Referentiality: Description

Brief project description

The Referentiality Project is a project of ‘type b’ in the DoBeS program, which does not document a language, but uses The Language Archive (TLA) for comparative research. It aims at contributing to the theory and typology of referentiality by exploiting documentations of endangered languages deposited in TLA. It will thereby

  1. enrich our understanding of referential operations in communication and of linguistic structures serving them
  2. demonstrate the significance of hitherto essentially unknown languages for empirical and theoretical linguistic work
  3. establish a structured, multi-language data-base of annotated texts and develop the tools and protocols for cross-corpus analysis, which will provide a resource for future research
  4. demonstrate that DoBeS documentations can fulfill one of the major purposes that the archive was set up for, viz. exploitation for corpus-based work in comparative linguistics.

Referentiality is a property of a linguistic sign that functions in a discourse. Linguistic signs in question are, in general, nominal expressions, including their reduced forms that result from pronominalization. The referentiality of a sign concerns the status of its designatum on two dimensions, individuation and discourse status:

  • Individuation is the operation of identifying a particular set of individuals on the basis of a concept. The entity designated by the linguistic sign may be individuated to different degrees.
  • Discourse status is the degree to which a referent is in the foreground of the awareness of the interlocutors.

The two areas of grammatical structure most commonly associated with this functional domain are determination and pronominalization. However, it is already known that referentiality may be coded by strategies outside both of these areas; and it is therefore assumed that there are more structural means (outside the nominal domain) to be found in other languages. The onomasiological approach starts from a characterization of the cognitive and communicative functions involved in referentiality and finds out whether and how these are coded in grammatical structure.

At the same time, the current theoretical account of referentiality will be enriched and modified by concepts found to be coded in the texts and the grammar of languages. The semasiological approach starts from an analysis of texts and asks which functions are fulfilled by the structural means identified. It therefore has the potential to discover new functions of language. The two approaches complement and cross-fertilize each other.

The referential status of an entity changes as it is introduced in the universe of discourse, attended to immediately and possibly resumed some time later. It thus develops in discourse and may therefore be empirically researched only by studying texts or even audio and video recordings.

The Referentiality Project primarily falls into the novel field of text-based typology. It takes its data from a selected subset of DoBeS corpora, contributed by 11 subprojects. The group of contributors consists of language documenters and typologists active in the DoBeS program and in a position to grant access rights to data in their projects. These corpus managers share a common goal of extending and enriching the potential of their own project through cooperation with other projects and agree on the main goals, principles and methods of typological research outlined in the Referentiality Project. Over several months of interaction and negotiation, culminating in a workshop held last October, a set of minimal requirements for the individual corpora were established regarding their accessibility, size and composition, representation of text at certain linguistic levels and availability of descriptive information.

The corpus managers will provide language-specific referentiality functions that emerge from a close analysis of their corpus. A team of typologists will investigate hypotheses like the following:

  • Every language has at least one strategy coding individuation and one strategy coding discourse status.
  • The structural means for reference-tracking are the simpler and the more grammaticalized, the more highly individuated the referent is.

In doing so, they will apply methods of database search, using query languages which allow searches across corpora and representation tiers, and methods of corpus linguistics, involving statistical analyses apt for the treatment of small corpora.

In sum, the Referentiality Project pursues a scientific goal which has become attainable only through the recent activity of language documentation. Cross-corpus linguistic research, using natural spoken language data, has never been undertaken at this scale before. In bringing together a significant subset of the DoBeS projects onto a common platform, it will prove that these documentations are amenable to comparative research, thus fulfilling certain expectations pinned on them.

Click here for the full project proposal as .pdf file.

Upcoming events

Meeting of the RP committee -
October 11th / 12th


A new project (Wichi) has joined the RP(corpora)

Articles added
"Type shifts and noun classes under determination in Teop"
(link section)

Presentation links added (link section)

A new project (Mapudungun) has joined the project (corpora)

Articles added
"1st and 2nd mention of referents Jan 2013"
and "Typology of ZERO"
(link section)