Welcome to the Website of the MaEiQCL Workshop

The emergence of data science has inspired a surge in interest in the application of quantitative and computational methods in comparative linguistics in the broad sense. By this we mean any kind of research studying features of several natural languages in parallel. High profile results touch upon three major topics:

• the study of deep history, both regarding reconstruction of past language stages and language change processes and of population history in general,

• statistical investigations of typological questions regarding, e.g., the (non-)universality of feature correlations,

• probing for - possibly causal - connections between linguistic properties and extra-linguistic variables such as language community size, climate, or diet.

These results are often met with a healthy skepticism within the linguistic community. It is tempting to discount the criticisms leveled against quantitative comparative linguistics — such as the insistence by practitioners of classical historical linguistics that historical linguistics must be based on the identification of sound laws — as inevitable side effects of a paradigm shift. However, computational and statistically minded comparativists do not agree among themselves regarding the standards of data quality, model validation, and model comparison. For instance, the debate in a recent issue of Theoretical Linguistics revealed that there is no consensus about some very profound issues pertaining the the nature and purpose of statistical models in computational historical linguistics. The open peer reviews on Dunn et al. (2011, Nature) in Linguistic Typology 15(2), 2011 revealed a similar demand for debate in typology which has not been conclusively settled so far.

The purpose of the workshop is to provide a forum for this methodological discussion. It will focus on:

• approaches to model validation and model comparison in statistical work on comparative linguistics,

• standards for data formats, data accessibility, and data sharing, and

• best practices for code sharing and code accessibility within an open science framework.

Call for Papers

• Date: February 24-26, 2021

• Location: Freiburg, Germany (short workshop at the 43rd Annual Conference of the DGfS)

• Contact persons: Gerhard Jäger (gerhard.jaeger@uni-tuebingen.de) and Johann-Mattis List (list@shh.mpg.de)

Workshop organizers

• Gerhard Jäger (Tübingen)

• Johann-Mattis List (Jena)

Call Deadline:

31-Aug-2020

Travel Grants

Please note that a limited number of travel grants will be available upon request.

Submission Requirements

We invite submissions for 20-minute oral presentations (+ 10 minutes discussion) in English. We are equally interested in contributions relating to data management and to data modeling. Abstracts should be anonymously submitted to: ...

Abstracts should be at most one page long, plus references on the second page, on A4 paper with 2.5cm margins on all sides, and must be set in Times New Roman font of at least 11 points. The deadline for submission is 31 August 2020; notification date is 15 September 2020.

Schedule

Day 1: Wednesday, February 24th, 2021

Time	Title	Speakers
16:30	Data in quantitative comparative linguistics	Johann-Mattis List
17:00	Models in quantitative comparative linguistics	Gerhard Jäger
17:30	Methods and models in historical comparative research on signed languages	Justin Power, Danny Law, and David Quinto-Pozos

Day 2: Thursday, February 25th, 2021

Time	Title	Speakers
9:00	Language contact in the evolution of linguistic features	Harald Hammarström
9:30	Partial cognate comparison and pre-settlement history of the Dogon ethnolinguistic group	Abbie Hantgan-Sonko
10:00	Why we need more study of methods, not data, in computational historical linguistics	Philipp Rönchen and Tilo Wiklund
11:15	Theoretical (in)compatibilities of the comparative method and cladistics	Erik Elgh
11:45	Lexedata: Tying existing software to CLDF Wordlists	Gereon Kaiping, Natalia Chousou-Polydouri, Kellen van Dam and Melvin Steiger
12:15	Controlling for geographical, areal, and family biases in typology	Matías Guzmán Naranjo and Laura Becker
13:45	Revisiting typological universals with Grambank	Annemarie Verkerk, Hannah Haynie, Russell Gray, Simon Greenhill, Olena Shcherbakova and Hedvig Skirgård
14:15	Towards richer multi-source machine-readable etymologies	Johannes Dellert

Day 3: Friday, February 26th, 2021

Time	Title	Speakers
11:45	Correlating borrowing eventts across concepts to derive a data-driven source of evidence for loanword etymologies	Verena Blaschke and Johannes Dellert
12:15	Modelling linguistic data in space using autologistic regression: The case of ejectives	Miri Mertner
12:45	Loanpy -- A framework for computer-aided borrowing detection	Viktor Martinovic
13:15	The effect of prior on tree topologies	Johannes Wahle
13:45	Final discussion	[all participants]

Location

It has now been decided that the event will take place in virtual space only.

Information

Detailed information on the workshop will be added about between a month and two weeks before the conference.