Why does a Serialist composition sound like a random one?

Question

Why can random notes and a structured Serialist composition sound the same? This Reddit comment cites Ligeti:

The irony, noted Ligeti, was that the kind of music pioneered by Boulez and Stockhausen was actually extremely structured -- but it was so structured that the resulting music came across as random to the listener because the structure was so dense that it was hard to grasp:

"Ligeti expressed that division with forensic detail in 1958, in an analysis of Boulez's infamous piece Structures Ia, one of the icons of the short-lived experiment of "total serialism", in which every aspect of music - not just pitch, but volume and rhythm as well - are subjected to systematic organisation. He revealed the limitations and contradictions of the technique, and the paradox that the results of total determinacy actually sound random, chaotic, and indeterminate.

"Boulez was incredibly angry after he read the article," Ligeti remembers. "He had been very nice with me when we had met before, but suddenly he did not speak to me any more. For 10 years, he was a complete enemy. [...]"

Prof. Dan Román MM DMA (Hartt School of Music) from Rethink: The Abyss of Music in the 20th Century: (Paraphrased by reason of poor audio quality)

It [a Pierre Boulez's serialist piano composition] sounds even more chaotic. I had a college professor in music who loved to demonstrate a kind of competition: He recorded a performer just improvising notes and rhythms at random. Then he would play the random sample and the Boulez piece back-to-back to a room of students. To the students, both examples sounded the same.

So we have here an example of a composition which is completely structured - all its music determined by a discreet series of numbers, and another piece of music that is virtually unstructured entirely, yet to the listener, they sound the same.

David Bruce MComp, PhD in Composition at King's College London in Music vs. Pattern:

09:02 Of course we can also take things
09:04 too far the other way. Around the same
09:06 time as those total serial pieces John
09:08 Cage was writing his music of changes. I
09:11 say writing, but he was actually using
09:12 chants based on the Chinese I Ching to
09:15 dictate the score. So you could say this
09:17 is the total removal of pattern. Now what
09:20 has struck many people is how similar
09:21 these two apparently polar opposite
09:23 approaches end up sounding. This is
09:25 Stockhausen's Klavierstück III,
09:27 the completely patterned
09:30; and this is Cage's Music of Changes, the
9:33 completely patternless.

Answer 1

This is a time where the "music as language" metaphor can be helpful. Think of a language you don't speak; let's say Hungarian. When you hear it, so much of it tends to sound random, no? It's not until you study the language and understand its grammar and sounds that you become more aware of the structure. (I also have my own thoughts on how not understanding a language tends to make people view that language as sounding "hostile" or "angry," and I wonder sometimes if the same is true in music.)

The same is true, I think, of total (or complete, or integrated) serialism. It may sound random at first, but it's in fact the opposite of random, and we can recognize that once we understand the grammar of the piece.

As for the little experiment in the quote, I wish they would have given a little analysis of the serialist piece and played it again to see if the listeners had a different reaction to it. Otherwise, the experiment is no different than someone playing you a recording of flawlessly spoken Hungarian followed by this or this, two examples of absolute gibberish that were designed to sound like English.

Answer 2

Let me attemt a mathematical answer and start by giving an analogy.

It is very hard to generate actual random numbers in a computer, since programs run deterministically. What you want is numbers that, over the long run, appear with equal likelyhood and that have no correlation to other numbers in the sequence. So what is done instead of actual randomness is to use pseudo-random numbers - you come up with an algorithm that runs through all integer numbers in a given range (let's say, 0 to 2^32) exactly once, and then starts again, and you choose it in a way that the next number bears as little obvious semblance to the current one as possible.

Now, in music, "normal" compositions impose a structure on the tones that are played. Certain notes are given special functions like the tonic or the dominant by how often and in what places they are played; there are correlations between consecutive tones and chords, like the dominant leading back to the tonic, etc.

Just like with random numbers, there are two ways of breaking up these structures: either by actual randomness - you just play whatever - or an algorithm that mimicks some aspects of randomness, like using all 12 tones exactly once before you can start again, which gives you exactly equal probablilities. So in that sense, serialist compositions have a strong pseudo-random aspect.

Answer 3

Why does a Serialist composition sound like a random one?

Most serialist compositions are essentially random in the way that humans perceive music.

A typical serial composition employs complex, dissonant harmonies without resolution. Listeners are deprived of understandable, consonant harmonies and constant strong dissonances are statistically what you get in random successions of tones.

And in trying to break new ground rhythmically, serialist composers either use a great variety of durations, replete with long pauses and rapid runs of notes, or serialize the durations along with the pitch classes.
In either case, it becomes much harder for the listener to detect any rhythmic patterns.

When both the tone sequences and durations in a musical piece do not appear have any perceivable structure, the piece will sound random.