by Leon Stassen
This map shows the distribution of the variation in the encoding of noun phrase conjunction. The basic distinction is between those languages which use a different marker for noun phrase conjunction and comitative phrases (so-called and-languages; Stassen 2000) and those languages in which the markers for noun phrase conjunction and comitative phrases are the same (with-languages). A clear example of an and-language is English. As the examples in (1) demonstrate, in this language the marker of noun phrase conjunction (and ) is different from the marker of comitative phrases (with). In contrast, the example in (2) shows that Nkore-Kiga (Bantu; Uganda) is a with-language: the item na can be translated as either 'and' or 'with'. An alternative way of stating this is to say that with-languages like Nkore-Kiga lack the structural option of noun phrase conjunction.
(1)
a.
John and Mary went to the movies.
b.
John went to the movies with Mary.
(2) Nkore-Kiga (Taylor 1985: 58)
n-ka-za-yo
na
Mugasho
1sg-rec.pst-go-there
and/with
Mugasho
‘Mugasho and I went there./ I went there with Mugasho.’
Accordingly, the following two values are shown on the map:
| Go to map | ||
|---|---|---|
| Value | Representation | |
 |
AND-languages: 'and' and 'with' are not identical | 131 |
 |
WITH-languages: 'and' and 'with' are identical | 103 |
| Total: | 234 | |
Among the languages which have the option of structural noun phrase conjunction, it is possible to subcategorize conjoined NP-structures on the basis of the linking device which they employ. First of all, there are languages with juxtaposition or zero-marking of such structures. Traditionally, the term asyndeton is employed to refer to such constructions. Quite a few languages can be shown to have this encoding possibility as at least one of their options. However, closer inspection reveals that asyndetic NP-conjunction is nonetheless a minor strategy. Obligatory use of this variant is very rare. This option can be documented in only a few languages, among which no significant areal cohesion can be defined. An example is Awtuw (Sepik; Papua New Guinea).
(3) Awtuw (Feldman 1986: 67)
Yowmen
Yawur
du-k-puy-ey
Yowmen
Yawur
dur-impf-hit-impf
‘Yowmen and Yawur are hitting (someone).’
By far the most prominent option in overt marking of noun phrase conjunction is the use of a medial connective, i.e. a conjunctional item which stands between the two noun phrases. This strategy is illustrated by the English item and (see1a). Noun phrase conjunction via a medial connective can be encountered all over the globe. In some areas, notably sub-Saharan Africa, the strategy is a minor one, but there is hardly any major linguistic area in which it is totally lacking.
In comparison to medial connectives, other overt conjunction strategies are fairly rare. All the strategies at issue feature postposition or suffixation of the connective item or items. Among these options, polysyndeton (i.e. the use of two conjunction markers) appears to be the most popular. This strategy can be found in a number of unconnected linguistic areas such as the Caucasus, northeastern Africa, Australia and New Guinea, southern India and northeastern Asia. In addition, we find isolated examples in the Americas, in West Africa, and in Burma. An example of the strategy is Tamil (Dravidian; southern India):
(4) Tamil (Asher 1982: 69)
akkaa-yum
tangkacci-yum
elder.sister-and
younger.sister-and
‘elder sister and younger sister’
Monosyndetic postposing of conjunction markers (whether to the first or to the second conjunct) is low in frequency. Moreover, languages which present this option typically also allow a construction of the polysyndetic type, so that these monosyndetic constructions are best regarded as variants in which one of the markers in the polysyndetic construction is optionally deleted. Given this, it will be clear that the two monosyndetic variants occur in roughly the same areas as the polysyndetic construction. An example of the “first-NP” subtype is Beja (Cushitic; Sudan), while the “second-NP” subtype is illustrated by NP-conjunction from Tubu (Nilo-Saharan; Chad and Niger).
(5) Beja (Reinisch 1893: 195)
a.
ani-wa
baruk-wa
1sg-and
2sg-and
‘you and I’
b.
mek-wa
laga
donkey-and
calf
‘a donkey and a calf’
(6) Tubu (Lukas 1953: 166)
a.
turku
ye
molofur
ye
jackal
and
hyena
and
‘the jackal and the hyena’
b.
wuden
arko
ye
antelope
goat
and
‘the antelope and the goat’
In the map, variation in the encoding of noun phrase conjunction has not been taken into account.
Although “dependent marking” (Nichols 1986) is the predominant encoding option in comitative constructions, there are some languages which deviate from this pattern. A “head-marking” strategy, in which the comitative marker is incorporated into the predicate, is encountered in Northwest Caucasian languages such as Ubykh.
(7) Ubykh (Dumézil 1931: 17)
go-u-ji-k'ä-qa
3sg-2sg-with-come-perf
'He came with you.'
(8) Selknam (Tonelli 1926: 58)
Kokoš
telken
okel-enen
Kokoš
boys
with-go
‘Kokoš goes with the boys.’
(9) Igbo (Welmers 1973: 369)
ha
sò
anyi
ga-á
Aba
they
accompany.stat
us
go.to-narr
Aba
‘They went to Aba with us.’
Again, this variation in the encoding of comitative phrases is not reflected on the map.
Although both for with-languages and for and-languages clear instances can be found, the typological status of these two types is probably not equally well established. In particular, there is a notable discrepancy in the stability of the types. In general, and-languages can be said to be stable diachronically and “pure” in their synchronic state: there is a sharp delineation between the two available strategies. On the other hand, “pure” instances of with-languages are relatively rare. For a considerable number of such languages, some process of “diachronic drift” of the comitative encoding can be attested. The general outcome of this process is to effectuate a shift from one-strategy to two-strategy status. These languages tend to introduce a secondary differentiation into the structural features of the comitative strategy, by changing one or more features of that strategy towards the features of the conjunctional strategy. In this way, the language acquires a two-strategy encoding of the domain, in which one of the strategies is still purely comitative, while the other is some sort of hybrid between the comitative and the conjunctional strategy.
The “shift” of the comitative strategy in with-languages is a gradual process. Moreover, the structural features which are changed in the course of this process are different for different groups of with-languages. Among the possible “paths” involved in the shift, we encounter cases in which
(i) The comitative NP can be moved from its canonical position to a position adjacent to the other conjunct, so that the two conjuncts now form a constituent. An example is Babungo (Bantoid; Cameroon).
(10) Babungo (Schaub 1985: 87)
a.
Làmbí
gè
táa
yìwìng
ghó
Ndùlá
Lambi
go.pfv
to
market
with
Ndula
‘Lambi went to the market with Ndula.’
b.
Làmbì
ghó
Ndùlá
gè
táa
yìwìng
Lambi
and/with
Ndula
go.pfv
to
market
‘Lambi and Ndula went to the market.’
(ii) There is differentiation in person/number agreement (mainly between singular and dual/plural). An example is Tolai (Oceanic; New Britain, Papua New Guinea).
(11) Tolai (Mosel 1984: 176)
a.
nam
ra
tutana
i
ga
rovoi
ma
ra
pap
dem
art
man
3sg
pfv
hunt
with
art
dog
‘That man went hunting with his dog.’
b.
Terengai
dir
rovoi
ma
ra
pap
Terengai
3du
hunt
and/with
art
dog
‘Terengai and his dog went hunting.’
(iii) There is doubling of the comitative marker when used as a conjunction, thereby overtly indicating the equality in rank of the two noun phrases. A case in point is Japanese.
(12) Japanese (Hinds 1986: 97, 94)
a.
Taroo
wa
Akiko
to
Nara
e
ikimashita
Taroo
top
Akiko
with
Nara
to
go.pst
‘Taroo went to Nara with Akiko.’
b.
Taroo
to
Akiko
to
wa
Nara
e
ikimashita
Taroo
with
Akiko
with
top
Nara
to
go.pst
‘Taroo and Akiko went to Nara.’
Since, however, in all these “shift”-languages the markers for coordination and comitatives remain identical, such languages have been coded as with-languages for the purposes of this map.
As the map demonstrates, the distinction between the two language types shows clear areal features. Concentrations of with-languages can be found in two large linguistic areas. First, we find such languages in East and Southeast Asia and the islands of Indonesia, Melanesia and Polynesia. Secondly, most of the languages of Africa in and below the Sahara (with the notable exception of Khoisan) have with-status.
Opposed to this, and-languages appear to be the rule in a macro-area which might roughly be called “Eurasia”. This area comprises all the languages of Europe, Central Asia and Siberia; in the east, it also includes some, though not all, of the Austronesian languages of the Philippines. To the southeast, the area stretches as far as India. Finally, it includes the Middle East and northern Africa. A second macro-area of and-languages is formed by (most of) the languages of Australia and the central highlands of New Guinea.
Outside of these two main and-areas, and-encoding can be found in several other places, in particular in the Americas, but the distribution of and-languages and with-languages seems to be rather haphazard here. North America seems to have a predominance of and-status, whereas at least the languages of the southern part of Central America and the eastern part of South America appear to tend towards with-encoding.