Alexis Dimitriadis

The use of databases in cross-linguistic studies

This book promotes the development of linguistic databases by describing a number of successful database projects, focusing especially on cross-linguistic and typological research. It has become increasingly clear that ready access to knowledge about cross-linguistic variation is of great value to many types of linguistic research. Such a systematic body of data is essential in order to gain a proper understanding of what is truly universal in language and what is determined by specific cultural settings. Moreover, it is increasingly needed as a tool to systematically evaluate contrasting theoretical claims. The book includes a chapter on general problems of using databases to handle language data and chapters on a number of individual projects.

On Clitics, Prepositions and Case Licensing in Standard and Macedonian Greek

Doubling in Romance languages is subject to Kayne’s generalisation, which states that a doubled object must be introduced by a preposition.

The Doubly Marked Reflexive in Chinese

We discuss an unusual reflexive construction in which the Chinese reflexive ziji appears twice, once before the verb and once after. We demonstrate that this is a distinct construct with its own rules of construal and interpretation; it is not, for example, a combination of a simple ziji reflexive and an adverbial intensifier. Notably, their locality properties are also different: Double ziji does not tolerate non-local readings. We argue that while ziji is (or can be) a logophor [1], double ziji is an ordinary Principle A anaphor with all the properties and restrictions that this implies.

Modeling double scroll time series

The ubiquity of strange attractors in nature suggests that nonlinear modeling techniques can improve performance in some signal processing applications. The authors introduce mixed state Markov models (MSMMs), a refinement of hidden filter HMMs, and apply both to a synthetic double scroll time series. Forecasts by HFHMMs diverge after a few steps. Using ad hoc procedures, forecasts by MSMMs, even models generated by crude methods without iterative optimization, can be made more stable. >

Towards a Linguist's Workbench Supporting eScience Methods

The domain of language resources is fragmented in many dimensions. Institutional fragmentation is currently being addressed by Grid projects, which will allow access to resources across institutional boundaries. While technical encoding and structural/format differences constitute significant challenges,, this paper focuses on the problem of the terminological differences encountered when researchers access resources from different projects and creators. We outline two projects that employ a bottom-up approach, and discuss potential extensions towards an eventual Service Oriented Architecture that will bring together all the different components required to overcome the various fragmentation boundaries and open the road to an eHumanities environment.

The typological database of the World Atlas of Language Structures

The World Atlas of Language Structures (often abbreviated as WALS) is primarily a book with 142 world maps showing the global distribution of structural features of language. It was put together by Martin Haspelmath, Matthew S. Dryer, David Gil and Bernard Comrie at the Max Planck Institute for Evolutionary Anthropology between 1999 and 2004, and published by Oxford University Press in July 2005 (Haspelmath et al. 2005). Over forty authors contributed to it, each structural feature (and thus each map) being the responsibility of a single author or team of authors. A sample map is shown in Figure 1.