Andrew Gordon Wilson

Rethinking Russia's post‐soviet Diaspora: The potential for political mobilisation in eastern Ukraine and north‐east Estonia

ONE OF THE MAJOR RESEARCH FOCI to emerge within post-Soviet studies has been the issue of the Russian diaspora.1 While work to date has drawn attention to the possibilities for local and regional destabilisation resulting from the presence of the 25 million Russians living in the 14 borderland states of the former USSR,2 it has shied away from developing either a theory to account for the relative passivity of the diaspora since 1991 or a conceptual framework that might aid comparative study. Our aim in this article is therefore to broaden conceptual horizons and offer some theoretical perspectives that might provide new insights into our understanding of these communities.

Multimodal Word Distributions.

Word embeddings provide point representations of words containing useful semantic information. We introduce multimodal word distributions formed from Gaussian mixtures, for multiple word meanings, entailment, and rich uncertainty information. To learn these distributions, we propose an energy-based max-margin objective. We show that the resulting approach captures uniquely expressive semantic information, and outperforms alternatives, such as word2vec skip-grams, and Gaussian embeddings, on benchmark datasets such as word similarity and entailment.

Parties and presidents in UKraine and Crimea, 1994

Several sets of elections in Ukraine in 1994 may have had the effect of turning that year into a watershed in the development of democracy in independent Ukraine. In the spring parliamentary elections the results were not markedly different from what they had been in the last elections to the Supreme Soviet. In the presidential elections, held in June and July, the incumbent was voted out on the basis of a poll marked by polarization along regional lines. In the election to the Crimean presidency, in January, the ethno‐linguistic divisions of Ukrainian society were even more acutely in evidence. While observance of procedures and acceptance of the results represents an important stage in consolidating democracy, the pattern of voting reveals sharp divisions that will require serious effort to overcome.

Nation-building in the Post-Soviet Borderlands: National history and national identity in Ukraine and Belarus

This chapter seeks to examine the relationship between historiography and the nation as an ‘imagined community’ in Russias two east Slavic neighbours, Ukraine and Belarus. The focus of the analysis is on the mythic structures of national historiography as a key influence shaping evolving national identities, and on the tug on identities exercised by rival narratives of the past, here classified for conveniences sake as ‘Ukrainophile’/‘Belarusophile’ and ‘Russophile’ or ‘pan-Slavic’. The former have come to the fore since independence in 1991, but Russophile myths have proved powerful and persistent, particularly in Belarus but also amongst the half of the population of Ukraine that is either ethnic Russian or Russian-speaking (see chapter 6). The historical, or historiographical, component of national identities in the region is therefore in transition. The single narrative of the Soviet era has given way not to monolithic new national alternatives, but to a fluid situation characterised by competing myths and dissonant voices. Whereas a potentially strong, if controversial, historiographical mythology is under construction in Ukraine, a key reason for the relative weakness of the Belarusian national movement to date has been its inability to displace hegemonic Russophile myths and anchor a new Belarusian identity firmly in a rival historiography. Russophile historiography, on the other hand, has so far failed to address seriously the fact of Ukrainian and Belarusian independence, and has remained content to recycle the myths of the tsarist and Soviet eras.

Nation-building in the Post-Soviet Borderlands: Language policy and ethnic relations in Uzbekistan

As elsewhere in the Soviet borderland states, an important watershed in the nation-building process in Uzbekistan was the adoption in October 1989 of the law ‘On the State Language’, which granted Uzbek the status of the sole state language within the Uzbek Soviet Socialist Republic. Lying at the heart of the new language politics were issues of power and status rather than communication, for, as Donald Horowitz has pointed out, language is a potent symbol of both new-found group dignity and status. Although the new law made Russian the ‘language of inter-ethnic communication’, it also required employees in the state sector as well as those serving the population to command enough Uzbek for the fulfilment of job responsibilities. Owing to material and organisational constraints, however, the pace of implementation of language legislation inevitably slowed. In December 1995, more than six years after the passage of the original legislation, a revised version of the Law on the State Language was adopted. The revised edition no longer made knowledge of Uzbek compulsory for public sector employees, yet it also abolished Russians special status, putting that medium on a par with all other ‘foreign’ languages. This chapter examines language policy in Uzbekistan and assesses how legislation has reconstituted ethnic relations between the titular group and key non-titular minorities. The first part outlines the general evolution of language policy in Uzbekistan since 1989, comparing the significant ways in which the first edition of the language law diverges from the revised edition.

Modelling input varying correlations between multiple responses

We introduced a generalised Wishart process (GWP) for modelling input dependent covariance matrices Σ(x), allowing one to model input varying correlations and uncertainties between multiple response variables. The GWP can naturally scale to thousands of response variables, as opposed to competing multivariate volatility models which are typically intractable for greater than 5 response variables. The GWP can also naturally capture a rich class of covariance dynamics --- periodicity, Brownian motion, smoothness, …--- through a covariance kernel.