1It is cross-linguistically common that the two core nominal terms of nominal predication constructions expressing referential identity between two nominal expressions can exchange their roles as argument and predicate, as in (1).
|
(1)
|
English
|
|
a.
|
JohnARG is my fatherPRED.
|
|
b.
|
My fatherARG is JohnPRED.
|
2An interesting particularity of some Mande languages is that a superficially similar phenomenon also occurs in inclusion statements (also known as class-membership statements), i.e., in clauses in which the difference in the referential properties of the two nominal terms excludes the possibility of an exchange of roles at the semantic level. For example, in the pair of Mandinka clauses quoted in (2), clause (b) does NOT mean ‘The one who is Mandinka is Laamin’, and can only be interpreted as expressing the same predicate-argument articulation as (a), from which it differs only in its discursive implications.
|
(2)
|
Mandinka
|
|
a.
|
Làamínì
|
mú
|
màndìŋkôo
|
lè
|
tí.
|
|
|
PRN
|
NomCOP
|
Mandinka.D
|
FOC
|
as
|
|
|
‘LaaminARG is MandinkaPRED.’
|
|
b.
|
Màndìŋkôo
|
lè
|
mú
|
Làamínì
|
tí.
|
|
|
Mandinka.D
|
FOC
|
NomCOP
|
PRN
|
as
|
|
|
‘HeARG is MandinkaPRED, Laamin that is.’
|
3The analysis of this phenomenon constitutes the main topic of the present article. §2 sets out the theoretical framework. §3 consists in a typological survey of non-verbal predication in Mande, with a special focus on nominal predication. §4 describes nominal predication in Mandinka and analyzes the phenomenon of argument-predicate reversal illustrated by example (2) above. §5 shows that the same phenomenon is found in several other Mande languages. §6 discusses the status of the focus marker in the nominal predication constructions in which the predicate phrase obligatorily combines with the focus marker. §7 puts forward some concluding remarks.
4The sample of Mande languages considered for this investigation and the sources that have been used are as follows:
5Bambara: (Dumestre 2003; Vydrin 2020) and personal documentation;
6Beng: (Paperno 2014);
7Bisa: (Naden 1982; Vanhoudt 1992);
8Boko: (Prost 1976; Perekhvalskaya 2017a);
9Bolon: (Zoungarana 1987);
10Busa: (Wedekind 1973);
11Dan: (Vydrin 2020) for Eastern Dan, (Makeeva 2017) for Kla-Dan;
12Dzuun: (Solomiac 2014);
13Gban: (Fedotov 2017);
14Guro: (Kuznetsova & Kuznetsova 2017);
15Jalkunan: (Heath 2017);
16Jalonke: (Lüpke 2005);
17Jeli: (Tröbs 1998);
18Jenaama: (Heath 2022a) (Cliffs Jenaama), (Heath 2022b) (Jenaama-Sorogaama of Djenné);
19Jula (Kong variety): (Sangaré 1984);
20Kagoro: (Creissels 1984);
21Kakabe: (Vydrina 2017);
22Koranko: (Kastenholz 1987);
23Koro: (Creissels 1986);
24Kpelle: (Konoshenko 2017);
25Loko: (Vydrin & Morozova 2017);
26Looma: (Mishchenko 2017);
27Mandinka: (Creissels & Sambou 2013) and personal documentation;
28Maninka: (Creissels 2009) (Kita Maninka), (Creissels 2013) (Niokolo Maninka), (Vydrin 2020) (Guinean Maninka);
29Mano: (Khachaturyan 2014);
30Manya: (Heydorn 1943; Heydorn 1949);
31Mauka: (Creissels 1982);
32Mende: (Innes 1971);
33Mwan: (Perekhvalskaya 2017b);
34San: (Perekhvalskaya 2017c);
35Seenku (aka Sembla): (Prost 1971; McPherson 2020);
36Soninke: personal documentation;
37Soso: (Touré 1994) and personal documentation;
38Tigemaxo: (Blecke 1996);
39Tura: (Idiatov & Aplonova 2017), Dmitry Idiatov (pers.com.);
40Vai: (Welmers 1976);
41Xasonka: (Koité-Herschel 1981);
42Yaure: (Kushnir 2017).
43This article is based on the delimitation between verbal and non-verbal predication put forward by Hengeveld (1992). In this approach to non-verbal predication, non-verbal predicative constructions can be defined as constructions giving rise to non-elliptical clauses analyzable as consisting of an argument phrase and a predicate phrase in which the property‑ or relation-denoting element that acts as the semantic nucleus of the predicate phrase is not a verb.
44For example, in (3a), kàràndìrìlâa ‘the teacher’ acts as the argument of the predicate phrase dǔntà búŋò kónò ‘entered the room’, whose nucleus (in the sense of role-assigning element) is the verb dǔŋ ‘enter’ (semantically, a two-place predicate). By contrast, in (3b), kàràndìrìláa ‘teacher’ acts as the nucleus of the predicate phrase té kàràndìrìláa tì ‘is not a teacher’, in the sense that it denotes a property (‘meeting the conditions to be categorized as a teacher’) predicated negatively on the referent of Làamínì.
|
(3)
|
Mandinka
|
|
a.
|
Kàràndìrìláa
|
dǔn-tà
|
búŋò
|
kónò.
|
|
|
|
teacher.D
|
enter.CPL
|
room.D
|
in
|
|
|
|
‘The teacher entered the room.’
|
|
b.
|
Làamínì
|
té
|
kàràndìrìláa
|
tì.
|
|
|
|
PRN
|
COP.NEG
|
teacher
|
as
|
|
|
|
‘Laamin is not a teacher.’
|
45In (4a), báakáŋkúŋò tó ‘at the riverside’ acts as a place adjunct in a clause whose predicative nucleus (the verb túlûŋ ‘play’) assigns the semantic role of player to its subject (díndíŋòlú ‘the children’), whereas in (4b), the role-assigning element in the construction of the predicate phrase is the locative postposition tó, semantically a two-place predicate. The postposition tó assigns the role of ground to its complement, thus creating the one-place predicate báakáŋkúŋò tó, which denotes the property of ‘being located at the riverside’ predicated on the referent of díndíŋòlú.
|
(4)
|
Mandinka
|
|
a.
|
Díndíŋò-lú
|
kà
|
túlûŋ
|
báakáŋkúŋò
|
tó.
|
|
|
|
|
child.D-PL
|
ICPL
|
play
|
riverside.D
|
LOC
|
|
|
|
|
‘The children generally play at the riverside.’
|
|
b.
|
Díndíŋò-lú
|
bè
|
báakáŋkúŋò
|
tó.
|
|
|
|
|
|
child.D-PL
|
AdvCOP
|
riverside.D
|
LOC
|
|
|
|
|
|
‘The children are at the riverside.’
|
- 1 As discussed in (Creissels 2016), in Mandinka (and other Manding languages), the comitative preposi (...)
46In (5a), níŋ dèenàanôo ‘with the baby’ is semantically a comitative adjunct in a clause whose predicative nucleus (the verb nǎa ‘come’) assigns the semantic role of comer to its subject (Fàatú),1 whereas in (5b), the role-assigning element in the construction of the predicate phrase níŋ kòlèyâa is the preposition nîŋ ‘with’, semantically a two-place predicate. nîŋ assigns the role of companion to its complement kòlèyâa, thus creating the one-place predicate níŋ kòlèyâa, which denotes the property of ‘having problems’ predicated on the referent of the first person pronoun.
|
(5)
|
Mandinka
|
|
a.
|
Fàatú
|
níŋ
|
dèenàanóo
|
nǎa-tà.
|
|
|
|
|
Fatou
|
with
|
baby.D
|
come-CPL.ITR
|
|
|
|
|
‘Fatou came with the baby.’
|
|
b.
|
Ŋ́
|
níŋ
|
kòlèyâa
|
lè
|
mú
|
|
|
|
|
1SG
|
with
|
problem.D
|
FOC
|
NomCOP
|
|
|
|
|
‘I have problems.’ lit. ‘I am with problems.’
|
47As discussed in detail by Hengeveld (1992: 27–30), non-verbal predicative constructions may involve verbs (copular verbs), but the verbs involved in non-verbal predication have no impact on the argument structure of the construction. For example, in the English equivalents of (3b) and (4b) (Laamin is not a teacher, The children are at the riverside) the verb be is syntactically an obligatory element of the predicate phrase, and can be analyzed as fulfilling the function of support for the expression of grammaticalized TAM values, but the argument structure of such clauses is entirely determined by the lexical meaning of a non-verbal word: the noun teacher in Laamin is not a teacher, the preposition at in The children are at the riverside.
48Depending on language-specific rules, clauses analyzable in terms of non-verbal predication may variously involve obligatory elements in addition to the non-verbal predicate and its argument. However, at least in some languages, words or phrases that can fulfill non-predicative roles in clauses whose predicative nucleus is a verb (nouns, adjectives, locative expressions) can also be found in predicate function, without necessitating any additional element, in clauses in which the non-verbal predicate is simply juxtaposed to the phrase representing its argument, as in examples (6) to (8).
|
(6)
|
Hungarian
|
|
János
|
orvos.
|
|
|
|
|
|
PRN
|
doctor
|
|
|
|
|
|
‘János is a doctor.’
|
|
(7)
|
Modern Standard Arabic
|
|
Zaydun
|
marīḍu-n.
|
|
|
|
|
|
PRN(M)
|
ill.SG.M-IDF
|
|
|
|
|
|
‘Zayd is ill.’
|
|
(8)
|
Russian
|
|
Ivan
|
teper’
|
v
|
Moskv-e.
|
|
|
|
PRN
|
now
|
in
|
Moscow-PRC
|
|
|
|
‘Ivan is presently in Moscow.’
|
49To summarize, in the approach to non-verbal predication adopted here, the notions of verbal vs. non-verbal predication and verbful vs. verbless clause should not be confounded:
50– a verbful clause is not necessarily analyzable as instantiating verbal predication, since words that are morphosyntactically verbs may act as copulæ playing no role in the assignment of semantic roles to the nominal terms of the clause;
51– conversely, a verbless clause is not necessarily analyzable as instantiating non-verbal predication, if it can only be interpreted in a context making it possible to restore an elided verb (for example, in the second utterance of a question-answer sequence such as –Who did you meet there? –My friend John).
52A crucial question for a general theory of non-verbal predication is the articulation between formal (morphosyntactic) and semantic types of non-verbal predication.
53Dryer (2007) puts forward the following three morphosyntactic types of non-verbal predication according to the morphosyntactic nature of the non-verbal predicate: nominal, adjectival, and locative. Such a classification can be taken as a basis for discussions of non-verbal predication, with, however, a revision concerning the definition of the third type. The point is that its characterization as ‘locative’ is too narrow, and locative predication is best viewed as a particular case of a broader notion of adverbial predication defined as follows: in adverbial predication, the role of predicate is fulfilled by a word or phrase that can also be found in clauses whose predicative nucleus is a verb with the role of oblique specifying the circumstances of the event denoted by the verb, as in English John is here / in the garden / with us / against this idea.
54As regards the marking of the argument-predicate relationship in non-verbal predication, four strategies can be distinguished: the copula strategy, the juxtaposition strategy, the predicative inflection strategy (i.e. the marking of non-verbal words in predicate function by means of a special inflection), and the combination of the copula strategy and the predicative inflection strategy. In the languages of the world, the copula strategy is particularly common. The juxtaposition strategy is also relatively common. The predicative inflection strategy is comparatively rare, and the combination of the copula strategy and the predicative inflection strategy even rarer.
55Morphologically, copulæ may be verbal or non-verbal, but the distinction is not always obvious, especially in languages with little inflection, since copular verbs are often defective and/or irregular. The distinction between verbal and non-verbal copulæ has to do with the historical origin of copulæ, since copulæ may grammaticalize from verbs (in particular, copular verbs resulting from the evolution of posture verbs are very common cross-linguistically), but also from non-verbal words, such as demonstratives or personal pronouns. Note, however, that originally verbal copulæ may loose their verbal characteristics, and originally non-verbal copulæ may acquire verbal characteristics.
56Languages with a single copula available for all types of non-verbal predication are very common. Another very common situation is the distinction between two copulæ, one for nominal (and possibly adjectival) predication, the other for locational predication (and possibly other subtypes of adverbial predication). Spanish illustrates the situation in which nominal predication and locative predication use two distinct copulæ that are in competition for adjectival predication.
57However, other patterns (including more complex ones) are not uncommon cross-linguistically. For example, Avar (Nakh-Daghestanian) has two formally unrelated copulæ with the following distribution: a non-verbal copula in present tense, and a verbal copula showing regular verbal inflection for all other TAM values.
58In addition to the nature of the word or phrase in predicate function and the expression of TAM distinctions, the factors that often condition the distribution of copulæ are negation, information structure, or the status of the clause as independent or subordinate.
59Negative copulæ formally unrelated to their counterpart in positive clauses are cross-linguistically common.
- 2 For a detailed discussion of this point, readers are referred to Hengeveld (1992: 30–46).
60Copulæ must not be viewed as necessarily ‘semantically void’. What really characterizes them (and distinguishes them from non-copular verbs) is that they do not contribute to the argument structure of the clause.2 For example, in Spanish, ser and estar in adjectival predication unquestionably meet the definition of copulæ, but do not just act as mere supports for the expression of TAM categories, since they also provide information about the permanent vs. contingent nature of the relationship between the adjectival predicate and its argument.
61Across the languages that make more or less productive use of the juxtaposition strategy, it is more frequent:
62in nominal predication than in adverbial predication,
63with third person arguments than with arguments representing speech act participants,
64in clauses expressing a TAM value of the type commonly labeled ‘indicative present’ than in clauses expressing other TAM values,
65in independent clauses than in subordinate clauses,
66in clauses that do not involve overt marking of information structure than in clauses in which a term is overtly focalized.
67Two variants of the predicative inflection construction can be distinguished. In the first variant, widely described in the literature, the predicative use of a non-verbal word or phrase is marked by the attachment of a marker not found in the non-predicative uses of the same word or phrase. The second variant also involves a dedicated morphological form of the non-verbal predicate (limited, however, to nouns and adjectives), but this dedicated predicative form stands out as morphologically lighter than any other form to be found in nouns or adjectives in non-predicative function, and consequently cannot be described as resulting from the attachment of a predicative marker. Cross-linguistically, this second variant of the predicative inflection strategy is much rarer than the first one.
68Given the topic of this article, the discussion of semantic types of non-verbal predication is limited to nominal predication. Two semantic subtypes of nominal predication are usually distinguished: inclusion (or class-membership) predication and identity predication.
69Inclusion statements express the relationship between a set and a proper subset thereof, as in Bats are not birds, or between a set and an element thereof, as in John is a doctor.
70Identity statements can be defined as statements in which the predicate phrase and the argument phrase are equireferential. They establish a correspondence between two descriptions of the same referent: John is Mary’s father, John is that one, Those men are the people I mentioned.
71Most languages do not use distinct constructions for inclusion and identity, but the relevance of the distinction is nevertheless confirmed by the fact that some languages use different copulæ for the two types of sentences, for example Thai (Stassen 1997: 104–105).
72Cross-linguistically, an important property of identity statements distinguishing them from inclusion statements is that, in identity statements (but not in inclusion statements), the nominal predicate and its argument can exchange roles, as for example John is our teacher / Our teacher is John. The possibility that the nominal predicate and its argument can readily exchange their roles in identity statements (but not in inclusion statements) is a natural consequence of their equireferentiality.
73Some authors (Stassen 1997; Dryer 2007; Roy 2013) argue that identity clauses should not be considered as properly predicative. However, this view is hardly reconcilable with the fact that an overwhelming majority of the world’s languages use the same constructions for inclusion and identity. Moreover, not all languages have a system of articles marking the distinction between, for example, That man is a teacher (inclusion) and That man is the teacher (identity). In fact, there is no reason to adopt a restrictive conception of predication excluding identity clauses from the notion of predication, quite on the contrary. The elaboration of the technical details can be left to formal semanticists, but from the point of view of a typologist, the mere fact that most of the world’s languages use the same constructions for inclusion and identity clearly supports a conception of predication making it possible to analyze an identity clause such as John is our teacher as predicating the property ‘being our teacher’ on the referent of John, and Our teacher is John as predicating the property ‘being John’ on the referent of our teacher.
74As already mentioned in the introduction, what makes Mande languages particularly interesting for a general typology of nominal predication is that, in some Mande languages, a phenomenon superficially similar to the exchange of roles that characterizes pairs of identity statements such as John is our teacher / Our teacher is John can be observed in inclusion statements too, in spite of the fact that, in inclusion statements, the nominal predicate and its argument cannot exchange their roles in the predicate-argument structure.
75Vydrin (2020: 77–78) provides a characterization of “the most remarkable peculiarities of non-verbal predicative constructions in the prototypical Mande languages” that can be taken as a starting point for the discussion in this section:
76“• connection between arguments and predicates is expressed, as a rule, by specialized elements, i.e. copulas. In other words, zero-connection is untypical;
77• most often, copulas in Mande languages have no inflection, and they usually cannot be classified as verbs;
78• it is typical of Mande languages to have more than one copula for different constructions (with different semantic types of non-verbal predicates);
79• copulas used in locational / oblique case constructions easily evolve into auxiliaries of imperfective verbal constructions;
80• Mande languages usually have negative copulas in addition to affirmative ones. Most often, there is only one negative copula in a language. In other words, in negative non-verbal clauses, semantic differences between types of predicates are not signaled by the use of distinct copulas.”
81In most Mande languages, as illustrated by Mandinka (see §4 below), the only fully productive strategy in non-verbal predication is the copular one. The juxtaposition strategy is productive in some Mande languages only.
82The juxtaposition strategy is productive for nominal predication in Dzuun, Bobo, Soso and Jalonke (examples are provided in §3.6 below).
83In Soso and Jalonke, the juxtaposition strategy is also productive in adverbial predication, in competition, however, with the use of an optional copula resulting from the grammaticalization of the adverb nàa ‘there’.
84In Beng, the use of the juxtaposition strategy is limited to a subclass of adjectives.
85In Maninka and Kakabe, copulæ can be optionally dropped in contact with a focalization marker, and the juxtaposition construction is productive in the predicative use of a subclass of adjectives, as in (9).
|
(9)
|
Maninka (Vydrin 2020: 90)
|
|
À
|
bɛ̀sɛ,
|
bɛ̀sɛya`
|
díman
|
à
|
ɲɛ́.
|
|
|
|
|
|
|
|
3SG
|
neat
|
neatness.D
|
pleasant
|
3SG
|
before
|
|
|
|
|
|
|
|
‘He is neat, he likes neatness.’ lit. ‘neatness is pleasant to him.’
|
- 3 This analysis is not explicitly formulated by Heath (2017), who simply designates the tonal copula (...)
86Jalkunan has a construction that at first sight seems to be an instance of the juxtaposition strategy with the linear order ARG PRED, but the argument NP undergoes a tonal modification that can be accounted for straightforwardly by positing a construction ARG=COP PRED with an enclitic copula consisting of a floating high tone.3
87As regards the predicative inflection strategy, no absolutely obvious and uncontroversial case is attested in Mande. It is, however, tempting to analyze nominal predication in Jenaama as an instance of the predicative inflection strategy, since it involves a marker nì which is consistently postposed to noun phrases in predicate function and is found in no other context.
- 4 Heath (2022a; 2022b) simply glosses nì as ‘it.is’, without discussing its status in a typology of (...)
88Jenaama has two distinct nominal predication constructions: a construction X nì ‘It is X’ (10a), in which the argument of the nominal predicate X is not expressed (and must be retrieved from the context), and a construction X gà Y nì ‘X is Y’ where X is the argument of the nominal predicate Y (10b). Moreover, gà also acts as a copula in locational predication (10c). In Mande languages, it is common that nominal predicates following a copula are flagged by an adposition, but nì is not attested as an adposition flagging oblique phrases in clauses whose predicative nucleus is a verb, and cannot be analyzed as an adposition in the construction illustrated in (10a). Consequently, the only analysis that does not encounter serious difficulties is that nì is a nominal predicate marker, a copula being required if and only if the argument of the predicative-marked NP is expressed.4
|
(10)
|
Jenaama (Heath 2022b: 340, 343, 344)
|
|
a.
|
Kɛ́ɛ́wú-yé
|
nì.
|
|
|
|
|
|
|
|
|
|
|
man-PL
|
NPRED
|
|
|
|
|
|
|
|
|
|
|
‘It’s (the) men.’
|
|
|
|
|
|
b.
|
Sèédù
|
gá
|
sùwɔ̀yá
|
nì.
|
|
|
PRN
|
COP
|
farmer
|
NPRED
|
|
|
‘Seydou is a farmer.’
|
|
c.
|
Wùláá
|
gà
|
júgúŋ
|
kùmà?
|
|
|
who
|
COP
|
tree
|
on
|
|
|
‘Who is up in the tree?’
|
89There is however a problem with the analysis of (10a) as an instance of the predicative inflection strategy, and of (10b) as combining the predicative inflection strategy and the copula strategy. The point is that, morphologically, nì cannot be described as a noun suffix. It is rather an enclitic attached to the last word of the noun phrase, and the word to which nì attaches is not necessarily the noun acting as the nucleus of the noun phrase, since in Jenaama, as in the other Mande languages, NPs are not head-final. Moreover, nì can be separated from the last word of the noun phrase by other enclitics such as the focus marker wò or the past marker tùn. Consequently, the decision to classify the constructions in question as involving predicative inflection of nouns or not entirely depends on the decision to adopt a broad definition of inflection allowing for the possibility of phrasal inflection, or a narrow definition of inflection restricting this notion to word inflection.
90Historically, the Jenaama nominal predicate marker nì may be related to the copula ni (whose tone alternates according to the context) that marks nominal predication in Soninke (see example (11) below), but synchronically, their distribution is very different.
91In Mande languages, non-verbal copulæ are particularly prominent, and the use of copular verbs is mainly conditioned by the expression of TAM values that cannot be expressed in constructions involving non-verbal copulæ, due to their incompatibility with the predicative markers that carry grammaticalized TAM distinctions in verbful clauses. The verbs used as copulæ often also have non-copular uses. For example, the copular use of a verb ‘do’ in nominal predication is attested not only in Manding languages, but also a.o. in Mano, Mwan, Gban, Boko.
92As regards the number of distinct non-verbal copulæ and their distribution, the configuration described for Mandinka in §4, with two non-verbal copulæ (nominal and adverbial) in positive clauses and a single negative non-verbal copula not related formally to the positive copulæ, is very common across the Mande language family, but other possibilities are attested.
93Soninke has two distinct non-verbal copulæ (nominal and adverbial) illustrated in (11), each one with its own negative counterpart. The same configuration is found in Vai and Kakabe.
|
(11)
|
Soninke
|
|
a.
|
Ké
|
yúgó
|
nì
|
tágén
|
ñà
|
yí.
|
|
|
DEM
|
man
|
NomCOP
|
blacksmith.D
|
FOC
|
POSTP
|
|
|
‘This man is a blacksmith.’
|
|
b.
|
Ké
|
yúgó
|
hètí
|
tágé
|
yì.
|
|
|
DEM
|
man
|
NomCOP.NEG
|
blacksmith
|
POSTP
|
|
|
‘This man is not a blacksmith.’
|
|
|
c.
|
Démbà
|
wá
|
kónpèn
|
dí.
|
|
|
PRN
|
AdvCOP
|
room.D
|
in
|
|
|
‘Demba is in the room.’
|
|
d.
|
Démbà
|
ntá
|
kónpèn
|
dí.
|
|
|
PRN
|
AdvCOP.NEG
|
room.D
|
in
|
|
|
‘Demba is not in the room.’
|
- 5 Maninka, Mauka, Kong Jula, Koro, Bolon and Kagoro all belong to the Manding dialect cluster, but in (...)
94Positive non-verbal copulæ used both for nominal and adverbial predication can be found in the following languages: Maninka, Mauka, Kong Jula (example (12)), Koro, Bolon (example (13)), Kagoro,5 Jalkunan, Jeli, Tigemaxo, Jenaama, Mende, Kpelle, Looma, Mano, Kla-dan, Mwan, Guro, Yaure.
|
(12)
|
Kong Jula (Sangaré 1984: 209, 213)
|
|
a.
|
Árí
|
bɛ́
|
mín?
|
|
|
|
|
|
|
|
|
|
|
|
2PL
|
COP
|
where
|
|
|
|
|
|
|
|
|
|
|
|
‘Where are you?’
|
|
b.
|
É
|
bɛ́
|
ń
|
tyɛ́rí
|
yé.
|
|
|
2SG
|
COP
|
1SG
|
friend
|
as
|
|
|
‘You are my friend.’
|
|
|
(13)
|
Bolon (Zoungarana 1987: 109, 151)
|
|
a.
|
À
|
mún
|
mùgɔ̀
|
yé.
|
|
|
|
|
|
|
|
|
|
|
3SG
|
COP
|
person
|
as
|
|
|
|
|
|
|
|
|
|
|
‘S/he is a person.’
|
|
b.
|
À
|
mún
|
ń
|
fɛ̀.
|
|
|
|
3SG
|
COP
|
1SG
|
at
|
|
|
|
‘S/he is at my place.’
|
95Some languages have two distinct copulæ in nominal predication depending on the fact that the argument of the nominal predicate is overtly expressed or left implicit: Bambara, Kong Jula (example (14)), Jalkunan, Looma.
|
(14)
|
Kong Jula (Sangaré 1984: 211, 213)
|
|
a.
|
Murú
|
lomú.
|
|
|
|
|
|
|
|
|
|
|
|
|
knife
|
NomCOP
|
|
|
|
|
|
|
|
|
|
|
|
|
‘This is a knife.’
|
|
b.
|
U
|
bɛ́
|
bínkɔ́nɔ́nɔ́fɛ́n
|
ne
|
yé.
|
|
|
3PL
|
COP
|
wild.animal
|
FOC
|
as
|
|
|
‘They are wild animals.’
|
|
- 6 Note that, in Kpelle, káá also acts as the adverbial copula, whereas ɓàà is only found in nomin (...)
96In some other languages, the choice between two possible copulæ in nominal predication correlates with variation in constituent order (ARG COP PRED vs. PRED COP ARG): Koro, Kpelle (example (15)), Tura.6
|
(15)
|
Kpelle (Konoshenko 2017: 311–312)
|
|
a.
|
Zààwòlò
|
káá
|
à
|
kólôkɛ́mùŋ
|
lɛ́lɛ́ɛ
|
|
|
|
|
|
|
|
|
|
PRN
|
COP
|
with
|
worker
|
good
|
|
|
|
|
|
|
|
|
|
‘Zaawolo is a good worker.’
|
|
b.
|
Ǹɛ̀ɛ̀nû
|
wáláa
|
ɓàà
|
à
|
yɛ́.
|
|
|
woman
|
strong
|
NomCOP
|
with
|
you
|
|
|
‘You are a strong woman.’
|
|
97Some languages depart from the typical Mande pattern in having no specific negative strategy for non-verbal predication (see §3.5 below).
98Finally, an atypical configuration has been signaled in San, with two copulæ equally used for nominal and adverbial predication, one of them encoding that the clause expresses “new information”, whereas the other is found in clauses referring to “general states of affairs”.
99As discussed by Creissels (2017), in Mande languages, the grammaticalization of the imperative of the verb ‘see’ into an ostensive marker, and further into a copula, is a common source of copulæ.
100The coincidence between copulæ and term focus markers in clauses whose predicative nucleus is a verb is widespread across Mande languages, and the languages in which a copula is optionally elided in contact with a focus marker, such as Guinean Maninka (Vydrin 2020) show that focus markers can be reanalyzed as copulæ. However, in the languages of the world, the grammaticalization of copulæ as focus markers is also very common. Moreover, as rightly pointed to me by Dmitry Idiatov, the evolutions involving copulæ and focus markers are not necessarily unidirectional and may go in cycles, viz. COP > FOC > COP or FOC > COP > FOC, and yet another possibility is that both a copula and a focus marker of the same form originate from the same source, such as a demonstrative, rather than from one another.
101The other possible scenarios for the creation of non-verbal copulæ are, on the one hand, deverbalization of copular verbs, and on the other hand, grammaticalization of demonstrative pronouns or adverbs into copulæ. Evidence of such evolutions can be found in some Mande languages, such as the grammaticalization of nàa ‘there’ as an adverbial copula in Soso, the grammaticalization of the verb ɗɤ́ in Eastern Dan (Vydrin 2020), and the grammaticalization of demonstrative adverbs into copulæ in Eastern Dan (Vydrin 2020).
102Specific negative strategies for non-verbal predication are the general rule in Mande. Dzuun, Jalkunan, Jeli, Soso and Jalonke, with negation markers shared by verbal and non-verbal predication, are exceptions. Example (16) shows that, in Dzuun, the same negation marker nā´ can be found in verbal and non-verbal predication.
103(16) Dzuun (Solomiac 2014: 258, 278)
|
|
a.
|
Ē
|
dzín
|
cī
|
fàkɔ́ɔ́n
|
rè
|
fàà
|
nā
|
|
|
2SG.COND
|
say
|
QUOT
|
madman.D
|
GEN
|
madness.D
|
NEG
|
|
|
ā
|
ré
|
fà, ...
|
|
2SG
|
GEN
|
madness
|
104‘If you say that the madman’s madness is not your madness, ...’
|
b.
|
À
|
nā
|
dzī
|
fú.
|
|
|
|
3SG
|
NEG
|
be.able
|
get.up
|
|
|
|
‘He was not able to get up.’
|
105In many Mande languages, the possibility of leaving unexpressed the argument of nominal predicates in constructions involving non-verbal copulæ distinguishes nominal predication from both adverbial and verbal predication, where nothing similar can be observed. In most Mande languages, in the absence of an overt argument phrase, the nominal predicate invariably precedes the copula, as in (17). However, the maintenance of the linear order COP PRED even in the absence of an overt argument phrase is attested in Seenku (example (18)).
|
(17)
|
Mandinka
|
|
a.
|
Tàmbàjáŋò
|
mú
|
ñêe
|
lè
|
tí.
|
|
|
fish(sp.).D
|
NomCOP
|
fish.D
|
FOC
|
as
|
|
|
‘The tàmbàjáŋ is a fish.’
|
|
b.
|
Ñêe
|
lè
|
mú.
|
|
|
|
|
fish.D
|
FOC
|
NomCOP
|
|
|
|
|
‘It’s (a) fish.’
|
|
(18)
|
Seenku, aka Sembla (Prost 1971: 47, 48)
|
|
a.
|
Mö
|
nī
|
kë
|
kwàasɛ̀rà.
|
|
|
|
1SG
|
father
|
NomCOP
|
farmer
|
|
|
|
‘My father is a farmer.’
|
|
b.
|
kë
|
sò.
|
|
|
|
|
|
NomCOP
|
horse
|
|
|
|
|
|
‘It’s a horse.’
|
106As already mentioned in §3.3, in some languages, distinct copulæ are used when the argument of a nominal predicate is left unexpressed. Bambara (example (19)) provides an additional illustration.
|
(19)
|
Bambara (Vydrin 2020: 82)
|
|
a.
|
Fántà Kúlibàli
|
yé
|
mùso
|
hákilima
|
yé.
|
|
|
PRN
|
NomCOP
|
woman
|
intelligent
|
POSTP
|
|
|
‘Fanta Kulibali is an intelligent woman.’
|
|
b.
|
Dúnan`
|
dòn.
|
|
|
|
|
|
fish.D
|
NomCOP
|
|
|
|
|
|
‘It is a stranger.’
|
107Nominal predication constructions with a single nominal term in predicate function implying an unexpressed argument whose identity must be retrieved from the context should be distinguished from ostensive constructions, i.e. from clauses whose function is to draw the attention of the addressee to the presence of some entity in the situation within which the speaker-addressee interaction takes place (speech situation). However, the distinction is not always easy to establish. The main criterion for establishing the distinction between ostensive markers (i.e., particles such as French voici or Russian vot) and copulæ in nominal predication constructions in which the argument of the nominal predicate is left implicit is that ostensive markers cannot occur in question-answer sequences such as ‘–Who/what is this? –This is (an) X’. According to this criterion, Bambara dòn in (19b) above cannot be analyzed as an ostensive marker, but only as copula in a nominal predication construction in which the argument of the nominal predicate is obligatorily unexpressed and must be retrieved from the context, cf. –Mǔn dòn ? –X dòn ‘–Who/what is this ? –This is (an) X’.
108Unfortunately, not all descriptions provide the data making it possible to establish with certainty the exact nature of markers that are just described as combining with a single NP into clauses that are simply glossed ‘This is N’. Moreover, the analysis can be made difficult by the fact that words or clitics that are originally ostensive markers may be engaged in an evolution tending to convert them into copulæ (as already mentioned, this is a very common grammaticalization path in Mande).
109On the possibility that some of the nominal copulæ found exclusively in constructions in which the argument is left unexpressed result from the fusion of a focus marker and a copula, see Creissels (1981).
110In Mande languages, it is common that focus marking involves syntactically optional markers that may attach to any of the terms of the clause and focalize the term to which they attach. In some Mande languages, focus markers that generally meet this characterization show a particular behavior in nominal predication. For example, in the Mandinka example (17a), repeated here as (20), the omission of the focus marker lè results in ungrammaticality (see §4 for more details).
|
(20)
|
Mandinka
|
|
Tàmbàjáŋò
|
mú
|
ñêe
|
*(lè)
|
tí.
|
|
fish(sp.).D
|
NomCOP
|
fish.D
|
FOC
|
as
|
|
‘The tàmbàjáŋ is a fish.’
|
111The same phenomenon, which will be further analyzed in §6, has also been signaled in Maninka, Kakabe, Soso, Jalonke, Soninke, and Gban.
112It is interesting to observe that relatively close Mande languages may contrast in this respect. For example, the constraint on the use of the focus marker in nominal predication observed in Mandinka is also found in Guinean Maninka (Vydrin 2020), but not in closely related Bambara, as illustrated by example (19a) above.
113A third property of nominal predication in Mande languages that has no equivalent in verbal or adverbial predication is the existence of relatively important variation in the order of the core terms and their flagging characteristics, sharply contrasting with the rigidity and uniformity of linear ordering and flagging patterns in verbal and adverbial predication across Mande languages. Seven distinct patterns, variously attested in the individual languages, can be distinguished.
114Constructions schematizable as ARG COP PRED+adp, in which a copula is followed by the predicate noun phrase flagged by an adposition, and preceded by the unflagged argument noun phrase, are particularly widespread across Mande languages. This type, illustrated in the introduction by example (3b) (repeated here as (21a)), is isomorphous with the construction formed by an intransitive verb, its subject and an oblique phrase, as in (21b), where the postposition flagging the oblique phrase is the same as that flagging the predicate phrase in (21a).
|
(21)
|
Mandinka
|
|
a.
|
Làamínì
|
té
|
kàràndìrìláa
|
tì.
|
|
|
|
PRN
|
COP.NEG
|
teacher
|
as
|
|
|
|
‘Laamin is not a teacher.’
|
|
b.
|
Làamínì
|
lôn-tá
|
kàràndìrìlâa
|
lè
|
tí.
|
|
|
PRN
|
know-CPL.TR
|
teacher.D
|
FOC
|
as
|
|
|
‘Laamin is known as a teacher.’
|
115The adpositions used to flag the predicate NP in the ARG COP PRED+adp construction are sometimes multifunctional adpositions, which makes it difficult to decide which of their meanings may explain their use in nominal predication, but in general, they are also found in verbal clauses with a functive or comitative function (i.e., as the equivalent of English as or with). A locative postposition ‘in’ is also attested in the role of nominal predicate flag in Jeli. The use of a comitative preposition to flag noun phrases in predicate function is found among others in Kpelle (example (15a), repeated here as (22)).
|
(22)
|
Kpelle (Konoshenko 2017: 311)
|
|
Zàawòlò
|
káa
|
à
|
kólôkɛ́mùŋ
|
lɛ́lɛ́ɛ.
|
|
|
|
|
|
|
|
|
PRN
|
COP
|
with
|
worker
|
good
|
|
|
|
|
|
|
|
|
‘Zaawolo is a good worker.’
|
116The use of comitative adpositions in nominal predication and the connection it suggests between comitative and functive functions is at first sight an intriguing aspect of Mande adposition systems, since ‘X is with Y’ implies referential disjunction between X and Y, and it is difficult to imagine how a comitative marker could extend its use to functive function, or vice-versa. However, a plausible explanation of the comitative-functive connection is that locative markers are a possible source of both comitative markers and functive markers.
117Some languages have a construction formally identical to that described in §3.6.3.1, but in which the unflagged noun phrase that precedes the nominal copula is the predicate, and the flagged noun phrase that follows it is the argument: PRED COP ARG+adp. This construction is the only one mentioned in the sources I have been able to consult on Manya, Koranko, Mende, and Loko.
|
(23)
|
Mende (Innes 1971: 51)
|
|
Mɛndémɔ̂
|
lɔ
|
a
|
ngé.
|
|
|
|
|
|
|
|
|
|
Mendeman
|
COP
|
with
|
1SG
|
|
|
|
|
|
|
|
|
|
‘I am a Mendeman.’
|
118As already mentioned in the introduction, there are also Mande languages in which the ARG COP PRED+adp and PRED COP ARG+adp patterns, with the same adposition flagging the predicate phrase in the first variant and the argument phrase in the second one, can be used to express the same predicate-argument articulation, with just a difference in information packaging. This phenomenon, which constitutes the main topic of this article, and will be described and analyzed in more detail for Mandinka in §4, has been illustrated in the introduction by example (2).
119Soso and Jalonke have a construction schematizable as PRED ARG+adp. No copula is involved in this construction, but the argument phrase is flagged by the comitative postposition rá, and in independent positive clauses, the predicate phrase is obligatorily followed by the focus maker nán.
|
(24)
|
Soso (pers.doc.)
|
|
Xàràndîi
|
nán
|
jí
|
dímèdíi
|
rà.
|
|
|
|
|
pupil
|
FOC
|
DEM
|
child
|
with
|
|
|
|
|
‘This child is a pupil.’
|
120As illustrated by example (25), a copulative construction ARG COP PRED(+adp), in which the flagging of the predicate NP seems to be conditioned by the distinction between inclusion and identity, is found in Jeli.
|
(25)
|
Jeli (Tröbs 1998: 174, 177)
|
|
a.
|
Naa
|
jeli
|
si
|
kɔŋ.
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1SG.COP
|
Jeli
|
FOC
|
in
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
‘I am Jeli.’ (answer to the question ‘What is your ethnic group?’
|
|
b.
|
Na
|
seŋ
|
sa
|
Bilari.
|
|
|
|
1SG
|
father
|
FOC.COP
|
PRN
|
|
|
|
‘My father is Bilari.’
|
121As illustrated by example (18a), repeated here as (26), a copulative construction ARG COP PRED with no flagging of either nominal term is found in Seenku.
|
(26)
|
Seenku, aka Sembla (Prost 1971: 47, 48)
|
|
Mö
|
nī
|
kë
|
kwàasɛ̀rà.
|
|
|
|
|
|
1SG
|
father
|
NomCOP
|
farmer
|
|
|
|
|
|
‘My father is a farmer.’
|
122The ARG PRED COP pattern with no flagging of either nominal term, and the copula in final position, is found in Beng.
|
(27)
|
Beng (Paperno 2014: 96)
|
|
Lɛ́ŋ́
|
gɔ̄ŋ̄
|
yā̰á̰
|
sɔ̰̀ŋ̀
|
jàté-lí
|
bɛ́ɛ̄
|
dō
|
ɛ̀
|
|
child
|
man
|
this
|
person
|
respect-AG
|
big
|
one
|
NomCOP
|
|
‘This boy is very polite.’ lit. ‘... is a big respecter of people’
|
123In Vai, there is variation between the PRED COP ARG+adp and ARG PRED COP patterns, the latter expressing topicalization of the argument.
|
(28)
|
Vai (Welmers 1976: 74)
|
|
a.
|
Kàì
|
sókɛ̀
|
mù
|
kò’à lɛ́ɛ̀
|
’à.
|
|
|
man
|
job
|
COP
|
weaving
|
POSTP
|
|
|
‘Weaving is a man’s job.’
|
|
b.
|
Kò’à lɛ́ɛ̀,
|
kàì
|
sókɛ̀
|
mù.
|
|
|
|
weaving
|
man
|
job
|
COP
|
|
|
|
same meaning as (a)
|
124This suggests that constructions schematizable as ARG PRED COP with no flagging of either of the nominal terms and the copula in final position may emerge from routinization of argument topicalization in a construction in which the argument NP occurs in second position, and in fact, this hypothesis is supported by Soso and Jalonke data. The point is that, in Soso and Jalonke, the routinization of argument topicalization in the PRED (FOC) ARG+adp construction presented in 3.5.3.3 has resulted in the emergence of a construction ARG PRED COP with a copula nâará. Historically, nâará is the contraction of the sequence nân (focus marker) + à (3rd person pronoun resuming the topicalized argument) + rá (the comitative postposition that initially flagged the NP in argument function), but this decomposition is not relevant anymore, since nâará is also used with first or second person pronouns in argument role, as in (29).
|
(29)
|
Soso (pers.doc.)
|
|
Ń-tàn
|
xámɛ́ɛ
|
nâará.
|
|
|
|
|
|
|
1SG-EMPH
|
man
|
NomCOP
|
|
|
|
|
|
|
‘I am a man.’
|
125Finally, a juxtaposition construction ARG PRED with no flagging of either nominal term is productive in Dzuun (30) and Bobo (31).
|
(30)
|
Dzuun (Solomiac 2014: 257)
|
|
Síbírí
|
fɛ̄ɛfābáá.
|
|
|
|
|
|
|
|
PRN
|
farmer
|
|
|
|
|
|
|
|
‘Sibiri is a farmer.’
|
|
|
|
|
(31)
|
Bobo (le Bris & Prost 1981: 52)
|
|
Mɛ́
|
diàmū
|
Sànɔ̄n.
|
|
|
|
|
|
|
|
1SG
|
surname
|
PRN
|
|
|
|
|
|
|
|
‘My surname is Sanon.’
|
|
|
|
|
126Mandinka has no productive pattern of non-verbal predication involving mere juxtaposition of noun phrases and/or adpositional phrases. With some marginal exceptions (on which see (Creissels & Sambou 2013: 152–153)), non-verbal predication in Mandinka relies on the use of copulæ. Copular verbs are mainly used to compensate for the impossibility of expressing the grammaticalized TAM values that characterize verbful clauses in the constructions involving non-verbal copulæ.
127Mandinka has two non-verbal copulæ in positive clauses. Each of them corresponds to a synonymous copular verb.
- 7 The variant bí of the adverbial copula bé is commonly used (in free variation with bé) in combin (...)
128The non-verbal adverbial copula bé (with variants bí, bá, jé and jí)7 marks adverbial predication, i.e., constructions in which the predicate is an adpositional phrase or adverb otherwise found in the role of adjunct in verbal clauses.
|
(32)
|
Mandinka
|
|
Fàatú
|
bé
|
fàrôo
|
tó.
|
|
|
|
|
|
|
PRN
|
AdvCOP
|
rice.field.D
|
LOC
|
|
|
|
|
|
|
‘Faatu is at the rice field.’
|
129The corresponding copular verb is tàrá, also used transitively with the meaning ‘find’, as in (33).
|
(33)
|
Mandinka
|
|
Ŋ́
|
ŋá
|
kèê-lú
|
tàrá
|
dómóròo
|
lá
|
sìtôo
|
kótò.
|
|
1SG
|
CPL.TR
|
man.D-PL
|
find
|
eating.D
|
POSTP
|
baobab.D
|
under
|
|
‘I found the men eating under the baobab.’
|
130Quite regularly, tàrá can also be used intransitively with the meaning ‘be found’. However, tàrá used intransitively is mainly found as a copular verb whose function is to combine adverbial predication with the expression of TAM values that cannot be expressed in clauses involving non-verbal copulæ, as for example ‘potential’ in (34b).
131(34) Mandinka
|
a.
|
Í
|
màŋ
|
búwàa-lú
|
jè,
|
bàrí,
|
ì
|
bé
|
í
|
|
|
2SG
|
CPL.NEG
|
sorcerer.D-PL
|
see
|
yet
|
3PL
|
AdvCOP
|
2SG
|
132‘You did not see the sorcerers, yet they are close by you.’
|
b.
|
Í
|
tè
|
búwàa-lú
|
jé-là,
|
hání
|
ì
|
|
2SG
|
COP.NEG>AUX
|
sorcerer.D-PL
|
see-INF
|
even.if
|
3PL
|
|
sè
|
tàrá
|
í
|
dáalà.
|
|
POT
|
be.found
|
2SG
|
close.to
|
133‘You will not see the sorcerers, even if they may be close by you’, lit. ‘...even if they may be found close by you.’
134The non-verbal nominal copula mú marks nominal predication.
|
(35)
|
Mandinka
|
|
Làamínì
|
mú
|
kàràndìrìlâa
|
lè
|
tí.
|
|
|
|
|
|
PRN
|
NomCOP
|
teacher.D
|
FOC
|
as
|
|
|
|
|
|
‘Laamin is a teacher.’
|
135The corresponding copular verb is ké, used transitively with the meaning ‘do’. In its intransitive use, ké expresses the meanings ‘happen, occur, become’, but also acts as a substitute for the nominal copula, if the speaker wants to combine non-verbal predication with the expression of the TAM distinctions carried by predicative markers in verbful clauses, as illustrated by kà expressing habituality in (36b).
|
(36)
|
Mandinka
|
|
a.
|
Ñǐŋ
|
mòô-lú
|
mú
|
tìlìbòŋkôo-lú
|
lè
|
tí.
|
|
|
|
|
DEM
|
person.D-PL
|
NomCOP
|
Easterner.D-PL
|
FOC
|
as
|
|
|
|
|
‘Those people are Easterners.’
|
|
b.
|
Jàmáajàmáa,
|
ñǐŋ
|
mòô-lú
|
kà
|
ké
|
tìlìbòŋkôo-lú
|
|
|
often
|
DEM
|
person.D-PL
|
ICPL
|
be
|
Easterner.D-PL
|
136‘Often, those people are Easterners.’
- 8 For example, with jàmfá ‘move away’, à jàmfá-tà may be understood as ‘he/she/it moved away’ ( (...)
137Moreover, the completive form of ké can be interpreted as fully equivalent to the nominal copula, due to the fact that, with some Mandinka verbs (including ké), a purely stative reading of the predicative markers for which the label ‘completive’ is used here is possible.8 As illustrated by example (37), the use of the negative form mâŋ ké not only with the meaning ‘did not become’, but also as a mere equivalent of the negative copula in nominal predication, is particularly common.
|
(37)
|
Mandinka
|
|
Ŋ́
|
tòo
|
mâŋ
|
ké
|
Músáa
|
tì,
|
Áamádù
|
lè
|
mú.
|
|
1SG
|
name.D
|
CPL.NEG
|
be
|
PRN
|
as
|
PRN
|
FOC
|
NomCOP
|
|
‘My name is not Musaa, it’s Aamadu.’
|
|
138In its use as a copular verb, ké differs syntactically from the non-verbal nominal copula in the following two respects: with ké, the argument of the nominal predicate is obligatorily expressed, and no argument-predicate reversal can occur.
- 9 The variant tí of the negative copula is found exclusively in adverbial predication, with the same (...)
139The same non-verbal negative copula té (with variants tí and ǹté(ŋ))9 is used in nominal and adverbial predication.
|
(38)
|
Mandinka
|
|
a.
|
Fàatú
|
té
|
fàrôo
|
tó.
|
|
|
PRN
|
COP.NEG
|
rice.field.D
|
LOC
|
|
|
‘Faatu is not at the rice field.’
|
|
b.
|
Làamínì
|
té
|
kàràndìrìláa
|
tì.
|
|
|
PRN
|
COP.NEG
|
teacher
|
as
|
|
|
‘Laamin is not a teacher.’
|
140From a purely syntactic point of view, the non-verbal copulæ can be analyzed as fulfilling a role similar to that of intransitive verbs in the construction of clauses, and one might consider analyzing them as irregular and defective intransitive verbs. However, non-verbal copulæ differ from regular intransitive verbs not only in their incompatibility with the predicative markers whose paradigm characterizes verbal predication, but also in the following respects:
- 10 On the infinitive and other dependent forms of Mandinka verbs, see Creissels & Sambou (2013: 125–13 (...)
141they do not have dependent forms;10
- 11 On the possibility of using the verbal lexemes of Mandinka as event nouns, see Creissels & Sambou ( (...)
142they cannot be used as event nouns;11
143they cannot serve as the input of the derivational operations to which verbs lend themselves.
144In Mandinka, as in most Mande languages, the same constructions are used for inclusion and identity. As illustrated in (39), there is in general no overt indication of the distinction between identity (39a) and inclusion (39b-c).
|
(39)
|
Mandinka
|
|
a.
|
Ñǐŋ
|
kèê
|
mú
|
ŋ̀
|
ná
|
àlìmáamòo
|
lè
|
tí.
|
|
|
DEM
|
man.D
|
NomCOP
|
1PL
|
GEN
|
imam.D
|
FOC
|
as
|
|
|
‘This man is our imam.’
|
|
b.
|
Ŋ́
|
bárímmáa
|
mú
|
àlìmáamòo
|
lè
|
tí.
|
|
|
1SG
|
uncle
|
NomCOP
|
imam.D
|
FOC
|
as
|
|
|
‘My uncle is an imam.’
|
|
c.
|
Sìtôo
|
mú
|
yíròo
|
lè
|
tí.
|
|
|
baobab.D
|
NomCOP
|
tree.D
|
FOC
|
as
|
|
|
‘Baobabs are trees.’
|
145However, in negative and interrogative contexts, the distinction can be made apparent by the possibility of contrasting the presence of the determiner ‑ò with its absence, as in (40).
|
(40)
|
Mandinka
|
|
a.
|
Ñǐŋ
|
kèê
|
té
|
ŋ̀
|
ná
|
àlìmáamòo
|
tí.
|
|
|
DEM
|
man.D
|
COP.NEG
|
1PL
|
GEN
|
imam.D
|
as
|
|
|
‘This man is not our imam.’
|
|
b.
|
Ŋ́
|
bárímmáa
|
té
|
àlìmáamì
|
tí.
|
|
|
|
|
|
|
|
|
|
|
DEM
|
uncle
|
COP.NEG
|
imam
|
as
|
|
|
|
|
|
|
|
|
|
|
‘My uncle is not an imam.’
|
146The absence of any manifestation of the distinction between identity and inclusion in positive assertive clauses expressing nominal predication is due to the fact that nouns in predicate function behave exactly like nouns fulfilling other functions with respect to the use of the determiner ‑ò. In positive assertive clauses, common nouns in predicate function, like common nouns in other functions, are normally in the default ò-form, whatever their referential value, whereas in negative or interrogative contexts, common nouns in predicate function with no referential content, like common nouns in other functions, may be in the bare form, as in (40b).
147Nominal predication is characterized by a special behavior of the focus marker lè. In verbal or adverbial predication, the focus marker is never obligatory, and when present, it can attach to any term of the clause. By contrast, in nominal predication, as described in more detail in the remainder of this section, it may be obligatory, and it can only attach to the predicate NP.
148As already mentioned, a property that sharply distinguishes nominal predication from both verbal and adverbial predication in many Mande languages is that, in nominal predication marked by a non-verbal copula, the argument of the nominal predicate (i.e., the entity whose categorization or identification is expressed by the nominal predicate) can be left unexpressed. This is the case with Mandinka, where two varieties of nominal predication marked by a non-verbal copula must therefore be distinguished: X mú/té, with a single nominal term X in predicate function, and X mú/té Y tí, with two nominal terms X and Y, and the postposition tí flagging the nominal term that follows the copula.
149In the construction X mú/té ‘It is (not) X’, X is a predicate NP whose argument is not expressed, and must be retrieved from the context. This construction is distinct from the ostensive construction X félè ‘Here is X’, where the ostensive marker félè is etymologically the imperative of the verb félè ‘look’.
- 12 Note that, when the focus marker lè immediately precedes the copula mú, the sequence lè mú may (...)
150In independent positive clauses instantiating the X mú/té construction (either assertive or interrogative), the focus marker is obligatorily present. It may be found either between the predicate NP and the copula (41a),12 in which case it carries no particular semantic implication, or immediately after a genitival modifier within the predicate NP (41b-c), in which case it has its usual focalizing effect.
|
(41)
|
Mandinka
|
|
a.
|
Músáa
|
lá
|
mùsôo
|
lá
|
kódòo
|
lè
|
mú.
|
|
|
PRN
|
GEN
|
wife.D
|
GEN
|
money.D
|
FOC
|
NomCOP
|
|
|
‘It’s Musaa’s wife’s money.’
|
|
b.
|
Músáa
|
lá
|
mùsôo
|
lè
|
lá
|
kódòo
|
mú.
|
|
|
|
|
|
|
|
|
PRN
|
GEN
|
wife.D
|
FOC
|
GEN
|
money.D
|
NomCOP
|
|
|
|
|
|
|
|
|
‘It’s MUSAA’S WIFE’s money’ (not someone else’s money).
|
|
c.
|
Músáa
|
lè
|
lá
|
mùsôo
|
lá
|
kódòo
|
mú.
|
|
|
|
|
|
|
|
|
PRN
|
FOC
|
GEN
|
wife.D
|
GEN
|
money.D
|
NomCOP
|
|
|
|
|
|
|
|
|
‘It’s MUSAA’s wife’s money’ (not someone else’s wife’s money).
|
151In the corresponding negative clauses, it is impossible to introduce the focus marker.
|
(42)
|
Mandinka
|
|
a.
|
Músáa
|
*(lè)
|
mú.
|
|
|
PRN
|
FOC
|
NomCOP
|
|
|
‘It’s Musaa.’
|
|
b.
|
Músáa
|
(*lè)
|
ǹté.
|
|
|
PRN
|
FOC
|
COP.NEG
|
|
|
‘It’s not Musaa.’
|
152In the construction X mú/té Y tí, tí is a postposition also used productively in verbal predication with a functive meaning, as in (43).
|
(43)
|
Mandinka
|
|
Ŋ́
|
ŋá
|
à
|
lóŋ
|
ŋ̀
|
ná
|
àlìmáamòo
|
tí.
|
|
|
|
|
|
|
1SG
|
CPL.TR
|
3SG
|
know
|
1PL
|
GEN
|
imam.D
|
as
|
|
|
|
|
|
|
‘I knew him as our imam.’
|
153In the construction X mú/té Y tí, as in the construction X mú/té, the focus marker is obligatorily present in independent positive clauses but cannot be introduced in the corresponding negative clauses, and when present, it can only attach to the predicate NP, or to a genitival modifier within the predicate NP. Moreover, the presence of the focus marker correlates with an important difference in the behavior of the construction with respect to the coding of the nominal predicate and its argument.
154When the focus marker is absent, the predicate NP can only be encoded as the second term of the construction, i.e. as the term flagged by the postposition tí, as in (44). This is consistent with the functive meaning expressed by tí ‘as’ in verbal predication.
|
(44)
|
Mandinka
|
|
Ñǐŋ
|
té
|
kúu
|
kùmmáa
|
tì.
|
|
DEM
|
COP.NEG
|
issue
|
important
|
as
|
|
‘This is not an important issue.’
|
155In Mandinka, the focus marker cannot be introduced in relative clauses either, and, quite consistently, in relative clauses instantiating the construction X mú/té Y tí, as in negative independent clauses, the predicate NP can only be encoded as the second term of the construction, as in (45).
|
(45)
|
Mandinka
|
|
mîŋ
|
mú
|
kàràndìrìlâa
|
tí
|
|
|
REL
|
NomCOP
|
teacher.D
|
as
|
|
|
‘the one who is a teacher’
|
- 13 Note that two occurrences of the focus marker in the same clause are not possible. ‘It is Laamin wh (...)
156By contrast, when the focus marker is present, the constructions ARG COP PRED+lè+tí and PRED+lè COP ARG+tí are equally possible. The focus marker lè consistently attaches to the predicate phrase, but it is always the NP in second position that is flagged by the postposition tí, regardless of its semantic role as predicate or argument.13
|
(46)
|
Mandinka
|
|
a.
|
Làamínì
|
mú
|
màndìŋkôo
|
lè
|
tí.
|
|
|
PRN
|
NomCOP
|
Mandinka.D
|
FOC
|
as
|
|
|
‘LaaminARG is MandinkaPRED.’
|
|
b.
|
Màndìŋkôo
|
lè
|
mú
|
Làamínì
|
tí.
|
|
|
Mandinka.D
|
FOC
|
NomCOP
|
PRN
|
as
|
|
|
‘HeARG is MandinkaPRED, Laamin that is.’
|
157The only thing that matters for the denotative meaning and logical structure of the clause is that the focus marker must attach to the predicate phrase or to a genitival modifier within the predicate phrase. However, the choice between the two options (argument first or predicate first) has discursive implications. In this respect, as reflected in the translations of example (46), the argument-first option suggests a relatively neuter topic-comment information packaging, whereas the predicate-first option rather suggests the relatively marked comment-afterthought information packaging.
158A crucial observation is that the argument-first option is the only one attested in gnomic sentences such as (47), whose terms have no link with referents already introduced in the discourse or present in the context of utterance.
|
(47)
|
Mandinka
|
|
Jáwúyàa
|
bóoròo
|
mú
|
jàmfàñóolòo
|
lè
|
tí.
|
|
|
enmity.D
|
remedy.D
|
NomCOP
|
drifting.apart.D
|
FOC
|
as
|
|
|
‘The remedy to enmity is to drift apart from each other.’
|
159The fact that the argument-first variant is the only possibility in the absence of the focus marker, and is systematically used in gnomic sentences, suggests analyzing it as basic. More precisely, from the point of view of information structure, the PRED COP ARG variant can be analyzed as a marked option signaling that the argument fulfills the discursive role of afterthought (or antitopic). Moreover, the analysis of the argument-first variant as basic is consistent with the fact that, semantically, the functive marking that characterizes the predicate in the argument-first variant and the argument in the predicate-first variant makes sense for the noun phrase in predicate function, but not for the noun phrase in argument function.
160Cross-linguistically, it is common that variation in information packaging is encoded by constituent order alternations. It is therefore important to stress that the alternation analyzed here cannot be analyzed as a case of constituent order flexibility, since the flagged nominal term is invariably the second one. An analysis in terms of exchange of semantic roles between the two NPs fulfilling the roles of predicate and argument is also ruled out, since this phenomenon is observed in clauses expressing inclusion, in which such an exchange of roles simply cannot be conceived.
161Moreover, the fact that the alternation observed in pairs of inclusion statements such as (46) should not be confused with the possible exchange of roles between the predicate phrase and the argument phrase in identity statements is confirmed by the observation that, in Mandinka, both phenomena can combine in identity statements, giving rise to the four possibilities illustrated in (48). In this example, the role of predicate is fulfilled by ‘the chief of our village’ in (a) and (b), by ‘Laamin Siisee’ in (c) and (d). The (a) and (b) variants coincide in predicate-argument structure but differ in information structure, and the same holds true for (c) and (d).
162(48) Mandinka
|
a.
|
Làamíní Sìisée
|
mú
|
ŋ̀
|
ná
|
sàatée
|
àlìkáalòo
|
lè
|
tí.
|
|
|
PRN
|
NomCOP
|
1PL
|
GEN
|
village.D
|
chief.D
|
FOC
|
as
|
|
|
‘Laamin Siisee is the chief of our village.’
|
|
b.
|
Ŋ̀
|
ná
|
sàatée
|
àlìkáalòo
|
lè
|
mú
|
Làamíní Sìisée
|
tì.
|
|
1PL
|
GEN
|
village.D
|
chief.D
|
FOC
|
NomCOP
|
PRN
|
as
|
|
‘He is the chief of our village, Laamin Siisee.’
|
|
c.
|
Ŋ̀
|
ná
|
sàatée
|
àlìkáalòo
|
mú
|
Làamíní Sìisée
|
lè
|
tí.
|
|
1PL
|
GEN
|
village.D
|
chief.D
|
NomCOP
|
PRN
|
FOC
|
as
|
|
‘The chief of our village is Laamin Siisee.’
|
|
d.
|
Làamíní Sìisée
|
lè
|
mú
|
ŋ̀
|
ná
|
sàatée
|
àlìkáalòo
|
tí.
|
|
Laamin Siisee
|
FOC
|
NomCOP
|
1PL
|
GEN
|
village.D
|
chief.D
|
as
|
|
‘It is Laamin Siisee, the chief of our village.’
|
163To summarize, the alternation illustrated in (46a-b), (48a-b) and (48c-d), for which I propose the term ‘argument-predicate reversal’, does not involve an exchange of semantic roles in argument-predicate structure, but an exchange of coding characteristics between the argument phrase and the predicate phrase in a predicative construction in which the predicate is consistently marked as such by a marker acting as a focus marker in other constructions.
164Unfortunately, the available grammars of Mande languages do not always provide sufficient data about inclusion statements to establish with certainty whether argument-predicate reversal in the sense given to this term here is attested or not in the languages they describe. In this section, I just briefly mention languages for which I have been able to find pairs of sentences attesting the possibility of an alternation similar to that described for Mandinka in §4.4.
- 14 However, as pointed to me by Dmitry Idiatov (pers.com.), argument-predicate reversal is probably no (...)
165Among the Manding varieties other than Mandinka, the possibility of argument-predicate reversal in nominal predication is explicitly discussed for Bambara and Guinean Maninka by Vydrin (2020: 89), who states that, in this respect, a situation identical to that of Mandinka is found in Guinean Maninka (example (49)), but not in Bambara.14
|
(49)
|
Guinean Maninka (Vydrin 2020: 89, 90)
|
|
a.
|
Músa
|
yé
|
kàranden
|
nè
|
dí.
|
|
|
|
PRN
|
COP
|
student
|
FOC
|
as
|
|
|
|
‘Musa is a student.’
|
|
b.
|
Kàranden
|
nè
|
yé
|
Músa
|
dí.
|
|
|
|
student
|
FOC
|
COP
|
PRN
|
as
|
|
|
|
same meaning as (a)
|
166Argument-predicate reversal is also attested in Niokolo Maninka (Creissels 2013: 67–68) and in Xasonka (Koité-Herschel 1981: 202–203).
167Example (50) shows that a similar alternation is found in Kakabe.
|
(50)
|
Kakabe (Vydrina 2017: 59, 60)
|
|
a.
|
À
|
mù
|
fúlá
|
lè
|
là.
|
|
|
|
3SG
|
NomCOP
|
Fula
|
FOC
|
POSTP
|
|
|
|
‘He is a Fula.’
|
|
b.
|
Déndén
|
ɲák!kɔ́ɛ́
|
lè
|
mù
|
ílè
|
là.
|
|
|
child
|
cursed
|
FOC
|
NomCOP
|
2SG.EMPH
|
POSTP
|
|
|
‘You are a cursed child.’
|
168Example (51) illustrates the possibility of argument-predicate reversal in the inclusion statements of Soninke.
|
(51)
|
Soninke (pers.doc.)
|
|
a.
|
Múusá
|
nì
|
tàgàndàanà-ŋàná-n
|
ñà
|
yí.
|
|
|
|
PRN
|
COP
|
mason-skilled-D
|
FOC
|
POSTP
|
|
|
|
‘Muusa is a skilled mason.’
|
|
b.
|
Tàgàndàanà-ŋàná-n
|
ñà
|
nί
|
Múusá
|
yì.
|
|
|
|
mason-skilled-D
|
FOC
|
COP
|
PRN
|
POSTP
|
|
|
|
‘He is a skilled mason, Muusa that is.’
|
169Kpelle has already been mentioned as a language using two distinct copulæ in nominal predication, depending on the linear ordering of the argument phrase and the predicate phrase. In both variants of nominal predication, the second term of the construction is introduced by the comitative preposition à, irrespective of its role of predicate or argument. As illustrated by example (15) (repeated here as (52), both constructions are possible in inclusion statements, and, consequently, Kpelle qualifies as a language having argument predicate-reversal in the sense given to this term here.
|
(52)
|
Kpelle (Konoshenko 2017: 311–312)
|
|
a.
|
Zààwòlò
|
káá
|
à
|
kólôkɛ́mùŋ
|
lɛ́lɛ́ɛ.
|
|
|
|
|
|
|
|
|
|
PRN
|
COP
|
with
|
worker
|
good
|
|
|
|
|
|
|
|
|
|
‘Zaawolo is a good worker.’
|
|
b.
|
Ǹɛ̀ɛ̀nû
|
wáláa
|
ɓàà
|
à
|
yɛ́.
|
|
|
woman
|
strong
|
NomCOP
|
with
|
you
|
|
|
‘You are a strong woman.’
|
|
170As in the languages mentioned so far as having argument-predicate reversal, the same preposition (in Kpelle: à ‘with’) is used to flag the predicate phrase in the ARG COP PRED+adp pattern and the argument phrase in the PRED COP ARG+adp pattern. However, in contrast to the languages mentioned so far, the distinction between the two alternating patterns is not made apparent by the position of an obligatory focus marker, but by the choice between two distinct copulæ. It is interesting to observe that the copula káá is also used in adverbial predication, but not ɓàà. This is consistent with the fact that, in adverbial predication (contrary to nominal predication), ARG COP PRED is the only possible order.
171Example (53) illustrates the possibility of argument-predicate reversal in the inclusion statements of Looma, with the same preposition gà ‘with’ flagging the predicate phrase in the ARG COP PRED+adp pattern and the argument phrase in the PRED COP ARG+adp pattern. As in Kpelle, the constructions showing this alternation do not involve an obligatory focus marker.
|
(53)
|
Looma (Mishchenko 2017: 403)
|
|
a.
|
Gáá
|
gà
|
lɔ̀ɔ̀mà.
|
|
|
|
|
|
|
|
|
|
|
|
1PL.COP
|
with
|
Looma
|
|
|
|
|
|
|
|
|
|
|
|
‘We are Looma.’
|
|
b.
|
Lɔ̀ɔ̀mà
|
kà
|
gà
|
zè.
|
|
|
|
Looma
|
COP
|
with
|
1SG
|
|
|
|
‘I am Looma.’
|
|
172Example (54) illustrates the possibility of argument-predicate reversal in the inclusion statements of Tura, with the same functive-comitative postposition flagging the predicate phrase in the ARG COP PRED+adp pattern and the argument phrase in the PRED COP ARG+adp pattern. The constructions showing this alternation do not involve an obligatory focus marker, but two distinct copulæ are used in the predicate-first and argument-first variants of inclusion statements. According to Dmitry Idiatov (pers.com.), the argument-first pattern is used in non-thetic contexts, and the predicate-first pattern in thetic contexts. The terminal marker lè found in (54b) “is a clause-final marker primarily used when there is focalization somewhere in the clause, but it is also the default terminal marker used in the presentative construction when no deictic distinction is deemed relevant” (Dmitry Idiatov, pers.com.).
|
(54)
|
Tura (Dmitry Idiatov, pers.com.)
|
|
a.
|
È
|
tɪ̀ɪ̀nɛ́=a̋.
|
|
|
|
|
|
|
|
|
|
|
|
|
3SG.COP
|
orphan=as
|
|
|
|
|
|
|
|
|
|
|
|
|
‘He is an orphan.’ (non-thetic context)
|
|
b.
|
Tɪ̀ɪ̀nɛ̋=ɛ̋
|
à
|
ga̋
|
lè.
|
|
|
|
|
|
|
|
orphan=COP
|
3SG
|
as
|
TM
|
|
|
|
|
|
|
|
‘He is an orphan.’ (thetic context)
|
173Within the limits of the documentation I have been able to gather, argument-predicate reversal in nominal predication is only attested in the western half of the Mande area. However, some of the languages spoken in the western part of the Mande domain (for example, Soso and Jalonke) do not have this alternation. Given the gaps in the documentation, and the possibility that argument-predicate reversal might be found in languages whose available descriptions do not mention its existence, this is a question that would require further investigation.
174In several of the Mande languages in which argument-predicate reversal is attested, the noun phrase in predicate role in the constructions lending themselves to this alternation is not marked as such by its position with regard to the copula or its flagging characteristics, but by the attachment of an obligatory focus marker. Among the nominal predication constructions that are not involved in such an alternation, Soso and Jalonke also have, in similar conditions, an obligatory focus marker attached to noun phrases in predicate function (see §3.6.3.3).
175In other words, in the nominal predication constructions of some Mande languages, in some conditions, a marker otherwise characterizable as a focus marker acts as a nominal predicate marker. In this perspective, the nominal predication construction of Soso and Jalonke with the focus marker nán obligatorily attached to the predicate NP could be analyzed as an instance of the predicative inflection strategy, and the Mandinka-style constructions combining a copula and a focus marker obligatorily attached to the predicate NP could be analyzed as an instance of the strategy combining predicative inflection with the use of a copula.
176In fact, the only problem with this analysis lies in the fact that the focus marker acting as a nominal predicate marker does not consistently attach to the noun that constitutes the nucleus of the noun phrase in predicate function, or to the last word of the predicate phrase. Its default position is at the right margin of the predicate phrase (possibly separated from the noun by modifiers of various types), and (at least in Mandinka) it still has some mobility within the noun phrase in predicate function, since it may also attach to a genitival modifier within the predicate noun phrase. Consequently, as in the case of the nominal predicate marker of Jenaama briefly discussed in §3.1, the decision to classify the constructions in question as involving predicative inflection of nouns or not entirely depends on the decision to adopt a broad definition of inflection allowing for the possibility of phrasal inflection, or a narrow definition of inflection restricting this notion to word inflection.
177In this article, I have shown that, in several Mande languages, nominal predication is characterized by the possibility of an alternation between two constructions for which I have proposed the term ‘argument-predicate reversal’.
178This type of alternation has so far not been discussed in the general literature on nominal predication, and the question arises whether it is attested at all outside of the Mande language family. Significantly, a question about the possibility of Mandinka-style argument-predicate reversal was included in the questionnaire sent to the contributors to a volume on the typology of non-verbal predication I am co-editing with Pier Marco Bertinetto and Luca Ciucci, and the answers we got showed not only that nothing similar occurs in any of the languages of our worldwide sample, but also that linguists that are not familiar with the Mande languages where argument-predicate reversal is attested may have difficulties in conceiving its very existence.
179Argument-predicate reversal in the sense given here to this term, although superficially reminiscent of the cross-linguistically common exchange of roles between the argument phrase and the predicate phrase in identity statements (as in English John is my father / My father is John), does not lend itself to the same analysis, since it also concerns inclusion statements, in which the argument phrase and the predicate phrase are not equireferential, and consequently cannot exchange their roles in the argument-predicate articulation.
180A crucial aspect of argument-predicate reversal as attested in Mandinka and in the other Mande languages listed in §5 is that this alternation cannot be described as a mere instance of constituent order flexibility, since it concerns nominal predication constructions in which one of the two nominal terms is flagged by an adposition, and the same adposition flags the predicate phrase in one of the two variants of inclusion statements, and the argument phrase in the other.
181Although relatively uncommon cross-linguistically, the flagging of one of the two terms of nominal predication constructions by means of case markers or adpositions is not totally unknown. For example, in Slavic languages, the instrumental case can variously be used to flag noun phrases in predicate function. However, in the inclusion statements of Slavic languages, instrumental flagging is reserved to the predicate phrase. For example, in Polish, it is possible to have Karp jest rybą ‘A carp is a fish’, with ‘carp’ in the nominative case and ‘fish’ in the instrumental case, but not *Ryba jest karpiem, with ‘fish’ in the nominative and ‘carp’ in the instrumental, as would be the case if Polish had Mandinka-style argument-predicate reversal.
182What is special in the Mandinka-style argument-predicate reversal is that, in the two variants of inclusion statements, the same adposition is used to flag the predicate in one of the two variants, and the argument in the other, resulting in pairs of sentences that are literally something like ‘A baobab is with/as a tree’ and ‘A tree is with/as a baobab’, but do not differ in the predicate-argument articulation they express: ‘A baobab is a tree’. This undoubtedly constitutes a typological rarum, if not rarissimum.