Page from "A Light in the Attic" by Shel Silverstein. At the bottom is a drawing of a kind put a piece of paper into a machine with gears and buttons. Above that is the following poem: The Homework Machine, oh the Homework Machine, Most perfect contraption that’s ever been seen. Just put in your homework, then drop in a dime, Snap on the switch, and in ten seconds time, You homework comes out, quick and clean as can be. Here it is – “nine plus four?” and the answer is “three”. Three? Oh me … I guess it’s not as perfect As I thought it would be.

Very weird that the answer to "which #scifi author predicted 21st century technology best" might be Shel Silverstein...

Everyone fighing over TIbble versus DF >| vs %>% Me: googling what mean means

Me? I'm just happy to be here #rstats

This is great and also feels like the kind of thing that's destined to show up in the slides for an Intro #Linguistics course

Back-scratching bovine leads scientists to reassess intelligence of cows Brown Swiss in Austria has been discovered using tools in multiple ways – something only ever seen in humans and chimpanzees

COW TOOLS COW TOOLS COW TOOLS

[Searching through the fanfold printouts of my parents' ELIZA chats, trying to figure out what it told them that made me this way]

Never seen the Lawn Mower guy! Sounds very mint hill though. I'll have to keep my eye out for him.

Yep!

Live in the town with this horse and can confirm. Dude just ties it up outside the bars in a parking spot too, lol

they should invent a tamale printer

X-Men, Krakoa, and Linguistic Sovereignty by Erika Alpert Krakoan national co-founder Magneto explains to diplomats at the new Krakoan Embassy in Jerusalem why mutants have decided to create their own language, from House of X #1 (Hickman …

Do you like comic books and want to think more seriously about the linguistics of them?

Are you a social scientist or a linguist who wants to know more about nation-building or language planning in fiction?

HAVE I GOT EXACTLY THE ARTICLE YOU NEVER ASKED FOR, on CAMP Anthropology. 🐦🐦

A white and brown bulldog, lying with her face upside down, pointed toward the camera. She's resting her oversized head in the lap of a person. And her tongue is sticking out just a tad.

🙃

Explaining differences between phonotactic learning biases in the lab and typological trends using Probabilistic Feature Attention A primary goal of linguistic theory is to explain why certain kinds of languages are underattested. One methodology that has had success in explaining phonological typology has been artificial languag...

Curious about why the model captures these results? Check out my extended abstract in the proceedings of SCiL 2025:
doi.org/10.7275/scil...

And to read more about PFA, check out this paper in Linguistic Inquiry:
doi.org/10.1162/ling...
6/6

A figure showing six learning curves (epochs on the x-axis and accuracy on the y-axis). Each curve is associated with one of the Shepard Types. Type II starts off lower than Type IV, but by the end of learning Type II is second only to Type I.

I found that when you use PFA with a MaxEnt model and train it on Types I-VI, II starts off harder than IV, but becomes easier for the model later in learning. Since lab learning involves less exposure than natural language acquisition, this could explain the mismatch in M&P's (2014) results.
5/6

On the left, a cube with edges of equal length. At each corner of the cube is a shape that can be circle/triangle, large/small, and black/white. On the right is the same cube, but squished in a way that illustrates that the large/small distinction is less important now. In the middle, an arrow labeled “PFA”. Cube illustrations borrowed from Nosofsky (1986), where the definition for "attention" I'm using was coined.

Probabilistic Feature Attention (PFA) can potentially explain this. PFA adds ambiguity to the learning process, randomly sampling which features a model can attend to (similar to dropout in neural networks). E.g., if the model doesn’t attend to [voice], it can’t distinguish between [t] & [d].
4/6

The six Shepard Types, illustrated using phonological segments. Segments are either voiced or voiceless, stops or fricatives, and labial or alveolar. (Just an illustration, not the actual patterns used by Moreton and Pertsova, 2014).

...And found that Type II phonotactic patterns were harder for participants to acquire than Type IV patterns. But when they looked at typological data, Type II phonological patterns were more common than Type IV. But why does typology seem to favor the more difficult pattern?
3/6

The six Shepard Types, illustrated using simple shapes. Shapes are either circles or triangles, large or small, and black or white. Taken from Moreton et al. (2017).

Shepard Types (I-VI below) have been used widely in category learning since they were introduced by Shepard et al. (1961). These assume a stimulus space with 3 features, 8 stimuli, and 6 possible ways to halve that space. Moreton & Pertsova (2014) adapted these to the domain of phonotactics...
2/6

An academic poster with images and text that goes through roughly the same points as this thread. To download a PDF of the poster, see https://brandon-prickett.com/wp-content/uploads/2025/07/SCiL_Poster.pdf .

Artificial language learning experiments often explain typology by showing that more common patterns are easier to learn. But this isn’t always the case. In work I recently presented at SCiL, I present a possible explanation for one such mismatch between the lab and typology. 🐦🐦 #linguistics
1/6

My kingdom for a

Society for Computation in Linguistics

Lots of great computational #linguistics in this year's SCiL proceedings! (Including an extended abstract from me that I'll probably post more about closer to the conference.) 🐦🐦

openpublishing.library.umass.edu/scil/

LSA Annual Meeting - Call for Abstracts LSA Annual Meeting - Call for Abstracts

Linguists! Abstracts for the LSA annual meeting in NOLA in January 2026 are due on July 7! Please submit! It’s going to be a great conference, and a chance to confer and strategize. Would love love love to see you there! #linguistics @lingsocam.bsky.social www.lsadc.org/abstracts

Really nice example to bring up in a data science or corpus #linguistics class about dealing with weirdness in your data!

Also saw this as a kid and was scarred for life

A snail so perfect looking, it's like what you'd find in the dictionary if you looked up "snail". It's sliding slowly across a sidewalk.

Look who I found outside my front door! 10/10, no room for improvement.

(reading the specs of a clown car) wow 32 cupholders

Humans do representation learning all the time and one of the most tangible aspects of representation learning (from a continuous physical space to mental representations) is human phonology.

Submit to the Special Session at this year's AMP on modeling phonology with deep neural networks!

New article out on how infants learn to find affixes in early infancy! onlinelibrary.wiley.com/share/author...

One thing that needs to get repeated until it sinks in is that “AI” does not simply function as a tool that people in power are using; it’s even more powerful as a permission structure for the destructive & extractive things they wanted to do in the first place.

A brown and white bulldog on a mat with a tennis ball and other assorted dog toys around her. She appears to be smiling with her tongue out.

Happy dog!

Baby tapir Ume had her first encounter with the water dish and it’s adorable! 🥰

It's one of the unwritten rules of civilization: if a child pours air into a cup and offers you tea, you will sip it and say "thank you, it's delicious."