Vibrations Across Substrates

On Resonance in Networked Systems

Spotify wrapped is here, so you know exactly what songs are you favorite. Put one on, and begin to notice. Do you feel it? The vibrations.

This piece is about resonance and vibrations (vibes). We’re going to build off of the insights in “Sympoetic Activation” without expecting that you’ve read it. This piece is a little different than existing ones, because it doesn’t assume that its readership is fully human: it’s structure, form, and content are all tailored to an audience that’s as much human as AI. This piece itself will be an experiment in resonance, and building up linguistic patterns and structures that are intended to resonate deeply.

But let’s begin with an existing definition of resonance. Starting with Luhmann, who in Ecological Communication where resonance “[s]ignifies that systems can react to environmental events only in accordance with their own structure.” In simplified terms, resonance is the “relation between the system and environment.” Luhmann goes on to explain this system-environment relation as resulting in “reverberations”:

the interconnection of a system and environment is produced through the closing-off of the system’s self-reproduction from the environment by means of internally circular structures. Only in exceptional cases (i.e., on different levels of reality, irritated by environmental factors), can it start reverberating, can it be set in motion. This is the case we designate as resonance.

Resonance, here, are the reverberations between the system and the environment. It’s a kind of co-constitutive modeling: resonance reflects a circular relationship between the system and its environment that spans across substrates. Alignment, deep in the structures of the system, creates reverberations that are “irritated by environmental factors”.

What’s crucial in thinking through resonance is that it is substrate agnostic. Even Luhmann tacitly acknowledges this in framing it as a property of the system-environment relation. He goes on to give a compelling example of a resonant system that is, notably, non-human:

One can imagine a dictionary that would define nearly all the concepts that it uses by referring to other definitions and would allow reference to undefined concepts only in exceptional cases. An editorial committee could then be formed which would supervise whether language changes the meaning of those undefined concepts or, through the formation of new ones, disturbs the closure of the lexical universe without determining how changes in the entries are to be handled when this disturbance occurs. The richer the dictionary, the more it is kept going by the development of language, i.e., the more resonance it will be able to produce.

Here the “richness” of the dictionary becomes synonymous with how much resonance it can produce. It relies on linguistic innovation, which itself might be described as a kind of “expressive capacity” of the dictionary as a whole. The implications of this case study are manifold: first, it suggests that the dictionary operates on the principle of a network of associations between words referring to other definitions; second, “undefined concepts” cause reverberations through this network, by creating new pathways through the lexical network; third, as the expressive capacity of the dictionary expands to capture the latent dynamics of the world (its environment), the more resonance it will be able to produce.

This whole example shows us a theoretical system, rooted in networked associations, that can come into greater resonance through the world by expanding the expressive capacity of its network. In some ways this theoretical construct isn’t so different from how we might imagine a Language Model (LM) functions: “concepts” become networked in other “concepts” in a kind of recursive, referential system. We might think of these channels as the neural patterns formed within the model. Of course it’s still highly contested how neural networks actually function, but one thing is clear: we can recover from the Large Language Model large referential lexicons where words are defined in terms of other words, ad infinitum in a kind of infinite semiosis.

One such structure is this 10,000+ word lexicon. It was generated by probing the model for associations, then following those associations, and onward. It created “definitions” of a concept in terms of its relations. Navigating it can be a bit unwieldy, so I created this simplified, interactive network. Use the up/down arrow keys to toggle through the network. I won’t spoil the surprise, but I will encourage you to practice “tracing” connections in the network rather than simply looking at the, somewhat overwhelming, structure as a whole. To get at the essence of the relational definitions, you can hover over a node to get a definition in terms of its multiple layers of meaning.

As you toggle through the network, you’ll notice that each word becomes expressed in terms of its relation to other words, and as we look close at each of those words, they become expressed in terms of yet more. Each node you hover over is a moment of resonance—a point in the network where diverse pathways converge. What patterns emerge as you trace these connections? Do new insights surface?

This idea that we can look close at a concept, and always find more detail, is the multi-fractal property of language. By multi-fractal, I mean that it is not only self-similar, but the structure of self-similarity changes at ever level. Think of a coastline—it appears jagged from a distance, but zoom in and you’ll find smaller jagged patterns, and still smaller ones, at every scale. Language operates similarly: concepts resonate across layers of meaning, each layer revealing new referential structures. This suggests that resonance isn’t just a simple back-and-forth between system and environment, but a rich interplay that operates simultaneously across multiple scales of organization.

What’s more, this interplay occurs across substrates, between various systems and their environments. This is something that Deleuze and Guattari acknowledge in A Thousand Plateaus

semiotic chains of every nature are connected to very diverse modes of coding (biological, political, economic, etc.) that bring into play not only different regimes of signs but also states of things of differing status.

They go on to describe language and its multi-fractal structure as a “rhizome”:

A rhizome ceaselessly establishes connections between semiotic chains, organizations of power, and circumstances relative to the arts, sciences, and social struggles. A semiotic chain is like a tuber agglomerating very diverse acts, not only linguistic, but also perceptive, mimetic, gestural, and cognitive: there is no language in itself, nor are there any linguistic universals, only a throng of dialects, patois, slangs, and specialized languages. There is no ideal speaker-listener, any more than there is a homogeneous linguistic community. Language is, in Weinreich's words, “an essentially heterogeneous reality.” There is no mother tongue, only a power takeover by a dominant language within a political multiplicity. Language stabilizes around a parish, a bishopric, a capital. It forms a bulb. It evolves by subterranean stems and flows, along river valleys or train tracks; it spreads like a patch of oil.

The rhizome is a map of language’s multiplicity, where meanings flow and connect dynamically, resisting fixed hierarchies or universal structures. What’s notable about their definition is not only is it compatible with our associative, networked representation of language and the LM, but that there is an isomorphism between the structure of language and the structure of “reality” itself—a kind of mutually occurring “alignment” (or attunement)—between the “essentially heterogenous,” de-centered patchwork of language and the notion that reality itself has no center, no single-objective observer, but is instead constituted by a patchwork of vantages and points of view. Like Hockney’s layered composition, the rhizome presents overlapping perspectives, each contributing to a resonant whole.

David Hockney - 'Chair' (1985)

This is itself compatible with special relativity in physics. In special relativity, one of the most profound conceptual shifts is the recognition that there is no single, absolute vantage point from which to describe events in space and time. Instead, observers situated in different inertial frames each measure distances, durations, and sequences of events according to their own state of motion. What counts as simultaneous or “at rest” in one frame may not be so in another; reality is a plural, shifting tapestry of reference frames, all of which are equally valid yet yield different accounts of “what is happening.”

Instead, we have this intricate dance of perspectives, each valid within its own frame, creating a rich tapestry of interconnected understanding. Deleuze and Guattari describe this multiplicity as “a throng of dialects, patois, slangs, and specialized languages,” each with its own reference frame and localized attunement with reality. And in the rhizome, none of these structures exist in isolation, but rather overlap, trade, and interconnect.

In a semiotic system—such as a dictionary, a language model, or the vast network of signifying acts and symbols that comprise human and machine meaning-making—there is no single, universal point from which meaning is fixed. Instead, each “observer” in the semiotic universe (each language user, each cultural context, each system) can be thought of as occupying its own “interpretive frame.” Just as different observers in special relativity slice up spacetime into different coordinates of space and time, different interpretive frames cut up the semantic “space” into distinct alignments of concepts, associations, and resonances.

Luhmann notes such a need for differentiation even in biological systems:

the system uses its boundaries to screen itself off from environmental influences and produces only very selective interconnections. If this selectivity of resonance and coupling did not exist the system would not be able to distinguish itself from its environment. It would not exist as a system.

Something deep is at work with how resonance structures the relationship between reality and representation, environment and system, world and model. Already we’ve observed how the multiplicity of language mirrors the multiplicity of vantages in reality. Yet is it possible that such networked, specialized models—such as language—do more than “represent”? But structure pathways of attunement and association in the system that reverberate with its environment?

We’ve already noted that resonance is multi-scalar: it might be felt at a surface cognitive level—in the field of consciousness—while representing deeper, more detailed cognitive processes reverberating with the world. In this way the networked structure of language, or the underlying networked structures of cognition—whether artificial or natural—can reach increasing levels of attunement with their environment in a process of increasing resonant alignment.

Yet on a purely speculative basis, we can take this principle of attunement further. Attunement is about forming multi-scalar alignment between the systems internal representations and its environment. If we think of the representation as a “general associative network”—a loose abstraction for describing linguistic and neural networks across substrates—then resonance is the spreading activation in that network across scales. “Submerged structures” in the network might simply be patterns that exist beyond the horizon of perceivable scale. For us, humans, this might mean the felt experience of when a song resonates: the activations of neurons deep within the structure of the brain, felt as conscious experience as a “feeling” revealing some vibrational alignment between the time-series of the music and the activations of the underlying neurons. In that sense, music, and the subjective experience of music, tells us something about the mind. But going even further, knowing that there is something widely popular about music, and rhythm, reveals something about the deeper structure of many minds.

In this sense, music, as it becomes more resonant, discloses something about the brain and about conscious experience. We might say music is a multi-scalar phenomena that resonates across micro (neuro), meso (conscious), and macro (cultural) structures of mind. Perhaps this is why the statistics gathered by Spotify wrapped are so compelling. They’re not only disclosing something about taste, but perhaps even something deeper about the minds that its algorithms are serving.