Had fun writing this homage to Roy et al., 2015, who demonstrated that language learning is tied to spatial+temporal context. I touch on ecological validity, power, and boldness in study design in dev.psych. A slightly mad study showing just how cool science can be!
#DevSci #devpsy #langsky
Actually, both calibration and gaze-contingent attention getters are set via EB (and I believe you can change the type of calibration on the EyeLink computer itself). We used Python for online calculation of habituation thresholds.
These results challenge the idea that infant-directed speech (primarily) serves a didactic function for language learning, highlighting instead its social and emotional role in early parent-infant communication.
Data, code, and preregistration at osf.io/s8zrm
/end
That is, parents do not spontaneously adjust the acoustic features of words based on whether they think their infant knows them or not! /3
Analysing 38,000+ speech segments from Norwegian parents interacting with their 8–18-month-olds, we found that neither pitch, pitch variation, vowel length, nor vowel clarity differed depending on infants’ presumed knowledge of a word. /2
Do parents exaggerate the pronunciation of words they believe their infants don’t know yet? Our new paper suggests no. ❌ w/ @julienmayor.bsky.social & @nataliakartushina.bsky.social 🧵 1/ #DevSci #langsky #devpsy
Good question, we speculate about both, as well as overall social and cognitive maturation! But this is an avenue for future work for sure!
We invite others to explore and use the child-to-caregiver ratio in their own work via our open Shiny app calculator: 👉 socialnet.uiocloud.no/socialApp/ct... (warning; slow load!). Open data, code, and preregistration: osf.io/pjn8k/ /end
This shift from thinking in terms of birth order to household balance, from siblings as “resource competitors” to potential "language facilitators", offers a more holistic view of how family structure may shape early language development. 5/
As it turns out, the best predictor of infants' vocabulary sizes (over and beyond birth order) was when older siblings were considered "caregivers" from early adolescence, and (the twist): sisters 1–3 years earlier than brothers! 👧💬 4/
To make better sense of this, we introduced a new metric: the child-to-caregiver ratio, a data-driven measure of how many "caregivers" (parents and old-enough siblings) a child has relative to the number of children in the household. A lower ratio was associated with larger vocabularies. 3/
Using parent-reported data from 6,000+ Norwegian infants (8–36 months), we found that expressive and receptive vocabulary sizes tends to decrease with additional older siblings... but only up to a point. For later-borns, vocab. actually starts increasing again. 2/
Do older siblings help or hinder language development? And does it matter if they’re sisters or brothers? In our new paper we revisit the "birth order effect"... and took it a step further! w/ @julienmayor.bsky.social, Nora Serres and Natalia Kartushina 🙌 🧵 1/ #DevSci #langsky #devpsy
Consonants, like vowels, appear to be less distinct in IDS than ADS, thus reinforcing the interpretation that IDS may serve an attentional and/or affective aim, rather than a didactic purpose. /end
We did not find any association between features of VOTs and infants' consonant production and/or babbling as reported by parents. This held whether VOTs were operationalised as the IDS input itself or parents' adjustment/effort in IDS from their ADS. 4/
We found that, compared to adult-directed speech, voiceless stops in IDS had longer VOTs, while voiced stops had shorter, leading to overall less distinct consonant contrasts (/b-p/, /d-t/, /g-k/) in IDS than ADS. From 6 to 12 months, VOTs in IDS became more similar to ADS. 3/
We examined voice onset time (VOT)—the interval between a consonant's release and vocal fold vibration, which distinguishes voiced and voiceless stops (e.g., /b/-/p/)—in speech recorded during shared reading interactions in Norwegian parent-infant and parent-experimenter dyads. 2/
New paper out! Following from our work on vowels and prosody in infant-directed speech, we turned our attention to features of consonants and their role in language development. Great collab (as always!) w/ @julienmayor.bsky.social, Nina Varjola and Natalia Kartushina 🧵 1/ #DevSci #langsky #devpsy
Interested in doing a PhD on language acquisition in Oslo?
Check application process: www.jobbnorge.no/en/available... and feel free to reach out for more information!
PS. We have a nice growing community including Natalia Kartushina, Luca Onnis, @audunrosslund.bsky.social, and many more!
Data, code and materials to the study is of course openly available at OSF. Watch this space for whether and how these parents' acoustic properties relate to their infants’ language skills! /end
Our results show that IDS is both dynamic and static, and we suggest that future studies should investigate how the acoustic constants and variations influence the proposed attentional, socio-emotional, and linguistic functions of speech to infants. 7/
We also found that IDS was characterized by increased vowel variability and decreased vowel distinction, and more so in mothers and fathers, with no change with infant age. In other words, this highlights that IDS is not across-the-board ‘clear’ input. 6/
Analyses of vowel space expansion (vowel hyperarticulation) told a more complicated story; While fathers’ overall expanded their vowel space in IDS, mothers did not, perhaps as they appeared to shift from vowel space reduction to expansion with infant age. 5/
What did we find? Compared to their ADS, both mothers and fathers’ IDS had: Higher pitch, wider pitch range, slower articulation rate, and longer vowels. With infant age, pitch range widened and vowels shortened, while pitch and articulation rate remained relatively stable. 4/
We followed 69 Norwegian families for one year and five lab sessions, from infants were 6–18 months. Parents’ speech were recorded while reading the same picture book to their infant (IDS), and to an experimenter (ADS), ensuring no distortion from different linguistic content of speech. 3/
First things first, this was only possible because of the great team of Natalia Kartushina, @julienmayor.bsky.social, @acristia.bsky.social, Arun Singh, Roger Mundry, fantastic RAs and helpful reviewers! 2/
12 months of data collection, 317 lab sessions, 22,958 phrases, 54,594 vowels, and a gazillion supplementary tables later... proud to say that our longitudinal study on the acoustics of infant-directed speech is now out! 👶🧵 1/ #DevSci #langsky #devpsy
royalsocietypublishing.org/doi/10.1098/...
Fantastic! Would you be so kind and add me?
Developmental Cognitive Scientists by @deontbenton.bsky.social
Thanks for the kind words! I am unfortunately not well-versed in that literature, but I believe J. Dynia et al., did some work that could be of interest!