Word Pedometers: Another Really Dumb Tech Idea

You only have to get one "story at a glance" point into this article to know that this is going to be a freakin' disaster panda, and I have so many wuestions.

New devices can be worn by babies and toddlers to count the number of words they are exposed to each day.

It's a word pedometer, a sensor that you strap onto your child's chest that, well, records all the words tat show up in the area.

They are situated inside a small vest that keeps the tracker positioned on the baby’s chest, not to measure steps or stairs, of course, but to log words, sentences and entire conversations spoken to or overheard by the child.

This is not a new batch of baloney; here are 75 Providence Headstart families being used as a pilot program back in 2014. Because, of course, this is aimed at non-wealthy non-white families, and was mostly about "training parents." The reasoning back then is familiar. According to the program director, "Previous research has shown that the number of words that is heard between ages 0 and 3 is a good predictor for school success." It was backed by the LENA foundation, an outfit that's all about using tech to build early language skills and adult-child talk and pushing the word pedometer. LENA's president came from the aerospace industry, the senior director of research is degreed in linguistics and developmental psychology, and the director of growth strategy has experience taking projects to scale. That pilot was followed by a program called Providence Talks which is still tied to LENA.

This is all about closing the fabled "word gap," the notion that children in poor families hear fewer words. That goes back to a piece of 1995 research that has always been controversial (and has also proven resistant to replication). It's yet another example of confusing correlation with causation. One of my favorite word gap quotes comes from linguist Michael Erard:

Just as solving climate change isn’t about closing the polar bear gap, and preventing environmental degradation isn’t about closing the tree gap, you can’t increase children’s school readiness by closing the word gap.

If you want further examples of why the word gap is not a thing to be basing policy around (and there's agreement on this from many sides of the ed debates), you can look here, here, and here. (For balance, this is a nice, brief, clear debunking debunking.) There's wide agreement that early exposure to language matters a great deal, but the assertion that, somehow, poor folks don't talk to their kids is problematic.

Strapping a microphone to your bahy seems like quite a step to take, and the claims made by the tech seem, well, extreme. Researcher Kenneth Wong (Brown U) says the program is "designed to simply record things like word counts and 'conversational terms,' such as when a new topic is introduced." Anyone who has ever had a fight with Alexa or Siri or a talk-to-text program will have no trouble imagining all the ways this could fail. The article notes that this all sounds creepy, and that could b e because it is creepy.



The word pedometer, like the word gap concept itself, also plays to the absolute worst tendency of technocratic solutions-- it emphasizes reductive measurement that leads easily to focus on exactly the wrong stuff. The article cites one example of a mother who came home and read the real estate ads to her baby, just to get those words in . And the opening example compares word pedometers to fitbits, which is particularly apt, since a fitbit has no idea what kind of step you took--in fact, my old fitbit would give me credit for conducting a pit orchestra, which involved sitting on my butt and periodically waving my arms.

Trying to "fix" babies by exposing them to a bunch of words, as if language were a sort of virus you can catch, is not terribly sound. Giving your baby a microphone-embedded vest to wear (do they say "this conversation may be monitored for quality assurance"?) is just a shiny way to enable a bad approach that also takes the pressure off of things like fixing poverty and jobs and living conditions (all those things that can't be blamed on the poor).

Also, when you're imagining this, don't forget the suggestion is to strap the microphone onto toddlers. I'm pretty sure trying to count the words (as opposed to random noises) coming out of my toddlers as well as tracking their conversational topics would severely tax the limits of any software. This whole business is just a bad idea.