Following study of small-inventory classifier systems in a number of Indo-European, Turkic, Kartvelian and Semitic languages of the Araxes-Iran Linguistic Area, the paper presents an account of numeral classifiers in Udi, a Nakh-Daghestanian (Lezgic) language spoken in northern Azerbaijan. Being a peripheral member of the linguistic area in question, Udi possesses an even more reduced version of a small-classifier system, comprising one optional classifier dänä (Iranian borrowing, most likely via Azerbaijani) used with both human and inanimate nouns. A dedicated classifier for humans is lacking, although there is a word tan (also of Iranian origin) only used after numerals or quantifiers, but predominantly as a noun phrase head. The behaviour of dänä and tan is scrutinized, according to a set of parameters, in both spoken and written textual corpora of the Nizh dialect of Udi. Drawing in the data from the related Nakh-Daghestanian languages, the paper shows that among the languages of the family Udi may be unique in possessing classifiers (albeit as a result of contact), Khinalug possibly being the only other exception.

在研究了阿拉克斯-伊朗语言区的许多印欧语、突厥语、卡特维利亚语和闪米特语的小库存分类器系统后，本文介绍了乌迪语中的数字分类器，乌迪语是一种 Nakh-Daghestan (Lezgic)在阿塞拜疆北部。作为相关语言领域的外围成员，Udi 拥有一个更简化版本的小分类器系统，包括一个可选的分类器dänä（伊朗借用，很可能通过阿塞拜疆语借用），用于人类和无生命的名词。虽然有一个词tan（也起源于伊朗）只在数字或量词之后使用，但主要用作名词短语头部，但缺乏用于人类的专用分类器。dänä的行为和根据一组参数，在乌迪语 Nizh 方言的口语和书面语料库中仔细检查tan。从相关的纳赫-达吉斯坦语中提取数据，该论文表明，在乌迪语族的语言中，在拥有分类词方面可能是独一无二的（尽管是接触的结果），Khinalug 可能是唯一的另一个例外。