It’s okay to Karaoke (but why don’t you sing, instead?)

Today’s post focuses on the difference between memorising information and gaining knowledge. I don’t normally lead off with a descriptive sentence but I have to in this case because I’m going to talk about my previous life as a karaoke singer. I don’t want you to think I’ve accidentally posted on the wrong blog.

I quite enjoy singing and I have a reasonable karaoke voice – you’d never pay to hear me sing but you probably wouldn’t pay me to stop. When we used to sing in bars, when the machines first hit, you’d sing, people would listen, sometimes they’d clap and the assessment of the whole activity was based on how much you enjoyed it and, if you decided to compete, which of the dud t-shirts you won. For the record, my best placing was third place, a Yellow Cutty Sark t-shirt that made me look both jaundiced and leprous (a clever trick). I loved that t-shirt because, although third place was the best I ever achieved, but it was based on what the crowd felt and it was honestly earned. (I sang Prince’s Kiss in the style of Tom Jones, to complete this confession, so you may weigh the honesty for yourself.)

Of course, the fact that I can retrieve this fact, some 20 years down the track and well after the t-shirt was consigned to the rag pile, tells you a lot about the impression that this feedback had on me. However, the point of this anecdote is not that successful reinforcement is memorable (although that’s a good point) but that I wasn’t performing for points into a karaoke machine, I was using the technology to sing to subjective, analogue measuring devices – people, in other words. Now, the whole time I was singing, I had my own auditory feedback and (potentially) some crowd-based feedback but, being realistic, the only feedback I had was at the end if I won a t-shirt or if my friends either said ‘Yeah!’ or passed me a drink with a supportive expression. (Don’t judge me until you’ve attempted “Take On Me” by A-ha in a public space.)

Fast forward to today and you can have a karaoke machine in your home for the price of a PS2, PS3, X-Box or Wii. SingStar, Guitar Hero and Rock Band all provide you with the ability to play like a star and be adulated, and win awards, in the comfort of your own living room. Don’t like to rock out with pants on? Your lounge room – your rock and rules, baby! How, in this context, do we provide guidance that what you’re doing matches the song on screen? How do I award you a yellow Cutty Sark t-shirt?

In most cases, there is a ‘correct’ interpretation of the musical line – based on tone for singing and choice of input pad and timing for the other instruments in rock band. If you perform the right action (with varying degrees of tolerance) at the right time (again, with tolerance) you are recorded as having done the right thing. Based on this you get points, awards, more opportunities. There is nothing intrinsically wrong with this – this is a valid feedback mechanism. The problem arises when you don’t regard the score as assessing the activity, but the activity as a means for maximising the score.

I was slightly surprised the first time I saw Rock Band being played by a group and the singer, instead of singing, mumbled the tones into the microphone, at a volume I could barely hear, in order to exactly match what was going across his screen without the risk of missing the note or stumbling by getting a word wrong. Sure enough, excellent score. I’ve seen people ace SingStar while singing out of tune because they completely repeated every phrasing, every tonal movement and every pause, but stayed within the tolerance, despite it being nearly painful to the human ear. Why were they still singing? Look at the score! I’m doing it right! I bet I can make a higher score if I sing the same song another 10,000 times!

At the same time, I’ve seen professional singers get canned by SingStar and given terrible scores, despite putting together a beautiful performance of the song. What’s going on?

Anyone who’s played any of these games probably knows why this is: SingStar (and RockBand to a great extent) reward memorisation and recall, not interpretation. It’s the degree to which you can recite the song that matters, not your knowledge of how the song is constructed and how it can be rearranged, while still being that song – or better.

Musically speaking, it takes a reasonable amount of experience and knowledge to be able to start interpreting songs in a way that sounds good. Anyone can hit vibrato and wave around the song like a 90s boy band on a roller-coaster – it takes talent to harmonise, built, manipulate cadence and involve the audience in something that makes them breathless. It takes even more talent to know that you don’t do that all the time but in the right place and at the right time. SingStar has no easy way to assess anything outside of the norm so it will reward you if you hit note X at time Y within tolerance Delta. If you decide to mess with the cadence or add some colour, which are both demonstrations of a sound knowledge of the underlying work and the techniques that are valid for manipulating it, you won’t do well. This makes sense from an electronic game designer’s perspective – you can’t encode all of musical theory into the game but you can easily check for conformity.

(To its credit, Rock Band does allow some room for free-styling on certain instruments and at certain times. However, I think it’s probably good that Keith Moon is dead or he would have personally picked up every kit and eaten it after his first encounter with that particular game.)

It should come as no surprise that I believe that we often fall into the trap of requiring memorisation and recitation, effectively in lieu of demonstration of knowledge. Recited answers are easy to mark, manually and automatically. Knowledge requires the marker to be knowledgeable – subtleties abound and there are fewer templates. Electronic systems often lean in this direction for exactly the same reason as SingStar. A simple script to check the output of something is far easier to write than a detailed analyser with all human knowledge in this area.

Knowledge has to be built upon a strong foundation of information – there have to be core facts to provide a basis for your learning (such as the multiplication tables or the names of human anatomy, as two obvious examples). But at the higher level, at the 2nd and 3rd year level of University, we shouldn’t be rewarding people purely for mumbling our own words back to us in the microphone. We’re after knowledge and the demonstration of knowledge and that means more than training parrots.

A bit of karaoke never hurt anyone, but we in higher ed should be training our students to interpret to the fullest, to learn everything and to be able to show us what they know in ways that surprise, delight and, yes, challenge us.

They should be singing.