This post was inspired by a really great question that a blog reader named Susan posted on the Felix Forum this morning. She pointed out that though it’s easy to talk about overimitation as a humorous quirk of children’s learning, it’s not really clear that we as adults are immune to the same kinds of errors. I couldn’t agree more: when it comes to the seemingly “obvious” mistakes that overimitation can introduce, adults definitely don’t have the last laugh.

It’s funny (if not altogether shocking), but the anecdotal evidence on this point seems to come primarily from situations in which adults are interacting with computers. Actually, there was a blog post to exactly this effect just a day or two ago from an IT professional named Anders Aspnas in Finland. On his Efficiency Limit blog, Aspnas comments that:

“I guess [overimitation] occurs a lot in IT… Part of what is done has no real relevance to the problem at hand. The procedures just happen that way as a result of accidents being replicated… just because someone happened to originally do it that way. Somebody may finally realize that some parts of procedures are not bringing any value and improve, but that seems to take ages.”

My own experience definitely mirrors what Aspnas is pointing out. In the process of converting my parents to using Macs, for example, I’ve had many instances of noticing odd idiosyncracies in the procedures that they’ve learned, many of which turn out to be attributable to some incidental thing that I did in the process of demonstrating for them. Blog reader Susan makes a similar comment here, about how her mother “overimitated” the ritual of closing all the windows open on the desktop before launching a web browser. Having seen Susan herself do it, she was convinced that Firefox just couldn’t get off the ground any other way.

In all honesty though, this phenomenon really isn’t limited to novice users. I may have degrees in computer science and AI from Oxford and MIT, but I’ve definitely still done my fair share of computer-related overimitation. On the left is an example screen capture taken from some of the development tools we used at MIT. Click on the thumbnail for a larger view, then look at the last checkbox down toward the bottom – the one that says “invoke real build tools instead of faking them.”

What on earth does that mean? I haven’t the faintest idea.

I can tell you though that if I’d seen a more senior and experienced team member check that box in the course of debugging a project, forever after I’d have made certain that it was checked whenever I had to troubleshoot something myself. I did this kind of “overimitating” constantly at MIT, and I’m sure I wouldn’t have been able to accomplish nearly as much while I was there if I hadn’t. Using other’s knowledge as a catalyst for your own learning (even when we don’t fully understand the details) isn’t just a good strategy, in my opinion -it’s an important aspect of human nature. When we need to extend our own understanding, the knowledge of others is often the first resource that we call on.

What do you think? Do examples like this count as overimitation in the same way that children inefficiently opening the Puzzle Objects does? Please do take a moment to comment! You can do so right here on the blog through the “Comments” link below, or on the new Felix Forum area of the site.

Related posts

• No related posts.

Yesterday proved to be an exciting milestone for me, as the first portion of my overimitation research was officially published by the Proceedings of the National Academy of Sciences. It’s been a long time coming, and I’d really like to thank my co-authors and friends Andrew Young (now in his first year of the PhD program at the University of Wisconsin-Madison) and Professor Frank Keil for sharing all the hard work that helped to make it happen. I’d also like to thank Frank for graciously footing the additional cost of publishing the article as an “open access” submission. This means that anyone, anywhere can download the complete article from the PNAS website, without the need for a (very expensive) subscription. You can also check out the extensive online supporting information that we submitted to supplement the print article, including some really great video footage of children participating in the experiment. The videos are quite entertaining if I do say so myself, and they really help to convey what overimitation is in a way that’s hard to do in just words, so be sure to check them out.

You can access the PNAS article right here.

Also, if you’ve been reading this blog but haven’t yet ventured over to see the rest of the site, I hope you’ll take an opportunity to do so now. I’ve been working overtime to get the site up and running in time for the article’s release, and it’s been really fun to see the first wave of visitors rolling in as the overimitation story percolates its way across the internet. One of the real highlights so far has been having the story picked up by Carl Zimmer, the prominent science journalist and frequent contributor to the New York Times, among other places. I had the pleasure of meeting Carl for the first time about two years ago, when we wrote an article for the Times on the start of my research. It’s been fun then for me to see things come full-circle with his latest blog post about this work. You can read Carl’s comments on overimitation at his science blog, The Loom.

Returning to thinking about this site, as I’ve discussed in a few other contexts, my vision for Hello Felix is to create a trustworthy, authoritative source of news and information for anyone interested in learning more about the science of childhood development and learning – especially parents. Cognitive scientists are doing so much fascinating work on how kids think, I hope that this site will help to increase the level of interchange between the people who study kids in the research lab and the people who, well, raise them. It’s an exchange that would be a real benefit for everyone involved, I know, and a lot of fun at that. So please check it out!

Over time the site will be expanding to cover lots more topics and varieties of developmental psychology research, but as a starting point I’ve put together a nice guided tour of overimitation and imitative learning that I hope people will find useful.

As always, I’d love to hear any comments or feedback you might want to provide. Please feel free to comment on articles on the site (look for the links at the bottom of each page), or to send me an email directly. Also, I’m also pleased to say that the brand new Felix Forum is now up and running! I hope you’ll stop by to see what people are discussing, or better yet to start a discussion of your own. I’m online posting and answering questions each day, so all new user submissions are sure to get a quick response!

Related posts

• No related posts.

In my last couple of posts I’ve been talking about a curious quirk of children’s learning that I call overimitation. Basically, when kids see an adult operating an unfamiliar object in a noticeably roundabout, inefficient way (for example, mixing unnecessary, nonsensical actions in with the actions that are actually needed), the kids are much more likely to copy those unnecessary actions than to edit them out. They overimitate by copying things that they could just as easily ignore and still accomplish their goal. Though it’s easy at first to chalk this persistent copying up to the fact that kids just enjoy imitating others, the sheer persistence with which most kids overimitate quickly becomes surprising. In fact, as I talked about in some of my last posts, the majority of preschool-aged kids will precisely copy actions that are so inefficient that even chimpanzees know to ignore them!

So here’s the big question: does overimitation tell us anything new and important about the way children learn? Or do kids overimitate just because they find it fun to copy others, even when the actions that they are copying aren’t very useful?

When my colleague Andrew Young and I started pondering this question, we decided that there was a simple and direct way of telling these two possibilities apart. Namely, we’d simply teach kids not to overimitate. If kids could get the idea right away, and start imitating only the useful parts of an adult’s demonstration, this would indicate that overimitation was something kids did just for fun. On the other hand, if kids couldn’t get the idea, and were pretty much unable to stop copying the unnecessary parts of an adult’s demonstration even when they were trying to do so, this would suggest that overimitation was a lot more important than psychologists had realized.

So we decided to put our idea to the test with what is arguably the most important tool in the developmental psychologist’s tool belt: plastic dinosaurs. In both our lab at Yale and in several dozen preschools all around Connecticut, we introduced preschoolers to a rather hapless adult psychologist, who needed help figuring out how to best get plastic dinosaurs out of a bunch of different household containers, shown below. With the jar item in panel (A), for example, the adult got the dinosaur out by first tapping on the jar with the feather (an irrelevant action), and then unscrewing the lid (a relevant action). The adult then asked the child to help troubleshoot his strategy, asking them questions like “Did I have to tap on the jar to get the dinosaur out? Or was tapping something extra?” Kids were asked these questions about both steps in the procedure, both the unnecessary feather tap and the necessary removal of the lid.

Unsurprisingly, by the age of 4 or so most children are able to easily identify the feather tap as “silly.” Actually, they often find the question easy enough so as to be rather amusing, and are very good at explaining to the clearly confused adult why the feather tap is not at all necessary for getting the dinosaur. We went through this whole procedure about seven times in total with each child, giving them a chance to observe the adult open each of the training objects shown above, and then to point out which parts of his actions were silly.

Right then – so preschoolers are clearly able to distinguish between things someone has to do to accomplish a goal, and things that are extra. At least, that is, when accomplishing the goal involves manipulating an object that they are familiar with. But what about objects that they are less familiar with?

When each child finished the dinosaur game, they then moved on to playing the closely related turtle game. Just as in the dinosaur game, in the turtle game kids first watched as the adult retrieved a toy turtle from inside an object. This time, however, the object in question wasn’t a familiar household container, but rather a spectacularly unfamiliar Puzzle Object such as the ones shown here.

Now here’s the important part: the exact way in which the adult retrieved the turtle was very inefficient, incorporating lots of superfluous, unnecessary, and generally “silly” actions. For example, on the Puzzle Box the adult first went to great pains to remove the red wooden bolt, then tapped carefully in the visibly empty top part of the box, and only then actually did the one thing that was obvious from the start, namely removing the shiny red door on the front of the object.

So children saw the adult retrieve a turtle from the Puzzle Object, but in a way that included lots of the same kinds of “silly” unnecessary actions they had just been trained to be on the lookout for. Then the moment of truth arrived: explaining that she needed to check on something in the other room, the adult left the child alone with the object. On her way out, she casually suggested to the child, “If you want, you can get the turtle out while I’m gone. You can get it however you want.”

How did children operate the puzzles for themselves? Remember, just moments before playing with the puzzles they had gone through a long (very long, if you’re four years old!) teaching exercise in which:

• They were shown over and over that the adult frequently does things that are rather silly, and not actually necessary for her goals.
• They were given lots of praise and encouragement for identifying those unnecessary actions, and explaining exactly why they were silly and did not really need to be done!

The amazing thing is, despite all of this teaching, the vast majority of the preschoolers that we tested continued to overimitate the adult’s unnecessary actions. Rather than just getting the turtle out in a simple direct way, they expended considerable time and energy meticulously copying all of the adult’s actions, even those that should have been clearly unnecessary.

This result was very unexpected to us – and extremely interesting. It suggested that overimitation wasn’t something that kids were doing “just for fun,” but rather – it seems – something that they really might not be able to avoid doing, even when they wanted to! It’s as though having watched the adult perform her unnecessary actions on the puzzles, kids became firmly rooted in the notion that those actions were actually necessary to get the turtle out.

Related posts

• No related posts.

In my post yesterday, When is a Chimp Smarter Than a Child?, I introduced a curious result from a recent comparative psychology study. To recap: both children and chimps were shown an extremely simple, transparent Puzzle Box, and then watched as an adult retrieved a prize from inside it. Rather than getting the prize out in the obvious direct way though, the adult first performed a number of superfluous actions such as tapping on the top of the box and unnecessarily removing a sliding bolt. When the chimps and kids later had an opportunity to get a prize out for themselves, only the kids wound up copying those irrelevant actions; the chimps, contrastingly, succeeded in ignoring the unnecessary steps and just going straight for the prize. On the surface then, it’s the chimps who seem to have unexpectedly come out on top in this task.

So what does this mean? The dissertation work that I’m currently completing at Yale began with this exact question. What does this odd situation, in which kids suddenly seem to be less clever than chimps, have to tell us about childhood cognition?

One possibility, as scientists and parents alike have pointed out, is that it may not tell us very much at all. That is, there are a number of plausible, deflationary explanations for children’s behavior in this task, none of which would be any great shakes scientifically. For example:

• The children may have assumed that they were “supposed” to copy the adult; they may have thought that they were playing an imitation game in which the whole point was to copy what the observed.
• They may have been reluctant to contradict an adult authority figure by ignoring her actions and choosing a more efficient route to the prize.
• They may have wanted to please the adult by copying her actions.
• They may have felt that the adult understood something non-obvious about the structure of the Puzzle Box, and thus that she had a reason for performing the seemingly unnecessary actions, even if it was not obvious.
• Since children learn so much from imitation, it may have been less effortful for them to simply copy what was observed rather than critically analyzing each step.

Unfortunately, Horner and Whiten’s original study didn’t rule out any of these reasonable explanations. Because of this, many psychologists assumed that there wasn’t much to the chimp/child contrast in this case. It was assumed that children overimitated, or copied the adult’s unnecessary actions, not for deep cognitive reasons, but instead because for one or more of the comparatively uninteresting social reasons outlined above.

It seemed to me though that while these deflationary explanations were indeed plausible, more interesting alternatives also existed. In particular, I wondered if perhaps observing the adult’s actions was changing the child’s understanding of the object in a more concrete, cognitively significant way. Perhaps somehow after watching the adult retrieve the prize inefficiently, children were really getting stuck thinking that one had to perform all of the observed actions to open the Puzzle Box? Essentially I wondered if children might be copying the adult’s unnecessary actions because of a “causal illusion” not unlike an optical illusion such as the Muller-Lyer effect illustrated below:

Consider the two horizontal line segments in the image. Excluding the arrow points on their ends, which horizontal segment seems to be longer? For most of us, it’s hard to resist the strong intuition that the bottom line segment is slightly longer than the top. In fact though, they are exactly the same length. (See later in this post for a link to a possible explanation). However, even once you’ve measured the line segments for yourself, it’s still impossible to look at the image without feeling as though the bottom segment is definitely longer.

I wondered if the children in the original Horner and Whiten study might be experiencing a similar kind of cognitive illusion, this one of the causal variety. That is, perhaps after observing the adult, they were left with the strong (false) intuition that all of the adult’s actions – even the ones that the chimps identified as unnecessary – were really somehow important for getting the prize out. Just as the visual system is tricked by the Muller-Lyer illusion, I wondered if children might be being misled by the way their brains processed and encoded the adult’s actions on the Puzzle Box.

Working with my colleague Andrew Young (now in the PhD program at the University of Wisconsin-Madison) I set out to answer this question by running a new study, this one involving more than 100 preschool-aged children from all over Connecticut. As in the Horner and Whiten study, the basic structure of the experiment was pretty simple: the kids watched as an adult retrieved a prize from a simple puzzle object, and then later had a chance to retrieve a prize for themselves. However, our study also included an important twist, one that was critical for addressing the deflationary explanations that I discussed above. Specifically, Andrew and I deliberately structured the task to make it as clear as possible to children that they were not supposed to exactly copy what the adult did. Instead, our task biased them to be on the lookout for “silly” unnecessary actions, and then to avoid reproducing them when acting on the puzzle objects themselves. In psychology jargon, we introduced implicit task demands that would encourage children to ignore what the adult did, and instead to operate the puzzle object in the most efficient way possible.

In my next blog post, I’ll talk about the methods that we used in more detail, and then introduce our initial results as New York Times science journalist Carl Zimmer described them. Mr. Zimmer’s article on what we discovered is extremely informative and entertaining, so stay tuned!

In the meantime, here’s a link to some more information on the Muller-Lyer illusion that I discussed above.

Related posts

• No related posts.

So here’s an interesting question: when is a chimpanzee smarter than a child? Such situations, it turns out, do occur, and under surprisingly everyday sorts of circumstances. A particularly striking recent example comes from the research of two comparative psychologists from the University of St. Andrews in Scotland, Dr. Victoria Horner and Professor Andrew Whiten. Horner and Whiten showed that when it comes to learning how a simple “Puzzle Box” works by watching an adult, it’s easy to make chimps appear far cleverer than kids. All you have to do is make the Puzzle Box very, very simple.

For example, take a look at this Puzzle Box that my colleague Andrew Young and I built, based on Horner and Whiten’s model.

There’s not too much to it – it’s basically just a Plexiglas cube with a divider in the middle. Now imagine that you’ve been told that there is a prize inside the box; where is it and how would you get it out? Since there’s only one place in the box that is opaque (the blue tube in the lower compartment) it seems pretty clear that that is exactly where the prize has to be. And that big red door on the front just begs to be pulled off, so that seems like an immediately obvious and expedient way of finding out what’s inside.

All of that is true, but when Horner and Whiten showed a Puzzle Box just like this one to both chimpanzees and preschool-aged children, they added a twist. Before the chimps or kids could retrieve a prize, they saw an adult get one out for herself. Rather than just getting the prize out directly though, the adult did it in a very circuitous, roundabout way. Take a look at this picture and you’ll see what I mean:

The adult began by using a small wand to (1) tap on the bolt on top of the box, (2) push the bolt out of its frame, and then (3) tap on the floor of the empty upper compartment. Only after all of this rigmarole did the adult do the thing she actually had to do to get the prize, namely moving the door and sticking the wand inside the opaque compartment.

After the adult had her prize, both the kids and the chimps both had a chance to get a prize out for themselves. How did they choose to go about it?

The chimps, it turns out, weren’t impressed with the adults roundabout technique. Rather than imitating what the adult did , the chimps just cut right to the chase – they skipped over the unnecessary steps (the ones shown in yellow) and instead just did the necessary things to get the prize out (the steps shown in blue). Score one for our chimps relatives.

Based on what the chimps did, I think most people would expect – as I certainly did – that kids would take a similarly smart approach. But that’s not what happened at all. Instead, fully 80% of the children copied all of the adult’s actions, including the unnecessary steps that the chimps couldn’t be bothered to reproduce. The kids, in other words, approached the task in a way that was a lot less clever than the chimps.

Now to be sure, there are a number of simple explanations for this result. For example, maybe the kids just thought it would be fun to use all the parts of the box. Maybe they thought that they were supposed to copy the adult, and felt sheepish about ‘contradicting’ an authority figure by ignoring part of the display. Or maybe the kids are just in the habit of imitating, and simply copying all of the steps seemed like less work to them than thinking it through.

All of these are possible, and even plausible, explanations. They are pretty much what Dr. Horner and Professor Whiten themselves said about this curious result. But when I first came across this work, I wondered: might something deeper be going on? Is it possible that the kids aren’t copying the adult because the want to or think that they should, but instead because they are confused in some interesting way about how the Puzzle Box works? I’ll return to this question tomorrow as I begin describing the first in a series of experiments I did to get to the bottom of this mystery.

Related posts

• No related posts.