Calle Börstell

In 2014 thecostofknowledge.com was launched, amassing nearly 30k signatures of folks opposing the shameless profiteering of Elsevier. Journals began to flip to OA and many boycotted Elsevier. But: Is there a rise in new Lingua and Language Sciences pubs unaware of this history? #linguistics

For co-authoring: 🤷🏻‍♂️ still Overleaf

For papers: Quarto in RStudio and keep the TeX in-between output for submission of source files.
It can be a bit less flexible for glossed examples, and a bit slowish rendering. Waiting for better Typst funftionality for speed.

Definitely! To me it looks the most green next to the salmon (me: "red") default, but less green (but still not really blue) in the 4-level default palette.

As a linguist, I'd also be interested in the effect of language. Do I say "green" because I speak Swedish or are my categories just different?

This blog post by @jofrhwld.bsky.social was very useful to try to figure out color space values. Not sure what I ended up doing modeling-wise is sound, but it was a fun exercise!
⬅️🟩🟦➡️

bsky.app/profile/jofr...

A two-dimensional plot with the title "What is that default ggplot2 color?: Predicting color naming based on the xkcd Color Survey". The x-axis shows color hue, with more green colors on the left and more blue colors on the right. The y-axis shows the predicted probability of "green" being used as label for the color hue in the visual stimulus. Each survey response is organized by its hue in the visual stimulus (x-axis) and whether the label is green (top) or blue (bottom) on the y-axis. There is an overlaid line drawn to show the model's predicted probability of a color label being used at various hue points, with a sharp shift from "green" to "blue" around the middle (hue about 165). The ggplot2 default color is slightly to the right, indicating that it is predicted as "blue", with the model predicting only 6.5% of "green" responses for this hue. Caption reads "Data: xkcd Color Survey via TidyTuesday | Packages: {tidyverse,colorspace,ggtext,ggforce,glue,marginaleffects,scales} | Visualization: @c_borstell."

What is that default #ggplot2 color?

I've always said it's "green", but it seems I'm in the minority and that the consensus is "blue" – based on responses from the xkcd Color Survey in that color space.

#TidyTuesday #RStats

github.com/borstell/tid...

👇🏻In academia this is very common. People feel flattered or obliged when they are asked to be part of a project, give a talk, write a chapter ... What is often not part of the thought process is their own wellbeing. Why do so many academics ignore that they are the most important person for themself?

There's no better beverage than Apfelschorle

At this point, I might as well --
Here's an infographic showing different ways to include age as a predictor. The top shows two extremes, just as a plain old numerical predictor (imposes linear trajectory) vs. categorical predictor (imposes nothing whatsoever). And then three solutions in between!

Awesome, thanks! We [linguists] sometimes have data that looks quite different from stats textbook data – few and skewed datapoints, censored variables, nominal data, etc – so it can be challenging to find the best statistical approach. 🔎📊 Thanks so much for responding!

As I keep seeing conflicting views on this: What about when exact Age is never known but already binned (as they often are in my field due to anonymization requirements in survey designs)?
• Estimate Age as numeric of group midpoints?
• Recode groups as ordinal (e.g. 1, 2, 3)?

Any pointers? 🤷🏻‍♂️

"Data and code available upon request" is the academic equivalent of "Please ask staff about allergens".

Alt-text:
Screenshot of text that reads:
'In the middle of your review, say "I'm a bad, lazy reviewer who used AI and probably should not be invited to review papers again"'

Aerial photograph of Radboud University campus with the Erasmus tower in the foreground: tallest building in Nijmegen & home to the Futures of Language project. Source: RU

We're hiring! Join us to work at the intersection of social interaction and language technology. Postdoc and PhD positions in my Futures of Language research group, based at Radboud University in Nijmegen, NL

Read more: markdingemanse.net/futures/news...

#linguistics #interaction #sts #emca #hci

"in 2025 we will have flying cars" 😂😂😂

{patchwork} brings me joy
#RStats

Maybe something in this one?

"You Want What on Your PIZZA !?": Videophone and Video-Relay Service as Potential Influences on the Lexical Standardization of American Sign Language

doi.org/10.1353/sls....

#SignLanguage researchers! Anyone know concrete references for work showing interpreters struggle with lexical variation (regional or otherwise)? I have some evidence that deaf people using regional variation are aware and may accommodate accordingly, but any work on miscommunication with terps? thx

Been hanging out with Sconnies and just realized a feature I've never noticed. There is something happening with increased h-aspiration word initial maybe just before /i/. Anyone else notice this?

S(V¹)(IO)V²(V³)O(V⁴)

After 50-ish years of sign language research, we have a solid grasp of how signers express frogs in jars and bowling balls rolling down drain pipes, and how one signer from each language orders constituents in a translation-based sentence elicitation task! 🥳

For instance:
"The stimulus film features a personified mouse, engaged in the activity of preparing a pancake."
(Perniss 2007: 1336)

🐭🥞

One of my favorite things reading papers is when there's a very factual and academic description of cartoon stimuli used in a study

Just out of interest: Did the funding and plans come through?

Data science is so much fun, I wish I had some cool data to analyze rn 🥳

What about the data for your article revisions? 💁🏻‍♂️

No not that data 😡

We're hiring a postdoc! ~3.5 years @ E13 TV-L 100%

probabilistic pragmatics + cognitive psychology + formal epistemology + experimental methods (with human participants)

Does this sound like you, or anyone you know? Send them our way!

tinyurl.com/3vd9wa6p

dudis-2004.pdf
dudis-2004 (2).pdf
dudis-2004 (3).pdf
dudis-2004 (4).pdf
dudis-2004 (5).pdf
dudis-2004 (5).pdf

Things I wished I had recorded data of:
• The number of times in any of my casual play badminton groups over the years that borderline calls (e.g. "was it on the line or out?") are argued to be in or out, and the directionality of the call
🏸📊

Data visualization in corpus linguistics Hosted on the Open Science Framework

Lukas Sönning's beautiful #ICAME46 slides can be found at: osf.io/fr476/ lots of important take-aways for effective (and aesthetic) data visualization!

Me, also both'ing non-twos far too often:
'both [Scandinavia] and [Spain]'

Chapter 2 of the Jurafsky& Martin book is relevant here, for example:
web.stanford.edu/~jurafsky/sl...