by Leon Stassen
This map shows the areal distribution of zero copula encoding for predicate nominals. That is, the map indicates whether a given language is like English, in which predicate nominals always require an overt copula (see 1), or rather like Russian, in which omission of the copula is allowed for at least some constructions (see 2).
(1)
a.
John is a sailor.
b.
*John a sailor.
(2) Russian (Maria Koptjevskaja-Tamm p.c.)
Moskva
gorod
Moscow
city
‘Moscow is a city.’
Thus, the following values will be shown on the map:
| Go to map | ||
|---|---|---|
| Value | Representation | |
 |
Zero copula is impossible | 211 |
 |
Zero copula is possible | 175 |
| Total: | 386 | |
To avoid misunderstandings, a couple of remarks on the use of the term zero copula may be in order. This term has been used traditionally in the literature on predicate nominal sentences (see, for example, Benveniste 1966b). It will therefore be familiar to many readers, which is why I have decided to employ it in this chapter. I realize, however, that the term may give rise to Euro-centric interpretations, in that it might suggest that languages which do not have an overt copula are somehow “defective”. On a more theoretical level, the term might suggest that, in languages with zero copula encoding of predicate nominal sentences, there is some phonologically null copula present in the syntax of these sentences, or, alternatively, that there is something corresponding to a copula in the conceptual/semantic structure underlying predicate nominal sentences in these languages. It should be stressed that none of these inferences is warranted. The term zero copula is used here as a strictly neutral technical label, in that it refers purely to a construction in which the relation between a subject and a nominal predicate is not marked by an overt item.
Furthermore, it should be noted that the scope of this chapter is restricted to stative predicate nominal sentences only. That is, the encoding of predicate nominal sentences which have a dynamic reading (commonly marked by some item meaning ‘become’) is not taken into consideration. Hence, when it is stated that a language has zero encoding in predicate nominal sentences, it should not be inferred that such languages cannot have an overt item in sentences which are similar to the English sentence John became a dentist; it only means that they lack an overt copular item in sentences that are similar to English John is a dentist.
Although, for the purposes of this map, zero copula encoding is defined as a yes/no-parameter, it is important to realize that this parameter in effect forms a continuum, with two extremes and a number of in-between cases. At one extreme, we have languages like English, in which zero copula encoding is never allowed. Other examples of this type are Finnish, Kabyle (Berber; Algeria), and Thai.
(3) Finnish (Fromm and Sadeniemi 1956: 115)
ystävä-ni
on
pappi
friend-my
cop.3sg.pres
vicar
‘My friend is a vicar.’
(4) Kabyle (Hanoteau 1906: 88)
netsa
d’
agellid’
en
temourth
agi
he
cop
king
of
country
this
‘He is king of this country.’
khāw
pen
nág:rian:
he
cop
student
‘He is a student.’
At the other end of the continuum we find languages in which zero copula encoding is mandatory for any construction that involves predicate nominals. Sinhalese (Indic; Sri Lanka) and Tubu (Western Saharan; Libya) are cases in point. A concentration of languages of this type is found in Australia and New Guinea. Pitjantjatjara (Pama-Nyungan; Western Australia) and Asmat (Trans-New Guinea; Papua, Indonesia) are examples.
(6) Sinhalese (Gair 1970: 145)
unnaehee
hungak
presidde
kene-k
he
very
famous
person-nom
‘He is/ was a very famous person.’
(7) Tubu (Lukas 1953: 170)
sígən
līfi
3sg.emph
orphan
‘He is/was an orphan.’
(8) Pitjantjatjara (Douglas 1959: 55)
wait
nglayayala
man
doctor
‘The man is/ was a doctor.’
(9) Asmat (Voorhoeve 1965a: 168)
no
ów
akát
1sg
man
handsome
‘I am/was a handsome man.’
On the intermediate positions on the continuum can be situated languages in which the use of the zero copula is limited by a condition of some sort. A well-known, and frequent, condition in this respect involves a split between present and nonpresent tense. Russian and Maltese are examples of languages in which a zero copula is used in the Present, whereas a full copula is mandatory for all other tenses.
(10) Russian (Maria Koptjevskaja-Tamm p.c.)
a.
ona
vrač
she
doctor
‘She is a doctor.’
b.
on
byl
učenik-om
he
be.m.pst
pupil-instr
‘He was a pupil.’
(11) Maltese (Albert Borg, p.c.)
a.
Albert
tabib
Albert
doctor
‘Albert is a doctor.’
b.
Albert
kien
tabib
Albert
be.3sg.m.pst
doctor
‘Albert was a doctor.’
An even more radically limited use of the zero copula is illustrated by Hungarian. Zero copula encoding in this language is restricted not only to the present tense, but also to sentences with third person subjects. In all other cases, use of the full copula vagy is obligatory.
(12) Hungarian (Ginter and Tarnói 1991: 76, 78)
a.
én
tanár
vagyok
1sg
teacher
be.1sg.pres
‘I am a teacher.’
b.
ö
diák
3sg.m
pupil
‘He is a pupil.’
For the purposes of the map, only languages of the English type have been rated as type 1. In other words, a language is considered to be a member of type 2 if it allows the use of a zero copula, however minimal – as in the case of Hungarian – this use may be.
As shown extensively by Stassen (1997: 76-100), not all items that are used as copulas in nominal predication have the same morphosyntactic status. A well-known class of copular items is formed by the verbal copulas, which have (by and large) the same morphosyntactic properties as verbs: English be, Spanish ser and Russian byt’ are illustrations of this class. In addition to this, however, large areas of the world show the use of nonverbal copular items. Particularly prominent is the use of a pro-copula, i.e., a demonstrative or personal pronoun which serves as the linker between subject and predicate nominal, and which is obligatory in nominal predication. We find this option especially in northern and central Asia, in North Africa and the Middle East, and in eastern Indonesia and Melanesia. Hebrew, Motu (Oceanic; New Guinea), Turkish, and Beja (Cushitic; Sudan) are examples of the pro-copula strategy; in Turkish and Beja, the pro-copula takes the form of a suffix.
(13) Hebrew (Li and Thompson 1977: 428)
Moše
hu
student
Moshe
3sg.m
student
‘Moshe is a student.’
(14) Motu (Lister-Turner and Clark 1930: 54)
lau
na
tau
ia
be
hahine
I
dem/cop
man
she
dem/cop
woman
‘I am a man, she is a woman.’
(15) Turkish (Lewis 1967: 98)
Türk-üm
Turk-1sg
‘I am a Turk.’
(16) Beja (Tucker and Bryan 1966: 543)
barú:k
ha’á-bwa
2sg.m
sheik-2sg.m
‘You are a sheik.’
In addition to pronouns, quite a few other nonverbal items are attested in copular function. Such particle copulas have their origin in a variety of markers of discourse-oriented phenomena such as topicalization, backgrounding, or contrastive focus for subjects or predicates. An example of such a particle copula has been given in sentence (4), where the Kabyle copula d’ can be identified as a general conjunctional/contrastive marker; for one thing, it is used as the conjunction ‘and/with’ between noun phrases. Another example is from Awtuw (Lower Sepik-Ramu; Papua New Guinea), where the item po is a general focus marker. Unlike in verbal sentences, the item is obligatory with nominal predicates. A similar situation holds in Vai (Mande; Liberia).
(17) Awtuw (Feldman 1986: 60, 138, 148)
a.
Yen
w-æy-re
2sg
nonfact-go-fut
‘You will go.’
b.
wan
po
‘w-æy-rere
1sg
foc
nonfact-go-desid
‘I am the one who wants to go.’
c.
wan
po
rameyæn
1sg
foc/cop
human being
‘I am a human being.’
(18) Vai (Welmers 1976: 116; Koelle 1854: 92)
a.
mànjáā
mù
ná’à
à
fɛ’ɛ
chief
foc
I
him
see.perf
‘It was the chief whom I saw.’
b.
mō
kóro
mú
nda
man
old
foc/cop
I
‘I am an old man.’
It must be concluded, then, that full copulas do not form a homogeneous morphosyntactic class. For the purposes of this map, however, differentiation between various classes of full copulas has been ignored. As a result, any language that has a full copula has been rated as a member of type 1, regardless of the morphosyntactic status of the copular item (provided, of course, that the zero option is inadmissible in the language).
Regarding the distribution of zero copula encoding across the world’s languages, the first thing to be noted is that there appears to be no major linguistic area in which the use of a zero copula is completely excluded. However, the areal distribution of zero copula encoding is not completely random either. The map shows a number of linguistic areas in which zero copula encoding is highly conspicuous. In particular, a zero-strategy is used almost uniformly in the Pacific region (Australian, Papuan, and eastern Austronesian languages), and it is highly prominent in South America, as well as in the northern part of Africa. Opposed to this, we find areas in which a zero copula constitutes a minor or marginal option. This is the case in Europe and the Middle East (covering Indo-European, Semitic, Turkic, Caucasian, and Uralic languages), northern India and the Himalayas, sub-Saharan Africa, North and Central America, and East Asia (covering China, Korea, Japan, and the Philippines). It should be added that, quite often, the distribution of zero copulas within a language family or area is rather unpredictable. Languages which are closely related areally or genetically may differ considerably in the extent to which they allow zero encoding. For instance, while Austronesian languages typically opt for zero encoding, a full copula is mandatory in a closely related group of three Austronesian languages from northern and central Vanuatu (Ambrym, Big Nambas, and Paamese).