Research‎ > ‎

Post-doctoral position in Paris: Statistical modelling for Historical Linguistics

Environment


We are an interdisciplinary research group hosted at the Department of Cognitive Sciences (DEC) of the Ecole Normale Superieure in Paris and the Centre de Recherche en Mathématiques de la Décision (CEREMADE) at Université Paris-Dauphine. The main local investigators are:


Robin Ryder: Computational statistics and applications to Linguistics

Emmanuel Chemla: Psycholinguistics

Ewan Dunbar: Machine learning and human speech

Philippe Schlenker: Formal semantics


The postdoc will be a member of both DEC and CEREMADE and will be expected to interact with members of both laboratories.


Project and tasks


The postdoc will be expected to take part in the two following projects. The allotment of time between these projects will depend on the candidate’s profile and preferences, with the expectation that project 2 will represent at least 20% of their time.


  1. Phylogenetic modelling of the history of human languages


This project will be mainly supervised by Robin Ryder, with interactions with members of DEC.


Since the original phylogenetic analysis of lexical data by Gray and Atkinson (2003), there have been a large number of systematic quantitative studies applying statistical methods to answer questions from Historical Linguistics, usually through phylogenetic models of language change. We are interested in several approaches of these techniques: modelling questions, computational issues, and aspects related to statistical methodology. We have several projects related to Bayesian computational modelling of lexical and morphosyntactic data that the postdoctoral fellow could take part in, but are also open to suggestions from the successful candidate.


  1. Reconstructing the evolutionary history of monkey communication systems


This project will be supervised jointly by Emmanuel Chemla, Ewan Dunbar and Robin Ryder, and will involve phylogenetic statistics applied to recordings of monkey calls. This promises to offer a window into the evolution of form and meaning in simple systems of animal communication. We have a growing database of recordings of monkey calls (currently thousands of annotated recordings from more than 10 phylogenetically diverse species), and we hope to make inferences about the evolution of communication systems over an unprecedented time depth of millions of years.


The main task of the post-doctoral fellow will be to apply machine learning tools to develop this research program: represent the calls in a space that allows comparisons of calls within and across different species. This task will involve the use of standard tools and will also require creativity. The post-doctoral fellow will be the driving force, but we will contribute our relevant expertise at all times in the process. The post-doctoral fellow will be expected to interact with colleagues from a variety of academic disciplines, including Linguistics, Zoology, Machine Learning and Mathematics.


We expect that the main tasks will take a significant part of the working time, but it comes with no administrative or teaching requirements beyond scientific interactions with colleagues (e.g., in the form of active participation to relevant seminars). There should thus be ample time for the post-doctoral fellow to pursue their own research program within the team. The Department of Cognitive Sciences is a major centre for psycholinguistic research in speech processing, language acquisition, and semantics of natural languages. CEREMADE is a major centre for research on Bayesian statistics, computational statistics and nonparametrics, amongst other specialities.


Profile


The successful candidate should have a PhD or equivalent experience in Statistics or Machine Learning applied to Linguistics or Phylogenetics.


Practical aspects


- Deadline: 23 August - We realize that this announcement comes at a short notice, we will take this into account in reviewing applications. Please get in touch with us if this could create any difficulty.

- The application file should contain

    - a CV

    - a document up to 2 pages stating the relevant expertise in Statistics (or any other relevant expertise) and explaining the candidate's motivation and research interests

- up to 3 letters of reference

- The application file and requests should be sent to: Stat.HistLing@gmail.com

- The post-doc will receive roughly €2500 per month. Expected starting date: October 2017, or as soon as possible thereafter. Initial contract is for twelve months; there may be opportunities for renewal. Post-doc will be located at the Ecole Normale Supérieure and Université Paris-Dauphine in Paris, France.

Comments