What We’re Doing Now

We are proceeding to analyze the first 157 weeks of videotaped interactions of an infant and her family. The foci are object explorations, language development and social interactions. The 30 minute weekly recordings have been divided into clips by changes in activity focus, typically 8 clips per videotape. Those clips have been digitized, organized, and made pubic on an internet accessible website. Our intention is to share this material with others interested in understanding more thoroughly development of human minds. The material is expected to be useful to people with different interests: researchers, new parents, students of developmental psychology, and developers of machine learning faciities.

How we are preparing the video material for use (the current primary activity)
We are creating text transcriptions of the activities and speech of the IPS videos; we are nearing completion of those in Peggy’s first year. The clips contain multiple actors: always Peggy and usually Bob, the cameraman (her Dad); her Mom, older brother and sister, and the family dog.
The interactions are developed in dialogue form, by speech when available and by composed action descriptions when appropriate (for the infant and the dog). Local context when relevant is added by the Analyst (originally Bob the cameraman and active participant). These transcriptions are Clip Dialogues. (For detail on this process see: Unpacking Video to Text for Analysis)

The Clip Dialogues are converted into Video Super-titles added to re-rendered copies of video clips installed in analysis-oriented, detailed Clip Notes-pages for each clip. The primary text of each Clip Notes page is the created dialogue, segmented into episodes with time-stamps referring back to moments in the video clips.

Two AI generated files are appended to each page of Clip Notes; they are generated by chatGPT in response to these requests:
AI contents:
Referring to individuals by their names, list in episodic order (with alphabetic and time stamped headers), without categorizing and bullet points, the behaviors, events and social interactions relevant to child development described in the family video transcript in the string “dialogue”
This results in a discursive text covering the same material as the dialogue, as interpreted by current notions of what is relevant in the material.
AI narrative:
Write a summary narrative for new parents and students of infant development of the infant behaviors and social interactions described in the text of the video transcript text string “dialogue”
This results in a short essay on the content, its interpretation and value of the material to two classes of potential audience for the study.

In sum, the content of each clip is represented in five different forms:
1. the original video clip
2. the video clip as augmented by super-title speech and action specification, synchronized onto a copy of the original video clip
3. the video clip dialogue, segmented into episodes for definitive precise reference
4. a discursive text format of the relevant video clip content as interpreted by chatGPT
5. a short essay of the video clip dialogue’s presentation of behaviors and social interactions as interpreted by chatGPT
We estimate the number of video clips from the first year (week 18 through 52) as 210.
Current count of completed Clip Note pages is 187.

Analyst
6/18/2026