Apart from the traditional CG field of corpus-oriented tagging and parsing, there is a growing body of applicational work, where CG provides the backbone of end user-oriented systems in various areas of language technology, such as spell and grammar checking, comma correction, ICALL, machine translation, lexicography and others. We therefore invited workshop contributions both regarding basic grammatical research and corpus linguistics on the one hand, and CG-based applications on the other hand. CG has always elicited a strong interest from researchers working on less-resourced languages, and we therefore expected papers targeting minor languages, such as the Sami languages, Greenlandic, Faroese, Tibetan and the Celtic languages.
However, contrary to expectations and in stark contrast to the 2015 workshop, neither minority languages nor applications, but another target area - methodological research - turned out to completely dominate this year’s submissions. Thus, papers explored topics like automatic rule creation and rule optimization as well as what one might call CG typology - the expressivity and power of the formalism as such. Another methodological issue was cross-platform compatibility for CG dependency corpora. Where these papers did touch on language, examples were drawn - with the notable exception of Basque - from only a few major languages, English, Spanish and Portuguese. Given the de-facto wide language-spread of ongoing CG work, this underrepresentation of languages represents a challenge to future CFP’s. In the same vein, it is also an interesting question, whether theoretical issues are more paper-motivating simply because they involve 100% research, while researchers with application-oriented work like machine translation will have to choose between either presenting in a CG forum or (as evidently happened this year) in a workshop or conference section for the relevant application itself.
On behalf of the workshop organizers
Eckhard Bick & Trond Trosterud