Will Future Generations Listen to Music?

Answer: probably not; it seems destined for obsolescence

Section 1. Many artists see their art as a way of achieving a kind of immortality: by creating something that survives time and is appreciated by future generations, the artist “lives on” through his or her work in the memories of such individuals. Having dabbled in future studies, [1] I’ve often wondered what future music will be like – will there be genera of music that are completely new and distinct? It’s as hard for me to imagine such genera as it is to think of a color that’s outside the color spectrum. (I wonder if any musicians in the sixties could possibly have anticipated rap music, or whether Mozart could possibly have imagined a style of music like jazz.)

Even more, I wonder whether music has a future at all. One could, I believe, make a good case that future generations simply won’t be interested in – except maybe for historical reasons – the sort of music that humans currently listen to. The argument is basically this:

1) music is very much a human (or species-relative) phenomenon, for both non-phenomenological and phenomenological reasons.
(2) There are good reasons for thinking that our descendants won’t be human. Therefore,
(3) there are good reasons for thinking that our descendants won’t appreciate music the way we do, or will have some new form of sound-based expression (call it “music” with scare quotes) that takes the place of music. I will begin with the second premise, and then move on to the first.

Section 2. There are two broad phenomena that could result in our descendants being so different from us that we’d be inclined to call them posthuman rather than human.

The first may be called cyborgization. This refers to the merging of biology and technology, organism and artifact. The result is a biotechnological hybrid, or cyborg, which stands for “cybernetic organism.” While talk of cyborgs might conjure up fantastical stories told by sci-fi writers, the fact is that our contemporary world is full of cyborgs. I would argue that every contemporary human is already a cyborg, in some nontrivial sense, although some of us are more cyborgized than others. Similarly, the cognitive scientist and philosopher Andy Clarke argues that humans are all “natural-born cyborgs.” That is, since our debut on the evolutionary stage, humans have used inventions ranging from natural language to the iPhone to literally extend our minds beyond the “arbitrary boundary” of “skin and skull.” (For more, see The Extended Mind hypothesis.)

A simple example of cyborgization is the use of clothes – a technology – to extend our skin, thus making it easier for our bodies to thermoregulate.[2] As Daniel Dennett writes in Consciousness Explained, clothes “are part of the extended phenotype of Homo sapiens in almost every niche inhabited by that species. An illustrated encyclopedia of zoology should no more picture Homo sapiens naked that it should picture Ursus arctus – the black bear – wearing a clown suit and riding a bicycle.” Another simple example is the use of glasses (or contacts) to mediate our perception of the external world. According to the philosopher Don Ihde, glasses are “embodied” such that they become self rather than not-self in a phenomenological (or experiential) sense.

Other examples include the use of pacemakers to regulate heartbeat, cochlear implants to restore hearing, deep brain stimulation to treat dystonia and Parkinson’s disease, and prosthetic limbs that are myoelectrically controlled. In all these cases, the individual’s phenotype (or observable features) consists of both biological and technological components that are cybernetically merged together to create a single, unified “organism.” (Note the Greek and Latin etyma from which this term derives, which literally mean “instrument” and “artifact.”)

Music provides many examples of cyborgization. The contemporary musician performing on stage, for example, sings into a microphone that amplifiers her voice; in the process, effects like reverb and even pitch-correction may be added. The result is a sound that’s both biological and technological in nature. She may also play an instrument that she straps to her body and which produces sounds that neither the instrument nor the musician alone could create. The human is thus a fleshy cog embedded in a larger machine of cyborgizing technologies – many of which engage musicians in embodiment relations (following Ihde). But music has always been technologized. Indeed, there are musical instruments like flutes made out of bone that have even been found with Neanderthal remains. [3]

If trends of cyborgization continue, the earth will someday be crowded with individuals who are only faintly recognizable as human – or not recognizable at all. Some may contain increasingly more technology until there’s no biology left and the cyborg becomes an android. Perhaps the most important changes will be those that affect the cognitive-emotional and perceptual capacities of our descendants. (Indeed, humans have already created machines that can perform all sorts of physical activities, such as moving across space and lifting heavy objects, far better than we can. But such properties aren’t considered to be close to what makes humans unique – cognition, emotion and perception are. This is one reason that many theorists find the Turing Test compelling: it requires a computer to linguistically express thoughts and feelings just like a human. If the computer consistently succeeds in doing this, we have some reason for attributing to them a mind.) In terms of nomenclature, technologies that augment our mental abilities are cognitive enhancements, while those that expand our perceptual abilities are perceptual enhancements.

There are a range of radical enhancements currently being researched and developed today. Such enhancements include pharmaceuticals (or nootropics) that could significantly increase our memories; neural implants (or BCIs, for brain-computer interfaces [4]) that could enable us to access information by searching Wikipedia simply by thinking; and genetic modifications that could significantly enhance our ability to learn. Cognitive enhancement may also be facilitated by uploading one’s mind to a computer (assuming an almost universally accepted theory of mentality discussed below), since simulated brains would be easier to manipulate and modify than the squishy pink glands suspended in a fishbowl of cerebrospinal fluid that nature provided us.

If such technologies are fully realized, the result may be a posthuman with either weak or strong superintelligence. A weak superintelligence is a being whose mental capacities are quantitatively superior to ours. Such a being might – for example – be able to think significantly faster than we could possibly think in principle, given that computer signals travel at the speed of light and neural signals must cross synapses between neurons (the “rate-limiting” step of neural transmission). In contrast, a strong intelligence is a being whose mind is qualitatively different from ours. Such a being might be able to grasp concepts that are in principle beyond our epistemic reach, just as the second law of thermodynamics and the concept of a fermion are permanently beyond the reach of a dog.[5]

By analogy, there may be future beings whose perceptual abilities are either quantitatively or qualitatively different from ours as well. For example, a weakly super-perceptual being might have much better visual acuity than us, or be able to see frequencies of electromagnetic radiation outside the color spectrum. A strongly super-perceptual being, in contrast, might have sensory modalities that are qualitatively different from the ten or so that are native to the human sensorium (not just sight, sound, smell, taste and touch, but modalities like thermoception, equilibrioception and nociception). A being of this sort might have an echolocatory sense or the sense of magnetoception, and thus be able to make perceptual contact with the world in ways that are in principle off-limits to us.

The second phenomenon that could result in a world full of posthumans involves artificial intelligence (AI). There are already computational devices, e.g., the calculator, that can far outperform humans in narrow domains. But no device has yet come close to equaling our general cognitive capacities. The failure of computer scientists to directly program an artificial being to possess general intelligence has led some theorists to see artificial evolution as the best approach to AI (see David Chalmers, for example). The idea is simple: even if we can’t figure out how to create an intelligent mind, we can be certain that it’s possible. After all, here we are! Maybe artificial intelligence is a problem so difficult that only an excessively “dumb” mechanism like natural selection can solve it.[6]

This line of reasoning resulted in a discipline (or “methodology”) called evolutionary robotics, which has yielded some extraordinary results. Without going into much detail, one strategy in this field involves creating a complex simulated environment for artificial beings to live in. (See Karl Siims’ video.[7]) Just like in nature, this environment probabilistically selects for those individuals within a population whose variant features (e.g., higher intelligence) are best suited for survival and reproduction. The next generation is then derived from the selected individuals plus random mutations in their (artificial) genomes to generate further intra-populational variation.

If enough generations are created, the simulation could produce an “organism” with cognitive abilities that exceed our own, or one whose mind is different in fundamental ways from ours (which was shaped by millions of years of contingent environmental conditions). Thus, the weak/strong superintelligence distinction applies here as well, although it seems to me that artificial evolution is more likely than cyborgization to result in a strong superintelligence. The reason is that cyborgization begins with humanity’s biological core while artificial evolution does not – that is, a posthuman cyborg would be phylogenetically related to humanity in a way that an artificially evolved AI wouldn’t be. The cyborg would thus be more likely to share attributes (or derivative attributes) with contemporary humans than the AI.

An advantage of artificial evolution is that thousands of generations could be simulated in a relatively short period of time. In other words, it wouldn’t take 3.5 billion years for a being with human-level intelligence to arise.[8] A problem with artificial evolution is that we currently lack the computational power necessary to simulate a sufficiently complex and rich environment for artificial beings to evolve in. On the up-side, progress in the domain of computer hardware seems to be following an exponentially exponential curve (see Moore’s law).

To summarize so far: there are fairly good reasons for expecting our descendants to be quite different from us in the most significant ways possible – that is, in terms of cognition, emotion and their perceptual abilities. This should not come as a surprise, since one of Darwin’s most important contributions was showing that evolution is a non-teleological process: there are no final states, only moments in-between other moments. It follows that, as the transhumanists make explicit, humanity is a half-baked work-in-progress (although one need not agree with normative transhumanism to accept this as true).

Section 3. The implications of these conclusions for music seem to be profound. In this section, I will explore a few of these implications. It may be helpful to divide my comments into those that are non-phenomenological (or quantitative) and those that are phenomenological (or qualitative) in nature. Taking these in order:

- The human ear can hear a range of sound frequencies – mechanical disturbances of molecules at different rates in a medium such as air. This range extends from about 20 Hz to 20,000 Hz. Frequencies below 20 Hz can be felt (as a rumbling) but not heard, while those above 20,000 Hz can be detected by animals like dogs but not us. The human ear is also most sensitive to frequencies between 500 and 4,000 Hz (which corresponds to the speech band). It follows that the music that we create won’t typically contain any frequencies outside this range.

But future posthumans with cochlear implant-like devices could potentially detect sounds above and below the human range. This would, of course, expand the realm of tonal possibility that a composer could explore. For example, an instrumental solo occur around 30,000 Hz or higher. And melodies played by bass instruments could be heard in the 10 Hz range rather than above 20 Hz (usually around 80 Hz). There could also be harmonies stacked into upper octaves that our human ears can’t hear. And so on.

Thus, it’s not hard to conceive of a future “music” listener – e.g., a cyborg whose sensorium is technologically augmented – looking back at our music and thinking of it as narrow and circumscribed with respect to the tones that it uses. They might therefore find our music to be impoverished, just as we would find music today impoverished if musicians only wrote songs with notes falling between an octave above and below middle C.

- A similar thing could be said about the dynamics of musical compositions. The human ear can detect sounds just above zero decibels (although this depends in part on the frequency of the sound). The sound of someone breathing normally is about 10 decibels, conversations occur at about 60 decibels, trains rumble by at about 100 decibels, rock concerts peak around 120 decibels – the pain threshold for humans – and, above this, a rocket engine produces roughly 180 decibels of intense sound.

The point is that it’s not hard to imagine an auditory apparatus that would enable one to hear sounds above the human pain threshold without being damaged. It follows that “music” of the future might not only contain greater tonal variation than our music, but it could exploit a wider dynamic range as well. Again, future “music” listeners might listen to our music and find it dynamically flat and boring.

- The human auditory system consists of two sensory inputs (the ears); our hearing is thus stereophonic in nature. This organization enables us to localize the source of a sound in our environment; for example, a sound that originates from the left of an individual will be slightly louder in the left ear than in the right, and it will arrive at the left ear sooner (by milliseconds) than at the right. These are called the interaural level and time differences. Mixing engineers can exploit both localization strategies to give the human listener the impression that a sound originates from one or the other side of the sound field. For example, by panning an instrument to the right, the sound produced in the right speaker becomes louder than that produced in the left. This makes the sound seem like it’s coming from the right. The engineer can also keep the decibel level between the two speakers fixed but make the right speaker produce sound slightly before the left one (by milliseconds). This too gives one the sense that the sound originates from the right side in the sound field.

The point is that our brain has two auditory inputs and differences in how these two apparatuses are stimulated determines how we interpret sounds in space. One could, therefore, imagine a future being whose auditory system consists not of two input devices but of (many) more; headphones wouldn’t fit on such a creature’s necktop. Consequently, all sorts of interpretive effects might be producible by tweaking the “interaural” levels and timings of sounds reaching different sensory inputs. The combinatory possibilities would, indeed, grow exponentially with each additional input.

- On a related note, humans have evolved to interpret reverberation as a property of sounds that originate in different places along the front-rear axis. Adding reverb to a sound is one way for mixing engineers to move that sound back and forth in the sound field. But the fact that we experience “wet” sounds as being distant and “dry” ones as being close is a product of our contingent evolutionary history in our particular world. A population of AIs living in a simulated environment could potentially evolve to have a quite different interpretation of such stimuli properties. These beings might therefore approach mixing songs quite differently than we do.

Moving on now to the more abstruse issue of phenomenology: It’s a peculiar fact about human beings (and other organisms with sufficiently complex nervous systems?) that there is something it is like to be one of us.[9] That is to say, humans aren’t just complex assemblies of stimulus-response mechanisms; creatures whose behavior is determined by classical and operant conditioning. We have inner mental lives that are qualitative/phenomenal in nature and completely subjective. Philosophers refer to this aspect of mentality as “consciousness,” “experience” or (putting them together) “conscious experience.” Phenomenology is a field dedicated to studying conscious experience – the “what it is like to be” a given organism, such as a human or even a bat.

How can one be sure that, for example, another person’s inner experience of red is the same as one’s own? What if the “raw feel” of seeing red for person A is the same as the “raw feel” of seeing green for person B? There isn’t any way of settling this issue: one cannot simply peek into another’s mind to compare states of consciousness (or to even confirm that others have consciousness at all! Maybe some of the people around you are “philosophical zombies,” or beings whose behavior is entirely indistinguishable from a non-zombie but who lacks any sort of inner experience at all).

Nonetheless, a large majority of contemporary philosophers and scientists hold that there is good reason for believing that we all have similar experiences. The thesis behind this view is called functionalism (which was alluded to above). There are two types of functionalists: reductive functionalists hold that phenomenal states depend upon and can be reduced to functional states instantiated by the brain, whereas non-reductive functionalists hold that phenomenal states cannot be reduced to functional states but that such states nonetheless depend upon functional states instantiated by the brain (or any other suitably organized physical system). The point is that both positions maintain that there exists a dependency relation between types of conscious states and types of functional states, even if the latter cannot be characterized entirely in terms of the former (which is what reduction involves).

The point is that if future beings have cognitive systems (or “brains,” in scare quotes) whose functional organization is sufficiently different from ours, then their conscious experiences of phenomena like music may be different as well. As a result, such beings might find certain chord combinations pleasing to their ears – chord combinations (say, B and Fm in contrast to G and C) that strain our ears. Or they may have a preference for musical scales that contain, say, 16 rather than 12 pitches in an octave (such as this).

They may also find certain sounds euphonious that we find unbearable to hear. Consider that there may be some causal-evolutionary explanation for why the sound of fingernails on the chalkboard produces such an unpleasant experience in human beings,[10] but there isn’t any phenomenological reason. By analogy, there isn’t anything objectively gross about the sight and smell of a rotting human corpse. For many animals, a human corpse would produce a distinct sense of delight. But our species has evolved over millions of years to find the sight and smell of dead bodies repugnant, since dead bodies are a source of disease and illness. (Thus, individuals who didn’t avoid rotting flesh would have been more likely to get sick and die before procreating.) The idea here is that, phenomenologically speaking, detecting a dead body through vision or olfaction is compatible with completely different internal experiences, i.e., there isn’t anything about decomposition itself that’s experientially disgusting (or pleasing).

To summarize these points, future beings with significantly different cognitive systems might listen to our music and find it phenomenologically unenjoyable – i.e., our music might not produce the same sort of pleasurable experiences in them that it produces in us. Similarly, we might hear their “music” and have the same reaction that a dog has to listening to Mozart.

Section 4. There appears to be good reason for thinking that our descendants either will have no desire to listen to music – that is, without scare quotes – or will be unable to appreciate it. Thus, the effort to immortalize oneself through one’s musical composition is very likely futile. Music as we know it today will one day obsolesce.

 

NOTES

[1] See…

[2] To mention only one function of clothing.

[3] Although there is some debate about how to interpret such findings.

[4] Lots of great research on BCIs at Duke; see Miguel Nicolelis.

[5] Thus, Chomskian mysteries for us may be mere problems for them. See also Colin McGinn on “cognitive closure,” or Jerry Fodor on “epistemic boundedness.”

[6] As one theorist has put it, what’s more “anti-intellectual” than natural selection itself? Thus, biological design is unintelligent design.

[7] See also Nick Bostrom’s paper on the simulation hypothesis – the idea that if we do start running such simulations, it’s significantly more likely that we’re in a simulation than not.

[8] I don’t mean to imply that the rise of humanity was in any way necessary – indeed, I just mentioned that we arose as a result of environmental conditions that were contingent. As some biologists have noted, if life on Earth were to start over again, there’s no reason whatsoever to think that anything like Homo sapiens would evolve. The key with artificial evolution is that we would be able to direct the phylogeny of the AIs by controlling the environments in which they evolve.

[9] This famous characterization of phenomenology comes from Thomas Nagel.

[10]  See this fascinating article, for example.