Blokový kurz: Multifactorial feature analysis: A quantitative method for qualitative linguistics (5 kreditů)

Začátek:01/01/1970 - 00:00

Časové pásmo:Europe/Prague

Multifactorial feature analysis.

A quantitative method for qualitative linguistics

Dylan Glynn

dglynn [at] univ-paris8 [dot] fr

www.dsglynn.univ-paris8.fr

Course Organisation:

the course will take place in 8 sessions, two sessions per day (11:00 - 13:00 and 14:00 - 16:00) on September 16.
–
17. (room P217) and 19.
–
20. (room P104) , note that the last day is Saturday
it will be conducted in English
it is primarily intended for PhD and advanced MA students of English Linguistics, but if space allows, BA students and students of other fields are also welcome
we are currently not registering any new participants
PhD and MA students may gain 5 credits, BA students 4 credits (classified as PVP credits for students of English Linguistics)
if you have any question about the organisation of the course, let us know at ondrej [dot] tichy [at] ff [dot] cuni [dot] cz

Course Description:

The empirical turn in linguistics brings with it a range of quantitative tools developed in sociology and psychology over the last half of a century. This entirely practical course is designed to show how some of these tools can be applied to traditional linguistic research questions in Functional Linguistics, Discourse Analysis, and Cognitive Linguistics. The course introduces one method termed multifactorial usage-feature analysis (also the profile-based approach). This method is currently extremely popular in the fields of semantics / pragmatics and conceptual / functional analysis, but also in variationist / sociolinguistic research and research in morpho-syntax and construction grammar.

There are two steps to this methodology. The first is the “usage-feature” analysis and the second is the multivariate statistics that are needed to interpret the results of the feature analysis. Linguists are always concerned that the statistics will be difficult. This is not the case at all. The statistics are merely a tool to enable us to improve the research that we already do. The linguistics research questions that you all work on remain difficult. The statistics will allow you to identify things would not otherwise see and to make falsifiable claims in ways that you otherwise could not, but the linguistics remain the point and the most difficult part of research.

Before we start, I will put on line materials for you. I will also ask you to send me a 1-2 line summary of your research interests to help me better orientate the content of the course. The actually course will be based on a topic that the group chooses. It will be broken down into 8 sessions of two hours, which will be held over the course of 4 days. Over the course of the workshop I will also discuss how these methods are applicable to your own work. After the workshop, I will be able to help you apply these techniques in your own projects.

Requirements

No prior knowledge of statistics is needed. A computer with internet access is needed – preferably your own laptop.

Aims

To understand and apply multifactorial usage-feature analysis. To run the basic functionalities of R, the statistical programming environment. To be able to apply and interpret basic tests for statistical significance, and run multivariate statistical analyses on complex data. To be able to apply and interpret binary logistic regression modelling of complex categorical data.

Course Outline

Session 1. Introduction (16.9. 11.00-13.00, room P217)

A summary of the method, its strengths and weaknesses and possible applications.

In this session you sit and listen.

Session 2. Research Question (16.9. 14.00-16.00, room P217)

Choice of research topic, collection of data, and development of a coding schema.

In this session we work together to obtain data to learn the method.

Session 3. R (17.9. 11.00-13.00, room P217)

The statistical program for doing statistics.

In this session we introduce the program R and learn how to load data and to run some basic tests.

Session 4. Correspondence analysis (17.9. 14.00-16.00, room P217)

A multivariate statistical technique for identifying complex patterns of association in the data.