbs as locational predicators.8
4.3. Rough equivalents of existential predication in languages devoid of
dedicated existential predicative construction
4.3.1. Introductory remarks
There cannot be exact equivalents of existential predication in the languages that
have not grammaticalized the selection of the ground as the perspectival center in
the encoding of figure-ground relationships. However, rough translational
equivalence between languages with and without dedicated existential predicative
constructions can often be established on the basis that, when the figure is selected
as the perspectival center, it is typically topical and definitive, whereas it is typically
non-topical and indefinite when the term selected as the perspectival center is the
ground. Consequently, in the languages that have not grammaticalized the selection
of the ground as the perspectival center in the encoding of figure-ground
relationships, information structure marking and definiteness marking in locational
clauses may result in distinctions functionally comparable (at least to some extent)
to those expressed in other languages by the choice between plain locational and
existential predication.
This is particularly obvious in languages with relatively flexible constituent order,
in which mere permutations of constituents are available to express modifications of
information structure implying the de-topicalization of the term that has the greatest
propensity to be interpreted as the topic. This situation has already been illustrated
above for Russian. Basque Ex. (13) and Finnish Ex. (14) provide additional
examples.
(13) Basque (pers.doc.)
a. Parke-a
park-SG
ibai-ondo-an
river-side-SG.LOC
dago.
be.PRS.3SG
By systematic use of postural verbs as locational predicators, I mean the routinized use of postural
verbs in contexts in which location constitutes the relevant piece of information, and the particular
posture of the figure is just the default posture for the entity fulfilling the figure role in the spatial
configuration referred to.
8
b. Ibai-ondo-an
river-side-SG.LOC
parke
eder
park
bat
lovely
one
dago.
be.PRS.3SG
on
boy
be-PRS.3SG
piha-lla.
yard-ADESS
on
yard-ADESS be-PRS.3SG
poika.
boy
TOP/SUBJ
table
ue
GEN
GEN
ue
top
ni
at
top
ni
aru.
at
be.PRS
(wa) hon ga
TOP
book
SUBJ
aru.
be.PRS
4.3.2. Locational predications in which movement of the figure phrase to the position
immediately before the locational predicator expresses de-topicalization of the figure
In Basque and Japanese ex. (13) and (15) above, FIG GR Pred, with the ground
phrase immediately before the predicator, is the default constituent order in
locational predication, expressing the unmarked information structure in which the
figure is topical, but the figure phrase can move to the position immediately before
the predicator, being then interpreted as forming part of the comment.
Among the languages of my sample devoid of dedicated existential predicative
construction and for which I have been able to find the relevant data, detopicalization of the figure in locational predication by means of movement of the
figure phrase to the position immediately before the locational predicator is attested
in the following languages:
Avar (Nakh-Daghestanian)
Basque (isolate Spain and France)
Buryat (Mongolic)
Georgian (Kartvelian)
Hayu (Sino-Tibetan, Tibeto-Burman)
Lezgi (Nakh-Daghestanian)
Paez (isolate Colombia),
Palula (Indo-European, Indo-Aryan, Dardic)
Slave (Athabaskan)
Tsamakko (Afro-Asiatic, Cushitic)
Udihe (Tungusic)
Zhaba (Tibeto-Burman, Qiangic)
All of these languages have basic OV type of constituent order, and in all of them,
the position immediately before the locational predicator is the default position of
the ground phrase in locational predication.
This situation is probably much more widespread among OV languages than
suggested by this relative short list of languages. The point is that most of my
sources on OV languages are vague as to the possible function of constituent order
variation in locational clauses, but interestingly, one of them only mentions a rigid
constituent order in locational predication, and in many others, the examples
provided, although not sufficient to conclude, suggest a situation of this type. We
will see in Sections 4.3.3 and 4.4 that the situation is markedly different for VO
languages.
4.3.3. Locational predications in which movement of the figure phrase to the position
immediately after the locational predicator expresses de-topicalization of the figure
In Finnish ex. (14) above, FIG Pred GR, with the ground phrase immediately after
the predicator, is the default constituent order in locational predication, expressing
the unmarked information structure in which the figure is topical, but the figure
phrase can move to the position immediately after the predicator, being then
interpreted as forming part of the comment.
Among the languages of my sample devoid of dedicated existential predicative
construction and for which I have been able to find the relevant data, detopicalization of the figure in locational predication by means of movement of the
figure phrase to the position immediately after the locational predicator is attested
in the following languages:
Czech (Indo-European, Slavic)
Estonian (Uralic, Finnic)
Finnish (Uralic, Finnic)
Kabyle (Afro-Asiatic, Berber)
Latvian (Indo-European, Baltic)
Lithuanian (Indo-European, Baltic)
Romanian (Indo-European, Romance)
Nganasan (Uralic, Samoyedic)
With the only exception of Nganasan, all of these languages are VO languages in
which the position immediately after the locational predicator is the default position
of the ground phrase in locational predication.
It is however important to emphasize that this alternation is far from general
among VO languages that have not grammaticalized a dedicated existential
predicative construction. The data I have collected unquestionably show that,
contrary to a widespread opinion, a minority of them only have locational clauses in
which the flexibility of constituent order provides a way of expressing the detopicalization of the figure.
The uncommon alternation observed in Nganasan by Wagner-Nagy (2009), with
basic FIG GR Pred order in plain locational clauses and the alternative order GR
Pred FIG expressing the de-topicalization of the figure, can be explained by the
influence of a dominant VO language (Russian) on a severely endangered language
that originally was a consistent OV language.
4.3.4. Constituent order variation in locational predications devoid of overt predicator
Among the languages of my sample devoid of dedicated existential predicative
construction and in which locational predication consists simply of the juxtaposition
of the figure phrase and the ground phrase, the order FIG GR in basic locational
clauses with the possibility of expressing de-topicalization of the figure by means of
the alternative order GR FIG is attested by the following language:
Nyangumarta (Pama-Nyungan)
whereas GR FIG in basic locational clauses with the possibility of expressing detopicalization of the figure by means of the alternative order FIG GR is attested by
the following language:
Maori (Austronesian, Oceanic, Polynesian)
This particularity of Maori is consistent with the fact that, more generally, Maori can
be described as a predicate-initial language (Chung and Ladusaw 2001: 31)
4.4. Languages devoid of dedicated existential predicative construction, and
with rigid constituent order in locational clauses
Among the languages of my sample that have not grammaticalized a dedicated
existential predicative construction, a relatively high proportion have locational
clauses with a rigid constituent order that excludes the possibility of de-topicalizing
the figure by moving the figure phrase to the position occupied by the ground
phrase in the default constituent order. All of them have basic VO constituent order,
with the only exception of Retuar, and in all of them without exception, the figure
phrase precedes the ground phrase in locational predication.
This finding contradicts the common opinion that existential constructions in
basic SVO languages (or their equivalents in languages that have not
3-3SG-be inside
a-bogan.
LOC-box
snon
LCOP
tree
y.
ma.
on
LCOP
there
There is wine.
For some languages, for example Tigemaxo (Mande), locational predication with
such a default ground phrase is explicitly mentioned as the usual way to express a
meaning similar to that encoded by means of the verb exist in English ex. (18).
(18) Tigemaxo (Blecke 1996: 205-206)
a.
DEM
ye
ga
PL
COP
Kuntoolo.
Kuntoolo
b. Ala ga
God
COP
g.
there
the
chiave sul
key
is on.the
tavolo.
table
thereexpl-is a
key
on.the
tavolo.
table
Ewe has the particularity that the default ground phrase in locational predication is not a locative
expression, but a third person singular pronoun. This is however consistent with the fact that this
third person singular pronoun can also be used anaphorically in the same position to refer to an
already mentioned location Felix Ameka, pers.com.
In some Italo-Romance varieties, for example, Genovese ex. (20), in addition to the
existential marker, the existential construction includes a third person masculine
expletive subject clitic.
(20) Genovese (Bentley and al. 2013: 16)
Sta
stay.2SG.IMP careful
that
in
this
frta u
fruit 3SG.Mexpl
gh
thereexpl-be
tanti
many
ossi.
seeds
in-DEF-office.GEN
b. Huna ka
thereexpl
rajulu-n
man-INDEF
f-l-maktabi.
in-DEF-office.GEN
fq i-ara.
on
DEF-tree
b. Fmm
thereexpl
bum
owl
fq i-ara.
on
DEF-tree
f-h
uld
fi-d-dr.
child.PL in-DEF-house
In Nahuatl, the prefix on- added to the locational verb cah- in existential
predication is used with other verbs as an andative marker (i.e., as a marker
encoding movement towards a place distinct from the deictic center) ex. (24).
(24) Nahuatl (Launey 1981)
a. N-on-no-tltia.
1SG-AND-REFL-hide
I am going to hide.
b. Nicn on-cah
tl.
here
ANDexpl-be
water
There is water here.
5.2. Loc-existentials in the languages of the world
Among the languages of my sample, existential predication of this type is found in
the following languages:
Arabic, classical and vernacular varieties (Afro-Asiatic, Semitic)
Danish (Indo-European, Germanic)
Dutch (Indo-European, Germanic)
Emrillon (Tupi-Guarani)
Eviya (Niger-Congo, Benue-Congo, Bantu)
English (Indo-European, Germanic)
German (Indo-European, Germanic)
Italian and other Italo-Romance varieties (Indo-European, Romance)
Mwotlap (Autronesian, Oceanic)
Nahuatl (Uto-Aztecan)
Palauan (Austronesian, Malayo-Polynesian)
Samoan (Autronesian, Oceanic, Polynesian)
Yiddish (Indo-European, Germanic)
Zaar (Afro-Asiatic, Chadic)
This type of existential predication is not very frequent at world level, but due to its
presence in some major Germanic and Romance languages, there is a huge literature
devoted to the discussion of the syntactic status of the figure phrase and of the
existential marker in such constructions. Depending on the individual languages, the
existential marker may share some formal properties with canonical subjects (this is
quite clearly the case of there in the English existential construction, but not of
Italian ci). As regards the figure phrase, in the languages in question it
uncontroversially fulfills the subject function in plain locational predication, but its
behavior in existential predication shows variation, in particular with respect to the
control of verb agreement, and consequently such existential constructions figure
prominently in discussions about impersonality. For a recent and well-informed
discussion, cf. Gast and Haas (2011).
languages that do not have subject-verb agreement, and in which the deletion of the
noun phrase preceding the transitive verb of possession is the only thing that
differentiates existential predication from possessive predication.
(25) Greek (pers.doc.)
a. Ta
the
villages
NEG
ner.
have.PRS.3PL
water.ACC
Germanus
sto
b. Den eche
NEG
have.PST.3SG
Germans.ACC
chri.
in.the village
There were no Germans in the village. (also interpretable as He/she did not
have Germans in the village. in an appropriate context)
(26) Vietnamese (pers.doc.)
a. Ti
I
sch.
have book
I have books.
b. C
have one
CLS
fly
bt
in
canh ca ti.
bowl soup of
me
have
aynesa karukri.
some
fish
ahakwa
in
money
un.
water
10
predication involves transitive coding, with the possessor and the possessee encoded
like the agent and the patient of typical transitive verbs, respectively.
In the typology of existential predication I am putting forward, the languages in
which the same predicator is found in (locational/)existential predication and in
possessive predication are classified as having a trans.poss-existential construction if
and only if the possessive use of the predicator in question can be viewed as an
instance of transitive coding, but this is not always easy to evaluate. This question is
made difficult by the fact that transitive verbs of possession are identified as such by
the coding frame through which they express their argument structure, but are
rarely well-behaved transitive verbs in other respects.
Practically, I decided to recognize a trans.poss-existential construction whenever
there is no obvious difference between the coding of the possessor and the possessee
and that of the agent and the patient of a typical transitive verb, but I cannot
exclude that perhaps a better knowledge of some of the language I have classified as
having trans.poss-existentials would lead to reconsidering this decision. This
problem arises mainly for languages of South East Asia (including Formosan and
West Malayo-Polynesian languages), and most of the descriptions I consulted for the
languages of this zone that are more or less problematic in this respect do not take a
clear stand on this issue.
6.3. Trans.poss-existentials in the languages of the world
Contrary to a widespread opinion according to which transitive verbs of possession,
and consequently trans.poss-existentials, are rare (if not totally inexistent) outside
Europe,11 transitive verbs of possession and trans.poss-existentials are not rare in the
languages of the world, and Europe is not the only area in which they are
particularly common. Among the languages of my sample, trans.poss-existentials can
be recognized (subject to the caveat expressed in Section 6.2) in the following
languages:12
Akan (Niger-Congo, Kwa) 13
Albanian (Indo-European)
Alemannic (Indo-European, Germanic)
Bulgarian (Indo-European, Slavic)
Cantonese (Sino-Tibetan, Chinese)
Cape Verdean (Portuguese-based Creole)
Early Italo-Romance varieties14 (Indo-European, Romance)
The origin of this opinion can be traced back to the theory of language change elaborateded in the
first half of the 20th century by Indo-Europeanists such as Meillet cf. in particular Meillet (1924).
12
Edelman (1975) states that an existential use of the Iranian verb of possession *dar- has also
developed in gilaki and other Iranian languages spoken in Iran, without however giving the details
that could allow checking the validity of this statement, and I consequently decided not to include the
languages in question in my sample.
13
The existential construction of Akan GR w FIG can be classfied as trans.poss-existential, since it
uses a predicator w that also occurs in the transitive positive construction POSSor w POSSee, but
Akan is quite atypical in that the same predicator also occurs in the locational construction FIG w
GR Boadi (1971). I came across no other language with a similar situation.
11
parti
in.the
di
parts
of
Grecia ebbe
Greece
have.PST.3SG
un
signore.
sir
PRF.3SG
woto.
car
hot
3SG.N have.PRS.3SG
Rssr
horses
voram
Hus.
in_front_of_the house
A similar situation is found in Pepel (Atlantic), a language with a noun class system
of the Niger-Congo type in which verb forms include a prefix indexing the subject.
With the verb have denoting possession, this prefix indexes the noun class of the
possessor, whereas have as an existential predicator shows an invariable expletive
prefix of a non-human class.
A more radical distinction between the possessive use of a transitive verb of
possession and its use as an existential predicator is found in some West African
languages belonging to the Atlantic family. In these languages, the general rule is
that, either subjects are obligatorily indexed on the verb, or subject indexes attached
to the verb are obligatory if the subject is not expressed as a noun phrase, but
trans.poss-existentials are an exception to this rule, and the absence of the subject
index distinguishes the impersonal use of have in the function of existential
predicator from its possessive use Ex. (31).
(31) Joola Banjal (Bassne and Creissels 2011: 294-295)
a. Na-baj-
CLa-poss-CMP
j-ba.
CLji-knife
poss-CMP
j-ba
CLji-knife
n
in
-pc
CLe-bag
yay.
DEF
have.PRS.3SG
jedna
cura
u sokaku
one.SG.F.NOM
girl.SG.NOM in street.SG.PREP
jedna
krma
b. Ima
have.PRS.3SG
mome.
my.SG.M.LOC
u planini.
one.SG.F.NOM
inn.SG.NOM in mountain.SG.PREP
jednu
krmu
c. Ima
have.PRS.3SG
one.SG.F.ACC
inn.SG.ACC
u planini.
in mountain.SG.PREP
This change, which probably started with the re-analysis of accusative forms
homonymous with the nominative, reinforces the distinction between the existential
construction and the transitive possessive construction from which it developed.
However, in spite of its possibility of nominative marking, the NP representing the
figure in the existential construction with imati illustrated by Ex. (24a) cannot be
analyzed as an inverted subject in a more or less canonical intransitive construction,
since in the plural, the genitive must be used in conditions in which intransitive
subjects (even in postverbal position) are normally in the nominative, and the verb
does not show plural agreement Ex. (33b).
(33) Serbo-Croat (Creissels 2013: 467-468)
a. Ima
have.PRS.3SG
lijepa
pretty.SG.F.NOM
djevojka
u ovoj
girl.SG.NOM in this.SG.F.PREP
have.PRS.3SG
lijepih
pretty.PL.GEN
djevojaka u ovom
girl.PL.GEN
in this.SG.N.PREP
kui.
house.SG.PREP
selu.
village.SG.PREP
that can also be viewed as involving a ground-figure relationship as in This well has
water ~ There is water in this well.
6.6.2. Other possible scenarios
The impersonalization of a transitive verb of possession is however not the only
scenario that may yield existential constructions synchronically analyzable as
trans.poss-existentials. One may also imagine that the coding characteristics of a
possessive construction that originally did not involve a transitive verb of possession
(in particular, a possessive construction with locative marking of the possessor (at
Possessor is Possessee) have changed in such a way that the possessor and the
possessee have acquired coding characteristics identical to those of the agent and the
patient of typical transitive verbs, which automatically triggers the reanalysis of the
existential construction as a trans.poss-existential conclusion.
I am aware of no language for which this scenario would be historically
documented, but several languages show that possessive constructions that
originally derived from locational predication may acquire coding characteristics
typical of the transitive construction, and that consequently, this is at least a possible
source of transitive positive constructions. For example, the predicative possessive
construction of Finnish clearly derives from a locative inversion construction
(something like at Possessor is Possessee), but in the possessive construction of
present-day Finnish, pronouns in possessee role are usually in the accusative case,
which excludes analyzing them as inverted subjects in the locational construction
Creissels (2013: 469). This phenomenon, sometimes designated as have-drift, is
analyzed by Ziv (1982) and Zuckermann (2009) for Israeli. Stassen (2009) argues
that such an evolution occurred in Cornish, Ahland (2009) analyzes similar changes
that have affected the Amharic possessive construction, and Bar-Asher (2011)
discusses evidence that the transitive possession verb of Akkadian im have
derived from an existential predicate *i.
To summarize, classifying an existential construction as trans.poss-existential in a
synchronic typology of existential constructions does not necessarily implies
historical derivation from a construction involving a transitive verb of possession.
7. Incorp.poss-existentials
7.1. Definition and illustration
Incorp.poss-existentials are existential constructions in which the figure is treated
like the possessee in an incorporating possessive construction, i.e. a possessive
predicative construction in which the noun referring to the possessee cannot be
analyzed as the head of an NP in a construction including two slots for NPs (as in
the other types or predicative possession), and must be analyzed as converted into a
one-place predicate meaning be an N-owner by a proprietive operator.
For example, Kalaallisut (aka West Greenlandic) has a suffix -qar converting
nouns into intransitive verbs be an N-owner (proprietive verbs) that assign the role
of possessor to their argument, encoded as a noun phrase in the zero case (alias
taana
that
illu-qar-puq.
house-PROPR-IND.3SG
b. Nillataartarfim-mi tallima-nik
fridge-LOC
five-INSTR.PL
manne-qar-puq.
egg-PROPR-IND.3SG
pera
ang bata.
money
NOM
child
tao
sa
bahay.
person
LOC
house
What rules out the analysis of (35a) as a transitive construction is that, in Tagalog,
as illustrated by Ex. (36), both arguments of a transitive verb must be introduced by
a proclitic case marker.
A priori, a possible analysis is that -qar is in fact a bivalent verb have that obligatorily
incorporates its second argument. However, due to my lack of familiarity with Eskimo languages, I
am not in a position to discuss the validity of such an analysis.
16
GEN
child
NOM
branch
In the transitive construction, one of the two core arguments must be marked by ang
(nominative) and the other one by n(an)g (genitive), and the choice of the
nominative-marked argument is correlated to the choice of a voice marker which
constitutes an obligatory element of the verb form. By contrast, in the possessive
construction, may includes no voice marker, and the noun representing the possessee
is invariably devoid of any case marker. Consequently, the noun referring to the
possessee cannot be analyzed as the head of a noun phrase in a predicative
construction with two syntactic slots for noun phrases representing the arguments of
a two-place predicate. The only possible analysis is considering may as a proprietive
operator that converts the noun it precedes into a monovalent predicate whose
meaning can be glossed as be an N-owner, which implies analyzing the existential
construction in (34b) as an incorp.poss-existential construction.
7.2. Incorp.poss-existentials in the languages of the world
Incorporating possessive constructions are not rare in the languages of the word, in
particular among Amerindian languages, but within the limits of the data I have
been able to collect, existential constructions with the figure treated like the
possessee in an incorporating possessive construction are attested in three languages
only:
Kalaallisut, aka West Greenlandic (Eskimo-Aleut, Eskimo)
Tagalog (Austronesian, West Malayo-Polynesian)
Yupik (Eskimo-Aleut, Eskimo)
8. Poss/loc-existentials
8.1. Definition and illustration
Poss/loc-existentials share with trans.poss-existential the use of a predicator also
used in a transitive possessive construction, and with loc-existentials the presence of
an expletive locative.17 This type of possessive predication has already been
illustrated in Catalan Ex. (1). Occitan Ex. (37) provides an additional illustration.
thereexp
has
un
a
dog
in
the-garden
The existential predicator of French il y a belongs to this type, but in addition to the
expletive locative (y) and a third person singular form of avoir have, it includes an
expletive subject clitic of third person masculine (il).
8.2. Poss/loc-existentials in the languages of the world
All the languages of my sample in which a predicative construction of the poss/locexistential type can be identified belong to the Romance family:
Catalan (Indo-European, Romance)
French (Indo-European, Romance)
Occitan (Indo-European, Romance)
Sardinian (Indo-European, Romance)
Historically, the present form of the Spanish existential predicator hay originates
from such a poss/loc-existential construction, since it can be decomposed as ha third
person singular of haber < Latin habere have plus -y reflex of an expletive locative,
but synchronically, hay can only be analyzed as a dedicated existential predicator,
since in Spanish, haber has been completely replaced by tener (< tenere hold) in the
expression of possession.
9. Com-existentials
9.1. Definition and illustration
Com-existentials are defined as existential predications in which the figure is
encoded like the phrase representing the companion in comitative predication. Comexistentials are common among Bantu languages, and there is a clear relationship
with the possessive use of comitative predication in Bantu languages.
In languages in which locative phrases have access to the syntactic function of
subject, like Swahili, com-existentials can straightforwardly be described as
comitative predications with the ground in the syntactic role of subject: GR is with
FIG. Example (38a-b) illustrates the possessive use of the Swahili comitative
predication with plain NPs in subject function, and (38c-d) illustrate the same
construction with locative phrases in subject function.18
18
For a detailed analysis of locational and existential predication in Swahili, cf. Marten (2013).
(CL1)Hamisi
CL1
na
with
kitabu.
CL7.book
b. Kisima ki
CL7.well CL7
na
with
maji.
CL6.water
lit. The well it (is) with water. The well has water.
c. Kisima-ni
CL7.well-LOC
CL18
na
maji.
with
CL6.water
na
kitabu.
lit. at-the-well there (is) with water. There is water in the well.
d. Meza-ni
(CL9)table-LOC
pa
CL16
with
CL7.book
lit. at-the-table there (is) with book. There is a book on the table.
In Southern Bantu languages, where locative phrases have lost the ability to be used
in the syntactic function of subject, the com-existential construction involves an
existential predicator consisting of a an expletive locative index and the comitative
predicator, as in Tswana go na (le) there is, lit. thereexpl is.with Ex. (39).19
(39) Tswana (pers.doc.)
a. Ga
NEG
ke
1SG
na
mathata.
be.with problems
go
na
mathata.
For a detailed analysis of comitative and possessive predication in Tswana, cf. Creissels (2013).
Hausa also has a dedicated existential predicator kwai, already illustrated by Ex. (3), Section 2.1.
3SG.M.CMP
yanka d
wu.
fensr.
c. D
dli li.
with reason
There is a reason.
d. D
sasshen
with enough
shy?
tea
10. Id-existentials
10.1. Definition and illustration
Id-existentials are defined as existential constructions characterized by the presence
of either a dedicated identificational predicator, or an identificational/locational
predicator accompanied by a non-locative expletive element also used in
identificational clauses equivalent to English This/that is an N. Icelandic illustrates
this type of existential predication Ex. (41).
(41) Icelandic (Neijmann 2001: 22, Freeze 2001: 949)
a. a er
that
is
kirkja.
church
That is a church.
b. a eru
that
are
mys bakerinu.
mice in bathtub
There are mice in the bathtub. litt. That are mice in the bathtub.
This type of existential construction is interesting in a theoretical perspective, since
it emphasizes the semantic relationship between existential predication in the sense
of inverse locational predication and identificational predication: in some sense, the
existential perspective on figure-ground relationships is tantamount to identifying an
entity present at a given place. This connection is even more obvious in the variant
of the id-existential type found in Tahitian (Polynesian). In most Polynesian
languages, the figure phrase in existential predication is introduced by an expletive
DEM
pape
te-i
man
b. E
ICOP
water
ART-at
pape
te-o
ter vhi.
DEM
place
b. E
ICOP
water
ART-of
ter vhi.
DEM
place
grammaticalizations from the same source, such as for example that of auxiliary
verb.
Ex. (43) illustrates the use of the distinction between locational predication and
existential predication involving a dedicated existential predication in Turkish.
(43) Turkish (pers.doc.)
a. Otel ehir-de(-dir)
b. Otel ehir-de
hotel town-LOC(-be)
hotel town-LOC
c. Bu
DEM
ehir-de
town-LOC
bir
one
NEG(-be)
otel var.
hotel
deil(-dir)
EXIST
d. Bu
DEM
ehir-de
town-LOC
otel yok.
hotel
EXIST.NEG
Malayalam (Dravidian)
Mari (Uralic, Permic)
Madi (Central Sudanic)
Mixtec (Oto-Manguean
Mori Bawah (Austronesian, West Malayo-Polynesian)
Mosetn (Isolate Bolivia)
Nheengat (Tupi-Guarani)
Nias (Austronesian, West Malayo-Polynesian)
Portuguese (Indo-European, Romance)
Quechua, Ancash variety (Quechuan)
Quechua, Imbabura variety (Quechuan)
Rapanui (Austronesian, Oceanic, Polynesian)
Russian (Indo-European, Slavic)
Sakha aka Yakut (Turkic)
Soureth (Afro-Asiatic, Semitic, Aramaic)
Spanish (Indo-European, Romance)
Tobelo (West Papuan)
Tsou (Austronesian)
Tukang Besi (Austronesian)
Tupuri (Niger-Congo, Adamawa)
Turkish (Turkic)
Tuvaluan (Austronesian, Oceanic, Polynesian)
Udmurt (Uralic, Permic)
Vaekau-Taumako aka Pileni (Austronesian, Oceanic)
Xamtanga (Afro-Asiatic, Cushitic)
Dedicated existentials predicators are unquestionably widespread in the worlds
languages. A caveat is however in order. In some languages for which the source I
have used describes an existential predicator without mentioning a possible
etymology, it may well be that a better knowledge of the languages in question
would have allowed me to identify another type of existential construction. In
particular, several Oceanic languages in my sample are described as having an
existential predicator that seems to be cognate with the locative expletives found in
the existential construction of Samoan (i ai) or Mwotlap (a), but the sources I have
used do not discuss the possibility of such an etymology, and I am not in a position
to decide whether the languages in question should be re-classified as having locexistential constructions, or their classification as having dedicated existential
predicators is correct.
11.3. The origin of dedicated existential predicators
A priori, existential predications belonging to any of the types presented in the
previous sections may undergo evolutions that moves them apart from their source
construction, resulting in the conversion of existential constructions initially
analyzable as belonging to one of the types analyzed in the previous sections into
existential constructions involving dedicated existential predicators.
EXIST.IMPF.3SG
un
problem
very serious
un
have.IMPF.3SG
problem
very serious
predicator in the present. As a result of this evolution, the use of jest has become
very marginal in locational and identificational predication, whereas jest has been
maintained as an existential predicator. By contrast, in the past, be has been
maintained in locational/identificational predication, and consequently no dedicated
existential predicator has emerged Ex. (45).
(45) Russian (pers.doc.)
a. Derevnja
village
za
behind
goroj.
b. Za
hill.INSTR
byla
be.PST.SG.F
za
behind
goroj.
hill.INSTR
behind
goroj
est
derevnja.
behind
goroj
hill.INSTR
byla
be.PST.SG.F
derevnja.
village
still
ima
dobrih
ljudi.
According to Geoffrey Heath (pers.com.), there is no consensus about the origin of the element hethat distinguishes the existential predicator hebn from the copula bn in Kurdish, but one
reasonably plausible suggestion is to connect it to the postposed demonstrative element ha(n), which
is quite similar to German postposed particle da in things like das Buch da that book (there).
21
b. Bilo
be.PST.SG.N
je
AUX.3SG
dobrih
good.PL.GEN and
loih
bad.PL.GEN
dana.
person.PL.GEN
be.PRS.3PL
jeszcze wolne
still
free.PL
miejsca.
place.PL
have.PRS.3SG
ju
wolnych
miejsc.
ju
nikogo.
13. Is the mere juxtaposition of the ground phrase and the figure phrase a
possible type of dedicated existential construction?
It follows from the data presented in the previous sections that there are languages
with a locational predicative construction consisting of the mere juxtaposition of the
figure phrase and the ground phrase and no possible contrast between this
construction and a dedicated existential predicative construction. In such languages,
clauses including no overt predicator may constitute the translational equivalent of
the existential clauses of languages that have a dedicated existential predicative
construction.
There are also languages in which a locational predicative construction consisting
of the mere juxtaposition of the figure phrase and the ground phrase contrasts with
a dedicated existential construction involving an overt predicator.
The question examined in this section is whether existential predicative
constructions including no overt predicator can contrast with locational
constructions involving an overt predicator.
Whatever the construction from which an existential construction derives
historically, changes in the source construction may result in the emergence of
dedicated existential constructions whose structure maintains the original shape of
the source construction. According to this principle, it should be possible to find
existential predications involving no overt predicator in languages in which a
locational predication that did not necessarily involve an overt predicator was
replaced by a locational construction with an obligatory overt predicator.
References
Abraham, Roy Clive. 1940. The principles of Tiv. Gregg International.
Ahland, Michael. 2009. From topic to subject. Grammatical change in the Amharic possessive construction. Studies in
Language 33(3). 685-717.
Amha, Azeb. 2001. The Maale Language. Leiden: CNWS.
Anderson, Neil. 1989. Folopa existential construction verbs. In Karl Franklin (ed.), Studies in
componential analysis. Data Papers on Papua New Guinea Languages 36. Ukarumpa: Summer
Institute of Linguistics. 83-105.Bally, Charles. 1932. Linguistique gnrale et linguistique franaise. Paris: Leroux.
Asher, R. E. and T. C. Kumari. 1997. Malayalam. London / New York: Routledge.
Atoyebi, Joseph Dele. 2010. A reference grammar of Oko. Kln: Rdiger Kppe.
Aziz, Yowell Y. 1995. Existential sentences in Arabic-English translation. Meta : journal des
traducteurs / Meta: Translators' Journal 40(1). 47-53.
Baker, Philip. 1972. Kreol. A Description of a Mauritian Creole. London: C. Hurst and Company.
Bar-Asher, Elitzur. 2011. From typology to diachrony: synchronic and diachronic aspects of
predicative possessive constructions in Akkadian. Folia Linguistica Historica 32. 4388.
Bassne, Alain-Christian and Denis Creissels. 2011. Impersonal constructions in Jola-Banjal. In Andrej
Malchukov and Anna Siewierska (eds.) Impersonal constructions: a cross-linguistic perspective. John
Benjamins. 285-306.
Becuwe, Jacques. 1982. Elments de phonologie et de grammaire du lobiri (parler de Bouna, Cote
d'Ivoire. PhD thesis, Universit de la Sorbonne Nouvelle, Paris III.
Beffa, Marie-Lise and Roberte Hamayon. 1975. Les figures du corps. Parie: Socit dethnologie.
Bentley, Delia, Francesco Maria Ciconte, and Silvio Cruschina. 2013. Micro-variation in subject
agreement: The case of existential pivots with split focus in Romance. Rivista di Linguistica 25.1.
15-43.
Berghll, Liisa. 2010. Mauwake reference grammar. PhD thesis. University of Helsinki.
Besnier, Niko. 2000. Tuvaluan, a Polynesian Language of the Central Pacific. London / New York:
Routledge.
Beyer, Klaus. 2006. La langue pana (Burkina Faso et Mali): description linguistique, lexique, textes. Kln:
Rdiger Kppe.
Blackings, Mairi and Fabb, Nigel. 2003. A Grammar of Madi. Berlin: Mouton de Gruyter.
Blanc, Jean-Franois. 1985. Le verbe urhobo. PhD thesis. University of Grenoble.
Blau, Joyce and Veysi Barak. 1999. Manuel de kurde: kurmanji. Paris: LHarmattan.
Blecke, Thomas. 1996. Lexikalische Kategorien und grammatische Strukturen im Tigemaxo (Bozo, Mande).
Kln: Rdiger Kppe.
Boadi, Lawrence A. 1971. Existential Sentences in Akan. Foundations of Language 7. 19-29.
Borschev, Vladimir and Barbara Partee. 2002. The Russian genitive of negation: theme-rheme
structure of perspective structure?. Journal of Slavic Linguistics 10. 105-144.
Boyeldieu, Pascal. 2000. La langue bagiro (Rpublique centrafricaine), Systmatique, textes et lexique.
Frankfurt am Main: Peter Lang.
Broadwell, George Aaron. 2006. A Choctaw reference grammar. University of Nebraska Press.
Brown, Lea. 2005. Nias. In K. Alexander Adelaar and Nikolaus Himmelmann (eds), The Austronesian
languages of Asia and Madagascar. Routledge. 562-589.
Brown, Penelope. 2006). A sketch of the grammar of space in Tzeltal. In S. C. Levinson, and D. P.
Wilkins (Eds.), Grammars of space: Explorations in cognitive diversity. Cambridge: Cambridge
University Press. 230-272.
Bugaeva, Anna. 2011. A diachronic study of the impersonal passive in Ainu. In Andrej Malchukov
and Anna Siewierska (eds.) Impersonal constructions: a cross-linguistic perspective. John Benjamins.
517-546.
Campbell, Lyle. 1985. The Pipil Language of El Salvador. Berlin: Mouton de Gruyter.
Carlin, Eithne. 2004. A Grammar of Trio. Duisburger Arbeiten Zur Sprach- Und Kulturwissenschaft.
Christiansen-Bolli, Regula. A grammar of Tadaksahak. Kln: Rdiger Kppe.
Christie, J. J. 1970. Locative, possessive and existential in Swahili. Foundations of Language Vol. 6,
No. 2. 166-177.
Chung, Sandra. 1987. The syntax of Chamorro existential sentences. In Eric Reulang and Alice ter
Meulen (eds.), The representation of (In)definiteness. MIT Press.