https://commons.wikimedia.org/wiki/File%3AJive_Langella_Moshenska_1107.JPG

How learning science might inform dance: A cognitive psychologist responds

Dr Yana Weinstein (@DoctorWhy) responds to my speculative post about cognitive science and ballroom/latin dance.

I’ve been a professional cognitive psychologist for 10 years and an amateur Latin dancer for 2, so I couldn’t not get over-excited when I saw this post. Unfortunately, none of the cognitive psychology research has been carried out with dancers (at least, to our knowledge), and I’m currently suppressing a very impractical urge to go in this direction! (I’ve already gotten myself in trouble with a project on music memorization – you have no idea how hard it is to branch out beyond generic cognitive tasks to this kind of applied work).

So all we have are the rather well-informed speculations discussed here. Let me speculate further.

1. Chunking – I wonder if it’s not an accident that my teacher always breaks up new choreography into 15-30 second chunks for us. I’m sure she’s not aware that this is the duration of working memory, but maybe she naturally defaulted to this duration.

For this part of the suggestion – “dancing with a partner or being aware of others watching you may be factors that overwhelm or impede learning” – there is also some relevant research. For example, it has been suggested that certain types of test anxiety may lead to off-task thoughts that take up part of working memory resources.

2. Interleaved vs. Massed practice – so much to say here. First of all, don’t confuse spacing with interleaving (it is extremely easy to confuse the two, and honestly, I still do myself sometimes even though I co-wrote a chapter with a large section on each one ). For example, in the main point, “you’ll end up with stronger memory of routines if you interleave dances” is about interleaving, but in the suggestion, “it has been shown that by giving ourselves a bit of time to forget an idea then […] the long-term strengthening of memory of this approach is much greater” is about spacing.

Spacing is more about how you would distribute practice over a period of time. If your performance is in a week, should you practice for an hour every day, or 7 hours the day before? I think we all know the answer.

Interleaving has been most commonly studied with math problems: should you do 10 of the same type of problem, or mix it up? What’s tricky and confusing about the two is that interleaving naturally adds in spacing, so it is hard to have a pure measure of the effectiveness of interleaving without also implicating spacing. However, this paper  attempts to hold spacing constant and still finds a long-term benefit of interleaving.

I don’t know if this has been explicitly specified, but my strong suspicion is that the reason interleaving is helpful is that it enables you to practice doing a problem “cold” (i.e., without the carry-over effect of having just done a similar problem). It’s important, though, to consider what your goal is. If you tend to go to dances where you’re expected to dance salsa and then bachata, interleaved (as I do), then yes – practice them interleaved. But if the two dances tend to be danced in separate parties (e.g., salsa and tango), then I’m not sure interleaving is going to be as necessary, because you will always get to warm up and then after that you’re in a massed testing situation. The grain size of what you are interleaving also matters. You could take this to the absurd: do you need to interleave practicing one rumba dance with doing one yoga pose? Probably not.

Also, I suspect that level of familiarity with the material is also a factor, and some massing is necessary before interleaving can be useful. Have you noticed how a teacher will always introduce a small chunk (cf., chunking), and then you will mass-practice it until it feels comfortable, and only then practice it with whatever comes before and after? It seems obvious now that I think about it, but clearly what this is doing is getting the information from working memory into long-term memory! Think about how the first few times you run through a new chunk, you’re really just copying the teacher in the moment, so you’re not really using your long-term memory. So maybe the ideal practice protocol includes some massing to transfer information to LTM, and then lots of interleaving. But does this mean you should scramble the order of chunks within a choreography and practice them in different orders? This is the same as the rumba vs. yoga pose question, and I’m honestly not sure.

3. Transfer – the suggestions are great, but the main point is more about context than transfer. When we talk about transfer, we usually mean to a new problem (rather than the same problem in a different environment). For example, if you know salsa, how well will you dance bachata? Probably pretty badly until you get taught at least the basic step and some basic moves – and then you’ll probably pick it up faster than a novice dancer, showing some transfer. A more subtle example of transfer might be: say you were taught musicality with a particular song (e.g., here’s a good part of the song to do a dip). An example of transfer would be noticing those moments in a different song.

4. Still images and attentional focus – the point about learning from still images better surprises me, but I must not know the research on moving vs. still images. I just can’t imagine that any better way of demonstrating a dance move could involve still images, but I’m willing to consider the possibility. I suspect, though, that some research (that I either do not know, or cannot recognize from the description) is being misapplied because the learning outcome (what we call “criterion test”) of the study was not mimicry of a physical movement. The idea of attentional focus during learning is also fairly unfamiliar to me, although it definitely makes sense, and perhaps could be addressed by having the dancer wear bright red shoes or bright red gloves depending on whether they were demonstrating feet or arms??

5. Self-explanation – this seems reasonable, and on a totally personal level, I have to say my instructor doesn’t really let us ask questions. I mean, of course she clarifies specific points, but I tend to ask a lot of questions, and she has more or less shut me down on some occasions because it was slowing down the class. Now, as a teacher I completely appreciate the frustration of tangential, time-consuming questions, but I like to think (probably erroneously!) that my questions are relevant. It just feels as though she has a certain amount of material to get through each class, and any elaborative interrogation (and it’s really not that elaborative – I’m usually just trying to figure out exactly how something fits together) cuts into that time. Perhaps a discussion session after class would be valuable.

6. I don’t know the research on putting two concrete examples side-by-side, but I feel that this point also addresses the value of feedback. To which point, maybe it’s useful to mention that the jury is still out on whether immediate or delayed feedback is best for learning.

Overall, though, here’s my take on learning dance: I feel like we are generally pretty well aligned with the cognitive principles. Most importantly, I think – and David didn’t mention this in his post, possibly because it’s too obvious – we make constant use of the testing effect. We learn by doing the dance, over and over, from memory. We don’t sit for an hour listening to our dance teacher describe the moves and taking notes; we don’t practice by watching the teacher’s performance video over and over again (though we do consult it in between attempts to get it right); and the teacher doesn’t just call on one student at a time to do one small part of the routine while others watch passively. Why, then, is the same not true in a typical school classroom? Why aren’t our kids repeatedly practicing retrieval of the information the teacher is trying to teach?

And here is a dress rehearsal of my latest routine with Leah’s Chicas! 🙂

3 thoughts on “How learning science might inform dance: A cognitive psychologist responds”

  1. I am curious as to your take on “muscle memory” in this context. That is, are there other factors (other than cognitive psychology) that come into play when learning muscle movement? To me, while interesting to compare the mental aspects of dance to cognitive psychology, I wonder if there may be additional items to keep in mind here since we are dealing with movement (I am under the impression that movement kind of learning happens a bit differently than thinking kind of learning)?

  2. Oh! I would say that “muscle memory” is what we call “implicit memory” in cognitive psychology. So, for instance, you may not be able to describe the movements, but you can do them. For instance, the famous amnesic patient H.M. could learn mirror-drawing (an implicit task), but could not explicitly retrieve any new information. Whether he could learn dance choreography is sadly a question we will never know the answer to, since he is now dead.

    1. Interesting! Perhaps this suggests then that implicit learning occurs in a different part of the brain (this seems reasonable to me)? So in effect, when dancing and learning choreography, you are actually utilizing two brain functions (one for encoding the implicit task, and one encoding the new information)? Perhaps this is where the myth that ‘moving while learning is helpful’ ?

Leave a Reply