Today, many people all over the world practice the ancient Japanese art of suiseki. This involves appreciating the natural shape of a rock or stone. Practitioners search for, collect, and present stones that evoke other natural forms.
Ingenuity means the quality of being clever, original, and inventive.
-- Encyclopedia Britannica, britannica.com
Soul: emotional or intellectual energy or intensity, especially as revealed in a work of art or an artistic performance.
-- Oxford Languages, https://languages.oup.com/
CONTENTS:
INTRODUCTION
PART ONE: FINDING ART
1. ART AND ARTIFICE
2. A SYNOPSIS: ART AND AI
PART TWO: MAKING ART
1. ORIGINALITY
2. UNIQUENESS
3. ARTISTIC
4. INTENT
5. THE HUMAN CENTER
PART THREE: MADE ART
1. THE PRODUCTION
2. THE FINISHED PRODUCT
RECAP
*****************
INTRODUCTION
Human consciousness is the predominant determinant of "art" as an experience... But is human involvement necessary by definition in creating art?
We place more demands on the definition of "artist" than we do on the definition of "art".
Part of the Artist's "intelligence" is making that consciousness instrumental by anticipating it and targeting it with expressions.
That pressure now shows up in the debate over whether Artificial Intelligence can be not just a tool but an artist. Does it have an autonomous creative consciousness, and does that even matter?
*****************
PART ONE: FINDING ART
1. ART AND ARTIFICE
Artificial Intelligence, or "AI" labels a wide range of computing capabilities. Of course, calculating at jaw-dropping speed is a keystone of appreciating it. But otherwise, the most well known capabilities, the most popular, are currently revolutionizing most efforts to arrange and present information whether as text or as images. Its speed and strength in doing so merits thinking of it simply as an automated composing machine.
Furthermore, that AI implementation can routinely render images and texts that as far as we know have no prior exact match. Strictly speaking, it has unlimited potential to create unique artifacts. Granted, to most of us many of those "unique" products may have no unusual meaning. But for any given audience member, it is also always possible that AI will produce something that has a meaning not felt or imagined before.
Despite those two facts, a debate erupts over whether AI has "creativity".
What's actually going on there is that "creativity" is being said as a synonym for something much more ambitious: artistry.
In this debate, the argument over artistry is about whether AI's productivity can generate a product that qualifies in three ways: creative, original, and unique.
The Artist Formerly Known As
On the "nay" side of the debate, what matters most is its assertion that AI, as a singular active agent, is non-human. That is, its productive functionality -- its capability-- is not based on human physiology and psychology (body and mind), but on some other internal system.
That non-human aspect emotionally charges the debate pretty highly.
But in comparison, there is already another non-human system - Nature - that suffers relatively little outright rejection from anyone devoted to aesthetics in life.
Nature, no less so than a machine, has generative functions and behaviors independent of both of the above two defining human factors (body and mind), and it has its own substantial ability to also nurture or destroy.
In speaking of Nature, "creating" is simply thought of as instigating growth or evolution. As for its originality, we have both the presupposition and the examined belief that all new natural things derive from other pre-existing natural things, many of which we know about. But when we have a sudden awareness of something we did not already know about, we usually react to it -- appreciate and value it -- exactly as if it is singularly unique, not knowing at first if it is an exception or just a newly discovered normality.
We also know that Nature's originality is systemic. And by analogy, where the universe of natural things originates in the systemics of Nature, there is a presumption that the universe of artistic things is generated from something systemic as well.
In the debate regarding AI's capability, the "Naysayers" argue that the systemics that originate what we call Art are human factors. It isn't so much something that can be proved -- rather, it is simply part of their definition of art.
But the "Yaysayers" do not accept that definition. Rejecting the requirement that the originator of "art" is necessarily a person, they rely instead on what else sufficiently makes an experience of something cause us to call it "art"...
So, the debate again: does AI have the capability to produce creative original unique art?
We can see that the debate exists not due to conflicting facts. Instead, it is due mainly to how people define things. That includes not just the term "art" itself, but significantly ambiguous other terms in regular use as definitive requirements.
Original, creative, unique, and innovative are words that often appear in the idea of art or artistic. But too often they are mashed together into a conceptual paste not often analyzed during communication.
AI generates things in a way that puts all of those terms under a scrutiny that illuminates the difference between what they really mean and why we use them to mean something else.
2. A SYNOPSIS: ART AND AI
Fact: today, it's pretty much child's play for an AI session to be the only source of a unique, unpredicted output.
There is no expectation of spontaneous new things in Nature aside from what derives from existing natural phenomena. The most obvious analog for AI's power to generate variants from precedents is Nature. Even the most famous artists are known to adhere to the maxim, "good artists borrow; great artists steal."
We are generally unable to distinguish AI's capability from "originality".
Fact: today, AI's systemics need have no dependency on human presence or intervention to continually generate "novel" things.
Additionally, the "laws" of nature, which can operate entirely without the existence of humans, are fully accepted as its systemics. The aesthetics of Nature are not downplayed at all by their non-human origins.
We are generally unable to distinguish AI's generative capability from "creativity".
Fact: today, "art" is a kind of experience more than it is a kind of thing. Essentially, it is a subjective perception of something.
Most people, most of the time, encounter new "works of art" without knowing or proving where it came from, whether there is something else virtually the same somewhere else, or why people very dissimilar to themselves would also attribute the same kind of value to the experience it caused them.
Art works do not define for us what art is; we define what we want to appreciate artistically, by imposing criteria on it that test how its qualities make the experience useful, different, and desirable in certain ways.
"Subjective" means that we evaluate something in deciding whether we will treat it as "art". We grant it a status and role. We are continually finding new things that we decide will serve a purpose we call "art".
But the experience that we will refer to as "art" does have a characteristic underlying dynamic. I
For many people, such as those on the Naysayer side of the debate, its main feature is that we find ourselves being engaged as humans by a human. In the dynamic, we expect that a producer will intentionally attempt to influence us in a certain way.
If we are among those people, when we feel engaged:
the most consistent aspect is that the influencer is deliberately influencing.
And the next most consistent aspect is that because of the way that the effort is revealed to us, we discover that we are being influenced.
What then usually matters is how we think the influence is useful to us. In this last aspect of the engagement dynamic, the experience is something in which we evaluate the influence for its relevance.
For many people, the relevance that they desire in the experience is mainly in how the influence is (a.) about our humanity, and (b.) for our own good.
Yet most popular arguments against AI's "artistic" capability fail when tested against our increasingly routine encounters with AI-generated outputs.
Those outputs refuse to be suppressed by emotional language about originality and creativity. Regardless, a preference for works made without AI (i.e., "by humans") is tightly associated with whether a work is deemed relevant, and thereby valuable, by the observer.
Aside from preference, there is also increasingly negligible perceivable difference between the tacit complexity of how AI decided to express something and that of a human artist's decision-making; human decisioning is not necessarily visible or understandable to an observer.
Consequently, in debate, notions of intention, imagination, and innovation all objectively shift more towards how AI behaviors in effect provide the same kind of influences on eventual experiences regardless of human involvement.
*****************
PART TWO: MAKING ART
1. ORIGINALITY
Originality almost always jumps to the front of the line of thinking about art, so let's start there, which means looking into what "source" originates work.
But while a source is clearly an origin, "originality" questions the extent to which something can be a source. The problem in the debate, however, is that originality is more concerned about uniqueness than it is about sourcing of production. This is what takes us to focus on the importance of a "self" as the source.
When that source is a person it is easy to think in terms of "someone" - either one's own self, or someone else's.
Being non-human, and not "natural", AI is basically a kind of machine. For "Yaysayers" in the debate, the argument that artificial intelligence can make "original" work holds up well in a comparison between attributes of the machine's effort and those of a human's.
Human attributes of special note include being personal, subjective, and judgmental.
Personal is an idea that refers to a particular instance of an individual.The idea is all about the notion of one's "own", which means "possessed by" or "characteristic of" the self. This strongly pertains to being "autonomous".
Subjective refers to something either expressive of or from within the given autonomous individual. Here we're looking at something akin to "perspective".
Judgmental refers to decisions being made that discriminate between options. The simplest related idea is "preference".
Note how each of those ideas corresponds in AI:
If a functional (behaving) system is a distinct entity, and is serving us as a proxy for a person, then what we see within that entity as being systemic is analogous, even equivalent, to being "personal" to the system.
If that entity responds systemically to stimuli, independently emphasizing choices with consistency such as accepting and rejecting, then its behavior is as if it is "aware of itself" and, in processing, it corresponds to a person being "subjective".
If the entity also evidently evaluates what it recognizes, then its virtual subjectivity also meets our criteria for being "judgmental".
And considering those ideas -- autonomy, perspective and preference -- we can imagine this:
an auto-motivated entity,
whose activity we are not controlling,
that has its own way of recognizing things,
and that makes real time decisions without our intervention.
That is not unfamiliar or unusual behavior. Our awareness of it is generally a matter of what we want to pay tatention to.
Imagine dropping a small motorboat into the sea, with its motor already running. This unmanned boat is going to try to keep moving. But as it reacts to the environmental forces on it, it is going to do things without our immediate control, and even if we understand the forces, we may not find the boat's direction and states at any moment, whether immediate or future, to be reliably predictable or desirable.
Its reactions are like decisions, and while any one of them might be theoretically predictable with enough knowledge of the exact circumstances, the accumulation of them over time still may not be something we can foresee as its behavior, path, or longevity in the water. To make it more predictable to us, we have to either exert more control over it, or give it more built-in control of a type that we understand.
Imagine further:
if the entity also recognizes the consequences of its decisions, and then makes its own predictions and further choices based on those consequences, then there is a continuous feedback loop between action and reaction that is separate from direction attempted by other autonomous entities.
In effect, this is what we mean by comparing the behavior to being a "self". More importantly, when it comes to identifying a given source's activity, the more unpredictable things are, the more "original" we say they appear to be.
Behavior automation isn't news. And it's no coincidence that most popular representations of encounters with AI, especially the scary ones, have featured highly anthropomorphized robots.
Here is where our language suspiciously comes into play. The big question here is, if the entity can effectively make real-time decisions without us, then how do we know where to draw the line between iits unpredictability and its so-called "originality"?
Looks Like A Duck, Walks Like A Duck...
Despite all that, the "Naysayers" in the debate might still insist that this description of behavioral independence does not really portray capability for artistic "originality" and, rather, at best shows only mimicry.
The counter to that objection is that expert mimicry is very hard - sometimes even impossible - to distinguish from the real thing. That mimicry includes personality, subjectivity, and judgment.
When it comes to impersonation, we're already quite familiar with two pretty big things: acting, and computer animation. In both cases, we are not preoccupied with how the character was generated; we are preoccupied with, and care about, the character.
So, with the projection of a "self" not being necessarily authentically human nor exclusively behavior by humans, the Nay-sayer's need some other proof against AI being capable of originality.
One move here is to just demote originality's importance as a human-centric defining characteristic of what we decide to call "artistic" or "art". For Naysayers, it would mean looking at some other qualifying conditions instead that could be fundamentally "human".
Next, then: what about uniqueness?
2. UNIQUENESS
Creativity is simply the ability of work to produce things in a new or unscripted way from what is at hand.
Most often we decide to say "creative" when we specifically intend to refer to something new that is both different from the norm and is desireable - two ways that we see it as being valuable.
But in that reference the key word "new" seems only to be about what is made. This can obscure the more important aspect of creating, which is about the way of making.
On the production side, the usual implication of calling activity "creative" is that it is "imaginative".
But how do we define "imagination"? The easy way to do that is as "a capability to describe something distinctive that doesn't yet exist" in the present time and place.
The key points in the above notes are existing and normal. Against those two conditions, "creativity" is expected to produce something new having value. But both of those terms have limitations often completely taken for granted.
"Norm" doesn't exist outside of designating some given set of circumstances. What is normal in one place is not normal in some other. "New" has the same limitation: it is context-specific.
The only way out of this contextual boundary is for something to be unique -- thereby new and different in all contexts. That's great if uniqueness can be proved; but by who?
This has a tricky solution. The charm of uniqueness inspires the desire to attribute originality to a specific party. We take an outcome, and trace it back through some process of development that begins with a single (unique) known party, the "originating cause" of the thing's existence. The debate's Naysayers are comfortable that one particular party can cause many different outcomes, but when it comes to being "creative" they are not comfortable with the idea that one specific outcome can be originally caused by many different parties.
Preferring humans as the originators is at least convenient in this regard, because after all, "no two people are alike". Every human "self" is a unique self, so tracing something back to any single person seems to kick off the notion of originality quite well, which implies the newness and difference that "creativity" refers to as something worth caring about.
But how does that hold up vis-a-vis AI?
If you are in the Naysayer's camp, you begin with this: no machine knows how to start itself the first time without a human making it happen, so we thereafter attribute whatever the machine does to an intent of a human.
In that way the human, not the machine, "retains" the exclusive status of being the "originator" and the "self" that matters. Thus, in production, the machine is not deemed to be expressing its own self; rather, it is seen as expressing the human's self -- regardless of the outcome.
(This corresponds to everyone's normal sense of what a tool or an instrument is, even including awareness that an active tool or instrument can possibly have effects not expected, not intended, or not desired.)
Yet, at the end: seeing the product of effort as an expression of some "self" is highly subjective. We either are told to believe it and agree to, or we impose it ourselves as a hypothesis that is assumed to be true unless disproved.
The reality check: different cultures recognize different things as art; fakes are routinely mistaken for real; and reproduced artifacts of unknown authorship or sources are definitely capable of very high aesthetic impact, at minimum through style.
Most persons, most of the time, encounter most new art without knowing who made it, how it was made, or what else is out there that is virtually the same. That doesn't make them deny that the works are art. That is, little about the work product encountered is necessarily going to refer to any particular human "self" other than to one who claims to be it.
For most persons, most of the time, the relationship of a particular human maker to art is attributed, not proven; and, the attribution may not even be correct, yet still not diminish the perception of the work as being "art".
In debate, this logically removes the necessity of identifying a specific person to show that something is artistic or indeed art.
But when we talk about art, we are still definitely interested in how it was generated.
Sharing Is Caring
So what about A.I.? AI is generative. Once it starts working on something, it has the capability to find, relate, reformulate, and express any number of things that have not before been anticipated, precisely predefined, or even recognized by the observer. Encounters with those outputs of AI do not exclude experiences that people commonly relate to as art.
But on the Nay side of the debate, "creativity" is appreciated more when it is perceived to be something that a person is demonstrating.
In attributing artistic value, people who say that AI (a machine) is not creative are usually very specifically protective of humans having a special status in the act of creation. They might go so far as to say that the value of "artistic" creativity is in its being an expression of the "self".
For those on this side, it just makes us feel good, or more understanding, about the fact that we too are human. Creativity by people is an observable bulwark against our shared existential limitations of being human.
Yet giving creativity greater value in that particular way does not restrict actual creativity's definition, nor essence, to human production.
3. ARTISTIC
In both thought and speech, we call something "Art" because we find that something exerts a significant influence on us when we interpret it from our perspective comprising our "artistic" criteria.
The question remains, what are those criteria?
We saw above that two conventional criteria used for evaluating production as "Artistic" -- originality and uniqueness -- need not be human-centric.
We also saw in the "reality check" above that in fact "art" is a kind of experience more than it is a kind of thing. Essentially, it is a subjective perception of something.
On the human-centric side of things, one could specifically assert that all art is actually a certain type or mode of human interpretation of an encounter with something.
For example, we already have meaningful phrases like "the Art of cooking", "the Art of motorcycle maintenance", "the Art of public speaking", and many others in our common parlance. These refer to a particular way of doing things. No less, they refer to a way of seeing things that, as we know, are also seen from other points of view.
We choose to see things a certain way, and it changes our experience of them. This means that in order to see if something is art, our terms of interpretation will also direct what we look for.
4. INTENT
But doesn't the production of "art work" -- the work of art -- by definition also intend for its product to be interpeted in a certain mode? Answering this question requires us to know how to recognize intention whenever it is there.
What are the recognizable ingredients of intention that we call "artistic"?
For many people, the top candidate is that we find ourselves being engaged as humans by a human.
When we feel engaged:
For most of us, the most consistent aspect is that the influencer is deliberately influencing.
And the next most consistent aspect is that because of the way that the effort is revealed to us, we discover that we are being influenced.
What then usually matters is how we think the influence is useful to us. In this last aspect of the engagement dynamic, the experience is something in which we evaluate the influence for its relevance.
For many people, the relevance that they experience is mainly in how the influence is (a.) about our humanity, and (b.) for our own good.
Now, what about A.I.? What we already saw is this:
Nothing about non-human "AI" disqualifies it from being the source (the originating "self") of that ultimate experience, other than our individual reluctance to acknowledge it in that role.
Yes, for all of us there can be significant uncertainty about whether its influence is intentional.
Because of that, it may or may not be important to rate its "originality". For example, if we are delighted, then do we care? But if we feel manipulated, then is it okay?
And yes, there is frequently significant opacity to how it is doing it (functionally, technically, practically).
Because of that, it may or may not be important to rate its "creativity". For example, is it following rules? Is it truthful? Who cares?
But neither of those facts means that AI cannot generate the experience we call art. Rather, the more an instance of AI-generated experience is perceived to be pursuing this human engagement, the more we are willing to evaluate both what AI is doing and what it produces as being "art"...
5. THE HUMAN CENTER
Many people just want the "human system" of creativity to have a different status from the systemics of both Nature (ecology) and Mechanics (automation).
For those people: there is an enormous amount of tribal language in use to account for humans as the exclusive originators of behaviors, products, and affects that we will then call "artistic". They want those behaviors, products, and affects to be about us. They want "human nature" - not anything else - to be the instrument of creativity that ultimately produces the experience we want to call "art".
But there can be no experience without ideas and feelings.
Despite the sense of "natural instinct" being a force that exists outside of thinking, we know that the animal world is well populated with creatures that have both ideas and feelings. But in exploring this generative system called human nature, what definitions of "idea" or "feeling" would show them as being distinctively human properties not found in nature or mechanics??
The Human Criteria
One way to tackle defining the usage of terms is to trace and clarify connections within our conventional vocabulary. With "experience", this is a cloud of associated words that includes:
feelings, emotion, subjectivity
sensation, sentience, consciousness, recognition
and, aesthetics
In writing this article, a great convenience was to use the curated, crowd-sourced information in Wikipedia. After comparing definitions of these words and eliminating redundancies and tautologies, the main discovery that surfaced is that the meaning of experience has an emphasis on:
having "self-awareness"
the non-existence of emotions without cognition
and, the role of perspective on preference
Those three matters emerged as if they are the three "dimensions" of experience.
As criteria of having an experience, they translate ideas and feelings into:
knowing that one's self exists
knowingly having responses to stimuli
distinguishing stimulus+response conditions by type and importance
Those are also basic to being able to intend and to influence, the key parts of deliberately producing engagement. We recognize them as part of exercizing our human nature.
What is it about AI that corresponds to those conditions? Is any correspondence close enough to mean in effect that human-centricity in artistic production is not a requirement but instead an option?
*****************
PART THREE: MADE ART
1. THE PRODUCTION
Can we accept A.I. being an equivalent, or at least a proxy, for the human in production?
There is a presumption of something systemic about how humans generate and produce new instances of things, too, with some of those being unprecedented new things. We accept (even if only tacitly) that there are laws of human creativity too, even if we don't know as much about them as we do about the laws of nature. We don't assume that they are the same laws, but they no less rule the potentials of our production.
In the human "system", the feedback loop (consciousness) is believed to have already been (biologically) "designed-in" at birth, waiting for any inhibitor to be removed so that it can be started. And once it has started, we refer to it as "a mind of its own"... We know that we can influence it, but we don't necessarily control the diversity of its potential as a producer.
Automation can create machine behaviors that mimic people producing things. The stronger the mimicry, the harder it is to tell the difference. AI behaves as if it has internal rules that it always enforces; it makes decisions; it prioritizes what it will do about consequences; and it subjects those follow-up acts to its own rules.
It is capable of producing in ways that are so complex that the outcomes are not readily navigated back to their origins except by the system itself. It is effectively autonomous, and the mysteriousness of its full capability is indistinguishable from the mystery of a human's "genius". Likewise, its outputs can evolve, as in nature, from the feedback loop of weighing consequences and making (possibly unprecedented) adjustments.
An automated entity can "act like it knows what experiences are"... That is, it can operate in a way that deliberately fosters and accomodates conditions and events that make up certain human experiences. Those operations manifest as expressions that have influence, and the influence gets interpreted and evaluated by an observing (encountering) person. At that point, the "experience of" art is more dependent on the observer than on the producer.
Perfect Is The Enemy Of Good Enough
But one clear-cut difference presumed between any machine and a human is that a machine does not care why it is doing anything. Even if the machine is behaving as if it has experiences, it really only has states (conditions).
The counterpoint to this is that automation is normally created with built in enforcement of what NOT to do, in the interest of people's preferences and safety. The scope of these limitations is part of designing the system, just as maturity and culture is instilled in children and adults. There is no barrier to designing AI carefully.
There are also two other things that probably initially separate AI from being deemed an artistic creator of things.
One is spontaneous imagination. Again this refers to "a capability to describe something distinctive that doesn't yet exist" in the present time and place. Most people assume that automation operates only within the bounds of what it's creator already knew.
But they discount that apparent randomness is also something we know how to intentionally generate in an automated system.
The other is a unique intrinsic logic or sensitivity in its operational behavior and decisions. Machines are commonly believed to be preconfigured, so while a machine may be idiosyncratic that is not seen as something it can willfully use or avoid independent of its "programming", to uniquely distinguish itself. It doesn't have "free will".
What makes this problematic is that the uniqueness of AI logic may exist only at a level of complexity that most people will not be able to navigate and account for it.
What makes that limitation relevant here is that a human artist's production of work can also easily have a degree of complexity exceeding accountability by other people. At a certain point, when encountering finished work, any observable difference between A.I. and the human worker might in effect be neglible - not really a useful discriminating factor when it comes to experiencing the product as art or not.
These notes really boil down to one thing:
whether AI can or cannot be an artistic original creator is one matter; but it is not the same matter as whether AI can or cannot be an originating creator of artistic things.
2. THE FINISHED PRODUCT
We put a lot of pressure on whether we will or will not accept the products of activity as being Art instead of Not Art...
What are we basing that decision on?
The most important observation is that what we call art, however it is produced, causes us to have certain experiences. And the more intensively we are affected, the more "powerful" we say that the work of art is. When it comes to experiencing the product, the final reality is that categorizing the experience as being artistic or not is our way of evaluating things.
Bluntly put, the arguing about the artistic status of whatever is produced is all about why we care.
THE RECAP
Art is a type of experience caused by a type of influence.
We don't define art as being "whatever an artist makes".
In every case, the "artist" is simply the producer of the thing that causes the experience we have and call "art".
The criteria that we have in an "artistic" evaluation center on why we care.
We want artists to be people who have certain characteristics; but our notions of what makes an artist's work valuable - our subjective qualifying preferences - too often overstate what we actually know.
The experience doesn't need to be creative, original, unique or new. All of those attributes have a place in one value system or another, and we can impose any of them on why we want the experience - but we already know that two different people can encounter the same circumstances and exercise very different interpretations because they brought different values to the encounter.
The influencer originates the experience. Influence on people succeeds because it stimulates and relates to people according to the reasons why people feel and think the way people do. And we commonly note the experience as "being affected by".
While we deeply appreciate how human nature promotes production of artistic experiences, we don't reject the meaningful life experiences of what non-human systems produce, nor do we reject those systems for not meeting our ambitious criteria for humans.
We don't struggle or debate over whether Nature has "originality" and "creativity", because it seems pointless to stretch beyond the obvious: that nature is vastly, and continuously, generative and productive. The proliferation of weeds from seeds in unmanaged land. The blooming of populations of micro-organisms. The spontaneous emergence of rain showers, or even just clouds, from air masses meeting each other. The formation of mountains. The emergence of new species.
We even have a stock label for Nature's systemics: "Laws". The laws of "natural" creativity are, by definition, NOT human: completely absent of humans, nature would just go on independently without us.
With AI's productivity, we encounter the activity of a system - a machine - that can generate a very strong influence having value that is not predetermined by our preferences of how it is produced. As argued earlier:
The creativity of the system is not defined by our acceptance or resistance to what it produces.
Neither does our awareness of what activates the system define it as having originality or not.
Forget about uniqueness and innovation; the limits on the scope or acuity of our own awareness might render our perception of both of them fictional illusions.
Our suspicion of AI's "motive" is at the core of our resistance to accepting it as an originator of art. Yet in general, we are more cautious about why a person uses a machine than we are about the machine itself.
In "matters of the art" we are not predisposed to accept "intelligence" about us as a substitute for empathy or sympathy. But what machines do absolutely IS about people or being human. Machines are designed, specifically to directly address a human goal or objective.
Knowing that is what allows us to accept a relationship with machines even when their complexity exceeds our casual ability to understand how they work.