I am working on a book with this title, to be published in 2009. This book summarizes my research in reference done in the previous years. I rely on the notion of referential choice – the choice made by a speaker when s/he needs to mention a referent at a certain point in discourse. This is, for example, a choice between a full NP and an anaphoric pronoun. I am looking into the cognitive underpinnings of referential choice, linking the problem of referential choice to the cognitive domain known as working memory. I propose that a set of linguistic factors, taken in conjunction, determine a referent’s current level of activation in working memory, and this is what explains referential choice. When activation exceeds a certain thresholds, a reduced referential device may be used, and when activation is below such threshold, a full referential device is used; see Kibrik 2000 [pdf], Gruening and Kibrik 2005 [pdf]. I also provide a cross-linguistic overview of referential devices, differentiating between referential devices as such and referential aids. Referential devices are divided into full and reduced devices, and the latter fall into free pronouns, bound pronouns, and zeroes. Earlier representations of these ideas can be found in Kibrik 2001 [pdf].
One of the main stumbling blocks in the studies of reference is the multiplicity of factors potentially affecting the choice. I have developed an approach that addresses this problem. Multi-factorial models of referential choice for Russian and English narrative discourse have been proposed in Kibrik 2000 [pdf]. In each case, a set of activation factors (around ten) is postulated, each factor contributing to a referent’s current aggregate activation. Numerical weights were attributed to each value of each activation factor, so the aggregate activation can be estimated numerically as a sum of all relevant weights. Mathematical simplicity of this approach was somewhat amended in the study (Gruening and Kibrik 2005 [pdf]), based on the apparatus of artificial neural networks. Neural network is a self-training program that can find an optimal correspondence between a set of independent variables and a dependent variable (in this case, referential choice). A subset of activation factors were found to be particularly important for referential choice. A problem with all of these studies was a small dataset on which they were based. Now development of a large referential corpus (containing thousands of referential expressions) is underway. Each entry of a referential devices is annotated for many potential factors, and once this work is completed a large-scale statistic modelling will be undertaken. In this project I collaborate with Christian Chiarcos and Olga Krasavina.
This is a joint project of Vera I. Podlesskaya [link] and myself, together with our junior colleagues. We study a corpus of spoken Russian stories in which children and adolescents retell their night dreams. Originally we were particularly interested in the semantics of narrative discourse and in difference between stories by normal children and children with neuroses (Kibrik, Podlesskaya, Kal’kova, and Litvinenko 2002 [link]). Later we came to be focused on more general problems of understanding spoken discourse, including speaker’s planning of speech and self-monitoring, discourse segmentation, types of elementary discourse units, prosody, etc. We are currently developing the first systematic Russian discourse trascription (Kibrik and Podlesskaya 2003 [pdf]). Viewed from the angle of spoken language, Russian grammar looks very differently from the traditional account. Also, I believe that a serious understanding of language must involve knowledge of discourse prosody. A volume summarizing this work is is to be published in 2009. This volume includes an account of semantic, grammatical and prosodic phenomena found in the corpus, as well as the corpus itself.
Beginning from 1997, I have been studying Upper Kuskokwim, an Athabaskan language of interior Alaska. Uper Kuskokwim Athabaskan (UKA) is a highly endangered language, as the population speaking it numbers just a couple of dozens, and all fully-fledged speakers are older than 50. Most speakers of UKA reside in the village of Nikolai, Alaska. I have done fieldwork on UKA for two extended periods (four months each) staying in Nikolai with my family in 1997 and 2001. For an account of our experiences in Nikolai click here. I am working on an integral dsecription of the structure of UKA. For a preliminary sketch see Kibrik 1998 [pdf1], [pdf2]. This work is challenging for a number of reasons. First, Athabaskan languages are very different from a “typological average” in being almost exclusively prefixal, highly polysynthetic, and possessing very unusual morpheme ordering and intricate morphophonemics. Second, the community is not easily accessible in terms of distance, logistics, and cost. But I am planning to complete a grammar of UKA within the next few years.
This is the Russian Foundation for the Humanities [link] project (2008-2010) that I am implementing together with a group of colleagues. We are interested in the relationship between verbal, prosodic, and visual aspects of communication. Most linguists traditionally assume that language means a string of words, and words consist of phonemes. So the essence of linguistic form is thought to be of a phonemic nature. However, it is well known that there are other aspects of form that participate in message encoding, specifically prosody and body language. In some psychological studies it has been claimed that prosody and especially body language convey a lot more information than the verbal component; the contribution of the latter is sometimes estimated as low as 7% of overall information transmitted. In this line of research, an attempt is undertaken to approach this question empirically and see which of these polar views is correct. Preliminary results (see Kibrik and El’bert 2007 [ppt]) suggest that both are wrong. All three information channels are important. The verbal channel is the most informative but is far from being pre-eminent. Within the framework of this project, other issues will be addressed, such as discourse segmentation.
I have been teaching Discourse analysis in the Department of theoretical and applied linguistics [link], Moscow State University, since 1995; look here. Gradually I am composing a textbook in discourse analysis that would combine major results in this field with my own approach. According to this approach, three main issues are addressed in discourse analysis: taxonomy of classification of discourse types; discourse structure – both global and local; and influence of discourse factors upon smaller linguistic constituents (syntactic, morphological, and phonetic). History of discourse analysis, relationships with neighboring fields and used methodologies are also addressed in the textbook.
This is another course I am teaching in the Department of theoretical and applied linguistics [link]. The science of linguistic diversity consists of three main traditional disciplines: genealogical (comparative-historical) linguistics, areal linguistics, and linguistic typology. This course presents a mixture of all three, and also includes elements of sociolinguistics and other related sciences, such as anthropology, archaeology, genetics, etc. I provide an overview of the world’s languages, going from one macroarea to another. There are ten macroareas altogether. I am in the process of writing up a textbook on this topic; see Kibrik 2002 [pdf] for a preliminary prospectus.
This is my main duty in the Institute of Linguistics – I coordinate this longitudinal project that started back in the 1970s. I have co-edited a number of volumes of the encyclopedia: on Paleoasiatic, Turkic, Mogolic, old Indo-Aryan, Slavic and Baltic languages. Currently our working group is preparing up to ten further volumes focusing on various languages of the Old World, including Semitic, modern Indo-Aryan, Dravidian, Mande, relic languages of Europe and Asia, and Austroasiatic. To know more about the encyclopedia, click here.