See data and maps.
Nasledskova, P. and I. Netkachev (2021). “Ordinal numerals”. In: Typological Atlas of the Languages of Daghestan (TALD). Ed. by M. Daniel, K. Filatov, T. Maisak, G. Moroz, T. Mukhin, C. Naccarato and S. Verhees. Moscow: Linguistic Convergence Laboratory, NRU HSE. DOI: 10.5281/zenodo.6807070. http://lingconlab.ru/dagatlas.
@incollection{nasledskova2021,
title = {Ordinal numerals},
author = {Polina Nasledskova and Ivan Netkachev},
year = {2021},
editor = {Michael Daniel and Konstantin Filatov and Timur Maisak and George Moroz and Timofey Mukhin and Chiara Naccarato and Samira Verhees},
publisher = {Linguistic Convergence Laboratory, NRU HSE},
address = {Moscow},
booktitle = {Typological Atlas of the Languages of Daghestan (TALD)},
url = {http://lingconlab.ru/dagatlas},
doi = {10.5281/zenodo.6807070},
}Many East Caucasian languages express ordinal numerals by means of a construction with the participle of the verb ‘say’ or by means of an affix that is diachronically derived from ‘say’.
| q’a-no | āƛi-ru |
| two-dir | say-ptcp.pst |
| ‘second’ |
In Tsez (1), the ordinal meaning is expressed by the combination of a cardinal numeral with the past participle of the verb ‘say’.
| jaq’ud-pu | gada |
| four-ord | boy |
| ‘fourth boy’ |
| upu | gaf |
| say.pfv.ptcp | word |
| ‘said word’ |
In Huppuq’ Agul (2), the ordinal meaning is expressed by a dedicated ordinal suffix. Diachronically, this suffix is related to the verb ‘say’, namely it is a truncated form of the participle of ‘say’, compare (2) and (3). There are also languages which express ordinal meaning by markers that are not related to the verb ‘say’, either synchronically or diachronically.
Cross-linguistically this way of expressing ordinal meaning is rather rare, although it is not unique. Yine (Arawakan) uses a nominalized form of ‘say’ to express ordinal meaning (Hanson 2010: 74). Paraguayan Guaraní (Tupian) and Kannada (Dravidian) use dedicated markers, which, according to some scholars, were derived from verbs meaning ‘say’ (Dmitry Gerasmov, p.c. and Krishnamurti (2003: 266), respectively). East Caucasian languages are the only language family attested so far where forms of the verb ‘say’ are a widespread way of expressing ordinal meaning (Nasledskova 2018).
Our goal is to classify the languages of Daghestan according to the presence of the ‘say’ strategy of forming ordinal numerals.
The languages were divided into three groups: (i) those that have ordinals derived from ‘say’; (ii) those that have ordinals derived from ‘say’ but with a phonetic erosion of the form of the participle; (iii) and those where ordinals derived from ‘say’ are not attested.
The third group can be divided into sub-groups. Some languages have an ordinal affix borrowed from a neighboring language: Udi, for instance, has borrowed the ordinal marker from Azerbaijani (Schulze 2005: 281). Another language, Budukh, has borrowed a full set of ordinals, also from Azerbaijani (Gilles Authier, p.c.). Some Tsezic languages form ordinals by means of the genitive suffix: this is the case in Hinuq (Forker 2013: 402–403); Hinuq also has the ‘say’ strategy, which means that it has two ways of forming ordinals. Some other languages (e.g. Botlikh) use a participle suffix to form ordinals. Finally, it is not uncommon for languages of Daghestan to use a dedicated suffix with ordinal meaning.
Hinuq uses the infinitive form of ‘say’ instead of the participle. However, in this language the infinitive can be used attributively, so functionally the infinitive does not differ from participles in this case.
In our database there are two columns that contain information about ordinal markers in the languages of the sample: value and other functions of ordinal marker. The column value only marks the presence or absence of the ‘say’ strategy in a language. The column other functions of ordinal marker provides additional information, such as the exact form of ‘say’ that is used to express ordinal meaning in languages with the ‘say’ strategy and other functions of the ordinal marker in languages that do not employ the ‘say’ strategy. Also, if there is more than one way of expressing ordinal meaning in a language, this information is also shown in the other functions of ordinal marker column. If an ordinal marker has no other functions except for expressing ordinal meaning, it is marked as ‘dedicated’ in this column.
Examples of the ‘say’ strategy have been provided above (see (1) for an example from Tsez and (2) for an example from Agul).
Tsova-Tush lacks the ‘say’ strategy: as the example in (4) shows, it employs a specialized ordinal marker that is of an unknown origin.
| ši-lǧen |
| two-ord |
| ‘second’ |
The majority of the languages of the sample express ordinal meaning by means of the verb ‘say’: there are 18 languages that use a construction with ‘say’ and 5 more languages employ ordinal affixes diachronically derived from ‘say’. In 18 languages of the sample the ‘say’ strategy is not attested. As for the other functions of ordinal markers in the languages, the distribution of types are as follows: 23 languages use various forms of ‘say’ to express ordinal meaning. 17 languages have a synchronically dedicated ordinal marker. 3 languages use a genitive marker to express ordinal meaning. 2 languages use ordinal numerals borrowed from Azerbaijani. Finally, one language uses the participle of ‘become’ to express ordinal meaning. The sum of the languages in this paragraph is larger than the number of languages in the sample, because 5 languages employ more than one strategy of expressing ordinal meaning.
The distribution of the ‘say’ strategy clearly demonstrates an areal signal: the areas in which the ‘say’ strategy is absent lie in the areas contiguous with non-East Caucasian languages. The strategy is absent in the north of the area, i.e. in Nakh and most Andic languages; it is also absent in the East (but not West) Tsezic languages in the west, that are contiguous with Georgian. Khinalug, which is isolated from the rest of the family, also lacks the ‘say’ strategy. In the south of the area, the greater level of grammaticalization (phonetic erosion) is attested in the languages that are located the most closely to Azerbaijani.
It appears that genealogical relations between the East Caucasian languages play a role in the distribution of the ‘say’ strategy as well. All Lezgic languages (with the exception of Udi) have the ‘say’ strategy; this is also the case for Dargwa and Tsezic languages (with the exception of Hunzib). By contrast, Nakh languages lack the ‘say’ strategy altogether; this is also true of Andic languages (except Akhvakh).