Comparative Psychology – gaining understanding of human behavior by making comparisons to non-human agents — can be hard for students to appreciate, especially when the comparison is between humans and computers, a core element to the Cognitive Science enterprise. My senior Cognitive Science seminar students are not an exception here: all semester long they’ve been digging deep for ways to dismiss computer science/AI contributions to cognitive science as we delve into the pressing questions cognitive scientists ponder regarding the mysteries of the human condition. After all, if humans create computers and such, what are we really learning from them?

That’s been their reaction in a nutshell. That is, until last week when we discussed Pfeifer & Iida’s (c. 2004) “quadruped puppy project” (we read about it as cited in Andy Clark’s newly updated text Mindware: An Introduction to the Philosophy of Cognitive Science, but the project is discussed in many venues, such as the conference paper linked above or in this 2007 Science article). Here’s the gist of it. The puppy is a robot, designed to examine questions about self-organization and embodiment, key concepts in cognitive science. Puppy is designed simply, with aluminum, springs, sensors, and actuators. Very little top-down control exists in puppy’s artificial neural architecture, rather its design represents a “task distribution” set-up where actual locomotion emerges from interactions among the robot’s “brain,” its morphology, and its environment.

As I understand it, “puppy’s” first steps where shaky – the robot slipped and slid around on aluminum paddle-paws, leading its designers to reconsider the paw design. Flesh-and-blood puppies have pads on their paws, and for the real thing, the pads serve an important proprioceptive function. So, the robo-pup’s designers initially interpreted the slippage as a flaw and added pads to puppy’s paws and set it out for another romp around. Rather than romp though, robo-pup v.2 fell over.

Long story short: the initial slippage appeared to have initiated an imperative environment-organism feedback loop that enabled the organism to walk and run. Without a signal to adjust it’s position, the system as a whole was rendered dysfunctional. Too much top-down protective control didn’t allow the system to self-organize.

Wow.

My aim with sharing this here on the blog is not to rehash my in-class discussion (though that discussion was rich – indeed, it was one of those days that reminds me of just how lucky I am to have this job!). Rather, I want to share what first came to my mind when I read about this project: “Helicopter Parenting.”

Analogically, the puppy project can serve as a cautionary tale for sure: Does too much top-down control from parents render their little systems dysfunctional too? On the surface, the analogy fits. Click on the link above to see for yourself: Parenting literature and social commentary abounds with references to the damage done by over-protective parents. Too much control creates narcissism, anxiety, and emotional mayhem. Total system dysfunction might not be an overstatement.

As a psychologist, I get the message: authentic learning is key so parents need to stand back and let their kids learn by trial and error. True understanding requires contrast:

• What is success without failure?
• What is joy without sadness?
• What is friendship without isolation?
• What is winning without losing?

As a parent, I struggle with the message too. Culture here in the US exudes a positivity bias that permeates all that we do: we don’t want our children to cry, suffer, be lonely, and experience hardship! The puppy project brings this all into focus though: a little slippage is a good thing.

The real challenge then, in taking to heart this “robo-puppy project – parenting comparison,” lies in determining just how much slippage we should ask our children to endure. The backlash of the helicopter parenting publicity has led to what I call “hands-free” parenting instead:

• Dispute on the playground? Let kids hash it out themselves.
• Too many rules! Let kids play without constraining it.
• Playdate going south? Kids need to learn how to cooperate with others: let them work it out.

Indeed, a recent Atlantic Monthly article (The overprotected kid) characterizes the backlash nicely, with an incredible playground design sweeping Wales right now: the playgrounds are fenced-in former vacant lots with wood pallets, moldy mattresses, fire barrels, mud pits, and no rules! Kids love them – indeed — compared to the trend in the US where playgrounds are colorful, attractive, but mindlessly boring in their “safeness” these play-lots are downright dreamy.

Yet, is hands-free parenting (or no holds-barred playgrounds) the appropriate “take away” from “helicopter parenting” research and comparative robo-puppy research? Perhaps not. That is, hands-free parenting isn’t without its problems too. Kids do work it out…but not always in an optimal manner. Full blown hands-free parenting (and similar early-childhood-education classroom management strategies) can cause patterns to emerge that may, later, turn into bullying, relational aggression and/or social control tactics. That is, allowing too much slippage can backfire.

Moderation is a good thing. A little slippage is a good thing, but there’s a diminishing return with too much AND with too little. The robo-puppy works because of its simplicity – that is, there is some top-down control, but not a lot. The principle is “just enough.” In like manner, we parents (and by extension teachers too) need to figure out what “just enough” support is for our youth, so that we can enable optimal growth. The parenting literature in child psychology has a term for it: “authoritative parenting with respect for maturity demands,” which means, in normal language, that parents who show warmth, support, and reasonable control that reflects children’s changing maturational abilities promotes optimal growth.

The take-away from the puppy project as I see it, is the same for cognitive science as it is for parenting. Understanding intelligence (or intelligent action) requires close consideration of the interaction among the nervous system, the organism’s body, and the environment in which the organism acts. You can’t fully understand one element without consideration of the other two. Deceptively simple statement, that. For parents, what’s this look like? A little control, but not too much, and control that is sensitive and changeable with age, experience, and context, should do the trick nicely. Let’s let our kids slip, but not skid out of control. What do you think?

—

Note: the image above is from wikipedia, it is not the actual robo-puppy I reference here. Please link to the project author’s published work to see the critter as they designed it.