In Batsbi, a language of the Nakh group of the Nakh-Daghestanian language family, markers of gender-number agreement can occur simultaneously in several positions in a single verb form. Repetition of markers in this way has been called extended exponence. Until recently, it was widely believed that inflection, such as the gender-number agreement marker, was necessarily restricted to the edges of words, and that no language had repeated marking of agreement with a particular feature or set of features. However, such repetition is found in several of the languages of the Nakh-Daghestanian family, as well as in several in the Kiranti subgroup of Tibeto-Burman, and in other languages. The repetition of gender- number encoding raises questions of typological distribution (for example, do such patterns occur only in languages of certain types?), diachronic origin (how and why do such patterns originate?), and linguistic theory (how can we account for systems of this type?). The goals of this research include (a) characterization of variation possible in the pattern cross-linguistically, (b) a statement of constraints on the pattern, if any, and (c) a description of the origins of these patterns in the languages studied. One impact of this project will be in enhancing the status of the endangered languages to be studied, thus encouraging speakers to preserve them. The project provides support for linguistics and other academic fields in the Republic of Azerbaijan and in Daghestan, Russia. Collecting texts in Archi and Khinalug will lay the foundation for editing these texts in the future as a contribution to the documentation of these endangered languages

Project Report

Multiple exponence is the indication of any semantic feature such as [past] more than one time in a single word. There are no clear examples of this in English, but we may consider the difference between feel and felt. The past tense in felt is indicated both by the difference of the vowel and by the presence of the sound /t/ at the end (but see below on the cognitive status of this example). Cross-linguistically this topic is important for linguists to understand, since multiple exponence is counter to the principles of economy (the principle that language should require the least possible effort) and iconicity (the principle that conceptual complexity corresponds to complexity of form). Surveying many different languages for multiple exponence, we have found that all imaginable types exist, including multiple exponence signaling a new word vs. a form of an existing word, being optional or obligatory, and whether the markers are identical or not. We also found that there are few constraints on the occurrence of multiple exponence. We found that while lexically governed multiple exponence (feel ~ felt pairs only exist for a handful of verbs in English) seems to be rather common in the languages surveyed, systematic exponence (affecting all items in some grammatically determined environment) is less common and perhaps more interesting. Multiple exponence originates in several ways; among the ways that are best understood are grammaticalization and reinforcement. In grammaticalization, when two words that indicate the same features join to become a single word, in some instances both sets of feature indicators are preserved. For example, it is common for an auxiliary over time to become part of a main verb. If the two words both indicate, for example, the person and number of the subject, when they join as a single word, both indications of one subject may be preserved. (This seldom occurs in European languages, but in some other languages it is common for both main verb and auxiliary to indicate the subject.) In reinforcement, when a word part that indicates a category has become irregular or infrequent it is likely to be reinforced (added to) by a word part that is more regular or frequent, productive. Experiments show that multiple exponence does not function to improve word recognition, ease of judgment of grammaticality, or memory for verb forms; but, like any redundancy, it does make it more likely that a listener will discern the category that is doubly indicated. It is possible that this is responsible for the continued existence of multiple exponence in some languages. Experimental evidence suggests that forms like felt are processed as a whole, not broken down into component parts related to feel. In this sense we might argue that the pair feel ~ felt does not exemplify multiple exponence from a cognitive point of view. Some children go through a stage of creating multiple exponence in places where it does not occur in adult language. Thus in the language of some children we find redundant forms such as feets, gotted, and burneded. More work is needed on the acquisition of systematic multiple exponence and of multiple exponence in languages other than English. Having studied a wide range of examples of multiple exponence in many languages around the world, we are able to explain why this phenomenon is rare and, on the other hand, why languages do not eliminate it altogether. In terms of the broader impact, we (a) provided support for linguistics and other academic fields in Azerbaijan and Daghestan (Makhachkala), (b) enhanced the status of these small languages, thus encouraging speakers to preserve them, and (c) educated the graduate assistants in learning to conduct fieldwork and other research and coauthor papers. The impact of visiting scientists on a society should not be underestimated. The minority who speak the language being investigated recognize, often for the first time, that their language has value. This can help to raise the awareness of the importance of a language for both humanitarian and scientific reasons throughout the community. Awareness can provide an incentive for speakers to retain the language.

National Science Foundation (NSF)
Division of Behavioral and Cognitive Sciences (BCS)
Application #
Program Officer
Joan Maling
Project Start
Project End
Budget Start
Budget End
Support Year
Fiscal Year
Total Cost
Indirect Cost
University of Massachusetts Amherst
United States
Zip Code