LinguistMix VI

Posted on March 4, 2013 by

Thursday 7th March sees the sixth LinguistMix of the year, starring LEL’s Tine Breban and Goran Nenadic from Computer Science. Here’s the blurb from the events page:

Linguistmix is a great opportunity to have a chat with lecturers and find out more about the areas of linguistics you are (or could be) interested in. It’s also a great way to take advantage of free wine and nibbles.

**** students and lecturers from all departments are welcome ****

LinguistMix VI will have talks from Tine Breban (Linguistics and English Language) and Goran Nenadic (Computer Science)

Tine Breban (LEL)
Complex determiners in English

When we think about determiners in English, the usual suspects are the definite and indefinite articles, “the” and “a(n)”. Their function is commonly defined as signalling whether the referent they mark is given for “the” versus new for “a(n)” or identifiable for the versus non-identifiable for “a(n)”. On second thought, demonstrative and possessive determiners (e.g. “this”, “that”, “my”, “his”) enter the stage. They both mark the referent as identifiable, but in addition explain how identification can be achieved, on the basis of a demonstrative or possessive relation. This is why they have been called semantically complex in the Cambridge Grammar of the English Language (Huddleston and Pullum 2002). In this talk I want to show that it doesn’t end here and that there is a whole world of other determiners out there, which express similar complex identification relations, but which are traditionally not included in the category of determiners because they are multi-word expressions, combining a definite or indefinite determiner with a second element. Examples include “the same”, “the usual”, “a certain”, and many more. Using real-language data, I will discuss why these expressions should be seen as part of the determiner paradigm in current English, and offer a possible explanation as to why they developed historically.

Goran Nenadic (Computer Science)
How the world feels about it: sentiment mining of social networks

The number of tweets is growing: worldwide, there are 4,000 tweets per second, 340 million tweets per day, which is 7 times more than last year. Other social networks follow the same pattern. This data provides a huge opportunity for large-scale studies of human behaviour including gauging of (global) sentiment. What do people think about the markets? Can we spot people who have symptoms of depression? Or have self-harm intentions? In this talk we will briefly overview computational linguistic approaches to systematically identify sentiment from text and demonstrate some work in progress.