Thursday, October 11, 2007

Information theory for connecting with audiences

I just thought of a good way of understanding how to connect with audiences. It's funny, because instead thinking about audiences as a group, with various subgroups, as we have tended to so far, this approach is geared more toward the individual. It is to apply the psychological model of information theory. What better way is there to think about communicating with music than information theory, which deals with the communication of information?

The theory is that each communication consists of packets of meaning (either in word form or musical note form). Each word when communicated, gets your mind going, thinking about the meaning of the word, and then thinking about what might come after the word. Also, each word does not have to convey an equal amount of information. For example, take the word “God.” It has a great deal of meaning, bringing a bunch of associations in your mind. If you think about what might come next, there are so many possibilities. Then, say you add “save” onto it. Okay, so God is saving now. “Save” gives a significant amount of information that God is doing an action, namely saving, and also eliminates a lot of other possibilities. You would then probably think about who or what he is saying, and this would narrow your expectation to some noun. The next word is “the,” giving “God save the.” At this point, I think you would expect the next word to be “Queen,” based on your own personal knowledge of the world. If the next word is in fact “Queen,” then since you are already expecting it to be, it would not convey very much information. Suppose the next word had been “hippopotamus” instead. That would have foiled your expectations, surprising you, so “hippopotamus” would actually contain a lot of information. This is readily applicable to music, since there is so much setting up of expectation, resolving those expectations, prolonging them, or going against them.

And the second part to the theory is how amount of information relates to liking. If there is a lot of information in a communication, it is complex, and if there isn't, it is simple. If you imagine a graph where on the x-axis you have a range from simple (near the origin) to complex, and on the y-axis you having a range from not liking at all (near the origin) to liking a lot, the theory is that the graph relating these two parameters would be an inverted “U.” When a piece contains really little information and is simple, it will be boring so you won't like it. If the piece is really complex and you can't make heads or tails of it, you probably won't like it also. If the piece contains just the right amount of complexity for you, you will probably like it very well. Familiarity with a piece will also reduce the amount of complexity in a piece. This model explains why pop songs are immediately catchy but lose their appeal after repeated hearings. (It think it is telling that in today's society pop songs are designed this way. In the past, it was for one to immediately like the song when you heard it on the radio so you would go out and buy the CD. Now, an additional aim may be to have you impulsively buy songs off itunes. This culture of instant gratification not only makes the acquisition of music instantaneous, but also the emotional experience of music instantaneous.) It also explains why songs grow on you after repeated listenings: something that was too complex before is now more familiar and less complex. Also, we had talked in class about how the New World Symphony had introduced Shostakovich 9 and given the audience things to listen for before the performance, and how the audience really enjoyed it because they were able to hear some of the things they were listening for. The pre-program talk allowed the audience to have some expectation as to what was going to happen, and allowed them to make some sense of a complex jumble of notes. It also explains why atonal music is so difficult. It is very hard to predict what note will come next, unless you're a serialism expert, or you are already familiar with the piece. I think with atonal music, you expect to be surprised. My personal feeling is that liking atonal music means you are smugly surprised.

As for a general way of thinking about how to connect with audiences, it is important to recognize that the music we play is part of the Western music tradition. Our aim as classical musicians is not to just produce pleasant sounds to listen to. We want to communicate the meaning that the piece has for us. But the music only has meaning for us because we understand the expectations. When we talk about the progress of Western music, even something as radical as the development of atonality, it is all progress within the tradition. Because composers compose in dialogue with past composers, the tradition is continuous. The motivation for never straying too far from the tradition is, I believe, complexity. Let alone the listeners, the composer himself must understand the new music he is writing, and his taste in complexity is based on the Western tradition. It is obvious that what we think of as good music of the 20th century, is not palatable to much of audiences because it is too complex for them. They have not had as long a time as we to familiarize ourselves with the development of Western music. We have to be careful not to become too Western music-centric. I believe information theory provides a good way of judging what music audiences would enjoy and for judging the amount of preparatory work we have to do in order to let audiences enjoy more complex pieces.


Lea said...

hey Richard, this is quite challenging.I don't agree at all, although you make some good points. I'm just wondering how you can express something like aesthetic value through a coordinate system. And what about Mozart, how does his graph look? And what about a piano piece by... Taussig or Gruenfeld? These are composers of pieces with lots of notes and chromatic (very difficult), but not very interesting in general.

Richard said...

Yes, I agree that aesthetic value cannot be measured on a coordinate system. When I had learned about this before, I had thought the same, about how its measure of music is very unspecific. But I think in so far as measuring people's personal liking for music, which may not correlate with an absolute score of aesthetic value, I think it is valuable.

The graph would not be the same for every person for every piece. And the this is just a theory, because there is no way to measure how much information each person actually perceives in a piece of music. But it makes sense that one can get information from listening to a piece, and that to like something you have to be interested in it (meaning that there is enough information there for you to perceive). Take Mozart for example. I might register a lot more information when listening to Mozart than somebody who was not musically trained. I might notice that he ends his phrases with the same figure, or be surprised when he foils my expectation, or modulate to somewhere I don't expect, or how he plays around with sonata form. (The specifics of expectations in music as codified by Western music is explained by the study of music theory.) This information makes the piece complex for me and thus interesting enough for me to like it. Someone untrained might not perceive all that, so the piece will not be as interesting for them. If the piece holds less information for them, they might not like it as much.

As for Taussig or Gruenfeld, could they be an acquired taste? Could just not understand what they have to say, and will once we are used to their language? I think when we judge composers, we are judging them very much based on the Western tradition. Now this is a valid measure if the composers themselves choose to work solely within the Western tradition. When you say that they aren't very interesting, what would make them interesting? Is it that they don't provide enough surprising writing? By surprising I mean non-mediocre when judge against the rest of Western music. If they are repeating the same things, doesn't that mean they aren't saying much non-commonplace things according to the Western tradition, and thus you find them less interesting?

There was one finding though that leveled the difference between musicians and non-musicians though. That is the perception of tension in music. Beyond the musical notes, there is a human expression ,a "bluster," if you will, that is readily perceived by everyone because it is related to urge to communicate. Like you might communicate you are hungry without words by shouting, "Ungauh!" in a distressed way. (I highly recommend S.I. Hayakawa's "Language in Thought and Action" for more about this) Maybe the presence of this expressive intent in music can be one aspect of a universal criteria of aesthetic value?