Chapter 34: Adnominal distributive numerals
1. Feature description ⇫ ¶
Distributive numerals are special adnominal numerals that express distributive relations, as in German je drei in the sentence Die Männer trugen je drei Koffer. ‘The men carried three suitcases each.’ English lacks distributive numerals, because in a sentence like They carried three suitcases each, the numeral three does not form a continuous expression with the distributive word each, i.e. three...each does not qualify as a numeral.
Since the dividing line between distributive numerals and quantifiers is not always clear-cut, especially regarding ‘one’, examples with reduplicated adnominal numerals meaning ‘one by one’ or ‘every’ are also taken into account (see example 8).
This feature corresponds partially to WALS feature 54 (Gil 2005a). For reduplication in pidgins, creoles, and other contact languages, see also Kouwenberg (2003).
2. The values ⇫ ¶
We distinguish the following two values:
| excl | shrd | all | ||
|---|---|---|---|---|
 | Distributivity marked by reduplication | 20 | 0 | 20 |
 | No special adjacent distributive numerals | 50 | 0 | 50 |
| Representation: | 70 | |||
Value 1 (distributivity marked by reduplication) occurs in six Portuguese-based languages, in ten English-based languages, in Mauritian Creole, in Sango, in Fanakalo, and in Ambon Malay. The reduplication is generally total, but partial reduplication is found occasionally. The following examples illustrate partial reduplication:
In the following examples, the reduplication of the numeral is total:
Value 2 (no special adjacent distributive numeral) is the most widespread value. Several constructions are possible in order to express distributivity. In the Cape Verdean Creole of Brava, the numeral is reduplicated, but it is not adjacent to the noun:
In other languages, expressions like ‘everyone’ or ‘every man’ are used:
3. Distribution ⇫ ¶
Reduplication of the numeral for distributive functions (value 1) is totally absent from Spanish-based languages, Dutch-based languages, bilingual mixed languages, and also from most French-based languages (the only exception being Mauritian Creole); regarding the English-based languages, it is absent in 17 out of 27 languages.
Distributivity expressed by reduplication occurs in somewhat less than a third of our languages. It is present in the Atlantic as well as in Asia, from the Indian Ocean to the Pacific.
4. Comparison with WALS ⇫ ¶
Comparing this feature with the corresponding WALS feature (Gil 2005a), we can observe the following:
(1) The WALS languages show many more different types (affixation, preceding word, following word, other strategies) than the APiCS languages.
(2) Reduplication is the most widespread value in the WALS languages (84 out of 250 languages, i.e. roughly one third), whereas absence of adnominal distributive numerals is found in roughly one fourth of the languages. In the APiCS languages, reduplication is the only strategy found, with more or less the same percentage as in WALS; absence of adnominal distributive numerals is found in more than two thirds of the languages.
(3) In the WALS languages, reduplication is the predominant strategy in sub-Saharan Africa, in South Asia, and, to a lesser extent, in insular Southeast Asia and the Pacific, as well as in other areas. These areas correspond more or less to the sub- or adstrate languages of those APiCS languages which exhibit reduplication for this feature.
On the one hand, we can reasonably assume that the reduplication strategy in European-based APiCS languages is due to their substrate or adstrate languages. On the other hand, the high percentage of absence of adnominal distributive numerals seems to derive from the fact that the European superstrate languages Portuguese, Spanish, English, and French lack such numerals.