Ben Ambridge

Change tack

New paper! Accepted at Perspectives on Psychological Science. A textbook finding in infant psychology is that newborns - within ~1 hour of birth - visually track a simplified 'three blob' facelike stimulus further than its inverted control (see picture). This suggested inborn knowledge of faces!

PSA for the American (North, Middle, and South, I presume) collaborators: Europe just did the whole clock changing thing. Prepare for some meeting time confusion if you meet at fixed times

HOW TO SPOT A SPAM EMAIL: The sender is excited about your work rather than lukewarm or actively hostile

"Big" journals will spend hundreds of dollars on checking that the references are in the correct format, but literally zero dollars on checking that the analysis code runs, or that the study ever happened in the first place.

Congratulations to @liamblything.bsky.social, Anna Theakston, Silke Brandt, and @ambridge.bsky.social on the publication of their new paper on children's and adults' use of reflexive and non-reflexive pronouns in LDR! ldr.lps.library.cmu.edu/article/id/1...

📢 Two initiatives for displaced students and academics that deserve more publicity: 🇩🇪 www.maxminds.mpg.de - Max Planck mentorship scheme for displaced scholars 🇳🇱 www.dutchscholarsforpalestine.nl/scholarships... scholarships for Palestinians in NL If you know of more please add in reply

arxiv.org/pdf/2610.17538 “Everyone Was Right: Lexical Representations in Language Models Are Both Abstract and Holistic” - just as I argued (for kids!) in doi.org/10.1177/0142...

Probably only @drlearnasaurus.bsky.social will care about this but WARNING - Julia Call (for running mixed effects models from R) does NOT inherit contrasts specified in R. Just code as -0.5,0.5 manually!

Fantastic news! We did something similar a few years ago for specifically developmental language research ldr.lps.library.cmu.edu so do get in touch if you want to learn from our mistakes (or, even better, have any advice for us!)

When we fake smiles, are we smiling our way to happiness or misery? 🙃 For decades, emotion embodiment theorists have emphasized happiness boosts -- all while emotion labor theorists emphasized misery costs. In this Registered Report, we finally put these ideas side-by-side doi.org/10.1016/j.je...

New paper 🍼📄 ManyBabies 3 is out (open access, registered report in Developmental Science). 30 labs, 33 languages, 839 infants set out to replicate a textbook finding: 7-month-olds learn abstract rules from speech (@garymarcus.bsky.social et al., 1999). We didn’t find it. 🧵

Meta-analytic plot below: only 1 out of the 31 samples find evidence of the effect. Huge effort from the entire ManyBabies 3 team led by @ingmarvisser.bsky.social. Imo an important moment for reflection in infant research (esp. language development) given the centrality of this finding. 2/2

I decided it's time to practice what I preach and make sure I publish all my null/inconclusive results, particularly when my theoretical approach really predicts the effect will be found, (and it *has* been found in similar studies). Only this way can we hope to find out what's actually going on!

Fantastic news for all @manchester.ac.uk staff and students! You can now read @manchestermill.bsky.social for free at access-manchestermill-co-uk.manchester.idm.oclc.org

Understanding how learners conclude “X laughed Y” is incorrect is an age-old question, with several hypotheses, some of which have been ~impossible to disentangle! @tomyxw.bsky.social, @fredashi.bsky.social, and I use controlled rearing to shed light on this in our new EMNLP paper: 1/n

Last year, I publicly complained multiple times (see e.g. elevanth.org/blog/2025/07...) that the open access movement had been captured by publishers, and we are now worse off than before open access. Well people got mad at me for saying it. But I think things are still getting worse.

Fancy a weekend read? We've got a preprint for you! zenodo.org/records/2214...

In the UK, the “coffee maker”, is usually a tank of boiling water with a tap.

It's hard to believe that OECS has now published 200 articles!🎉 A huge thank you to everyone who has contributed to OECS—our authors, reviewers, readers, and fantastic section editors, and especially @mcxfrank.bsky.social frank.bsky.social and @hbaum.bsky.social Excited for the next chapter!

Nearly four years ago we asked ourselves, "what would it look like to have a field-leading, completely free reference work for cognitive science"? It's been so fun editing OECS with @asifamajid.bsky.social and @hbaum.bsky.social. I can't believe we have published 200 (!!) articles! oecs.mit.edu

Also great advice for academic writing!

I'm offering a PhD project under this position for anyone interested in studying the acquisition of a language other than English - www.jobs.manchester.ac.uk/Job/JobDetai... - though note that you need to already have the right to work in the UK (sorry!). Look me up & email me for more information!

Children master language either by internalizing innate grammatical rules or by building complex systems from stored examples of phrases they hear. Language Acquisition by Ben Ambridge https://doi.org/10.21428/e2759450.b4a2349c #CognitiveScience

100% this! I see so many between-participants studies with like N=50 when a proper power analysis would tell you you need like 500! Never between-subjects unless you have an EXTREMELY good reason. Remember, an underpowered study is often worse than no study as it can mislead!

I said it on T*****, and I'll say it again here! A mixed-effects model in lme4 with its default settings gives you outputs that LOOK LIKE "ANOVA-style" main effects, BUT THEY ARE NOT! www.sciencedirect.com/science/arti... Thanks @drlearnasaurus.bsky.social for saving my papers many times over!

Open Mind now has an IF of 2.9 after less than 9 years of existence, an amazing accomplishment. This is just short of Cognition (3.1)!

“The “strongest risk factor” is not the form most related to harm. It is the form least redundant with whatever else is in the model.”

New interactive blog! "Why Adjusted Regression Coefficients Are Less Descriptive Than They Look" rpsychologist.com/descriptive-...

Another one bites the dust - Delay of gratification and adult outcomes: The Marshmallow Test does not reliably predict adult functioning academic.oup.com/chidev/artic...

I want this quote on an inspirational poster in every office

If you want to contribute to any of our three projects (ManyTones, CLAPS, Many Uses), you can reach out to the project leads. You will find their contact information here: many-languages.com/projects.html

Big Team Science projects like @manybabies.org @themanybirds.bsky.social etc. are brilliant for more reasons that this short piece can cover, including their openness to researchers from all contexts and backgrounds - you don't even need to have access to the target population! #OpenScience

New interactive simulation in my online textbook illustrating publication bias in meta-analyses, and some techniques to model bias-adjusted effect sizes. None of these work perfectly. We need an unbiased literature to draw valid inferences. lakens.github.io/statistical_...

Does engaging with the arts help to slow the aging process? According to this new paper from Fancourt et al the answer is yes! Alas, the paper is actually unsuitable for such claims. Let’s examine by discussing how you'd establish whether arts engagement actually caused future health! 1/18 🧵

After years in preprint, our paper is finally out! We introduce LambdaG, a simple new authorship verification method that matches complex neural approaches. This supports the theory of grammar as a behavioural biometric (like a signature or gait). www.nature.com/articles/s41... #nlp #llm #forensic

To paraphrase Thomas Edison (or whoever actually said it) Statistical analysis is 1% Investigation and 99% Preparation. Most mistakes I see (and make!) result from rushing too quickly to model building!

This month’s #ManyLanguagesChallengeOfTheMonth comes from the world of **grammatical acceptability judgments** by the ML2 CLAPS team. @ambridge.bsky.social